CONVENZIONE

per la realizzazione del “Progetto di ricerca su cellule CAR-T per patologie ematologiche maligne e

per tumori solidi”

TRA

ALLEANZA CONTRO IL CANCRO (di seguito per brevità “ACC”) con sede in Roma, via Ribotta 5 CAP 00144 e CF 97262520584, legalmente rappresentata dal Direttore Generale, dott. Xxxxx Xx Xxxxx, nato a Udine il 18/11/1955 e CF DPLPLA55S18L483A

E

IRCCS CROB (Centro di Riferimento Oncologico della Basilicata, (di seguito per brevità “CROB ”) con sede in Rionero in Xxxxxxx,Xxx Xxxxx Xxx xx0 XXX 00000 X.X. 93002460769 e P.IVA 01323150761 legalmente rappresentato dal Direttore Generale, Dott.ssa Xxxxxxxxx Xxxxx, Legale Rappresentante pro tempore nato a Potenza il 12-12-1969 C.F. MCCCST69T52G942M

(ACC e CROB di seguito congiuntamente denominate “Parti”)

Premesso che

- Il combinato disposto normativo di cui all’art. 1, comma 523 della legge 30 dicembre 2018, n.145, all’art. 23-quater, comma 4, del decreto legge 23 ottobre 2018, n.119, convertito, con modificazioni dalla legge 17 dicembre 2018, n.136., nell’ambito di Alleanza Contro il Cancro, CUP E84119003590001, ha previsto finanziamenti, pari ad € 10.000.000,00 (Euro diecimilioni/00) ripartiti negli anni 2019 e 2020, in favore degli IRCCS afferenti alla Rete Alleanza contro il cancro per le nuove tecnologie CAR-T per la cura dei tumori;

- Il 15/04/2019 l’impianto progettuale è stato condiviso con gli IRCCS partecipanti alla riunione

tenutasi presso il Ministero della Salute;

- Il Ministero della Salute con Messaggio WF ID 2019013651 del 27/11/2019 ha comunicato ad Alleanza Contro il Cancro, in adempimento alle disposizioni di legge di cui sopra, l’erogazione di un importo pari ad € 5.000.000,00 (Euro cinquemilioni/00) comunicando inoltre le tempistiche e le modalità per la rendicontazione, la relazione scientifica e gli adempimenti conseguenti;

- In data 5 luglio 2019 sono stati raccolti i contributi al progetto dei singoli IRCCS e il medesimo è stato assemblato a cura del Coordinatore Scientifico, Xxxx. Xxxxxx Xxxxxxxxx dell’Ospedale Pediatrico Bambino Gesù, inviato nella sua versione finale al Ministero della Salute in data 25/11/2019 e presentato a tutti gli IRCCS partecipanti nella riunione del 20/12/2019 presso il Ministero della Salute;

- Alla realizzazione del progetto di cui sopra partecipano, oltre ad ACC, i seguenti IRCCS: Ospedale Pediatrico Bambino Gesù IRCCS - Roma, Fondazione Policlinico Universitario Xxxxxxxx Xxxxxxx IRCCS - Roma, Istituto Europeo di Oncologia S.r.l. - Milano, Fondazione Piemonte per l’Oncologia - Candiolo, Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori, IRST IRCCS - Meldola, Ospedale San Xxxxxxxx S.r.l. - Milano, Fondazione IRCCS Istituto Nazionale Tumori - Milano, Istituti Fisioterapici Ospitalieri - IRCCS - Istituto Nazionale Tumori Regina Elena - Roma, Istituto Nazionale Tumori IRCCS "Fondazione X. Xxxxxxx" - Napoli, IRCCS Centro di Riferimento Oncologico della Basilicata - Rionero in Vulture, Istituto Tumori "Xxxxxxxx Xxxxx XX", IRCCS - Bari, Istituto Clinico Humanitas - Humanitas Mirasole

S.p.A. - Rozzano, Istituto Oncologico Veneto IRCCS - Padova, Centro di Riferimento Oncologico di Aviano IRCCS - Aviano, Fondazione di Religione e di Culto “Casa Sollievo della Sofferenza” Opera di San Pio da Pietrelcina – San Xxxxxxxx Xxxxxxx e, come indicato nell’allegato 1 (Atti parlamentari), l’Ospedale X. Xxxxxxx – Fondazione Tettamanti di Monza, la società MolMed S.p.a. e l’Istituto di Biostrutture e Bioimmagini del CNR di Napoli (di seguito per brevità definiti “Partner”).

TANTO PREMESSO SI STIPULA E SI CONVIENE QUANTO SEGUE

Art. 1 Premesse

Le Premesse e gli Allegati sono parte integrante e sostanziale della presente Convenzione

Art. 2

Oggetto della Convenzione

La presente Convenzione regola la realizzazione del “Progetto di ricerca su cellule CAR-T per patologie ematologiche maligne e per tumori solidi” richiamato in premessa (di seguito per brevità “Progetto”) che, allegato alla presente Convenzione, ne costituisce parte integrante (allegato 2).

Art. 3 Durata

La presente Convenzione entrerà in vigore dalla data di avvio del Progetto identificata nel 27 novembre 2019 e terminerà i suoi effetti il 26 novembre 2021 alla conclusione del secondo anno di attività, salvo eventuali proroghe concesse dal Ministero della Salute.

Art. 4

Responsabili scientifici ed amministrativi

1. Per le finalità di cui alla presente Convenzione, le Parti designano:

• il Xxxx. Xxxxxx Xxxxxxxxx quale Coordinatore Scientifico del Progetto (di seguito per brevità

“Coordinatore Scientifico”);

• il Xxxx. Xxxxxxxx Xxxxxxxxxxxx quale Responsabile Scientifico per le attività di ricerca previste per IRCCS CROB nell’ambito del Progetto;

• il Dott. Xxxxx Xx Xxxxx quale Coordinatore Amministrativo del Progetto (di seguito per

brevità “Coordinatore Amministrativo”);

• la dr.ssa Xxxxxx Xxxxxxxxx Xxxxxxx quale Responsabile Amministrativo per l’IRCCS CROB.

Art. 5

Coordinatore Scientifico, Responsabile Scientifico e Comitato Tecnico Scientifico

1. Il Coordinatore Scientifico del Progetto, il Coordinatore Ammnistrativo di ACC ed i Responsabili Scientifici dei Partner leader di WP progettuali costituiscono il Comitato Tecnico Scientifico (di seguito per brevità il “CTS”) del Progetto, di durata pari a quella del Progetto medesimo.

2. Il CTS è presieduto dal Coordinatore Scientifico del Progetto e dal Coordinatore Amministrativo di ACC.

3. Il CTS del Progetto potrà essere convocato, a cura del Coordinatore Scientifico in relazione ad opportune verifiche sullo stato di avanzamento delle attività svolte dai Partner, sull’attinenza delle stesse agli obiettivi del Progetto, ovvero dal e/o dal Coordinatore Amministrativo in relazione ad opportune verifiche sullo stato dei rendiconti economici.

4. Il CTS si occuperà, nell’interesse comune, delle problematiche connesse all’eventuale mancato o ritardato invio al Coordinatore Scientifico delle relazioni scientifiche o al Coordinatore Amministrativo delle rendicontazioni economiche intermedie e finali, che rallenti, impedisca, o comunque incida negativamente sulla predisposizione dei rendiconti scientifici per il Ministero della Salute, comportando potenziali ripercussioni sull’erogazione dei relativi fondi.

Art. 6

Modalità di erogazione del contributo

1. Il finanziamento assegnato al Progetto dal Ministero della Salute è di € 10.000.000,00 (Euro diecimilioni/00) diviso in due annualità di pari valore, € 5.000.000,00 (Euro cinquemilioni/00) ciascuna.

2. Successivamente alla erogazione da parte del Ministero della Salute di ciascuna quota annuale ad ACC, questa provvederà al trasferimento all’ IRCCS CROB della quota spettante come da tabella Budget a pag. 98 del progetto allegato e limitatamente alla quota di spettanza approvata dal Ministero della Salute.

3. Il trasferimento degli importi avverrà, in presenza di attestazione di credito inviata da CROB ad ACC, mediante bonifico bancario sul c/c intestato a IRCCS CROB Centro di Riferimento Oncologico della Basilicata presso Banca Popolare di Bari – Filiale di Potenza - Viale Marconi n°194 85100 Potenza IBAN XX00X0000000000000000000000:

Causale di versamento “Progetto CAR-T”.

4. Il CROB si impegna a trasmettere alla Direzione Generale di ACC nota di rendicontazione con tutti gli estremi dei costi sostenuti, con le istruzioni e la modulistica che saranno comunicati dal Ministero della Salute.

5. In ogni spesa dovrà essere indicato il Codice Unico di Progetto.

6. Il trasferimento avverrà in regime di esclusione dal campo IVA, ai sensi del DPR 622/1972 e successive modificazioni, in quanto ricade nella gestione dei fondi stanziati per attività di ricerca e sperimentazione. Questa specifica destinazione ne esclude l’utilizzo per fini diversi da quelli stabiliti nel piano economico del progetto.

Art. 7

Relazione scientifica e rendicontazione economica

1. Per la realizzazione del Progetto, ACC e CROB sottoscrivono la presente convenzione che disciplina tutti gli aspetti di gestione del Progetto.

2. Il CROB provvederà all’invio delle relazioni scientifiche, redatte secondo il Template che sarà fornito dal Coordinatore del Progetto, e delle rendicontazioni economiche secondo il seguente schema:

a. Invio da parte di ogni Partner della relazione intermedia e finale al rispettivo Responsabile di Work Package come riportato nel Progetto entro le seguenti scadenze:

i. Relazione scientifica intermedia: 31 luglio 2020

ii. Relazione scientifica finale: 16 dicembre 2021 (20 giorni dalla conclusione del Progetto)

b. Invio da parte di ogni Responsabile di Work Package come riportato nel Progetto della relazione intermedia e finale al Coordinatore del Progetto entro le seguenti scadenze:

i. Relazione scientifica intermedia: 10 settembre 2020

ii. Relazione scientifica finale: 26 dicembre 2021 (30 giorni dalla conclusione del Progetto)

c. Invio da parte del Coordinatore del Progetto della approvazione della relazione intermedia e finale ad ACC entro le seguenti scadenze:

i. Relazione scientifica intermedia: 18 settembre 2020

ii. Relazione scientifica finale: 10 gennaio 2022 (45 giorni dalla conclusione del Progetto)

d. Invio da parte di ogni Partner del Rendiconto economico intermedio e finale al Coordinatore Amministrativo entro le seguenti scadenze:

i. Rendiconto economico intermedio: 10 settembre 2020

ii. Rendiconto economico finale: 10 gennaio 2022 (45 giorni dalla conclusione del Progetto)

e. Invio da parte di ACC della relazione intermedia e finale al Ministero della Salute entro le seguenti scadenze:

i. Relazione scientifica e rendiconto economico intermedi: 20 settembre 2020

ii. Relazione scientifica e rendiconto economico finali: 25 gennaio 2022 (60 giorni dalla conclusione del Progetto).

3. Per gli aspetti relativi alla rendicontazione economica intermedia e finale verranno forniti da ACC i modelli ministeriali che ciascuna Parte dovrà utilizzare e trasmettere via PEC alla Direzione Generale di ACC al seguente indirizzo: xxxxxxxxxxxxxxxxxxxxxx@xxx.xx.

Art. 8 Tracciabilità

Il Codice Unico di Progetto (CUP) è E84119003590001

Art. 9

Norme applicabili alla ricerca e autorizzazioni allo svolgimento delle attività

1. Il Progetto dovrà essere condotto secondo le vigenti disposizioni in materia e nel più scrupoloso rispetto del protocollo, dei principi etici e deontologici che ispirano le attività mediche.

2. Nel caso in cui per lo svolgimento delle attività previste dalla presente convenzione siano necessarie autorizzazioni/pareri da parte del Comitato Etico competente o di altro organismo di sorveglianza o controllo, ciascun partner, per le attività di propria competenza, si farà carico di ottenere tali autorizzazioni che saranno formalmente comunicate ad ACC e al Coordinatore Scientifico del Progetto.

Art. 10 Recesso

1. Ad ognuna delle Parti, ai sensi dell’art. 1373 c.c., è attribuita la facoltà di recedere dalla presente convenzione. Il recesso dovrà essere esercitato mediante comunicazione scritta da notificare con raccomandata A.R. e effetto decorsi trenta giorni dalla data di notifica dello stesso. Il recesso ha effetto per l’avvenire e non incide sulle attività già espletate in relazione alla Convenzione e al Progetto.

2. In caso di recesso, le Parti concordano fin d’ora di portare a conclusione le singole intese operative e gli impegni singolarmente assunti alla data di estinzione e perdita di efficacia della Convenzione.

Art. 11 Risoluzione

La presente Convenzione potrà essere risolta qualora una delle Parti dichiari e dimostri

l’impossibilità, per causa a quest’ultimo non imputabile, di proseguire la collaborazione.

Nell’ipotesi in cui una Parte si renda inadempiente, la Parte creditrice della prestazione inadempiuta dovrà intimare per iscritto all’altra Parte, mediante lettera raccomandata A.R., di adempiere entro e non oltre 7 giorni dal ricevimento della diffida, decorso il quale, la presente Convenzione si intenderà risolta.

Sono fatte salve le prestazioni rese fino alla data di notifica della diffida ad adempiere e l’eventuale

richiesta di risarcimento per i danni derivanti dall’inadempimento.

Art. 12 Modifiche

1. Le disposizioni di cui alla presente Convenzione potranno essere successivamente modificate solo in forma scritta dalle Parti, da persone munite di poteri di rappresentanza in nome e per conto delle Parti. Il presente atto viene firmato dalle Parti contraenti e sottoscritto per presa visione dal Coordinatore Scientifico del Progetto Xxxx. Xxxxxx Xxxxxxxxx.

Art.13

Proprietà intellettuale, utilizzazione e pubblicazione dei risultati

1. Ciascuna Parte resterà unica titolare dei diritti di proprietà intellettuale e industriale relativi:

- alle proprie conoscenze pregresse;

- alle proprie conoscenze parallele.

2. La titolarità di tutte le informazioni, invenzioni, cognizioni, ancorché non brevettabili, nonché dei brevetti e di ogni altro diritto di privativa industriale risultanti dal progetto sarà quella della comproprietà tra i partner in pari quota, salvo che si possa stabilire una diversa ripartizione della titolarità sulla base di una accertata diversità dell’importanza del contributo prestato da ciascuna Parte al conseguimento del risultato inventivo. Le Parti possono concordare con accordi successivi alla presente convenzione clausole specifiche relative a proprie situazioni particolari relative a proprietà intellettuale ed industriale.

3. Nel corso dello svolgimento delle attività, i responsabili scientifici dovranno comunicare reciprocamente i trovati suscettibili di protezione derivati dal Progetto.

4. Utilizzazione e pubblicazione dei risultati: trattandosi di progetto multicentrico, nell’ipotesi di risultati realizzati congiuntamente, le Parti si impegnano ad effettuare congiuntamente le pubblicazioni, purché tali pubblicazioni non compromettano la protezione dei risultati. Le pubblicazioni ottenute dai singoli IRCCS o ottenute congiuntamente dovranno riportare il nominativo di autori, la loro affiliazione e citare nel testo il nominativo del progetto e l’ente erogatore dei fondi utilizzati.

5. Le Convenzioni con i Partner di cui all’Art. 7 dovranno prevedere l’impegno di tutti i Partner alla stipula, entro e non oltre 6 mesi dalla data di avvio del progetto, di un apposito Consortium Agreement che disciplini tutti gli aspetti relativi alla gestione della proprietà intellettuale.

Art. 14 Assicurazioni e sicurezza

1. Il progetto rientra nell’ambito della normale copertura assicurativa prevista per l’attività di ricerca

in ciascuna Parte.

2. Ciascuna Parte provvede alla copertura assicurativa di legge del proprio personale che, in virtù della presente Convenzione, verrà chiamato a frequentare la sede di esecuzione delle attività.

3. Nel caso sia necessario richiedere l’approvazione di un protocollo di studio derivante dal Progetto da parte del comitato etico competente, la Parte provvederà a dotare la sperimentazione della specifica copertura assicurativa prevista.

4. Laddove il personale di ciascuna Parte dovesse recarsi per le finalità Progettuali presso la sede dell’altra, sarà tenuto ad uniformarsi ai regolamenti disciplinari e di sicurezza in vigore nelle sedi di esecuzione delle attività attinenti alla presente convenzione. Le Parti, al tal fine, si impegnano al rispetto della normativa vigente in materia di igiene e sicurezza sul lavoro di cui al D.Lgs 81/2008 e s.m.i.. Le Parti dovranno fornire dettagliate informazioni sui rischi specifici esistenti negli ambienti in cui il personale ospitato è destinato ad operare, sulla propria organizzazione della sicurezza e sulle misure di prevenzione e di emergenza adottate in relazione all’attività che verrà svolta, anche sulla base del documento di sicurezza elaborato dall’Ospedale, custodito presso quest'ultimo in base al D.lgs. 81/2008 e successive modificazioni e integrazioni. La responsabilità della osservanza di tali prescrizioni è in capo alla struttura ospitante.

Art. 15 Anticorruzione

1. Le Parti dichiarano di aver adottato misure di vigilanza e controllo ai fini del rispetto e

dell’attuazione delle previsioni del D. Lgs. 8 giugno 2001 n. 231.

L’IRCCS CROB dichiara di avere adottato il proprio Piano della Prevenzione della Corruzione 2020- 2022 e il Codice di condotta (Regolamento recante codice di comportamento dei dipendenti pubblici, a norma della art. 54 del decreto legislativo 30 marzo 2001, n, 165). Tali documenti sono consultabili sul sito internet dell’IRCCS CROB all’indirizzo xxxx://xxx.xxxx.xx nella sezione “Amministrazione Trasparente”.

Nel quadro di quanto sopra, ciascuna Parte s’impegna ad agire nell’esecuzione della presente Convenzione nel rispetto della normativa vigente con correttezza e trasparenza evitando comportamenti o azioni che possano configurarsi quale mala gestio con finalità corruttiva e più in generale che si pongano in contrasto con i principi, valori e regole di condotta etica tali da poter generare per l’altra Parte responsabilità da atto illecito. In tale contesto, inoltre, le Parti s’impegnano a collaborare in buona fede ai fini di facilitare la piena e corretta attuazione dei correlati reciproci obblighi.

Xxxxxxxx parte s’impegna a segnalare all’altra Parte qualsiasi violazione alle regole di trasparenza e correttezza di cui venisse a conoscenza con riferimento ai soggetti operanti per conto di quest’ultima affinché sempre quest’ultima possa adottare i conseguenti provvedimenti nei confronti dei responsabili ove la segnalazione risultasse fondata.

2. Le Parti riconoscono ed accettano reciprocamente che il puntuale rispetto degli obblighi previsti in tema riveste carattere essenziale e che qualsiasi violazione delle disposizioni di cui al presente articolo autorizzerà la Parte adempiente a tali obblighi a risolvere unilateralmente la presente convenzione ai sensi dell’art 1453-del c.c.

Art. 16

Riservatezza e Trattamento dei dati personali

Nell’esecuzione del presente accordo, le Parti e il personale impegnato nell’espletamento delle attività

afferenti il progetto di ricerca si impegnano ad osservare la massima riservatezza, a non divulgare, né

utilizzare per alcun scopo diverso da quello necessario per lo svolgimento delle attività previste dalla presente convenzione le informazioni di carattere sanitario, scientifico, aziendale e, più in generale, le informazioni di volta in volta qualificate confidenziali e/o riservate che siano state prodotte dall'altra parte nell'ambito delle attività progettuali, a non divulgarle a terzi e a utilizzarle esclusivamente per il raggiungimento delle finalità oggetto della presente convenzione.

L’obbligo della riservatezza non si applica alle informazioni:

- che le parti divulgatrici già detengono al momento della definizione del presente protocollo;

- che sono di pubblico dominio;

- che le parti ricevono in modo legittimo da terze parti senza essere soggette all’obbligo di

riservatezza;

- che le parti divulgatrici sviluppano o hanno sviluppato in modo autonomo al di fuori del presente contratto;

- che sono state esplicitamente esentate dall’obbligo di riservatezza dalla parte che le

comunica.

Le Parti nell’esecuzione delle attività previste dalla presente Convenzione si impegnano a trattare i dati personali, di cui vengano per qualsiasi motivo a conoscenza, nel rispetto degli obiettivi di cui ai precedenti articoli e in conformità a quanto disposto dal Regolamento (UE) 2016/679 (di seguito “RGPD”) del Parlamento Europeo e del Consiglio del 27 aprile 2016, nonché dalle correlate disposizioni legislative e amministrative nazionali vigenti, con le loro eventuali successive modifiche e/o integrazioni (di seguito, collettivamente, “Leggi in materia di Protezione dei dati”).

Le Parti si qualificano come autonomi titolari del trattamento ai sensi dell’art. 4 paragrafo 17) del

RGPD.

Laddove, nel caso di accordo tra le parti o se ciò sia previsto dal Progetto, le parti siano contitolari del trattamento oppure titolari e responsabili del trattamento, si applicheranno rispettivamente, gli articoli 26 e 28 del RGPD, con conseguente necessità di stipulare gli accordi richiamati dalle predette disposizioni.

Per le finalità del Progetto saranno trattati dati personali riferiti alle seguenti categorie di interessati: soggetti partecipanti al Progetto, loro rappresentanti e persone che operano per le Parti. Tali interessati sono informati sul trattamento che li riguarda a mezzo di idonea informativa. Per le finalità del Progetto saranno trattati le seguenti tipologie di dati personali: dati di cui all’art. 4 n.1 del RGPD; dati rientranti nelle categorie “particolari” di dati personali -e in particolare dati relativi alla salute e alla vita sessuale, dati genetici -di cui all’art. 9 del RGPD. Tali dati saranno trattati nel rispetto dei principi di liceità, correttezza, trasparenza, adeguatezza, pertinenza, necessità, minimizzazione e gli altri principi di cui all’art.5, paragrafo 1 del RGPD.

Il trattamento verrà effettuato, per lo svolgimento del Progetto, con la conseguente applicazione dell’art. 89 RGPD e tenendo conto di quanto stabilito, in particolare, nei “considerando” 33, 50, 112, 156, 159, 161 del preambolo del RGPD.

Tutti i dati di persone fisiche afferenti alle Parti, verranno reciprocamente trattati dai due titolari del trattamento in conformità al Regolamento 679/2016/UE, al D.lgs. 196/2003 come novellato dal D.lgs. 101/2018 e s.m.i. e per le seguenti finalità:

a) adempimenti di specifici obblighi contabili e fiscali;

b) gestione ed esecuzione del rapporto e degli obblighi contrattuali;

c) attività di ricerca e sperimentazione;

d) finalità connesse ad obblighi previsti da leggi, da regolamenti o dalla normativa comunitaria nonché da disposizioni impartite da Autorità a ciò legittimate dalla legge;

e) gestione del contenzioso;

f) finalità statistiche;

g) servizi di controllo interno.

Le previsioni di cui al presente articolo assolvono i requisiti di informativa di cui all’articolo 13 del

regolamento 679/2016/UE.

Le Parti dichiarano quindi espressamente di essere a conoscenza dei diritti a loro riconosciuti dagli articoli 15, 16, 17, 18, 20, 21, 22 del Regolamento 679/2016/UE in particolare del diritto di richiedere l’aggiornamento, la rettifica o la cancellazione dei loro dati personali.

Le obbligazioni e le previsioni del presente articolo continueranno ad essere valide ed efficaci anche successivamente al termine della convenzione dei suoi effetti, indipendentemente dalla causa per cui sia intervenuto.

11.11 Qualora una parte accerti una violazione dei dati personali, si impegna a comunicarlo all’altra entro 48 ore dall’accertamento della violazione, ferma restando l’autonomia della stessa nella valutazione della sussistenza delle condizioni e nell’adempimento degli obblighi previsti dagli artt. 33 e 34 del RGPD

Art. 17 Comunicazioni

per iscritto agli indirizzi e alle attenzioni di seguito riportati.

Gli indirizzi e le persone cui le comunicazioni devono essere indirizzate possono essere modificate da ciascuna delle parti previa comunicazione scritta all’altra parte secondo le modalità sopra riportate.

Riferimento

Indirizzo

IRCCS CROB

Dott. ssa Xxxxxx Xxxxxxxxx Xxxxxxx

Xxx Xxxxx Xxx xx0 00000 Xxxxxxx xx Xxxxxxx (XX)

tel. + 00 0000-000000

email: xxxxxx.xxxxxxx@xxxx.xx

Alleanza Contro il Cancro

Dott.ssa Xxxxxx Xxxxxxxx

Viale Regina Xxxxx, 299 – 00000 Xxxx Tel. x00 00 0000-0000

email: xxxxxxxx@xxxxxxxxxxxxxxxxxxxxxx.xx

Art. 18

Legge applicabile e foro competente

Le Parti si impegnano a risolvere gli eventuali conflitti concernenti l’applicazione, interpretazione,

esecuzione e risoluzione della presente convenzione, mediante bonario componimento.

In caso contrario, espressamente convengono di accettare la giurisdizione esclusiva del Tribunale di Potenza.

Art. 19 Registrazione e spese

1. La presente convenzione viene sottoscritta in forma digitale ai sensi dell’art.21 del decreto

legislativo 7 marzo 2005 n. 82 e s.m.i.

2. Conformemente al DPR del 26.10.1972 n. 642 e s.m.i. è previsto l’assolvimento dell’imposta di

bollo in modo virtuale a carico di ACC.

Art. 20 Disposizioni finali

1. Eventuali modifiche alla presente Convenzione potranno essere effettuate, previo accordo fra le Parti, solo tramite stesura di apposite modifiche scritte.

2. Le parti si danno reciprocamente atto che il contratto è stato negoziato in ogni sua parte e che non trovano pertanto applicazione le disposizioni di cui agli artt. 1341 e 1342, c.c.

Letto, confermato e sottoscritto con firma digitale, ai sensi dell’art.21 del decreto legislativo 7 marzo 2005,

n. 82 e s.m.i.

Roma, (data della sottoscrizione come quella dell’ultima firma digitale apposta)

per Alleanza Contro il Cancro Il Direttore Generale

Dott. Xxxxx Xx Xxxxx

per IRCCS CROB Centro di Riferimento Oncologico della Basilicata Il Rappresentante Legale pro tempore

Il Direttore Generale Dott.ssa Xxxxxxxxx Xxxxx

Il presente atto è sottoscritto per presa visione

Il Coordinatore Scientifico Xxxx. Xxxxxx Xxxxxxxxx

Allegato 1 Atti Parlamentari ‌

Allegato 2 Progetto

Italian Health Ministry ‌

Research project on CAR T cells for hematological malignancies and solid tumors

Conducted under the aegis of Alliance Against Cancer (ACC) network Project coordinator: Xxxx. Xxxxxx Xxxxxxxxx

Participating Institutions and their acronyms:

• IRCCS, Ospedale Pediatrico Bambino Gesù, Roma (OPBG)

• IRCCS, Fondazione Policlinico Gemelli, Roma (FPG)

• IRCCS, Istituto Europeo di Oncologia, Milano (IEO)

• IRCCS, Fondazione del Piemonte per l'Oncologia, Candiolo (TO) (FPO, Candiolo)

• IRCCS, Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori, Meldola (FO) (Meldola)

• IRCCS, Ospedale San Xxxxxxxx, Milano (OSR)

• IRCCS, Fondazione Istituto Nazionale dei Tumori, Milano (INT)

• IRCCS, Istituto Regina Elena, Roma (IRE)

• IRCCS Fondazione Casa Sollievo della Sofferenza, X. Xxxxxxxx Xxxxxxx (CSS)

• Alleanza Mediterranea Oncologia in Rete, Rete AMORe (IRCCS Fondazione

X. Xxxxxxx Xxxxxx, IRCCS Crob di Rionero in Vulture, IRCCS Istituto dei Tumori Bari)

• Istituto Clinico Humanitas, Milano (ICH)

• Istituto Oncologico del Veneto (IOV)

• Centro di Riferimento Oncologico, Aviano (CRO)

• MolMed Società per Azioni, Milano (MolMed)

• Xxxxxxxx Xxx Xxxxxxx, Xxxxx

• CNR Istituto di Biostrutture e Bioimmagini, Napoli (IBB- CNR)

Index

Abstract 4

Introduction 5-6

WP1 Description 7-11

WP2 Description 12-34

WP3 Description 35-52

WP4 Description 53-79

WP5 Description 80-93

WP6 Description 94-97

BUDGET 98

Abstract

The concept of combining the MHC-independent antigen recognition of monoclonal antibodies (mAbs) and the effector function of T cells has been a true revolution in the field of immunotherapy, leading to the development and clinical validation of chimeric antigen receptor (CAR)-engineered T cells. The advantages of CAR T cells over mAbs are multiple and mainly rely on the ability of CAR T cells to eliminate tumor cells even at low levels of antigen expression, on self-amplification upon activation and on better distribution into the tissues. Moreover, CAR T cells may enter the memory pool and provide long-lasting protection against tumor recurrence. CD19- directed CAR T-cell therapies for B-cell malignancies are currently the most advanced T-cell therapies tested in clinical trials. The results obtained in patients with B-cell lymphoproliferative disorders led to the recent approval by the Food and Drug Administration (FDA) and subsequently by the European Medicines Agency (EMA) of two CD19-directed CAR T-cell therapies. Despite significant efficacy, unfortunately, CAR-T cells have been associated with a suboptimal safety profile, since severe, sometimes life-threatening or even fatal toxicities have been reported in patients with B-cell neoplasms given CAR T cells. Moreover, the clinical outcome of the few patients with solid tumors or hematological malignancies other than B-lymphoproliferative disorders has been less encouraging. In light of these considerations, this collaborative multi- Institutional project will implement novel pre-clinical approaches to render CAR T-cell applicable to patients with solid tumors and other hematological malignancies, as well as to improve the safety and efficacy of the approach. Another important field of research will be that focused on making CAR-engineered cells universally usable; this will further expand accessibility to this kind of therapy. Finally, we will investigate strategies to develop an automatic production chain for CAR T cells and for manufacturing less costly and immediately available CAR cells.

Introduction

Immunotherapy of B-cell lymphoproliferative malignancies with the use of genetically modified T- lymphocytes redirected to the tumor target through the expression on the cell surface of a Chimeric Antigen Receptor (CAR) has obtained remarkable results. In fact, the use of CAR T-cells has been shown to induce complete remission (CR) in more than 80% of patients with B-cell precursor acute lymphoblastic leukemia (Bcp-ALL) either refractory or relapsed after conventional treatments, including allogeneic hematopoietic stem cell transplantation (allo-HSCT), and in about 50 % of patients with refractory/relapsed B-cell non-Hodgkin lymphoma (NHL). These remarkable results led to the market authorization by FDA, initially, and subsequently by EMA of two drug products (Kymriah®, marketed by Novartis, and Yescarta®, marketed by Gilead), directed against the CD19 antigen,. Both agencies approved the therapy with CAR T cells for pediatric/young adult patients (up to 25 years of age) with Bcp-ALL in second relapse of the disease or with refractory disease or in first relapse after allo-HSCT and for adult patients with diffuse large B-cell lymphomas (DLBCL) and primitive mediastinal large cell B lymphomas (PMBCL), either refractory or resistant to two or more systemic therapy lines.

However, in the context of other hematological malignancies (i.e. acute myeloid leukemia, AML) and of solid tumors, so far, the efficacy of the approach is much more limited. In particular, several hurdles to the development of an effective product emerged. The identification of a suitable target antigen, expressed preferentially or exclusively on the surface of the tumor cells and absent or at very low density on normal tissue, is often a challenge. Moreover, solid neoplasms are characterized by the presence of a tumor microenvironment (TME) that proved to be extremely hostile for CAR T cells, owing to an acid and hypoxic ambient and to the presence of immunosuppressive cells and soluble factors, ultimately leading to the inhibition of the activation and proliferation of CAR T cells. Notably, several clinical and preclinical models clearly showed that the in vivo activity of CAR T cells correlates strictly with their ability to traffic and home into the tumor site, to proliferate in the tumor microenvironment and to persist overtime.

Together with the remarkable efficacy of the treatment with CAR T-cells In B-cell neoplasia, however, some peculiar and sometimes severe, adverse events have emerged, raising the need for the optimization of the safety profile of the approach. In the studies reported to date, a significant proportion of patients developed severe adverse events, sometimes requiring an intensive care management. Exceptionally, toxic-related deaths have been reported, as well.

Beside the clinical and functional hurdles mentioned, other relevant aspects related to the manufacturing processes limit the broader diffusion of the approach to all the patients that could potentially benefit from its use. They include: a) the still limited manufacturing capacity, unable to provide individualized products in due time to satisfy the clinical need; b) the lack of standardization of the production processes among different production sites, making each experience isolated and not comparable; c) the lack of a network coordinating the distribution of products on the national territory; d) the relevant costs of production of CAR T cells. In order to overcome, at least partially, these hurdles, creating a GMP-grade production process with a high level of automation and using closed system could represent a significant improvement.

The present research project promoted by the Ministry of Health and developed under the aegis of Alliance Against Cancer (ACC) aims at dealing with all the reported limitations, leading to a common and coordinated management through the creation of a collaborative network joining the expertise of the different participating Institutions. In details, the research activities will be developed over 2 years according to the following working packages (WP):

WP1: Coordination; development of an automatic production chain for CAR T cells and of novel approaches for manufacturing CAR cells immediately available and/or less costly.

WP2: Identification of new targets selectively expressed on neoplastic cells and compatible with the clinical use of CAR T.

WP3: Preclinical development of CAR-killer cells against novel cancer targets in solid tumors and hematologic diseases other than B-cell neoplasms.

WP4: Novel strategies to increase activation, expansion, survival, tissue penetration, cytotoxicity and monitoring of CAR expressing cells.

WP5: Strategies to improve the safety profile of CAR T cells. WP6: Legal aspects and intellectual property.

WP1: Coordination; development of an automatic production chain for CAR T cells and of novel approaches for manufacturing CAR cells immediately available and/or less costly.

Leader: OPBG and ACC; Participants: MolMed; San Xxxxxxx

Rational

In the last few years, a new form of treatment of hematological malignancies has emerged: it is based on the use of genetically modified T-lymphocytes redirected to the tumor target, using proteins expressed on the lymphocyte cell surface and called Chimeric Antigen Receptors (CARs). In particular, CAR T-cells have been shown to be particularly effective in the context of B-cell lymphoproliferative diseases, including B-cell precursor Acute Lymphoblastic Leukemia (Bcp- ALL) or Non-Hodgkin's Lymphomas of B-lymphocyte origin (B-NHL). In the context of these diseases, CAR T-cells were directed against the CD19 antigen, expressed on the surface of malignant cells.

The use of CAR T-cells has been shown to induce complete remission (CR) in more than 80% of patients with refractory BCP-ALL or relapsed after conventional treatments, including allogeneic hematopoietic stem cell transplantation (allo-HSCT), and in about 50 % of patients with refractory/relapsed B-NHL. A significant proportion of patients who responded to treatment with CAR T cells obtained sustained CR, this observation suggesting that these patients can be considered cured through this approach. Thanks to the conduction of international multicenter studies, two types of CAR T-cells (Kymriah®, marketed by Novartis, and Yescarta®, marketed by Gilead) directed against the CD19 antigen have obtained market authorization initially by FDA, and subsequently by EMA, for the following indications:

- patients up to 25 years of age with Bcp-ALL in second relapse of the disease or with refractory disease or in first relapse after allo-HSCT;

- adult patients with diffuse large B-cell lymphomas (DLBCL) and primitive mediastinal large B- cell lymphomas (PMBCL) either refractory or resistant to two or more systemic therapy lines.

The present WP1a aims at coordinating the activities of the entire project, including also the development of shared research projects, identification and validation of innovative CAR cell platforms that will allow safe administration of allogeneic products. Such an innovation will be applied to either the autologous setting (T cells derived from each single patient) or to a third-party donor banks that can guarantee the availability, as quickly as necessary, of the genetically modified product, resulting in a reduction in waiting time for CAR cell manufacturing and production costs. Finally, through the development of specific and advanced research projects, the network aims at overcoming the main limitations of the present approach, which emerged from the studies reported to date, with the goal of further improving its effectiveness and treating an increasing number of patients. In particular, the project has the ambition to broaden the spectrum of disorders and number of patients potentially treatable with CAR cells through the development of a sophisticated screening system of molecules expressed on the surface of tumor cells for identifying new targets for CAR cells, and the subsequent creation and development of the relative engineered molecules. The network is also aimed at enhancing the anti-tumor activity of CAR constructs by implementing strategies that increase tissue distribution, penetration in solid mass and persistence of CAR T cells. Finally yet importantly, it aims at maximizing the safety of CAR cells, thus making this revolutionary therapeutic approach more easily accessible even to centers with less developed expertise in the field.

TASK1.1 The Coordination Program identifies and reports on each research activities carried out by the single WP of this project.

The Coordination Program will support the entire project achieving efficiency and guiding participants on how and when they must act to reach the milestones. OPBG will be in charge of the scientific coordination of the project. ACC will manage the budget of the entire project as well as the implementation of the project plan and the time schedul. The Environment Management of the project will facilitate the collaboration and integration of the WPs; it will help the single WP manage the change processes across the activities to reach the predicted deliverables. The Coordination program will also help identify potential drawbacks and risks before they affect the chance of success of the overall project.

TASK1.2 Facilitate the scientific collaboration between the involved groups.

The Coordination program will aim at optimizing the emerging scientific opportunities across the involved IRCCS/Company/Institutions and the WPs, indentifying public health challenges, or knowledge gaps that deserve special consideration. Such research benefits from conducting or supporting additional research that involves collaboration between IRCCS/academic institutions and the pharmaceutical company MOLMED, involved in the project in view of its relevant contribution in the field of CAR T cell/gene therapy in our Country.

TASK1.3 Management of the Communication Plan between ACC, Italian Health Ministry, Public and Patient foundations.

Communication between ACC, Italian Health Ministry, Parent and Patient associations will be instrumental to allow the dissemination of the information that will be generated across the WP.

TASK1.4 Management of the Communication Plan between IRCCS/Company/Institution Communication among the different involved units will be crucial to speed-up the activities of each WP, and to facilitate the dissemination of the data across the WPs.

Milestones

- Coordination of the activities of the entire project through the organization of virtual meetings of all the WPs every 2 months (1-24 months)

- Coordination of face-to-face meetings every 6 months involving all the WP leaders (1-24 months)

- Coordination of 2 annual meetings to update the Italian Health Ministry, Parent and Patient foundations and Hospitals on the progresses of the project (1-24 months)

- Coordination of the communication with the Scientific Advisory Board (2-24 months)

- Coordination of the interactions/collaborations with the private companies potentially interested in the different programs of the project (1-24 months)

TASK2: Definition of the manufacturing and release flow for CAR ATMP to reduce the costs related to the generation of CAR cells

Rational

Together with the remarkable treatment efficacy of CAR T-cells, some peculiar and severe adverse events were recorded. Cytokine Release Syndrome (CRS) and neurotoxicity represent the two most relevant side effects associated with CAR T-cells. Moreover, potential limitations to a wider use of CAR T-cells emerged; they are represented mainly by the limited manufacturing capacity (a specific CAR T-cell product must be created for each individual patient) and by the relevant costs approved in the US for the commercialized drug products. The extraordinary potential to change the therapeutic scenario of several hematological conditions with CAR T-cells is now evident, but there

is certainly the need to better understand and define the mechanisms that regulate efficacy and toxicity of this immunotherapy, as well as to potentiate the therapeutic approach to reproduce comparable results in patients with solid neoplasms. Furthermore, the development of production processes capable of implementing the manufacturing capacity and reducing the costs related to the generation of gene-modified CAR cells could increase the number of patients who might benefit from these therapies and optimize their economic sustainability in the context of a public national system, like that existing in Italy. Finally, the considerable interest that can be raised by an approach of such a clinical relevance could lead to a virtuous collaboration with pharmacological Companies, further enriching the panorama of potential opportunities related to the development of CAR T-cell-based therapies on a national basis. In fact, a potential virtuous synergism between academic institutions and Companies can be created with the goal of reproducing also in Italy the win-win collaboration model that certainly had an important role in the development of CAR T cells in the USA.

The present WP is aimed at dealing with several of the factors limiting the full potential use of CAR T-cells for the treatment of all patients who could potentially benefit from them. In particular, we will take in account that current production capacities are lower than expected and there is no network, to date, that allows the coordination and management of the distribution of products on the national territory. Moreover, the lack of standardization of production processes makes the experiences currently available in Italy isolated, non-homogeneous and, therefore, not comparable. Last, but not least, the production costs of CAR T-cells are significant and potentially limiting their use on a large scale. The creation of a national network dedicated specifically to the development of the CAR cell technology, under the aegis of the ACC, has the purpose of reducing these costs and, at the same time, of implementing the production capacity through the harmonization and automation of the processes. Moreover, the network can be open to the participation of pharmaceutical and industrial partners, upon expressions of interest.

TASK2.1. Generation of the automatic production chain for CAR T cells.

An automated production chain will be defined in compliance with the current GMP rules, with a high degree of automation and closure of the system itself. Through this implementation, it will be possible to: 1) facilitate multiple contemporary productions in the same environment; 2) reduce the costs related to the personnel employed for production; 3) carry out the production in unclassified environments, or class C; 4) limit the occurrence of biological contamination; 5) generate highly standardized procedures.

TASK2.2. Generation of an allogenic production for CAR cells.

A research project aimed at obtaining allogeneic products that can integrate or replace patient- specific productions of CAR cells will be carried out in OPBG and in San Xxxxxxx Units. The production of allogeneic CAR cells will allow the generation of banks for cellular products that can be readily sent to the requesting clinical center whenever needed, avoiding an ad hoc production for each patient. This approach would have several advantages, both managerial and clinical. For the first aspect, the creation of allogeneic CAR banks will solve the issues of the GMP facility capability, as well as it will reduce the impact of the costs of each treatment. From the clinical point of view, the main expected improvements include: 1) patients will not undergo lymphocyte- apheresis, avoiding the risk of obtaining low-quality T cells from heavily pre-treated patients; 2) the risk of production failures of the cellular product will be decisively reduced; 3) the product will be readily available, whenever needed, upon request, avoiding the waiting time of the production, sometimes incompatible with the rapid progression of the disease; 4) the possibility to design clinical studies that envisage the use of multiple doses of CAR cells.

Milestones

- OPBG will perform validation runs to obtain the automation of the CAR T platform (1-24 months)

- Molmed will optimize the large-scale viral production to achieve high expression of CAR in the target cells (1-24 months)

- OPBG will generate an allogenic CAR platform based on Natural Killer (NK) and γδ/T cells.

- San Xxxxxxx will generate an allogenic CAR platform based on cytokine-induced killer (CIK) cells from cord blood (CB) units.

Feasibility, risk assessment and contingency plans

The feasibility of the WP1 proposal is guaranteed by the leading role played by OPBG in the immunotherapy field and its long lasting experience in driving coordinative roles of several researche grants. In particular, OPBG has started and almost completed two Phase I-II academic clinical trials using CAR T-cells for treatment of paediatric patients affected by Bcp-ALL, B-NHL or neuroblastoma, enrolling more than 30 patients so far. Moreover, MolMed will provide its relevant contribution as Company to support the clinical translation of the project, focusing on research, development, manufacturing and clinical validation of cell & gene therapies for the treatment of cancer and rare diseases. Indeed, based on a dual business model leveraging on common technological assets, MolMed is able to manage from pure to clinical research, development, manufacturing and regulatory authorization, market access and pricing & reimbursement. Finally, the San Xxxxxxx Hospital Unit has developed and optimized an adoptive immunotherapy approach based on donor-derived CIK CAR cells employed in patients relapsing after HSCT and obtained with a non-viral platform, currently implemented in a Phase I clinical trial. Its contribution will be of a great value to guarantee the feasibility of the system optimization.

TASK1.1 The Coordination Program identifies and reports on each research activities carried out by the single WP of this project.

The Milestones and deliverables of each WP will be managed by each WP leaders who will refer to the Coordination program for the steps achieved or missed. The Coordination Leader will monitor and guide each WP leader to properly address major risks affecting the overall success of the WP. The Coordination Leader will act with a problem-solving strategy to ensure the achievement of the milestones.

TASK1.2 Facilitate the scientific collaboration between the involved IRCCS.

The involvement of different activities from each WP carries the potential risk of an inadequate integration amongst the WPs. Thus, the Coordination Leader will act to maximize the collaboration between the different Units, at both academic and industrial level.

TASK1.3 Management of the Communication Plan between ACC, Italian Health Ministry, Public and Patient foundations.

Since the goal of the project is to create novel therapeutic option for patients, it will be crucial to plan a correct communication of the results. Thus, the Coordination Leader, together with ACC, will provide a professional communication of the information to the Parent and Patient foundation/associations. Moreover, the coordination leader will represent with ACC the interface with the Italian Health Ministry for the constant monitoring of the project conduction.

TASK1.4 Management of the Communication Plan between IRCCS/Company/Institutions Considering the large number of Units involved in the project, the absence of a defined and well- established Communication Plan would negatively influence the project success. Thus, the Coordination Leader will schedule and organize meeting to facilitate the dissemination of the data/results obtained across the WPs.

TASK2.1. Generation of the automatic production chain for CAR T cells.

The generation of an automated process for the manufacturing of CAR T is an ambitious task that also involve the contribution of the expertise of medical engineers, as well as of informatics in order to develop dedicated software to manage novel instrumentation. If the milestone will be judged not feasible to be reached in 2 years, OPBG will focus on the validation of the best and more automated procedure, even if not completely and fully automatic.

TASK2.2. Generation of an allogenic production for CAR cells.

The previous experience of OPBG and San Xxxxxxx will ensure the feasibility of this specific aspect of the project. We do not envisage relevant issues. The only issue might be related to the large-scale validation of the allogenic products, due the lack of adequate bioreactors. If it will be the case, we will consider a medium scale of production using culture bags.

WP2: Identification of new targets selectively expressed on neoplastic cells and compatible with the clinical use of CAR T

Leader: Fondazione Policlinico Gemelli (FPG); Participants: Istituto Europeo di Oncologia (IEO), Ospedale Pediatrico Bambino Gesù (OPBG), Fondazione Casa Sollievo della Sofferenza (CSS), Istituto Nazionale dei Tumori (INT), Fondazione del Piemonte per l'Oncologia (FPO), Istituto di Biostrutture e Bioimmagini- Consiglio Nazionale delle Ricerche (IBB-CNR)

Rational

To date, the clinically most relevant CAR T cell target molecules are expressed on the surface of neoplastic elements, but also on healthy normal cells. CAR T cells directed against CD19 or CD22 are very effective in Bcp-ALL or B-NHL (Xxx et al., 2018; Xxxxx et al., 2014; Xxxxx et al., 2018; Xxxxxxx et al., 2017; Xxxxxxxx et al., 2017). However, both surface molecules are also expressed by normal B lymphocytes. Fortunately, the patient survives this off-target activity without substantial or unacceptable clinical risks as the off-target killing of B-cells can be compensated through regular immunoglobulin replacement therapy. We are still facing the unmet medical need to expand CAR T-cell therapies to other cancers, including solid tumors, such as advanced colorectal cancer (CRC) or glioblastoma (GB) and other leukemias including acute myeloid leukemia (AML) or T-cell acute lymphoblastic leukemia (T-ALL).

For efficiently targeting tumors, the identification of molecules preferentially, or, at best, selectively present only on neoplastic cells represents a goal of primary importance to extend the use of CAR T cells. The research groups participating in WP2 of the project will contribute to the identification or validation of targets expressed on the surface of tumor cells and to the generation of new CARs targeting these targets. According to the expertise of the participating groups, priority will be given to the identification of targets that are expressed by brain tumors, sarcoma, T-ALL and AML (OPBG), carcinomas of different origin (INT), leukemia-initiating cells (LIC) in leukemic cell subsets in T-ALL (CSS), and cancer stem cells (CSCs) of solid tumors, such as GB and metastatic CRC (FPG).

The research groups participating in this project have contributed to the characterization of LICs (Xxxxxxx et al., 2018; Xxxxxxx et al., 2015; Xxxxxxx et al., 2012) and discovered the existence of CSCs in many solid tumors including CRC and other tumors (Xxxxx et al., 2008; Xxxxx-Xxxxxxx et al., 2007; Xxxxx-Xxxxxxx et al., 2008; Xxxxxx et al., 2010). Furthermore, biobanks of molecularly annotated primary spheroids and organoids from different tumors, such as GB and metastatic CRC (Manic et al., 2018; Xxxxx-Xxxxxxx et al., 2008) were established. These primary tumor cultures of GB and metastatic CRC readily generate orthotopic xenografts in mice that reproduce histological phenocopies of the original patient tumors. The groups participating to this project aim at developing known antibodies raised against tumor-associated antigens (INT and OPBG) or at identifying novel targets expressed on the surface of tumor cells (FPG, IEO and CSS). For the de novo identification of antigens different complementary approaches will be used: the first methodology is based on the de novo generation of monoclonal antibodies (mAbs) recognizing CRC and GB stem cells. Because tumor associated antigens (TAAs) are often lost by long-term established cell lines (XxxxXxxxx et al., 2018; Xxxxxxxxx et al., 2018), mAbs raised against antigens expressed by primary CSC cultures enriched in tumorigenic cells represent strong tools for the comprehension and targeting of solid tumors (Xxxx et al., 2017). The immunization with primary tumor cells can give rise to mAbs binding virtually all possible tumor-specific epitopes, including tumor-specific protein undergoing modifications, such as proteolytic processing, glycosylation or conformational changes that are not detectable by genomic or proteomic (i.e. LC-ESI-MS/MS) approaches. In fact, many known tumor-selective mAbs bind oligosaccharides (i.e. SSEA-3), posttranslational modifications (i.e. CD133, CA 19-9) or structural epitopes (i.e. CA-125) (Xxxxxxx et al., 1983; Xxxxxx et al., 2010; Xxxxxx et al., 1998). Two hybridoma libraries producing

antibodies against CSCs of different solid tumors are already available. These libraries were obtained by immunization techniques of immunocompetent mice with primary CSCs. Subsequent screening was established by the research group of the FPG (Xxxx et al., 2017) and allowed for the isolation of mAbs able to recognize target cells, but not normal cells derived from healthy human tissues. These mAbs will be used to generate the antigen-binding portions of the CAR construct (i.e. scFv) and thus the targeting domain of CAR T cells. In addition, new tumor-specific antibodies will be developed to increase the repertoire of targetable surface structures. A second screening methodology that will be used by researchers of the IEO to select tumor-specific scFv is based on a library of phages expressing recombinant human scFv. A platform, that allows the in vivo selection of scFv specific for human cells transplanted into immunocompromised mice, has been developed. The library is initially subjected to a procedure, which ensures the elimination of all scFv that can react against healthy cells. The most specific antibodies for both CSCs and the bulk of proliferating tumor cells are, then, selected in an unbiased manner with respect to the target antigen. The most promising scFvs are identified by next generation sequencing and their tumor affinity and specificity are determined in silico and validated with multi-parameter flow-cytometry techniques. In parallel with the discovery activities, a screening platform combining flow-cytometry and immunohistochemistry on normal tissues, primary tumors and metastases will be optimized, to test the specificity of the new targets that will be further validated upon FDA-approved healthy tissue microarrays. Furthermore, the CSS group will perform single-cell RNA sequencing in Wnt-active LIC-enriched cell populations in T-ALL to identify markers, exclusively expressed in leukemia cell subsets, resistant to conventional therapeutic treatments. The aim of this part of the project is to explore the cellular heterogeneity of acute leukemia in selected cell subsets associated with LIC activity. Through approaches of single cell RNA-Sequencing (scRNA-Seq) and computational reductions, researchers at the CSS will identify novel targets, specifically expressed in LIC-enriched subsets of T-cell ALL. These markers can then be used to develop scFvs in collaboration with the research groups of IEO in order to generate CARs targeting this highly malignant cell population. In addition to the screening for new markers, in this project also previously identified antigens will be considered. At the INT, researchers identified as cancer biomarker the Xxxxx Y (LeY or CD174) antigen overexpressed on several human carcinomas (ovary, colon, breast, prostate, lung, melanoma), with relatively limited expression in normal tissues thus representing an attractive target for immunotherapy. Starting point is a mAb, MLuC1 (Xxxxx et al., 1992), directed against XxX. Unlike others anti-carbohydrates mAb, MLuC1 revealed a good strength of binding. A scFv with the VL-linker-VH structure was constructed and used to produce one of the first CAR-T construct, named T-body (Mezzanzanica et al., 1998). However, although the T-body showed the ability to recognize LeY positive tumor cell targets, its limited affinity and the absence of a co- stimulation reduced the efficacy of such a reagent. Taking advantage of the previous experience on the target biology, new technologies will be employed to improve the performance of the MLuC1 scFv and obtain an optimized CAR targeting of LeY-positive tumors.

As mentioned above, there is so far an unmet medical need to extend CAR T cell therapy to patients with AML. As compared to Bcp-ALL, this type of leukemia is by far the most prevalent one in the adult population. In this project, the OPBG group will develop research strands aiming at identifying strategies that can make CAR T-cell treatment extensible also to patients with AML, both as a preparation therapy approach for transplantation, and as a rescue in case of post-transplant relapse. More specifically, we will aim at identifying constructs able to guarantee an action of CAR T cells preferentially expressed on leukemia blasts with respect to normal hematopoietic progenitors. Furthermore, constructs capable of attacking molecules that are not currently investigated as potential targets for CAR Ts will be explored.

In summary, the integration of existing competences within the various groups participating in the research project, together with the strong expertise of the FPO in IHC analyses for target expression validation, will facilitate the identification of these new targets. The availability of primary cell lines deriving from different human tumor histotypes and their sharing among the various groups

that participate in this project will help ensure the development of these new constructs. The clinical potency of the novel CAR T will be validated in mouse models specifically developed for this purpose in WP3.

Preliminary data

The part of the project proposed by FPG is based on low-passage primary tumor cultures enriched in CSCs derived from metastatic CRC and GB. Using GB and CRC CSCs, two different hybridoma libraries were already generated (see FPG Figure 1). A combined FACS and IHC workflow (Fig. 1a) was established to identify additional tumor-specific surface markers and will be further optimized in this project. Interestingly, in preliminary experiments, all mAbs showing a high specificity for tumor tissue by IHC (Fig. 1b) and flow-cytometry analysis (Fig. 1c) preferentially bind to primary cultures of CRC when compared to long-term ATCC cell lines (Fig. 1d). Although the mAbs were raised and screened using patient-derived primary tumor cells, the expression of the antigens was also present in the metastatic CRCs and will be further developed as CAR target in this project (Fig. 1d).

CRC (Patient 2) Liver mets (Patient 4)

CRC (Patient 1) Liver mets (Patient 3)

Pancreas Lung Small Intestine

Adrenal

gland

Prostate Lymph node

Liver

Gallbladder Breast

Thyroid

gland

Kidney Spleen

Figure 1. Identification of antibodies binding to primary CRC-CSCs and tumor-tissue. (a) Workflow for the hybridomas isolated from the large-scale screening for mAbs binding to CRC-CSCs: After the initial screening (Fig. 1a), we established the IHC protocol on frozen sections from embedded inclusions of CRC-CSCs. Then, the functional mAbs were tested on frozen tissue slides from CRC patients and on normal colon mucosa. The mAbs were then tested for functionality in IHC performed with formalin fixed paraffin embedded (FFPE) tissues. (b) Representative IHC analysis of FFPE CRC tissue versus normal mucosa from the same patient: staining was performed with the mAbs isolated as described in (a). The normal mucosa was derived from the same surgery and examined by an experienced pathologist. (c) Representative FACS dot plots from the analysis of mAbs binding on freshly dissociated CRC tissue versus normal mucosa: The upper row shows the binding of the mAbs on normal mucosa, the lower row on cancer tissue from the same patient. Please note that we used the marker EpCAM to distinguish between epithelial (mucosa or CRC) and tissue derived from the microenvironment. (d) Analyses for the expression of the 13C4, 19D10 and 20G4: CRC cultures either obtained from ATCC or primary cultures from patient CRCs (primary tumor or liver metastases).

(e) Green: Representative TMA for 12 out of 18 healthy tissues analyzed. Red: Representative IHC analyses for primary CRC tissues and liver metastases from patients with metastatic CRC.

In collaboration with the FPO, the mAbs were used to stain tissue microarrays constructed with a large series of normal tissues, and either very weak or no immunoreactivity was detected on all 18 healthy tissues investigated. In contrast, both primary and metastatic CRC samples showed widespread staining in cancer cells (Fig. 1e). Thus, these antigens are not lost during the formation of metastases and may be exploited to target metastases in patients that all appeared to be positive for the antigen. These mAbs will provide the starting point for the in-depth characterization of the tissue binding (WP2 Task 2) and the generation of recombinant scFvs to target CAR T cells.

Maternal mAb

Determination of VH/Vk b)

XX0

Xx XX0

XX0

CD8

CD28

CH2

4.1BB

CH3

CD3ζ

Design of scFv (VH-Vk and Vk-VH)

(1)

(2)

(3)

Lentiviral CAR construct

(pLenti6)

Cloning of 2nd and 3rd gen CAR

Cloning of VH-Vk-Fc Cloning of Vk-VH-Fc

Test for expression and binding

(1) Modular cloning of VH/VK fragments in scFv

(2) Expression of Fc-tagged scFv (pcDNA3.1)

(3) Shuttling of functional scFvs in 2nd and 3rd generation CAR constructs

BSA

scFv-

d) e)

13C4

20G4

flow wash

eluate

Coomassie

IgM (positive control)

+ XX000 XX000

+ SA115

13C4

IgM actin

f) g)

13C4

20G4

Anti hu-Fc

scFv-Fc

scFv- Fc

13C4

scFv actin

Figure 2: Established workflow for the generation of CARs using a highly versatile modular cloning strategy (a) General workflow used for the generation of functional CARs. (b) Representation of the vectors and cloning constructs used and enabling an efficient modular generation of several CARs containing VH-Vk and Vk-VH scFvs and two independent intracellular signalling domains (shown is Gen 3) in a lentiviral expression system. (c) Purification of the recombinant scFv-Fc protein. Shown is a representative SDS-PAGE after Coomassie staining and the immunoblot using anti-human Fc for detection. (d) Representative FACS analysis showing the binding of 13C4 and 20G4 scFv-Fc on the surface of primary CR-CSCs. The maternal 13C4 antibody was used as a positive control. (e) Immunoblot for the 13C4 antigen performed in parallel with the maternal mAb and an scFv-Fc using lysates of the indicated cells. (f) The model system used to test the CAR functionality. (g) Analysis of IL-2 secretion in the indicated cells as validation of CAR functionality.

Figure 2 shows the general workflow that was and will be applied to generate CARs targeting CRC: After extensive subcloning (n=3) of the hybridomas (Fig. 2a), the variable regions of the heavy (VH) and light chains (Vk) are sequenced. With this information and by adding an efficiently secreted Igkappa chain leader sequence (H5), as well as a flexible modified 4x(G3S) linker arm, scFvs are generated as artificial genes. A modular cloning strategy that enables easy swapping of VK and VH as well as the seamless shuttling of the complete scFvs in lentiviral vectors for the expression of the complete CAR or scFv-Fc constructs was designed (Fig. 2b). The-Fc constructs are expressed in HEK293T cells and purified for testing the affinity and potential functionality for immune blotting (Fig 2c-e). Then, the functional scFvs are transferred to the N-terminus of second- and third-generation CAR backbones containing transmembrane and intracellular signalling domains. We set up our system with CD8-hinge/TMD-4.1BB-zeta and CD8-hinge/TMD-CD28- 4.1BB-zeta, but we will discuss with the groups of OPBG, FPO and INT to define a generally used construct. Importantly, the modifications are achieved without the insertion of additional amino acids at the joining sites. This reduces the immunogenic potential of the constructs that may hamper the in vivo efficacy. Finally, the functionality was tested in a cell model using the T-cell line Jurkat, which produce IL-2 upon the activation of the TCR or the CAR (Fig. 2f). In these experiments, IL-2 production was detected in cells expressing two of the 13C4-based CARs only when co- cultivated with their target cells (Fig. 2g), indicating a functional signalling of our novel CAR constructs. These results show that the Jurkat cell model can be used to select for functional CARs before

starting more demanding cytotoxicity models with NK-cell lines (e.g., NK-92) or primary T cells. Furthermore, the results indicate the importance of the VH-Vk orientation in the context of the CAR-backbone and the need to test these combinations before deciding for the final CAR.

The group of Dr. Pelicci at the IEO has already applied a phage display platform containing > 1x109 human scFvs for the creation of novel tumor-targeting antibodies versus relapsed T-ALL grown in immunocompetent mice. The phAb libraries were first injected into healthy animals to deplete healthy tissue binders. phAbs were then recovered from serum, amplified, re-depleted versus normal mouse PBMCs and then injected into T-ALL engrafted mice (Fig. 3). phAbs bound to leukemic spleens were eluted, amplified, re-depleted and again administered into T-ALL engrafted recipients. These in vivo rounds of nested depletion followed by enrichment allowed for selection of phFiAgubres1.with the highest possible specificity for tumors grown in vivo.

in vivo

library depletion on WT healthy mice

ex vivo library depletion on WT splenocytes

b. Secondary Ab Control

T1H6

scFv

0.2% 36.7%

Healthy 1st round in vivo

mice library enrichment

on mT-ALL leukemic mice

T-ALL

leukemic mice

ETH2-Gold phage library

T-ALL

leukemic mice

Phage recovery from

2nd round in vivo

library enrichment

on mT-ALL leukemic mice

ex vivo library depletion on WT splenocytes

mT-ALL enriched library

0.01%

8.6%

CD45.2 positivity [%]

1x10 4

N3H3

scFv

48.1%

T2D3

scFv

57.6%

1x10 3

1x10 2

1x10 1

1x10 0

14.7% 15.8%

0 000 000 000 800 1000

myc-scFv positivity [%]

Figure 3: Schematic representation of in vivo selection of tumor binding phage were then selected as lead candidates.

Subsequent screening by multi-parametric flow-cytometry allowed for the selection of a cohort of phAb clones, which preferentially bound T-ALL blasts (both mouse and human) versus normal PBMCs. These phAb clones were then administered intravenously to T-ALL bearing mice and the capacity of the phage to specifically localize to and mediate a cytotoxic effect against T-ALL splenomegalies was evaluated. Clones that repeatedly diminished spleen weight and splenic murine T-ALL blast counts were identified. Three scFv molecules were subsequently cloned into fully functional human IgG frameworks and conjugated to human TNFa (IEOmAB1). When these TNFa fused immunocytokine (IC) proteins were given to human T-ALL cells in vitro, they killed blasts in a dose-dependent manner, while a non-binding scFv-IC failed to do so (data not shown). Thus, T- ALL blast cytotoxicity depended upon the antibody-antigen interaction to stabilize TNFa upon the cell membrane. When IEOmAB1 was then given to immunocompromised bearing a relapsed, human pediatric T-ALL PDX, it was able to significantly delay tumor growth and increase time-to- sacrifice for animals manifesting either minimal residual disease (Fig. 4a &b) or full-blown leukemia (Fig. 4c &d).

PBS

NCAB1-

hTNFa

XXXxXX0-

xXXXx

xXXXx xXXXx

PBS

NCAB1-

hTNFa

IEOmAB1-

hTNFa

PBS NCAB1- IEOmAB1-

hTNFa hTNFa

PBS NCAB1- IEOmAB1-

a. b.

Figure 4: Candidate Immunocytokines are therapeutically effective versus human T-ALL NSG mice bearing #03008 hT-ALL PDX tumors were treated every 2 days (12 injections total), starting at (a), (b) 2 days post- transplant, (c), (d) 12 days post-transplant; with intravenous Control-hTNFa (n=7), IEOmAB1-hTNFa (n=6) or PBS (n=4), and analyzed by (a), (c) flow cytometry for the presence of human CD45+ cells in PBMCS (at 33/38 days post-transplant (b), (d) and Xxxxxx-Xxxxx for overall survival.

As a parallel effort, the IEO group established rigorous screening methodologies to identify the TAAs being targeted by our IC proteins. By combining genetic gain-of-function methods based upon genome wide CRISPR gene activation screens with biochemical Mass spectrometry- immunoprecipitation approaches, we have been able to identify novel TAAs previously undiscovered by conventional methods. Using established methods for the screening of phAb- binders to human AML pdx specimens in vivo, the group at IEO aims to apply their platform to derive novel platform to create novel antibody molecules for the targeting of CAR-T/CAR-NKT vectors to AML within this project.

The group of Dr. Giambra at the Fondazione Casa Sollievo della Sofferenza (CSS) discovered that human T-cell ALL exhibit minor Wnt-active subpopulations with enriched LIC activity. Three patient-derived xenograft (PDX) leukemias were transduced with the 7TGC Wnt-reporter lentiviral construct. The 7TGC-transduced PDX cells were then isolated by FACS sorting and transplanted into immunodeficient NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) recipient mice (Xxxxxx et al., 2007). Flow-cytometry analysis of leukemic cells, harvested from moribund NSG mice, showed that only a minor fraction (<4%) of bulk leukemic populations was characterized by an active Wnt signaling. To assess whether the Wnt-active populations of PDX leukemias were also enriched with LICs, GFP+ and GFP− fractions were FACS sorted and transplanted into new recipients at limiting dilution (Fig. 5). Interestingly, the LIC frequency in the Wnt-active cells was more than 1 in 7.791 and significantly higher than the one observed in the Wnt-inactive population (1 in 22.894) as calculated using Extreme Limiting Dilution Analysis (XXXX) software (xxx.xxxxxx.xxxx.xxx.xx/xxxxxxxx/xxxx). These results suggest that the Wnt/β-Catenin pathway plays a critical role in the LIC activity of human T- ALL. We have also observed that the Wnt- active cells are also more proliferative than Wnt-inactive cells from the same tumor (data not shown). These ostensibly disparate observations raise the possibility that we may not be

appreciating heterogeneity within the Wnt-active and that the chemoresistance of Wnt-active cells might be a secondary consequence of replicative quiescence typical for stem cell populations

Figure 5. Leukemia-initiating cells (LICs) are highly enriched in the Wnt-active, GFP+ fraction in human PDX leukemia (a) Schematic diagram of experimental approach. PDX leukemias were transduced with the “7TGC” Wnt reporter lentivector, which is composed of 7 Tcf/Lef-binding sites upstream of a minimal promoter and GFP, followed by a separate SV40-Cherry selection marker (Giambra et al., 2015). The transduced cells were FACS sorted for hCD45+, Cherry+ cells 2 days later, and transplanted into NSG recipient mice in order to generate 7TGC-transduced human leukemias. At following, GFP+ and GFP− subsets were FACS sorted (also Cherry+ CD45+) and transplanted at limiting dilution into new recipients to assess LIC activity. (b) Survival of recipient mice after transplantation with FACS-sorted Wnt active (GFP+) or Wnt inactive (GFP−) subsets from three 7TGC-transduced PDXs. Each of 4 recipient animals was injected with the cell doses as indicated. The difference between GFP+ and GFP- subsets is significant (p<0.0001, chi-square test).

In order to address this hypothesis and to characterize the tumor heterogeneity of Wnt-active subsets of human T-ALLs, we have analyzed two 7TGC-transduced PDX leukemias using a high dimensional flow-cytometry approach. Taken together, these preliminary observations support the idea that novel information can be derived from highly dimensional analysis by flow-cytometry with important implications for understanding the tumor heterogeneity of human T-ALLs and in particular, how these heterogeneous Wnt-active populations might differently respond to conventional cytotoxic chemotherapy. The presence of different Wnt- active and chemoresistant subclones in T-ALL raises the question whether these cell subsets have any biologic or predictive significance. Former studies have reported that the prognostic impact of specific activating mutations or cytogenetic aberrations is depended on the therapeutic protocols (xxx Xxxxxx et al., 2006), suggesting that immunophenotypically defined subpopulations of Wnt-active cells may have different responses to the conventional drug treatments. In this proposal, we want to extend our data in order 1) to characterize cell subpopulations with greater resolution by employing similar bioinformatics procedures and 2) to identify novel cell receptors, specifically expressed in LICs of acute T-cell leukemia and suitable to be targeted with the CAR-T cell therapy.

Researchers at the Istituto Nationale dei Tumori (INT) identified the Xxxxx Y (LeY or CD174) antigen as cancer biomarker. It is overexpressed on several human carcinomas (ovary, colon, breast, prostate, lung, melanoma). LeY is a blood group-related molecule made up of difucosylated carbohydrate residues coupled to various proteins and lipids. It is not expressed on erythrocytes and its expression is independent of Xxxxx-A/B secretor status and there is relatively limited expression in normal tissues, thus making LeY an attractive target for immunotherapy. In early nineties researchers at the INT generated a monoclonal antibody (MAb), MLuC1 (Xxxxx et al. Int. J Cancer 1992), directed against XxX using the classical hybridoma technology. Unlike others anti carbohydrates MAb, that are usually IgM or IgD, MLuC1 is an IgG immunoglobulin revealing a good strength of binding. MLuC1 MAb recognized LeY expressed on a wide panel of tumor cell lines including ovary, breast, prostate, gynecological and lung cancer and melanoma. A scFv with

the VL-linker-VH structure was constructed and used to produce one of the first CAR-T construct (Mezzanzanica et al Cancer Gene Therapy 1998). The chimeric MLuC1 scFv-FcR gamma gene transfer was able to confer to this T-body an anti-LeY specificity. However, though the T-body showed the ability to recognize LeY positive tumor cell targets, its limited affinity and the absence of a co-stimulation reduced the efficacy of such a reagent. Here the group at the INT will take advantage of previous experience on the target biology and propose to apply the new technologies to improve the performance of the MLuC1 scFv and obtain an optimized CAR targeting LeY positive tumors.

The group at the Istituto di Biostrutture e Bioimmagini-Consiglio Nazionale delle Ricerche (IBB- CNR) has a longstanding expertise in developing and manipulating antibodies and antibody surrogate molecules, like Fabs and scFvs. It will then contribute to the project by generating scFv and Fab fragments in X. Xxxx starting from parent antibodies, and performing their characterization in terms of structure, stability, affinity and selectivity. The group has developed a method to generate bi- or multispecific Fab/scFv molecules that exploit the presence of a transglutaminase (TGase) consensus sequence on the C-terminus of the molecules. In Fig. 6A-D a representation of the type of monospecific or bi-specific molecules that can be generated is reported.

Figure 6. Representation of various antibody formats labelled with fluorescent dyes. Monospecific Fabs (A) or scFvs

(B) or bispecific scFv (C) or Fabs (D) site specifically labelled. Different colours indicate molecules with different specificities.

Following this approach, we have already prepared both scFvs and Fabs against Her2, a specific marker and therapeutic target for breast cancer that has been indicated as a potential target and biomarker also for CRC and GBM. Data on these molecules are shown here to demonstrate the methodology for their generation. scFvs and Fabs against Her2 were generated starting from the sequences of the therapeutic antibody Herceptin®, grafting the mAb CDRs onto a known scFv scaffold or reproducing the entire Fab sequence of the antibody. In Fig. 7A-E, the scFv sequence and the analytical data obtained with the purified domain are reported. Domain folding and stability was assessed by Circular Dichroism (CD) (Fig. 7F) observing the expected all-ß-sheet structure spectrum. The analysis provided the kinetic Kon, Xxxx and the affinity constant (KD) which are in the same range of that reported for the full-length antibody (KD was 0.7 nM for the scFv compared to 0.3 nM for the antibody).

HHHHHHESSGLFKREVQLVESGGGLVQPGGSLRLSCAASGFNIKD TYIHWVRQAPGKGLEWVARIYPTNGYTRYADSVKGRFTISADTSK NTAYLQMNSLRAEDTAVYYCSRWGGDGFYAMDYWGQGTLVTVSS G GGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCRASQDVNT AVAWYQQKPGKAPKLLIYSASFLYSGVPSRFSGSRSGTDFTLTIS SLQPEDFATYYCQQHYTTPPTFGQGTKVEIKGSGALQPTQGAMPA

kDa

260

140

Gln

25 kDa

2 3

Figure 7. Cartoon representation (A) and sequence (B) of the Herceptin scFv bearing a C-terminal glutamine for TGase mediated bioconjugation. The polypeptide contains an N-terminal His-tag (in red) used for affinity purification, which is then removed by treatment with a protease. The internal linker (G4S)3 is underlined in black. The C-terminal linker containing the TGase reactive glutamine (in red) is in magenta. In (C), (D) and € the SDS- PAGE, RP-HPLC and CD spectrum of the protein are reported to show that the final product is obtained highly pure and folded. In (F) the SPR (Biacore) binding analysis of the recombinant scFv derived from Herceptin to recombinant Her2 ectodomain. Analyses were performed at different ligand densities (Her2) and at increasing scFv concentrations. The experimental KD was about 0.7 nM. The recombinant Fab of Herceptin was similarly obtained.

The Pathology team of Fondazione del Piemonte per l'Oncologia (FPO) Candiolo Cancer Institute will contribute to the validation of the tumor specificity of the mAbs and scFv molecules identified by the FPG and IEO and to the comprehensive tissue analyses for tumor distribution. Technological and professional skills are mandatory requirements to afford diagnosis in cancer immunotherapy. This means that partnerships and collaborations among anatomical and clinical pathologists, molecular biologists, oncologists, informaticians and bioinformaticians need to be implemented for granting analytical validation of all procedures. This step is an obligatory prerequisite to clinical validation of immunotherapy. The Pathology Lab has established pipelines for derivation of cell and organoid cultures and for generation of PDXs from surgical and bioptic tumour samples that may be used as a first step of validation of the mAbs and scFv molecules. The Institute has already generated FFPE cytoclots from several colon cancer cell lines. In addition, the Pathology Diagnostic Unit deals with tumoral tissues or several origins (head and neck tumours, gastrointestinal carcinomas, breast cancer, ovarian cancer, urogenital tumours, sarcomas). Normal tissue can also be derived from surgical samples.

The Ospedale Pediatrico Bambino Gesù (OPBG) has already designed, developed and optimized several CAR constructs with high specificity for both haematologic diseases and solid tumors, including AML, sarcomas, neuroblastoma, diffuse intrinsic pontine glioma (DIPG), CRC and pancreatic carcinoma. Moreover, we have proved that the choice of specific costimulatory signalling results into significant amelioration of several CAR- T cell characteristics, including: 1) T-cell exhaustion, 2) basal T-cell activation, 3) in vivo tumor control and 4) CAR-T-cell persistence. OPBG has generated clinical-grade retroviral vectors, as well as lentiviral vectors, to reach a high level of CAR transduction in human polyclonal primary T cells. In particular, scFv specific for the GD2 antigen expressed by neuroblastoma cells has been used to generate two third generation CAR characterized by CD28/4.1bb or CD28/OX40 co-stimulation (Xxxxxxxxxxx et al., 2018). As shown in Figure 8, in a representative flow-cytometry analysis, OPBG has developed specific tools to obtain clinical grade retroviral vectors characterized by high degree of transduction capability. OPBG will evaluate mAbs to test their suitability for generating novel CAR T construct able to target different antigens expressed on the cell surface of a variety of hematological and solid neoplasias.

Figure 8. Characterization of CAR.GD2 T cells. (A) Representative flow-cytometry analysis of NT (left panel), CARGD2.28.OX40zeta (middle panel) or CARGD2.28.4-1BBzeta (right panel) T cells derived from healthy donor (HD) 15 days after transduction and culture in a medium containing IL2. CAR expression was detected using a specific anti-idiotype antibody (1A7). (B) CD3+ CAR.GD2 T cells present in the CAR T-cell production 15 days after transduction in the presence of IL2. Data from 7 HDs are expressed as average and standard deviation (SD)

General research plan Specific tasks and milestones

Task 1: Identification of new targets for chimeric antigen receptors (FPG, IEO, CSS).

Task 2: Validation of the tumor specificity of the mAbs and scFv molecules identified and comprehensive tissue analyses for tumor distribution. (FPO, FPG, Candiolo, IEO, INT, CSS).

Task 3: Sequencing of the mAb variable regions (VH/Vk, VH/Vl) and generation of recombinant scFv and antibodies. (FPG, OPBG, IBB-CNR, IEO, INT).

Task 4: Identification of the antigens bound by the mAbs and scFv molecules. (FPG, IEO)

Task 5: Generation and early validation of chimeric antigen receptors (CARs) (FPG, OPBG, IEO, INT, CSS).

Task 1. Identification of new targets for CARs

Task 1.1 Generation and screening of hybridoma libraries producing mAbs targeting primary GB and metastatic CRC cells. We will immunize Balb/C mice to produce hybridomas generating mAbs recognizing surface antigens expressed on GB and metastatic CRC cells. We have improved the immunization protocol by using Titermax gold adjuvant. Using this novel immunization strategy, and the hybridoma libraries we produced before (Xxxx et al., 2017), we will select mAbs binding to tumor cells of at least 3 different patients, but not to peripheral blood lymphocytes (PBL). Based on our previous experience, we aim to obtain >200 hybridoma clones that will fulfil these criteria. Immediately after the hybridoma fusion, the cells will be plated in 10000 xxxxx/spleen to guaranty a maximum of 5 individual clones growing in each well. Several rounds of screening will be performed in a 96 well format (high throughput, >16,000 clones/screening).

Milestones Task 1.1 (FPG):

1.1.1 Production of hybridoma libraries with hybridomas producing mAbs recognizing surface molecules on GB and mCRC CSCs. Foreseen is the generation of at least 3 fusions/tumor type (a pool of approximately 30.000 individual hybridomas). (0-4 month)

1.1.2 Screening of 4000 hybridomas/tumor type, identification and selection of at least 200 hybridomas producing mAbs binding to the surface of GB or mCRC CSCs but not or very low to peripheral blood cells. (4-12 month)

Task 1.2 (IEO) Using a human scFv phage display library to identify accessible tumor antigens in PDX models. We will engraft nod/scid IL2Rgamma null (NSG) mice with human AML pdx blasts. We will use at minimum 6 pdx specimens that give rise to fulminating leukemia in the peripheral blood. At least 3 mice per pdx will be used and once engraftment exceeds more than 50,000 human CD45+ per microliter of peripheral blood, mice will be injected with scFv- expressing phAb libraries. Ten minutes post-injection, animals will be sacrificed, and spleens will be explanted and processed. Human CD45+ cells will be sorted by FACS and bound phage eluted and amplified by infecting permissive bacteria. In parallel, two of the above human AML pdx samples will be infected with the Dox-off H2B-GFP reporter construct to generate stable transgenic cells, which will be engrafted in NSG mice. Mice bearing full-blown H2B-GFP+ AML will be fed a DOX-containing diet for 10 days. Subsequently, animals will be injected with phAb libraries and phAb-bound spleens will be processed. Cells will be sorted for human CD45-positivity as well as GFP-positivity. Phage will be eluted from both the tumor stem cell compartment (GPF-Hi) as well as the dividing cells (GFP-NEG).

Milestones Task 1.2 (IEO):

1.2.1 Selection of in vivo phAb binders to six human AML pdxs. Selection of phAb binders to the quiescent, tumor-initiating (H2B-GFP+) cell compartment of two human AML PDXs. (1-9 month)

Task 1.3 (CSS) Single cell RNAseq with Wnt-active LIC-enriched populations of PDX samples of T-ALL.

We will perform a cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq) (Xxxxxxxxx et al., 2017), which allows to obtain the transcriptional profile with a single cell resolution, simultaneously with immunophenotyping. Antibody-bound oligos, as well as the transcripts, will be indeed indexed with cell barcode and captured during the library preparation for single cell RNA sequencing (scRNA-Seq) (10x Genomics, Fluidigm). We will analyze about 5000 cells from FACS-sorted GFP+ (Wnt active) and GFP- (Wnt inactive) cell subsets from three PDX leukemias, transduced with the 7TGC Wnt/β-Catenin reporter, and subsequently expanded in vivo into immunocompromised mice (NSG). The cells will be stained with a panel, containing at least 30 oligo-conjugated antibodies against extracellular markers, selected to detect T-cells (e.g. CD1a, CD3 and CD4), signaling pathways and other biological phenotypes. The highly dimensional CITE- seq data will be subjected to diffusion based manifold modeling and dimension reduction algorithms including t-SNE and UMAP (Xxxxx xxx Xxxx, 2017.; Xxxxxxxxx et al., 2015). Informative phenotypes will be confirmed by a multiparameter flow cytometry assay with a 15- color panel and tested in PDXs and human T- ALL samples, without any in vivo passage in immunodeficient mice. The overall intent will be to identify cell receptors, specifically associated with LIC activity in T-ALL .

Milestones Task 1.3 (CSS):

1.3.1 Generating scRNA-seq and flow datasets of Wnt active/inactive cell subsets from three 7TGC-transduced PDX leukemias (0-8 months) and

1.3.2 analyzing data (8-12 months);

1.3.3 Validation of surface markers associated to LIC-enriched subsets in T-ALL by conventional flow cytometry-based analyses (9-12 months).

Task 2 (FPO, FPG, CSS, INT) Validation of the tumor specificity of the mAbs and scFv molecules identified and comprehensive tissue analyses for tumor distribution.

Task 2.1 (FPO) Setting up the conditions for comprehensive tissue analyses with the antigens identified within this project. To establish the experimental conditions, mAbs and scFvs will be tested by the Pathology team of the FPO on custom-made tissue microarrays (TMAs) produced using the Galileo TMA CK4500 platform. We will test multiple normal tissue and corresponding multiple tumor tissues arrayed on TMAs. We will also include metastases to ensure a highly selective binding on tumor tissue. If the results obtained on TMA will be heterogeneous and/or not consistent throughout experiments, we will switch for selected tissue to the analysis on whole section to best appreciate variability of staining in the different structures of the normal tissues and to best appreciate any phenomenon of heterogeneity in tumoral tissues. For the purpose of optimization and validation, the staining will be performed on automated immunostainers available at FPO (Ventana Benchmark Ultra, DAKO Autostainer, Leica Bond).

Milestone Task 2.1 (FPO):

2.1.1 Construction of a comprehensive library of human normal and tumoral tissues and optimization/implementation of a high-throughput platform for immunophenotyping and RNA-ISH procedures on human normal and neoplastic tissues (0-6 months).

2.1.2 Digital archive of gene/protein expression in a wide variety of normal and tumoral tissues. (0-24 months).

Task 2.2 (FPO, FPG, CSS, INT) Validation of the tumor specificity of the mAbs and scFv molecules identified

To test the tumor-specificity of the hybridomas selected by the FPG in Task 1, the candidate hybridomas will be subcloned to ensure monoclonality. The isotype of the mAbs will then be determined using an ELISA based assay. The functionality of the mAbs will then be tested by immunohistochemistry (IHC) on frozen sections from either GB or CRC PDX, followed by another round of selection on the corresponding normal counterpart. From previous experience, at this step, we expect to get ~30 mAbs/fusion. This number will be reduced to ~15 after selection of those working on formalin-fixed paraffin-embedded (FFPE) samples. In our previous screening at FPG, we have already identified 5 mAbs showing very promising binding patterns; therefore, we expect to be able to immediately start with the testing on comprehensive TMAs of normal and tumoral tissues. Furthermore, selective tumor-binding mAbs evolving from the mAb screening (FPG) or the phage display library (IEO), as well as the tumor specificity of commercially available Abs against surface markers highlighted with the proposed scRNA-Seq screening (CSS), will also be tested on multiple normal tissues, such as kidney, lung, liver, muscle, skin, bone marrow, spleen etc. We estimate that these experiments result in ~8 tumor-specific mAbs (FPG) and that can proceed for further investigation and be used for the development of scFvs (Task 3).

Milestones Task 2.2

2.2.1 (FPG/FPO): Validation of 5-7 mAbs and/or recombinant scFv-Fc molecules showing a very high degree of tumor specificity in FFPE tissue microarrays (0-6 months with existing mAbs).

2.2.2 (FPG/FPO): Selection of 30 mAbs specifically binding frozen tissue sections of CRC or GB specimens, while showing no reactivity with normal tissue (10-16 month).

2.2.3 (FPG/FPO): Validation of novel 5-7 mAbs and/or recombinant scFv-Fc molecules showing a very high degree of tumor specificity in FFPE tissue microarrays (15-21 months with novel mAbs).

2.2.4 (FPO/INT): The tumor specificity of the anti LeY MLuC1 mAb has been already investigated (see INT individual project table 1). The optimized MLuC1 scFv will be tested for specific recognition and binding on a wide panel of positive and negative tumor cell lines (ovary, breast, lung, gastric, prostate cancer and melanoma), on short-term cultures of ovarian cancer cells and on the TMA-platform mentioned above (15-21 months).

2.2.5 (FPO/CSS): Validation of LIC-associated cell receptors and characterization of in vivo tumor specificity through assessment of RNA and protein levels in different tissues from healthy donors (8-12 months).

Task 2.3 (IEO) In vitro validation of tumor specificity of the scFv/scFv-IgG clones and in vivo validation of the tumor specificity and cytotoxicity of the scFv-phage clones identified. Clonal phage-scFv will be administered to NSG animals bearing full-blown human AML pdxs. In comparison to negative control phage-scFv, the capacity of phage clones to accumulate in leukemic organs (spleen, bone marrow) will be measured. Clones demonstrating an in vivo AML localization, will be injected into leukemic mice and the levels of human AML blast levels in the blood and spleen will be monitored by flow-cytometry to identify scFv with the highest killing-tumor activity. In vivo validated scFvs and scFv-hIgG molecules will be produced in mammalian cell culture and their capacity to bind human AML cells will be determined by flow-cytometry and immunohistochemistry. Off-tumor binding and predicted toxicity will be determined immunohistochemically using FDA-approved paraffin-embedded healthy human tissue arrays.

Milestones Task 2.3 (IEO):

2.3.1 Characterization of the in vivo tumor specificity of up to 50 scFv- phage clones, from which we expect to obtain 10-15 bona fide tumor-binding and/or tumor initiating cell-binding scFv candidates. Upon the basis of protein expression levels and in vitro screening selection of 3-5 validated scFv molecules to be used for further validation (11-17 months).

Task 3 Sequencing of the mAb variable regions, scFvs encoded by phages and generation of recombinant scFv and antibodies.

Task 3.1 (FPG/IBB) Sequencing of the mAb variable regions (VH/Vk, VH/Vl) and generation of recombinant scFv and antibodies. Highly specific tumor binding of the mAbs selected in Task 2, have a good potential for the generation of CAR T cells. To identify the molecular target and to perform further tissue distribution analyses for the candidate scFv, the variable chains of candidate mAbs will be determined by cDNA sequencing. After subcloning (n=2) using FACS assisted single cell laydown, the isotypes of the candidate mAbs will be confirmed using commercially available isotyping strips (IsoStrips, Sigma). Then, we will use a set of primers binding the constant regions of the mAb cDNA and a set of 10 different VH and 12 different Vk (or Vl) specific oligonucleotides. The PCR products will be cloned in qCRII-TOPO vectors (Invitrogen) and 4 clones will be sequenced using Xxxxxx sequencing. If all clones result in the same sequences, we map the sequences with the database to determine the V(D)J chains used and continue with the construction of the scFvs. With the sequence information, the recombinant scFv constructs will be designed in silico and artificial genes will be used to generate scFv-Fc constructs taking advantage

of a pcDNA3.1-based vector (Invitrogen) encoding for the human IgG1 Fc portion. We already constructed a vector backbone that allows for easy swapping of VH and Vk/l chains, if needed to increase the affinity of the scFv. The protein will then be expressed in HEK293T cells, purified using affinity chromatography and tested for binding on the surface of the tumor cells using flow- cytometry. The same scFvs will be alternatively generated as soluble proteins in E. Coli (on a multi- milligram scale) without the Fc fusion tag (IBB), with a bioconjugation site on the C-terminus for selective introduction of fluorescent dyes or other tracers and for generating bi-specific scFv2s that may target two distinct tumor antigens. Recombinant scFvs and scFv2s will be used in all biochemical assays after a careful analytical and structural characterization.

Specifically, the IBB will perform aggregation studies with the scFvs which are reportedly highly prone to aggregation. Such studies will be performed by DLS and by SE-HPLC at various protein concentrations, assessing the occurrence of single peaks at the expected elution volume (SE-HPLC) and of the expected and mono-dispersed molecular sizes at DLS. We will study the folding properties by CD and DLS paralleled by thermal unfolding studies to determine their stability. Thermal stability will be estimated by unfolding studies collecting CD spectra at increasing temperatures (up to 95° C) and determining the melting points. Folding will be confirmed by the presence of the canonical beta-structures in the CD spectra. The new scFvs against GBM and CRC endowed with the highest stability, affinity and specificity will be suggested for transfer on CAR- Ts. Exploiting the presence of the TGase site the corresponding labelled variants will be generated conjugating on the specific glutamine a linker bearing fluorescent dyes or other suitable tags.

The recombinant scFvs can be further used for the biochemical identification of the antigen (Task 4) and to validate the mandatory extended tissue distribution analyses of the bound antigen (see Task 2). The rationale to perform the tissue specificity experiments with the scFv-Fc is that the binding specificity of the scFv and not of the intact murine mAb will determine the specificity of the CAR. Potential differences in binding patterns of recombinant scFv and mAbs could result in safety problems in the later development.

Milestones Task 3.1 (FPG):

3.1.1 Identification of the antigen-binding regions of 5-7 mAbs selected by the screening of the hybridoma libraries and generation of at least 5 functional scFv constructs (0-3 months with existing mAbs)

3.1.2 Identification of the antigen-binding regions of 5-7 mAbs selected by the screening of the hybridoma libraries and generation of at least 5 functional scFv constructs (15-21 months with novel mAbs).

Milestones Task 3.1 (IBB-CNR):

3.1.3 Preparation of at least 4 recombinant monomeric scFv with and without fluorescent labelling obtained from four mAbs developed by the Gemelli Unit 1-10 month)

3.1.4 Preparation of at least 1 bi-specific scFv2 also conjugated with a fluorescent label using two distinct scFv previously prepared against either CRC or GBM, accompanied by reports attesting full analytical characterization (1-12 month).

3.1.5 Preparation of at least a second bi-specific scFv2 conjugated with a fluorescent label using two distinct scFvs previously prepared against either CRC or GBM, accompanied by reports attesting full analytical characterization (18 month)

Task 3.2 (IEO) Sequencing of the selected scFv, identification of candidates and generation of clonal recombinant scFv-phage, scFv and antibodies. phAbs libraries generated post-panning will be amplified by PCR and conjugated to barcoded adaptors to allow next generation sequencing. The Complementary Determining Regions (CDR3) from both the Heavy and Light chains can be determined for each scFv molecule upon the MiSeq high throughput sequencer and scFv molecular diversity will be determined via our proprietary bioinformatics platform. Initially, scFv lists will be

filtered to remove molecules known to interact with healthy tissues or to be over-represented within the library. Clustering of related scFv amino acid species will be performed to account for sequence variance and the frequency of scFv will be determined within each sequence run. The list of scFvs common amongst the 3 biological replicates for each pdx will then be compared to one another to determine the conservation of scFv binding across different patient samples. scFv molecules identified as strong candidates will then be recreated using Xxxxxx cloning to graft the identified CDR variable regions into an invariant scFv framework. These scFvs will be cloned simultaneously in-frame with the pIII protein to allow for their expression upon clonal bacteriophage, with the human IgG1-Fc domain to allow for the creation of a fully functional human antibody or with the 6XHis/FLAG epitope tags to allow easy detection of the free scFv protein.

Milestones Task 3.2 (IEO)

3.2.1 Bioinformatic identification of hundreds of candidate scFv sequences. (6-9 months)

3.2.2 Selection of those clones with the most consistent cross-patient cell binding, with sequence frequency within the parameters we have established. (6-9 months)

3.2.3 Xxxxxx cloning of approximately 50 scFv candidates in multiple expression formats. (8-12 months)

Task 3.3 (INT): Engineering of the MLuC1 scFv in the new conformation VH-linker-VL to improve the affinity and specificity of binding to the LeY antigen. Starting from the MLuC1 hybridoma, the VH and VL cDNAs will be amplified and cloned into the phagemid pIT2 vector using specific primers to allow the scFv oriented as VH-Linker-VL or VL-linker-VH. Affinity maturation for the anti-LeY scFvs will be performed by phage display and epitope imprinting selection and a humanized version of the optimized murine scFv will be obtained as shown before (Xxxxxx and Xxxxxxxx, 2002). Expression of the developed soluble scFvs will be obtained in prokaryotic hosts, upon selection of the best bacterial strain and optimization of the production and purification processes. ScFvs will be biochemically characterized for integrity (SDS-PAGE and Western Blot) and for absence of aggregation properties (by size exclusion chromatography).

We will test and compare the developed scFvs for affinity and binding kinetics to the antigen LeY using the Biacore technology which employs the surface plasmon resonance We will test LeY specific recognition and binding on wide panel of positive and negative tumor cell lines (ovary breast, lung, gastric, prostate cancer and melanoma), on short-term cultures of ovarian cancer cells and on tissue macro arrays made available by the partners.

Milestones Task 3.3 (INT)

3.3.1 High affinity scFv generation including ScFvs construction in VH- Linker-VL and VL- Linker-VH orientation (month 1-3),

3.3.2 Affinity maturation by Phage Display (month 3-11)

3.3.3 Humanization by Phage Display (month 6-18).

3.3.4 ScFvs characterization and comparison of functionality/binding/affinity to LeY antigen (month 3-11).

3.3.5 ScFvs expression and optimization of production and purification processes (month 2-6; 11- 13; 18-20)

Task 3.4 (CSS) Generation of scFvs against the surface markers identified by scRNA seq. In order to generate the recombinant scFvs against the surface marker highlighted by the planned scRNA-Seq screening, we will test scFv-expressing phAb libraries in collaboration with the researchers of IEO. Specifically, the recombinant antigen of extracellular portion of cell receptor will be synthesized in vitro after cloning in expression vectors (pGEX) and transformation of E. coli. At the following, the recombinant antigen will be employed to screen the library of phages

expressing recombinant scFv. Once few antigen-specific clones are identified, a site directed mutagenesis will be also performed by PCR in order to increase the affinity of selected scFvs to the antigen. The CDR3 from both the Heavy and Light chains will be then determined for each scFv molecule by high throughput DNA sequencing and the specificity defined by bioinformatic approaches. The identified scFv will then be cloned into lentiviral vectors in frame with the human IgG1-Fc domain and epitope tags for the development of fully functional human antibodies. The specificity and efficiency of antibody will be determined in vitro by ELISA and flow-cytometry assays using the synthetic recombinant antigen, as well as cell lines with different genetic background for the selected surface marker, recognized by the antibody.

Milestones Task 3.4 (CSS):

3.4.1 Generation of recombinant antigen and screening of scFv- expressing phAb libraries (12-16 months);

3.4.2 development and validation of the human antibody against a LIC-associated cell receptor (16- 24 months).

Task 3.5 (OPBG) Cloning of novel scFvs. The OPBG has already performed a NGS approach to sequence the VH and VL regions of a specific monoclonal antibody derived from a murine hybridoma. The NGS platform could be applied to 20ng of DNA and it is based on a generation of library analyzed by Illumina system. Moreover, an approach to sequence B-cell receptor in single cell systems has been recently developed. This could represent a great advantage to reduce the time- consuming processes of subcloning. The sequenced VH and VL regions of the Ab will be assembled in the scFv conformation considering different hinge sequences to be able to optimize the antigen binding. OPBG will also perform humanization of the sequence of the murine Ab, comparing the binding specificity of wt scFv and the humanized scFv. OPBG will exploit the selection of novel scFvs to target pediatric tumors characterized by poor prognosis, including brain tumors, sarcoma, as well as AML and some subtypes of T-ALL.

Milestones Task 3.5 (OPBG)

3.5.1 Sequencing of VH and VL region in NGS approaches of a specific hybridoma generated by the different Units in Task 2. The number of the sequenced Ab is related to the activity of Task 2 and the provision of the DNA from the different involved IRCSS to OPBG. (0-6 months from the DNA or cell provided by the activity after Task 2).

3.5.2 Selection of scFv to target brain tumors, sarcoma, T-ALL and AML (at least one scFv for each pathology) (0-12 months).

Task 4. Identification of the new antigens identified by the Screening of the mAb and phage display libraries

Task 4.1 (FPG) Identification of the antigens bound by the mAbs and scFv molecules. The antigens targeted by mAbs showing high tumor specificity will be isolated by a modified immunoprecipitation method established in the PIs laboratories and identified by LC-ESI- MS/MS (Xxxxxx et al., 2018). This task will be performed in close collaboration with the IEO. First cells expressing significant amounts of the antigen will be selected. Then, the immunoprecipitation will be performed in triplicate, LFQ values will be determined using MaxQuant software and volcano plots will be used to visualize proteins significantly enriched in the mAb-containing sample. Further experiments aiming at isolating the antigen(s) targeted by mAbs will be performed using the recombinant soluble scFvs generated in Task 3.1 immobilized on solid supports packed in affinity columns. Proteins bound by the immobilized scFvs will be identified by LC-MS/MS. These proteins will then be compared to protein databases (e.g. Uniprot) and screened for potential surface proteins. In a procedure established in the PIs laboratories, the cDNA of 1-4 selected candidates

will then be ectopically expressed in cells with low surface binding. A candidate will be considered as being clearly identified as binding partner, if the transgenic cells appear positive by flow cytometry.

Milestones Task 4.1 (FPG):

4.1.1 Identification of the antigens recognized by at least 3 tumor specific antibodies (0-12 months for existing mAbs)

4.1.2 Identification of the antigens recognized by at least 3 tumor specific antibodies (18-24 months of novel mAbs).

Task 4.2 (IBB-CNR) Identification of target antigens and epitopes of selected mAbs/ScFv on the surface of cancer cells by affinity chromatography and mass spectrometry.

In close collaboration with FPG, we propose here to use some of the recombinant molecules generated and characterized in Task 3 as baits for the affinity capture of cell surface proteins (or even soluble, in case they are detected in cell supernatants) expressed by the tumour cells. After preparation and full analytical characterization, the new molecules (scFvs and/or Fabs) will be chemically immobilized on suitable solid supports, like Sepharose resins preactivated with NHS- ester. Immobilization will be checked by HPLC analysis of the supernatants achieving a density sufficiently high (at least 2 mg/mL resin) to process large amounts of cell lysate. Capture experiments will be performed by packing the resins into small columns, loading the samples in buffers containing different additives and washing/eluting under different stringency conditions to elute only high affinity binders. Bound protein fractions eluted from the columns will be neutralized, concentrated, extensively reduced and alkylated prior to trypsin digestion and protein identification by mass spectrometry using an Orbitrap mass spectrometer coupled to a nanoHPLC system. Proteins will be identified by database searching of tandem MS xxxxxxx. This approach will be complementary to the IP approach proposed by other Units and will be run in parallel to increase the prospect of capturing the right antigens. In parallel to antigen identification, we will also attempt to identify the epitopes recognized by antibodies. For this purpose, trypsin digested samples used for protein identification will be loaded on the same affinity columns. Captured protein fragments will be eluted by acidification and identified by LC-MS/MS mass spectrometry. Peptide sequences will be matched with those of the proteins as a double check of protein identification. The identity of the potential antigens and associated epitopes identified in this Task will be readily transferred to the Gemelli Unit for further validation and assessment by FACS, IHC, WB and other assays.

Milestones Task 4.2 (IBB-CNR)

4.2.1 Affinity capture experiments with at least two scFvs (or Fabs) and identification of at least 2 distinct antigen candidates, one for CRC and the other for GBM (month 0-10).

4.2.2 Affinity experiments with at least two scFvs or Fabs and identification/confirmation of at least 1 distinct epitope from each of the two antigens (one for GBM and the other for CRC) (month 0- 12).

4.2.3 Additional affinity experiments with at least one new scFv or Fabs and identification and confirmation of at least 1 new antigen and related epitope from either GBM or CRC (month 0-18).

Task 4.3 (IEO) Identification of the antigens bound by the scFvs. In parallel to the biochemical approach described in Task 4.1, genetic gain of function screenings will be performed in human cells negative for candidate scFv binding (Ex: 293T cells). Cells will be infected with a defective CAS9 (dCAS9) enzyme coupled to a potent transcriptional coactivator. In the presence of SAM guide RNA library infection (Xxxxx J et al Nat Protoc. 2017 12(4):828), cells that are able to express the targeted antigen will be fished out by FACS and the identity of the bound gene products will be determined by next generation sequencing. The union of the biochemical and genetic datasets will reveal a limited list of candidate antigens. Confirmation of scFv-antigen interaction

will be performed using a combination of biochemical assays (competitive flow cytometry, ELISA, TRIM-AWAY targeted proteolysis) while scFv-antigen specificity will be demonstrated genetically (siRNA knockdown or CRISPR gene knockout).

Milestone Task 4.3 IEO

4.3.1 Definition of antigen identity and antigen specificity for 3-5 candidate human AML-targeting scFvs (15-24 months).

Task 5. Generation of chimeric antigen receptor (CAR) T-cells targeting the antigen- expressing tumor cells.

Task 5.1 (FPG). Cloning on novel CAR. The scFvs with very high tumor specificity will then be placed on a scaffold carrying the hinge, transmembrane domain and intracellular signaling portions of functional 2nd and 3rd generation CAR constructs that are available in highly functional lentiviral vectors based on the proven pLenti6 system from Invitrogen (pLenti6T-IRES-GFP and pLenti6T-PGK- GFP). The functionality of the CAR T cell receptor will first be tested using Jurkat cells and measuring the production of IL2 upon co-culture with target cells or with CAR transgenic Jurkat cells carrying an IL2-promoter driven luciferase reporter. We will be able to initiate this work package right at the beginning of the project using 4 mAbs that have shown to display tumor specificity in our previous experiments. In the second year, we will perform the same experiments with the CARs based on newly identified scFvs.

Milestones Task 5.1 (FPG):

5.1.1 During the project we foresee the construction and preclinical testing of at least 3 functional CARs recognizing novel tumor antigens. We aim to construct 1-2 functional CARs with the existing mAbs (0-9 months)

5.1.2 1-2 additional CARs later in the project using novel mAbs (20-24 months)

Task 5.2 (IEO) Generation and early validation of chimeric antigen receptors (CARs) transformed cells. Validated scFv molecules will then be engineered to function as CARs for T, NK and NK-T cell adoptive cell therapy. Candidates will be Gibson cloned into a lentiviral vector allowing the fusion to a transmembrane domain linked to an intracellular activation domain (CD28/4-1BB/CD3). Constructs will be transfected into human acute lymphoblastic leukemia cell lines (CCRF-CEM, Jurkat), human NK cells (NK-92) and human NK-T cells (Hut78). Cytotoxic targeting of human AML pdx cells will be determined in vitro using co-cultures containing defined ratios of CAR-expressing killers to AML target cells. CAR activation will be determined by Luminex 200 multiparametric immunoassays to measure T/NK/NKT cell activation markers and AML cell viability will be determined by live/dead assay cell flow-cytometry measurements.

Milestones Task 5.2 (IEO):

5.2.1 Definition of the in vitro cytotoxic efficacy of at least 3 engineered CAR molecules in T, NK, and NKT cell lines (19-24 months).

5.2.2 Identification of one or more appropriate candidates for further translation into patient-derived primary T/NK/NKT cells (19-24 months).

Task 5.3 (INT) Generation of chimeric antigen receptor (CAR) T-cells targeting the LeY- expressing tumor cells. The most promising scFvs will be cloned in collaboration with FPG partner into a scaffold carrying the hinge, transmembrane domain and intracellular signaling portions of functional 2nd and 3rd generation CAR constructs that are available in highly functional

lentiviral vectors based on the proven pLenti6 system from Invitrogen (pLenti6T-IRES-GFP and pLenti6T-PGK- GFP). Jurkat cells will be infected and selected for anti-LeY CAR expression. We will characterize in vitro Jurkat cells expressing LeY CAR for: i) CAR expression by FACS analysis, western blot; ii) LeY antigen recognition and functional activation, by measuring cytokine production and IL-2 release (ELISA, FACS) and cell proliferation (CFSE assay) upon co-culture with LeY positive and negative tumor cell lines;

Milestones Task 5.3 (INT)

5.3.1 Construction and optimization of CAR gene cassette containing MLuc1 scFv (month 6-20)

5.3.2 Cloning the gene cassette into a vector for transduction of Jurkat cells, evaluation of transduction efficiencies (month 8-21)

5.3.3 In vitro assay of anti-LeY CAR-Jurkat for ability to recognize the antigen and evaluation of their functional activation (month 10-22)

Task 5.4 (OPBG). Generation of chimeric antigen receptor (CAR) T-cells targeting T-ALL, AML, sarcoma or brain tumor cells. The OPBG has already developed several retroviral and lentiviral clinical grade constructs in which classical or non-canonical costimulatory domains for CARs have been defined and could be used to generate novel CAR constructs based on the generation of new scFvs obtained in Task 3. Indeed, the designed and selected scFvs will be included in the DNA cassette of the vectors. Stable producer cell lines will be generated to obtain high title retroviral supernatant that will be used to genetically modify immunocells (T, NK or γδT cells) for functional evaluation. In particular, we will first generate primary donor-derived T cells genetically modified with retroviral vector carrying the novel CARs. Functional in vitro assays will be performed to assses basic parameters, including proliferation, activation and in vitro antitumor activities of CAR T cells. To this purpose, clinical grade Standard Operative Procedures (SOP) will be applied to generate CAR T cells, with the great advantage to easily translate the obtained data of this task to the clinical level. Providing the novel and selected scFvs to OPBG, we may support the different groups participating into WP2 in the design and cloning of the clinical-grade reagents for further clinical exploitation.

Milestones Task 5.4 (OPBG)

5.4.1 Generation of retroviral and lentiviral constructs for novel CAR sequences (6-12 months).

5.4.2 Generation of stable producer cell lines for the collection of high title retroviral supernatant (12-18 months).

5.4.3 Evaluation of transduction efficiency of the novel CAR construct on primary immune cells,

including T, NK and γδT cells (12-24 months)

Task 5.5 (CSS). The researchers of CSS will generate novel CAR-expressing T cells against a cell marker identified through the proposed scRNA-Seq screening and specifically expressed in Wnt- active LIC-enriched populations in T-ALL. Lentiviral vectors encoding CARs against the selected cell marker will be used to transduce human peripheral blood cells, which will be enriched and expanded in co-culture with anti-CD3/CD28 monoclonal antibodies (mAbs) and interleukin (IL)-2 as previously reported (Xxxxxxx-Xxxxxxxx et al., 2019). To validate the specific lytic function of de novo generated CAR T cells, we will perform standard cell killing assays and determine the T-cell cytokine productions. In addition, the efficiency of CAR receptors will be evaluated using Jurkat human cell line after lentiviral transduction of an IL2-promoter-driven luciferase reporter. The impact of antigen-stimulation on proliferation of redirected T cells will be also assessed with cell counting assays and 15-color panel flow-cytometry analysis to determine the different stages and/or blocks in the differentiation of modified T-cells.

Milestones Task 5.5 (CSS):

5.5.1 Generate novel CAR-vectors against one cell marker specifically expressed in Wnt-active LIC-enriched populations in T-ALL (18-22 months),

5.5.2 evaluating the in vitro efficiency, toxicity and fratricide effect of CAR-expressed T cells including human cell lines and PDX samples of T-ALL (20-24 months).

Feasibility, risk assessment and contingency plans

PART 1: Screening for new tumor markers

Besides the profound experience in the isolation of CSC from primary tumors and the generation of stable CSCs lines (Xxxxx et al., 2008; Xxxxx-Xxxxxxx et al., 2007; Xxxxxx et al., 2010), the research group at FPG successfully isolated therapeutically interesting mAbs generated against GB CSCs. To obtain these results, a screening approach very similar to the one proposed here was used (Xxxx et al., 2017). The group has access to laboratories equipped with facilities dedicated to automated cell culture as described in the separate proposal of the FPG. Similarly, the IEO campus has numerous facilities, which will be instrumental in the described experiments. The Flow Cytometry facility has 2 12-colour BD FACS Celesta machines and one 8-colour Miltenyi MACSQuant Analyzer flow cytometers which will be used for scFv screening, and a custom 16-colour BD FACS AriaFusion, a 12-colour BD FACS Jazz and a 9-colour BD FACS Melody cell sorters which will be employed in sorting of phage antibodies bound tumor cells. The IEO has a large tissue culture facility with dedicated areas for primary cell culture and high-capacity mammalian protein expression systems.

PART 2: Antigen specificity support with immunohistochemistry analysis of the mAb expression in tumors and normal tissue will be provided within the collaboration in WP2 by the high-profile pathology and TMA experts at the FPO. In addition, the IEO has an on-campus squad of dedicated pathologists and immunhistochemical technicians expert in the screening of antibody reactivity with tissue arrays and tumor specimens.

PART 3: Generation of recombinant antibodies and scFvs The IEO Antibody platform has more than a decade experience in antibody biochemistry and protein engineering (Massa PE et al. Blood. 2013 122(5):705.; Massa PE et al 2019 manuscript in submission) which have led to the creation of novel immunotherapeutic antibodies currently undergoing in-depth pre-clinical validation. Furthermore, the network will take advantage of the expertise of the IBB-CNR for the generation of recombinant mono- and bi-specific scFvs and Fabs (Xxxxxxxxx et al., 2017; Xxxx et al., 2015; Xxxx et al., 2019; Xxxxxxxxxxx et al., 2016a; Xxxxxxxxxxx et al., 2016b) and the OPBG in the context of targeting CAR T cells.

PART 4: Antigen identification The group at FPG has profound expertise in the biochemical identification of the antigens bound by the newly generated antibodies (Xxxx et al., 2009; Xxxx et al., 2017; Xxxxxx et al., 2019; Xxxxxx et al., 2018), the generation of recombinant antibodies and the design of new CAR constructs. In addition, the FPGs laboratory can count on collaborations that will assist in the optimal accomplishment of the proposal's goals. GB and metastatic CRC samples will be provided by the Gemelli hospital. At the IEO there is a dedicated facility for the rapid purification of antibody molecules in dedicated affinity columns using 2 AKTAExpress (GE HealthSciences) and a Mass Spectrometry facility with which Dr. Pelicci and Dr. De Xxxxx have a long-standing collaboration in order identify TAAs.

PART 5: Generation of CARs The quality of the CAR constructs that will be generated within this project will be ensured by the collaboration with the teams of CAR T cells experts at OPBG to ensure a productive exchange of translational and clinical issues concerning CAR T-cell therapies (Xxxxxxxxxxx et al., 2016; Xxxxxxxxxxx et al., 2018). Finally, we are convinced that the final part of the existing working pipeline for the identification of potentially therapeutic CARs will then seamlessly integrate with WP3 to obtain valid preclinical in vivo data.

RISK ASSESSMENT AND CONTINGENCY PLANS

Tasks	Risk	Proposed risk-mitigation measures
Task 1.1	Generation of tumor specific mAbs: High numbers of mAbs recognizing common human antigens / Identification of few tumor- specific mAbs / Bias towards generation of IgMs.	Performing a stringent screening and eliminating PBL binding hydridomas. / If needed increasing the number of xxxxx to be screened. / Optimization of the immunization protocol using TiterMax gold adjuvant.
Task 1.2	The unbiased phage- screening approach may result in the isolation of “sticky” phages. / Selection of unproductive phages.	The IEO has developed methods to efficiently deplete phages binding to normal tissues. / Protocols to minimize the necessity for bacterial amplification, permitting us to directly amplify tumor-bound phAbs, thus minimizing growth promoting phAb outgrowth or genetic drift in liquid cultures
Task 1.3	A technical limitation of current scRNA-seq technology comes as a result of data ‘dropout’, which is primarily caused by the low input amount of RNAs from a single cell. / Low numbers of LICs in patient samples.	To address this issue, we will apply the “MAGIC” algorithm, developed by Dr. Pe’er’s group to impute the dropout expression value. / Using high numbers of cells, from PDX samples we typically obtain ~200 million leukemia cells from one moribund mouse.
Task 2.1	Low availability of certain tissue types, either normal or from tumors.	Creating a network with reference centers for specific pathologies not presently treated at FPO- IRCCS. As backup the use commercially available TMAs of normal and tumoral tissues.
Task 2.2.1	Heterogeneity of CRCs and GBs: it is likely that mAbs binding only to the tumors of a subpopulation of patients will be identified.	We will develop our potentially therapeutic scFvs using thorough IHC analysis to test for tumor specificity and thus set up functional staining protocols. These protocols can then be used to determine the immunoreactivity as biomarkers.
Task 2.2.2	Non-specific mucosa.	staining	of	frozen	colon	mAbs showing an inconclusive staining pattern will be tested by flow cytometry with freshly dissociated colon mucosa.
Task 2.2.3-5	Problems with the immune-histologic analyses of tumor- and healthy tissue.	The groups will be supported by the IHC staff from the FPO, which provide their expertise to ensure a successful outcome of the experiments.
Task 2.3	It is likely that the majority of the identified scFvs will only bind to a subset of pdx specimens.	We will focus our efforts on the scFv clones that bind to the greatest number of pdx specimens possible.
Task 3.1	Grafting the CDRs of the new antibodies into the scFv framework may lead to the generation of molecules with insufficient affinity for the target nTAA or with poor stability. / Expression levels of the recombinant scFv and Fabs against the nTAA in the periplasmic space may be too low.	New scFv frameworks will be used for grafting and the residues adjacent the CDRs will be optimized. / Optimization of the expression using different vectors, different E. Coli strains and different expression conditions. Another solution could be the expression of the scFvs using eukaryotic systems.
Task 3.2	The unbiased phage- screening approach may yield thousands of potential scFv molecules. / A drawback of the use of phAb libraries is the generally poor affinity of the Antibody- antigen interaction	The IEO has developed statistical parameters to bioinformatically identify the most likely tumor- specific molecules. Including a large data set of scFv binding to healthy tissues, allowing us to automatically filter out clones with likely off- tumor reactivity. / scFv in vitro affinity maturation can be performed,
Task 3.3	The affinity of the VL-VH conformation of the MLuC1 is suboptimal	Performing chain shuttling and in vitro affinity maturation.
Task 3.4	Isolation of too many phages binding to the novel antigens	Optimization methods	of	the	depletion	and	selection
Task 3.5	Problems with the sequencing of the heavy and light chains of novel mAbs.	The OPBG established a NGS platform for scRNA seq of the BCR. Using this platform, the potential problems should be minimized.

Task 4.1	The identification of the novel antigens may be difficult due to: - Low antigen-mAb affinity - Limited amounts of biological material (primary cell lines).	We successfully reduced technical difficulties of the biochemical antigen identification by employing an improved “two-step affinity purification” method coupled to LC-MS/MS. The optimized CSC cultures enable the culture of up to 100 million cells that should be sufficient for the antigen isolation.
Task 4.2	MS identification of the antigens could be complicated by the occurrence of post- translationally modified protein fragments which may prevent fully elucidation of their structure.	The combined use of different tandem mass approaches (such as neutral loss for identifying phosphorylation) should give a contribution, though the elucidation of very complex structures may result inaccessible on the nanograms/micrograms scale.
Task 4.3	The identification of scFv-bound antigen from unbiased screens may bear caveats.	To facilitate the antigen identification, the IEO has pioneered a platform utilizing both genetic gain-of-function and high-throughput biochemical screens to identify the bound TAA.
Task 5.1	The proper expression and functionality of a CAR is dependent on many factors (scFv, linker, TMD and signaling domains) that are hard to predict.	The FPG has developed a highly functional modular cloning system to swap the domains. In addition to the Jurkat based testing system this will allow to generate and to test several CARs for each mAb before moving on to primary T cells.
Task 5.2	Due to the (potentially) limited affinities of the scFvs the function of the CAR has to tested thouroughly	The FPG will share the system for the generation of the CARs with the IEO. In addition, there will be project meetings with the experts in the INT and OPBG to discuss and solve potential problems.
Task 5.3	Low efficiency of the CAR	Construction of additional CAR constructs in collaboration with FPG and OPBG may show stronger anti-tumor efficacy.
Task 5.4	Unstable expression of the generated CAR could be faced an issue	OPBG has already demonstrated that the modification of the transmembrane domain and the hinge region may help stabilize the CAR expression.
Task 5.5	T ALL antigen may also be expressed on the surface of normal T cells leading to fratricide	CAR transduction of PB-derived mature NK cells or CD3+CD56+ Cytokine-induced killer (CIK) cells or genetic manipulation of the T cells could circumvent this problem.

WP3: Preclinical development of CAR-killer cells against novel cancer targets in solid tumours and hematologic diseases of non-B lineage.

Leader: FPO (Candiolo); Participants: FPG (Gemelli), INT, OPBG, FT/SGER (Fondazione Tettamanti/San Xxxxxxx), IRST (Meldola), IOV, IEO

Rational

Adoptive immunotherapy strategies based on CAR-T cells obtained extraordinary results in patients affected by relapsed/refractory hematological malignancies of the B-cell lineage. Unfortunately, similar results have not been achieved for solid tumors. Several limitations in this setting obstacle the efficacy, including: a composite and dynamic immunosuppressive microenvironment; the complex organization of the tumor, rich of structural barriers to T cell homing; the lack or heterogeneous expression of suitable target antigens; the challenging metabolic conditions of the tumor environment. Furthermore, a relevant issue is related to the CAR-specificity with risk of important “on-target” or “off-target” toxicities. For these reasons, effective and safe translation of CAR-based approaches to solid tumors and non-B cell lineage hematological malignancies represents a major unmet need and a timely challenge for the current oncology research. Innovative and informative preclinical models, representative of the realistic clinical challenges, are fundamental to explore new CAR-T strategies and to provide rational bases for innovative clinical trial design. With this aim and perspective, eight research groups will contribute ideas, expertise and preclinical models to define and factually explore a coordinated experimental platform, defining interconnected strategies that may maximize collaborations and synergies.

The main goal of this WP is therefore aimed at the optimization and exploitation of an experimental pipeline for preclinical testing and validation of new CAR strategies against solid tumors, with particular attention the issues with more relevance in the clinical perspective, like chemoresistant, relapsed and metastatic diseases. The project will also tackle acute myeloid leukemia, Hodgkin lymphoma and anaplastic large cell lymphoma, all characterized by a challenging immunosuppressive microenvironment.

The preclinical platform of WP3 includes four key components, corresponding to specific phases of the preclinical pipeline:

(1) Patient-derived cancer models.The antitumor activity of CAR-based therapies is not dependent on a given histotype but rather on the expression of the target antigen, being therefore able to work on tumours of different histotypes. Therefore, WP3 Partners will aggregate a wide collection of “pan-cancer” cell models, including hundreds of tumor cell lines, tridimensional 3D models with spheroids including enrichment in Cancer Stem Cells (CSCs) and organoids mimicking in vitro the complex tumoral architecture. Large resources are also already available relatively to patient-derived cancer in vivo models, with hundreds of subcutaneous tumor xenografts (PDXs), tumor cell line xenografts, CSC-derived orthotopic xenografts and human immune system-reconstituted (humanized) mice. Generation of a centralized database of all available models, including their corresponding gene expression profiles when available, will be a central goal of the project.

(2) Generation and characterization of CAR-effector cells. To develop CAR-effectors against solid tumors, the “conventional” CAR-T approach will be integrated by innovative CAR-engineering of different populations of killer lymphocytes endowed with intrinsic, HLA-independent antitumor activity, such as Cytokine induced killers (CIK), γδ T or natural killer (NK) lymphocytes. Such effectors may be characterized by a higher antitumor activity in challenging settings and with heterogenous expression of the target molecule, a common limitation in solid tumors. Moreover, the project includes the development of “universal CAR effectors”, such as

the FDA- and EMEA-approved NK-92 cell line, endowed with unlimited expandability potential, possible generation of clones and suitable for multiple engineering strategies, under controlled and reproducible conditions.

(3) In vitro preclinical assays. The wide and composite expertise of the participating Centers provides a valuable platform to explore the proposed CAR-based strategies, focusing on crucial and challenging issues in the perspective of clinical translation, such as recognition of tumor specific targets, migration towards cancer cells, CAR-driven specific killing and synergism with conventional therapies. 2D assays will be more suitable for large scale and multiple experimentations, to prioritize the best strategies and combinatorial approaches. Various 3D assays will be developed and/or exploited for testing more complex and dynamic aspects, like migration, infiltration and antitumour efficacy on multiple different tumors, including the more chemoresistant CSCs.

(4) In vivo preclinical tests. The final step of the preclinical validation will be represented by the evaluation of efficacy of CAR-effector cells in vivo, using xenografted mice. Research efforts will be dedicated to explore and therapeutically modulate in vivo: (i) CAR-lymphocytes homing/recruitment at tumor sites; (ii) tumour infiltration and effective killing of cancer cells;

(iii) synergism with conventional therapies previously validated in vivo; (iv) counteraction of immunosuppressive elements in the tumor microenvironment; (v) proliferation and persistence;

(vi) potential toxicities.

Sequential validation through these four phases is expected to provide robust preclinical evidence of efficacy for at least some of the proposed CAR strategies, solid enough to warrant subsequent testing in human patients.

To facilitate integration and sharing of CARs constructs, models, strategies and assays, three WP3- specific meetings are envisioned: a start meeting (M1-2), to discuss in detail the tasks and further possible synergies, after the general project kick-off meeting; a first interim meeting (M9-10), to coordinate the conclusion of the activities and reports for the Year 1; a second interim meeting (M17-18) for reciprocal updating and coordinating the research activities of the last six months of the project, towards the final WP3 report and general project meeting. Moreover, teleconferences will be organized whenever necessary on specific Tasks or subtasks.

Specific Activities

TASK 1. Patient-derived cancer models

1.1 FPO: Survey of available patient-derived preclinical cancer models. As WP3 Coordinator, we will run a survey across project partners to build a database of patient-derived preclinical cancer models available in the Consortium for the various tumour types of interest. In particular, we will survey: (i) cell lines established in conventional 2D culture; (ii) in vitro 3D culture systems;

(iii) established patient-derived xenografts; (iv) ex vivo culture systems for short-term assays. This will result in a comprehensive database from which each partner will be able to obtain the models of potential interest for preclinical testing. Ideally, the models should be also exploitable for the evaluation of the expression of genes of interest (e.g. CAR targets, immunomodulatory genes). When possible, this will be achieved by collecting expression profiles of the models, either from public databases/datasets or from the owning partner, and by generating an aggregated dataset to be deposited in databases such as Gene Expression Omnibus and/or ArrayExpress. In the long term, display of model expression data and metadata will be made available through a Data Portal, either the one of the EurOPDX European Research Infrastructure coordinated by the WP3 leader Xxxx Xxxxxx (xxxxx://xxxxxxxxxx.xxxxxxx.xx) or a dedicated one, depending on the amount and quality of available data.

Milestones Task 1.1 FPO:

1.1.1 First release of a database with patient-derived preclinical cancer models available from Project Partners (M1-6);

1.1.2 Second release and Data Portal (M7-18).

1.2 FPG: In vivo tumor models with primary CRC and GBM cultures enriched in CSCs. In this task, the capacity of newly established CRC and GBM CSC lines to give rise to subcutaneous and orthotopic tumors in immunocompromised recipient mice (NSG mice) will be assayed. For the orthotopic models, cells will be transduced with luciferase for imaging. For CRC samples the main focus will be on the generation of liver metastases by performing a spleen injection protocol. For GBM we will concentrate on intracranial injection of luciferase-labeled cells followed by IVIS measurement every two weeks. We will then select cells that readily give rise to tumors that reproduce histologic aspects of the patients’ tumor. Aim of these experiments is to enlarge the cellular model systems that can be employed to test the CAR function in vivo.

Milestones Task 1.2 FPG:

1.2.1 Establishment of the first 3-4 orthotopic tumor models (M1-12)

1.2.2 Establishment of additional 2-3 orthotopic tumor models (M12-24).

1.3 INT: LeY antigen-expressing patient-derived models. (i) Breast cancer-initiating cell (BCIC) line MS186, derived from a triple negative breast cancer (TNBC). MS186 can grow as sphere in non-adherent culture conditions. When serially transplanted in vivo, MS186 shows increased aggressiveness and metastatic ability at each passage. We plan to fully characterize MS186 cell lines by cytofluorimetry, ELISA, IHC, analyzing also its molecular profile (GEP/mRNAseq, WES, proteomics) to identify determinants of stemness and possible new targets for immunotherapy. MS186 will then be used to test CAR-T cells activity in vitro and in vivo. (ii) Ovarian cancer models. Starting from patients’ tumor and ascitic samples, a panel of primary ovarian cancer cells and derived short-term cultures have been established at INT and will be enriched with further samples. Short-term cultures and malignant ascitic cells will be characterized by cytofluorimetry and used as targets to test CAR-T cells activity in vitro.

Milestones Task 1.3 INT:

1.3.1 Full characterization of MS186 TNBC cell line; in vivo growth. (M1-12);

1.3.2 In vitro stabilization and full characterization of short-term cultures from tumor and ascitic cells of ovarian cancer patients and biobanking of primary cancer cells in their ascites from the same patients, in collaboration with Gynecological Oncology Unit (M2-24).

1.4 OPBG: Patient-derived models of paediatric tumours. OPBG will generate 2D and 3D patient- derived cell cultures of rhabdomyosarcoma, osteosarcoma, brain tumours including high grade GBM, DIPG, medulloblastoma, neuroblastoma. Cultures will be characterized to evaluate clonality and genetic alteration. OPBG will also establish PDXs for glioblastoma, DIPG, sarcoma or leukaemia. OPBG will characterize its large collection of patient samples for the expression of the novel antigens discovered using omics approach, validating the data at both mRNA level by qPCR and IHC. Intra-tumour heterogeneity will be explored in T-ALL, AML, Sarcoma or Glioblastoma cases by single cell RNA-seq. Finally, OPBG will generate orthotopic mouse xenografts of osteosarcoma, medulloblastoma, glioblastoma, DIPG, AML, HL, NHL and T-ALL. Humanized mouse models will be explored for a more "physiologic" tumour microenvironment.

Milestones Task 1.4 OPBG:

1.4.1 2D and 3D patient-derived cell cultures for rhabdomyosarcoma (n=3), osteosarcoma (n=3), brain tumours including high-grade glioblastoma (n=5), DIPG (n=3), medulloblastoma (n=3), neuroblastoma (n=8) (M1-24);

1.4.2 Collection and/or generation of gene expression data from at least 10 patient-derived samples in the AML, T-ALL and solid tumour model (M1-24);

1.4.3 At least one xenograft orthotropic mouse models for each of: CRC, PDAC, osteosarcoma, medulloblastoma, glioblastoma, DIPG, AML, HL, NHL, T-ALL (M1-24).

1.5 FT/SGER: Generation and validation of preclinical models for Acute Myeloid Leukaemia (AML). PDX models for AML will be generated and validated, together with AML cell lines, in preclinical cancer models; next-generation sequencing (NGS) on AML PDXs will be set up, to better characterize such a heterogeneous disease. Chemotherapy in xenograft mouse models will be optimised for the functional characterization of CAR-CIK cells against chemo- resistant/refractory AML cells.

Milestones Task 1.5 FT/SGER:

1.5.1 Patient-derived cancer models in the AML setting (M1-24).

1.6 IRST: Generation and characterization of in vitro and in vivo preclinical models for metastatic melanoma and pleural mesothelioma. (i) Generation of in vitro and in vivo models. The research group is already working with cells derived from PDXs of metastatic melanoma and pleural mesothelioma tumours (the latter engineered to constitutively express Firefly luciferase). The plan is to expand the different clones in vitro (low passage) and in vivo by inoculating them into NSG mice. (ii) Molecular characterization by RNA-Seq of the 3 mesothelioma histotypes (sarcomatoid, epithelioid and mixed biphasic subtypes) of the available PDX-derived cells.

Milestones Task 1.6 IRST:

1.6.1 Generation of a master cell bank for melanoma and pleural mesothelioma PDX-derived cells (M1-6);

1.6.2 Identification of genes expression by the 3 mesothelioma histotypes with a special focus on alternatively spliced isoforms for plasma membrane proteins (M6-12).

1.7 IEO: Ex-vivo selected NK cells obtained from haploidentical donors during haplo- hematopoietic stem cell transplant (HSCT). Prior clinical experience by several groups has clearly shown that donor lymphocytes infusion (DLI) following HSCT can eradicate residual leukemic cells in AML patients, but with a substantial risk of graft-versus host disease mediated by donor T cells reacting against normal host antigens. In our recent clinical experience, the combined approach of HSCT plus administration of donor-derived, ex-vivo selected NK cells on day 7 after HSCT is feasible and safe, without toxicity or GVHD. We here aim at enhancing the anti-tumour specificity of NK cells obtained from haploidentical donors by their ex-vivo engineering, to optimize the combined use of haplo-identical HSCT along with donor-derived NK cell administration in patients with refractory Acute Myeloblastic Leukaemia (AML) and lymphomas. PB cells collected from the haplo-identical donor selected for HSCT of each patient will be double purified with Clini-Macs column, to obtain 4-9 million NK cells/kg (expected purity: 50-97%). Alternatively, purified CD56+ve NK cells could be obtained from cord blood samples or from adult normal donors. Anti-tumour specificity and activity of NK cells obtained from haploidentical donors will subsequently be enhanced by their ex-vivo engineering.

Milestones Task 1.7 IEO:

1.7.1 Availability of adequate amounts of selected NK cells from haplo-identical donors during the haplo-HSCT procedures. Aliquots of selected CD56+ve cells will be stored for subsequent NK cell studies, under informed consent from donor and patient (M1-24).

TASK 2. Generation and characterization of CAR-effector cells

2.1 FPO: Generation and Characterization of CAR-CIK and CAR-NK92 against bone and soft tissue sarcoma, melanoma and colorectal cancer. We plan to engineer killer cells with anti- CSPG4 (collaboration with X. Xxxxx UNC, USA) and anti-GD2 against BSTS and melanomas, anti-mesothelin (MSLN) against selected retroperitoneal sarcomas and CRCs overexpressing MSLN, and anti-pan-CRC antigens, such as EPCAM, MUC1 and ERBB2. (i) CAR-CIK. Circulating PBMC precursors will be collected from patients with BSTS or melanoma, or healthy volunteer donors. CIK will be expanded within 2-3 weeks according to a standardized protocol. CIK will be engineered with retroviral/lentiviral vectors encoding 2nd or 3rd generation CARs. (ii) CAR-NK-92. The NK-92 cell line will be transduced with the same CAR- encoding vectors described above. Multiple clones of CAR-NK92, with variable CAR expression levels, will be obtained and expanded. (iii) NK-92 engineered for combinatorial antigen targeting of CRC. We aim to develop a combinatorial antigen targeting for CRC by engineering NK-92 cells with a synthetic Notch Receptor-based dual CARs expression system. In this system, the CAR targeting “antigen A” (e.g. MSLN) has intracellularly the synthetic Notch (synNotch) receptor domain plus a transcription factor domain, e.g. of GAL4. Upon biding with Antigen A, the TF is released and reaches the nucleus. GAL4-driven reporters allow monitoring synNotch activation upon target engagement. The second CAR, against “Antigen B” (e.g. EPCAM), is placed under GAL4 promoter, so that the CAR is expressed only upon engagement of the first CAR/synNotch. In this way, killing activity will occur only in tissues where both antigens are expressed. Notably, expression of Antigen A by all cells in the tumour is not necessary with this system. We will transduce effector cells with synNotch receptors and GAL4-driven CARs, and validate the system by co-culturing different ratios of effector/target CRC cells. (iv) Characterization of CAR-killer cells. For both CAR-CIK and CAR-NK92 we will assess expansion rates, CAR expression (by beads-based quantitative flow-cytometry), and extended phenotype, including immunoregulatory checkpoints, cytokine production (Th1 and Th2 type), and expression levels of NR4A TFs (associated with exhaustion). In selected cases, in particular of NK-92 clones, global RNA-sequencing will be performed for a complete assessment of possible functional deviations from the parental cell line.

Milestones Task 2.1 FPO:

2.1.1 Generation and characterization of CAR-CIKs from 15 BSTS patients and 10 melanoma patients (M1-18);

2.1.2 Generation and characterization of CAR-NK92 clones anti-CSPG4, anti-GD2, anti-MSLN or anti-pan-CRC (M1-12);

2.1.3 Generation and characterization of at least one NK-92 clone with a synNotch-CAR system against two antigens in CRC cells (M6-18)

2.2 FPG: Generation and characterization of CAR-T and CAR-NK cells against CRC and GBM. The CAR construct targeting novel antigens that were generated and optimized in WP2, will be further developed in this work package. The development includes the shuttling of the cDNA in retroviral backbones that are used in the clinical setting (provided by the OPBG). Using these systems, we will optimize the transfection of the producer cell lines to reach consistently high

virus titers (measured by rtPCR). We will test transient and stable models for the expression of retroviral particles. Furthermore, we will develop scFv-specific (idiotype-specific) recombinant antibodies (scFv-Fcs) using a comprehensive scFv-Phage display library in collaboration with the IEO. These recombinant antibodies will then be used to validate the expression levels of the CAR on the surface of the target cells (i.e., primary human T cells and NK-92). Finally, we will optimize and standardize the conditions for the isolation, transduction and expansion of primary T cells obtained from healthy donors. For every donor transduction efficiency will be monitored by flowcytometry and the cell expansion will be controlled by manual cell counting. In case the development of CAR-specific scFvs is more complicated than expected, we will also optimize the parameters described above to produce lentiviral CAR constructs carrying a GFP cassette and use these for the in vitro testing of CAR function (see Task 3), alternatively we will fuse an easily detectable exogenous sequence (i.e. deltaCD34 to the extracellular domain of the CAR).

Milestones Task 2.2 FPG:

2.2.1 Generation of CAR-effector cells for 2-3 CARs e.g. based on the mAbs 13C4, 20G4 and 19D10 (M5-12);

2.2.2 CAR-effector cells for additional 2-3 constructs, with mAb clones to be determined in the first year of the project (M12-20).

2.3 INT: Generation and characterization of CAR-T cells against LeY antigen. We would preferentially use a vector system in a certified GMP facility to produce “clinical-grade” engineered T cells whose functional characterization will be useful for regulatory purposes and for further clinical development. We will characterize the obtained LeY CAR-T cells for CAR expression, by FACS analysis and western blot, and for T-cell phenotype by FACS analysis of CD4/CD8 CAR-T cell ratio.

Milestones Task 2.3 INT:

2.3.1 Optimization of anti-LeY CAR lentiviral vector and human T-cell transduction (M12-19)

2.3.2 Selection and characterization of anti-LeY CAR-T cells phenotype (M17-24).

2.4 OPBG: Generation of CAR-T, CAR-NK, CAR-γδT and dual CAR effectors. (i) T cells: we will optimize clinical-grade protocols to genetically modify primary T cells with the generated CAR to obtain CAR-T cell anti-tumor activity even in the contest of high immunosuppressive environment like that of HL and NHL, as well as germinal tumours. Moreover, OPBG will generate CAR-T cells with a novel specificity against CRC/pancreatic cancers, and sarcoma. The production process will be mainly optimized to decrease exhaustion profile of the generated CAR-T cells, preserving their antitumor capability. (ii) NK cells: OPBG will perform a scale-up study to generate an allogenic CAR NK platform from peripheral blood cells of healthy donors. The process will be validated for the generation of CAR NK cells in the context of ALL, T- ALL, AML and solid tumours. (iii) γδ T cell: OPBG will generate a primary CAR γδ T-cell platform from peripheral blood cells of healthy donors. In particular, CAR γδ T cell activities will be evaluated in the context of ALL, T-ALL, AML and solid tumors. (iv) Dual CAR and other combinatorial strategies: OPBG will generate bispecific CARs to target CD19 negative relapses occurring in patients affected by Bcp-ALL; the same approach could be of value to treat patients that already at diagnosis could show highly heterogonous Bcp-ALL leukaemia. Moreover, bispecific CARs will be generated to target tumour antigens in combination with checkpoint inhibitors (including PDL1) in the solid tumor model.

Milestones Task 2.4 OPBG:

2.4.1 Validation of clinical-grade protocol for CAR T translation in the context of CD30+ lymphoma (M1-6);

2.4.2 Generation of a least two large-scale rounds of validation for the establishment of allogenic CAR NK cell bank (M6-12);

2.4.3 Generation of a least two large-scale rounds of validation for the establishment of allogenic CAR γδ T cell bank (M6-12);

2.4.4 Cloning a bispecific CAR (CD19-CD22) (M1-6);

2.4.5 After antigen screening, cloning a bispecific CAR for neuroblastoma and sarcoma patients (M6-12).

2.5 FT/SGER: Generation and Characterization of CAR-CIK against AML. (i) CAR-redirected cytokine-induced killer (CAR-CIK) cells will be obtained by the optimization of non-viral genetic engineering of CIK cells. Non-viral anti-AML CAR-CIK cells (e.g. vs CD33; CD123) will be then generated and in vitro phenotypically profiled (CAR expression, memory markers and CIK-related markers and receptors, etc..). (ii) Dual CAR and other combinatorial strategies. To limit on-target off tumour toxicity (e.g. vs CD33 and CD123), anti-AML Dual CARs exploiting trans-signalling strategy will be generated. Moreover, bi-specific CARs targeting both AML cells and the bone marrow immuno-suppressive microenvironment will be developed with the final aim to boost CAR T-cell therapy against AML.

Milestones Task 2.5 FT/SGER:

2.5.1 Generation and characterization of CAR-CIK cells against AML (M 1-10).

2.6 IRST: Generation and characterization of novel CAR-T against melanoma, pleural mesothelioma and T-ALL. We plan to transduce T-cells with CARs targeting unexplored antigens, if derived from scFvs obtained from antigen unbiased screenings, or classical antigens for melanoma (i.e. HER2), pleural mesothelioma and T-ALL (i.e.TRCB1), cloned into a novel transduction vector that harbours human NIS gene for in vivo cell monitoring. Transduced T cells will be assayed for CAR expression, phenotypically (Flow Cytometry), at the transcriptional level by using a panel of specific markers for several T cell features (CAR-T characterization panel, Nanostring) and for their activation in the presence of target tumour cells.

Milestones Task 2.6 IRST:

2.6.1 Generation of at least one validated CAR-T population with a working construct for melanoma, pleural mesothelioma and T-ALL, respectively (M6-M24).

2.7 IOV: Generation and characterization CAR-effector cells against prostate, gastric and nasopharyngeal carcinoma, and effector redirecting. (i) Bidirectional lentiviruses will be used to produce concentrated VSV-pseudotyped stocks of lentiviral particles to transduce CARs against BARF-1 for gastric and nasopharyngeal carcinomas (GC, NPC), or against PSMA, PSCA or dual PSMA/PSCA for prostate cancer (PCa). Human PBMCs will be activated in vitro using IL-2 and phytohemagglutinin or anti-CD3/CD28-coated beads, and subsequently incubated with lentiviruses in the presence of protamine sulphate. Similar conditions will be used to transduce NK-92 cells. For in vitro expansion, transduced T cells will be grown in IL-2 and re-stimulated once a week with either anti-CD3/CD28-coated beads or irradiated (60 Gy) BARF1+, PSMA+ or PSCA+ tumour cell lines. NK-92 cells will require just IL-2 as supplement but will undergo enrichment or sorting to provide a population at high CAR expression. At different time points, T and NK-92 cells will be checked for CAR expression by anti-c-myc or anti-HA tag mAb, and will be thoroughly phenotyped by flow cytometry. (ii) CIK cells will be obtained by stimulating PBMCs with IFN-γ and, 24 hours later, with OKT3 and IL-

2. CIK cell phenotype will be monitored using multicolour flow cytometry. In parallel, a series

of “immunotools” for redirecting CIK cells will be procured or produced from available

constructs, including full-length engineered mAbs anti-CD20 (rituximab and obinotuzumab) or anti-CD19 (4G7 clone, either as CD19-IgG1, or CD19-DE or CD19-FcKO, where all three antibodies share the same V regions, while the Fc domains are wild-type or engineered), bispecific antibodies (CD19xCD3 and CD20xCD3), and immunoligands (CD20xULBP2 and CD19xULBP2). All constructs will be transiently expressed in CHO-S cells, after which cell supernatant will be harvested and antibodies purified, dialysed and stored at 4°C.

Milestones Task 2.7 IOV:

2.7.1 Production of infectious viral particles for all CARs and transduction of T and NK-92 cells, validated by CAR expression analysis and immunophenotyping (M1-12);

2.7.2 Production of CIKs suitable for adoptive transfer, and immunotools for CIK redirecting (M1-12).

2.8 IEO: Engineering of activated NK cells with chimeric antigen receptors (CAR) specific for AML cells. Aim of this task is to genetically modify NK cells by CAR in view of their use as consolidation therapy following haplo-HSCT in patients with high-risk haematological malignancies such as MDS/AML. This will be carried out in collaboration with Xxxx. Xxxxx Xxxxxxx, Singapore University. Considering that higher transduction efficiency is achieved when NK cells are activated and expanded, donor-derived and cord blood (CB)-derived NK cells will be in vitro activated by exposure to the genetically-modified clinical grade K562- mb15-41BBL cell line (kindly provided by Prof. Campana), known to induce NK cell activation and expansion. Telomere length will be assessed before and after expansion. Emerging studies demonstrate that CD123 is overexpressed on AML cells, and represents a distinct marker of leukaemia stem cells (LSCs). Thus, retroviral-vector encoding an anti‐CD123 CAR will be used to transduce ex-vivo activated donor-derived and CB-derived NK cells. We plan to also explore alternative leukaemia-specific receptors, possibly identified in our laboratory (see IEO-WP2, Xxxx. Xxxxxxx, aimed to the creation of novel tumor-targeting antibodies versus relapsed acute leukaemia). Analogous engineering procedures will be applied to NK cells following down- regulation of their inhibitory receptors (see Task 3h). Expanded and activated NK cells will also be tested before and after transduction for in vitro ADCC against different tumours and using different therapeutic antibodies, including rituximab, obinotuzumab, daratumumab, cetuximab and trastuzumab. Differences between donor-derived and CB-derived NK cells will be explored.

Milestones Task 2.8 IEO:

2.8.1 Expanded and activated NK cells tested for in vitro ADCC responses against different tumours and using different therapeutic antibodies (M2-12);

2.8.2 Optimized transduction of ex-vivo expanded NK cells by retroviral supernatant for redirecting them against CD123-expressing leukaemic cells and fresh samples from AML patients (M2-12);

2.8.3 CAR transduction in NK cells following NK inhibitory receptor down-regulation (M6-18).

TASK 3. In vitro preclinical assays.

3.1 FPO: Antitumor activity of CAR-CIK and CAR-NK-92. (i) 2D cell cultures. Cell lines, patient- derived cells or PDX-derived cells of BSTS, Melanoma and CRC will be treated with CAR-CIK (autologous when available, or allogeneic) or CAR-NK92 (single CAR or dual synNotch/CAR) at various effector/target ratios (10:1 to 1:32) for different times (4 to 72 hours). We will evaluate residual viable tumour cells by flow cytometry and Viability CellTiter-Glow assays. Additionally, we will assess recovery kinetics and phenotype of surviving tumour cells (considering modulator checkpoints like PD-L1/2 and target molecules like MIC A/B and ULBPs), CAR-killer activation (Th1/Th2 cytokine production and TF/pathway changes upon CAR engagement), and activity against putative CSCs. In selected cases, antitumor activity will

also be tested under challenging metabolic conditions (hypoxia, low glucose/high lactate). In selected experiments, side by side comparisons with conventional CAR-T cells will be pursued. The above assays will also be employed to explore the combined activity of CAR-killer cells with currently available treatments (chemotherapy, molecular targeted drugs and checkpoint modulators). (ii) 3D cell cultures. The main experimental platform will employ BSTS, melanoma and CRC “tumoroids” embedded in two layers of type I collagen in 96-well plates, covered with cell culture medium. This platform is “permeable” to cytokines and drugs, and allows dynamic photographic documentation of tumoroids, to monitor directional recruitment of CAR-CIK and CAR-NK92, and their infiltration and killing activity against 3D tumoroids. Drug combinations with the most interesting results in 2D will also be tested in the 3D setting.

Milestones Task 3.1 FPO:

3.1.1 In vitro efficacy data in 2D/3D cultures for CAR-CIK/CAR-NK92 anti-CSPG4, -GD2, - MSLN or anti-pan-CRC against BSTS, Melanoma or CRC cells, alone or in combination with drugs (M3-20);

3.1.2 In vitro validation of at least one NK-92 clone carrying a synNotch-CAR dual system against CRC cells (M9-20).

3.2 FPG: CAR-killer activity against CRC and GBM. After the optimisation of the system (see task 2b), primary human T cells or NK-92 transduced with the functional CARs (i.e. 13C4(Vk-VH)- 3rd gen), will be tested by in vitro killing- and cytokine release assays (e.g., IFNγ and IL12). The killing assays will be performed by using luciferase-expressing target cells (e.g., CRC-CSC) and the killing will be monitored by a decrease in luciferase activity. This can be used to measure the viability of CSC spheres in suspension using different effector/target cell ratios. We will use several effector:target (E:T) ratios starting from 50:1 and scaling down to 1:1 and different duration times of the co-culture (24, 48 and 72 hours). The cytokine release assays will be performed with standard methods such as ELISA or intracellular FACS staining. Taking advantage of the CSC growth conditions, also the infiltration of CAR T cells in larger 3D tumor spheres can be analyzed by confocal microscopy. To achieve this aim we will either perform immunofluorescence microscopy using anti-CD3 antibodies or monitor the real-time infiltration of GFP-labeled CAR T cells.

Milestones Task 3.2 FPG:

3.2.1 In vitro tests on CAR-effector cells for 2-3 CARs e.g. based on the mAbs 13C4, 20G4 and 19D10 (M6-12);

3.2.2 Tests on CAR-effector cells for additional 2-3 constructs, with mAb clones to be determined in the first year of the project (M12-24).

3.3 INT: Characterization of anti-LeY CAR-T in vitro activity. We will evaluate: (i) CAR-T functional activation upon LeY antigen recognition by measuring cytokines (IFN-ɣ, IL-2 e GM- CSF) production and release (ELISA, FACS analysis) and cell proliferation (CFSE assay) upon co-culture with LeY positive and negative tumour cell lines, including MS186, as well as with malignant ascitic cells from ovarian cancer patients; (ii) tumour cell killing upon LeY antigen recognition by xCelligence technology, that allows monitoring the target killing in real time;

(iii) CAR-T activity in an ex-vivo test that we will set-up exploiting freshly removed ascitic fluids from ovarian cancer patients, containing both neoplastic cells and immune cells.

Milestones Task 3.3 INT:

3.3.1 Functional in vitro and ex vivo characterization of anti- LeY CAR-T cells: activation, cytokine release, proliferation, cytotoxicity (M18-24).

3.4 OPBG: Creation and exploitation of a portfolio of standardized assays to evaluate in vitro various aspects of the antitumor activity of CAR-killer lymphocytes. OPBG will evaluate: (i) the penetration potential of CAR-T/NK/γδT cells in 3D tumour models, including neurospheres from brain tumours, organoids from sarcoma, neuroblastoma, colonic-rectal carcinoma and PDAC. CAR-T cell migration toward tumour cells will also be evaluated in the presence of a complex microenvironment, including fibroblasts, stroma cells and endothelial cells. (ii) effector cells activation upon CAR expression, employing a flow cytometry-based multiplex cytokine panel (up to 60 cytokines associated to effector cell activation), ELISA assays and RNA-seq analysis; (iii) cytotoxic activity of CAR T cells, employing the standard cytotoxic 4hr assay plus long term co-culture assay, stressed co-culture upon to 30 days of in vitro CAR T- cell stimulation, cell killing activity in 3D cultures, and degranulation assay for NK CAR and CAR-γδ T cells; (iv) possible synergism with conventional treatments currently in clinical practice. In particular, OPBG will perform a high-throughput screening of drug library approved from FDA and EMA. The platform to evaluate drug synergy with CAR-T cells will be applied in several solid tumour models.

Milestones Task 3.4 OPBG:

3.4.1 Novel CARs tested in vitro for antitumor activities, as well as for their ability to proliferate and to be activated upon antigen stimulation (M6-12);

3.4.2 Discovery of novel synergic drugs able to maximize CAR-T cell activity (M6-24).

3.5 FT/SGER: In vitro activity of CAR-CIK against AML. (i) Target recognition. Studies of scFv binding characterization will be performed once a new CAR design is set up. Subsequently, CAR-CIK cell signalling transduction assays will be done in order to check the functionality of the CAR. Potential CAR design optimization will be envisaged for the ideal target recognition (epitope accessibility, spacer length, etc.). (ii) Killer cell activation. Functional in vitro assays of cytokine release and proliferation upon target antigen challenge will be done by means of intracellular staining and flow cytometry analysis. (iii) Target cell killing. The antitumor activity of CAR-CIK cells in vitro against AML cells will be performed by flow cytometry cytotoxicity assays by means of ANNV/7AAD apoptosis/necrosis assays.

Milestones Task 3.5 FT/SGER:

3.5.1 Standardized in vitro preclinical assays for the functional characterization of CAR-CIK cells against AML. (M 1-12)

3.6 IRST: CAR-T activity in 3D models of melanoma, prostate cancer, glioblastoma and mesothelioma. (i) For melanoma, GD2-expressing melanoma cells will be used to generate 3D organotypic skin melanoma culture (OMC) models, and test: (a) migration/infiltration ability. CAR-expressing cells (T, NK, CIK cells available in house or within the Consortium, e.g. anti- GD2 effectors from OPBG) will be seeded onto a pre-solidified gel in a Transwell, adherent to the dermal side. CAR cells will be tested also after engineering to express metalloproteases or chemokine receptors, designed to increase the degradation of ECM components and tumour penetration. (b) Killing ability. CAR-killer cells will be injected in the OMC, rather than seeded on top, at various ratios. Sequential IHC will be performed on OMC FFPE sections at different time points, assessing the positivity of the tumour cells for cleaved caspase 3. (c) CAR-killer activation. Gene expression profiles with the Nanostring nCounter “CAR-T Characterization Panel” will be generated on CAR-killer cells before and after 3D culturing. Additionally, the proliferation and activation status of CAR-effectors will be defined by multiparametric flow cytometry analysis, and supernatants collected from OMCs will be analysed. Organotypic skin cultures not containing melanoma cells will be used as negative control. (ii) For PCa, GBM and

mesothelioma, tumour spheroids generated in ultra-low-attachment plates and embedded in rat I type collagen or Matrigel will be adopted as 3D in vitro models, including melanoma cells used for OMCs for comparison. CAR-killer cells will be added at different ratio on top in medium or in Transwell 96-well permeable support placed into the spheroid microplate. Migration/infiltration and killer cell activation will be assessed as described for OMCs. To test killing in tumor spheroids, we will use dedicated CellTiter-Glo® Luminescent 3D Cell Viability Assay. Additionally, Luciferase reporter has been described as a convenient read-out of viable cells in cytotoxicity assay, therefore we will take advantage of in-house available mesothelioma cell lines engineered to express Firefly Luciferase. In both these two described 3D platforms, commercially available cell lines, short-term patient-derived tumour cell cultures as well as engineered tumour cells will be employed. Whenever possible, fresh surgical patient-derived material will be employed to generate primary organoid cultures to be used with autologous CAR-T cells, engineered in house or by the other partners.

Milestones Task 3.6 IRST:

3.6.1 OMC infiltration tests on 3 or more GD2 CAR-expressing cells (M3-8), and OMC infiltration tests on one or more CAR-expressing cells engineered to increase tumour penetration (M12-24);

3.6.2 OMC injection tests, with evaluation of cancer cell killing, CAR cell distribution and definition of optimal killer/target ratio and assay time (n=6; M3-M12).

3.6.3 CAR cell activation profile upon 3D tumour cell interaction defined (M15-M24). (iv) Derivation and testing of organoid cultures as described for OMCs, and comparative analyses (M6-24).

3.7 IOV: In vitro testing of CAR-effector cells against prostate, gastric and nasopharyngeal carcinoma, and of redirected CIK cells. (i) CAR-T and CAR-NK-92 (anti BARF-1-CAR, PSMA-CAR, PSCA-CAR or PSMA/PSCA-CAR) will be tested for lytic activity against different target tumour cell lines by standard 51Cr-release or calcein assays. The production of different cytokines upon CAR stimulation will be measured in standard ELISA tests, ELISPOT analysis or by intracellular staining. Additionally, a functional test (outgrowth assay) will be set up where scalar amounts of antigen-positive tumour cells are challenged with fixed numbers (from a 1:1 ratio) of CAR- T or CAR-NK-92 cells. This assay will provide a preliminary in vitro basis for the comparative assessment of the efficiency of effectors expressing different CAR but targeting the same tumour histotype (PCa). (ii) The activating and redirecting activities of immunotools, available and produced to enhance CIK cell-activity, will be evaluated for anti-tumor killing and compared to CIK cell activity. Multiplex cytokine profiling will be assessed in both effector populations upon challenge with CD19+/CD20+ relevant targets, and outgrowth assays will be carried out as previously detailed. ADCC will be assessed in standard cytotoxicity assays using as target cells CD19/CD20-positive B-ALL cell lines, as well as primary refractory B-ALL patient samples. By the end of project, these data will provide a proof of concept about the extension of such approach to non-B cell malignancies.

Milestones Task 3.7 IOV:

3.7.1 In vitro functionality tests on CAR-T and CARNK-92 cells against prostate, gastric and nasopharyngeal carcinoma cell lines (M6-18);

3.7.2 Demonstration that the concept of antigen-specific CIK cell retargeting is valid in haematological malignancies, and that immunotool-retargeted CIK cells are in vitro equally efficient as CD19 CAR-T (M6-18);

3.7.3 Demonstration that retargeted CIK cells can be efficiently applied to solid neoplasia and non-B cell tumours (M18-24).

3.8 IEO: In vitro NK cell activation by down-regulation of their inhibitory receptors. NK cells display receptors that upon stimulation may lower or inhibit anti-tumor activity. We will therefore generate reagents to inhibit transiently or permanently a series of inhibitor receptors in NK cells, including: (i) a collection of microRNAs targeting 12 NK inhibitory receptors, of which KIRDL1/2/3, CD94/NKG2 and LILR, bind MHC ligands, while XXX-X0, XXXX-0, Xxxxxx-0/0, xxx XXXXXX0 bind non-MHC class I ligands. Each receptor will be targeted with at least 4 miRNAs by electroporation of NK. microRNAs that deplete for at least 48h the expression of the receptor (total and on the surface of the cells) will be further explored. We will then test combinations of microRNA to get the most efficient multiple receptor downregulation.

(ii) RNA guides (CRISPR/Cas9 targeted genome editing) specific for each receptor that can be used to delete segments of the specific gene corresponding to the receptor that works better in terms of activation of NK cells in vitro before stem cell transplant. This part of the study will be performed in collaboration with Prof. E. V. Avvedimento at Xxxxxxxx II University, Naples. Modified NK cells will then be tested in a 4-hour cytotoxicity assays performed against luciferase-labelled K562 cells, a well-known NK target.

Milestones Task 3.8 IEO:

3.8.1 A set of 60-80 microRNAs that deplete for at least 48h the expression of the respective NK inhibitory receptor target (M2-8);

3.8.2 Sets of microRNAs that can be combined together to elicit concomitant downregulation of multiple NK inhibitory receptors. Duration of downregulation is critical, and we expect to get a sufficient inhibition in vitro for 3-4 days (M2-14);

3.8.3 Two to four selected NK inhibitory receptors deleted by CRISPR/Cas9 and characterized for the consequent changes in NK phenotype (M6-14).

TASK 4. In vivo preclinical tests.

4.1 FPO: In vivo antitumor activity of CAR-CIK and CAR-NK92. (i) Antitumor activity. The antitumor activity of CAR-CIK and CAR-NK92 (single CAR or dual synNotch/CAR) will be explored in vivo by their intravenous injection into NOD/SCID or NSG mice bearing palpable subcutaneous patient-derived BSTS, Melanoma or CRC xenografts. We plan to assess: (a) antitumour activity (tumour regression/growth inhibition, necrosis and metabolic response); (b) Tumour infiltration, assessed by IHC in explants; (c) Proliferation and persistence. Based on the results of previous in vitro experiments, we will also assess in vivo the efficacy of sequential or concomitant treatment with chemotherapy, targeted therapy or checkpoint modulators. (ii) Toxicity. We will check for possible side effects and toxicity by assessing general conditions and weight of treated mice, infiltration of CAR-CIK/NK-92 in healthy organs, and cytokines levels in peripheral blood.

Milestones Task 4.1 FPO:

4.1.1 In vivo efficacy/toxicity data for CAR-CIK/NK92 anti-CSPG4, anti-GD2. anti-MSLN or anti-pan-CRC against BSTS, Melanoma and CRC cells, alone (M7-18)

4.1.2 in combination with drugs (M13-24).

4.1.3 Functional validation in vivo of at least one NK-92 clone carrying a synNotch-CAR dual system against CRC cells (M15-24).

4.2 FPG: Preclinical mouse models with novel CAR T cells using xenografts derived from GBM and mCRCs. CAR-effector cells showing strong anti-tumor effects in vitro will also be tested for their function in the luciferase-expressing xenograft models described above. T cells will be isolated from human peripheral blood and Balb/C mice, transduced with lentiviral particles encoding for the CAR(s), expanded in vitro using anti-CD3/CD28 stimulation and injected into

tumor-bearing NSG mice (200,000 and 1,000,000 T cells/mouse). Alternatively, NK-92 cells expressing the novel CARs characterized before will be used at similar doses. The effect on tumor growth will then be evaluated using standard tumor measurement techniques (e.g., caliper) or small animal imaging (IVIS). Initially, we will test subcutaneous xenografts, which allow easy biopsy sampling of CAR T cell homing and cytotoxicity (~40 mice/target). The biopsies should be controlled for infiltration and activation markers of CAR-expressing T cells. Then we will take advantage of our orthotopic brain (GB) and liver (CRC) models (~90 mice/target). Mice will be monitored for weight loss and changes in the behavior that could be due to the treatment.

Milestones Task 4.2 FPG:

4.2.1 In vivo testing of CAR-killer cells for a total of 4 CAR constructs (M10-24).

4.3 INT: In vivo CAR-T activity against LeY expressing tumours. We will set up in vivo preclinical models to tests CAR-T therapeutic efficacy on LeY expressing xenografts (ovarian, breast, lung, gastric, prostate cancer lines and melanomas) and on xenografts from the MS186 line. Mice bearing LeY-positive and -negative xenografts will be injected with anti-LeY CAR-T and tumour growth will be monitored as well as toxicity.

Milestones Task 4.3 INT:

4.3.1 Functional in vivo characterization of anti-LeY CAR-T cells: anti-tumour efficacy and toxicity in mouse xenografts and PDXs (M18-24)

4.4 OPBG: In vivo testing of CAR-T/NK/γδT cells. OPBG will apply bioluminescence in vivo approach to monitor CAR-effector cells homing to the tumours. Moreover, homing to the specific tissue will be also proved by IHC approach. By using bioluminescence in vivo approach, we will monitor the anti-tumour activity of CAR-effectors towards several tumour models, including ALL, AML, NHL, HL, neuroblastoma, glioblastoma, medulloblastoma, colon carcinoma and pancreatic carcinoma. Drugs selected in vitro will be further evaluated in vivo. Next, we will evaluate tumour-intrinsic and microenvironment-related factors associated to CAR-effector cell immunoregulation, in ALL, AML, neuroblastoma, CRC and PDAC models. With the aim to optimize CAR-T cell in vivo proliferation and persistence, OPBG will generate an animal model mimicking patient tumour relapse and challenging CAR-T cells for their ability to expand in the presence of target cells. In terms of toxicity, OPBG will evaluate the potential on-target off-tumour toxicity of novel CARs in a humanized mouse model. In addition, the activity of suicide genes to control potential CAR-effector cell toxicity will be exploited in an animal model.

Milestones Task 4.4 OPBG:

4.4.1 One in vivo experiment to evaluate the activity of a novel CAR in the context of sarcoma (M9-12);

4.4.2 One in vivo experiment to evaluate the activity of a novel CAR in the context of glioblastoma (M9-12);

4.4.3 Humanized model to evaluate on-target off-tumour activity of the novel CARs (M12-24).

4.5 FT/SGER: In vivo activity of CAR-CIK against AML. (i) Anti-AML activity. The in vivo anti- AML activity of CAR- CIK cells will be characterized by means of in vivo detection of leukemic cells within the hematopoietic compartments (bone marrow, spleen and peripheral blood) after treatment of AML-engrafted mice with CAR-CIK cells and comparison with the untreated cohort. (ii) Synergism with conventional treatments. The in vivo potential activity of CAR-CIK cells on chemo-resistant/refractory AML cells will be explored by using the

chemotherapy xenograft model which mimics the AML relapse after induction therapy. (iii) Resistance to immunosuppression. The main players involved in AML immunosuppression within the bone marrow niche will be characterized in order to identify novel targeted therapies to boost CAR-CIK cells against AML. (iv) Proliferation and persistence. To understand the potential limitations of proliferation and persistence of CAR-CIK cells within the AML setting, a deep characterization of checkpoint inhibitors or T-cell exhaustion markers and comparison between CAR-CIK cells expanded in vitro and CAR-CIK cells derived ex vivo from the treated mice will be done. (v) Toxicity. The potential toxicity of anti-AML CAR-CIK cells will be characterized in terms of (a) myelotoxicity (b) employment of CRS mouse models to check the toxicity profile of anti-AML CAR-CIK cells (single vs dual CARs).

Milestones Task 4.5 FT/SGER:

4.5.1 Standardized in vivo preclinical assays for the functional characterization of CAR-CIK cells against AML (M 10-24).

4.6 IOV: In vivo testing of CAR-effector cells against prostate, gastric and nasopharyngeal carcinoma, and of redirected CIK cells. (i) For both imaging and therapeutic purposes, experiments will be performed using six-to-eight-week-old NSG mice, housed in our Specific Pathogen Free (SPF) animal facility. The research project may exploit the property of adopted LV constructs that are already designed to contain firefly luciferase, to allow longitudinal tracking of CAR-killer cells in vivo. Biodistribution and tumour-homing properties of different CAR-killers will be assessed by bioluminescence on s.c. or orthotopic tumour-bearing mice, at different timepoints after in vivo transfer. Xxxxxxx employed in vivo will be engineered to express a different luciferase (renilla) using a different substrate, thus allowing to carry out co- localization studies aimed at visualizing both the tumour and the infiltrating transferred CAR- killers. Whole-body images will be obtained and analysed with Living Image 2.50 software (Caliper). For therapeutic experiments, tumour-bearing mice will be repeatedly treated by i.v. injections with CAR-T cells or lethally irradiated CAR-NK-92 cells, and therapeutic activity will be assessed by monitoring tumour growth and by recording survival. (ii) In vivo functional analysis of immunotools-redirected CIK cells as compared to CD19 CAR-T cells. For the in vivo assessment of therapeutic efficacy, NSG mice will be inoculated either i.v. with luciferase- transduced B-ALL cell lines, or intra-femorally with primary samples from patients affected by refractory CD19-expressing B-cell malignancies. Upon treatment with immunotool-redirected CIK cells or with anti-CD19 CAR-T cells, tumour burden and response to therapy will be estimated by bioluminescence or PET scanning, and mouse survival will be recorded.

Milestones Task 4.6 IOV:

4.6.1 In vivo assessment of tumour homing and therapeutic efficacy of CAR-killer cells against prostate, gastric and nasopharyngeal carcinoma (M12-24);

4.6.2 Demonstration that the concept of antigen-specific CIK cell retargeting is valid in haematological malignancies, and that immunotool-retargeted CIK cells can efficiently replace CD19 CAR-T (M12-24).

4.7 IEO: In vivo tests for NK cell persistence and biodistribution. We will set up a sensitive imaging procedure, applicable to human patients, whereby NK cells are ex vivo labelled for in vivo imaging. We will explore the MRI-based approach to obtain anatomical and disease diagnostic information without ionizing radiations, using Fluorine-19 (19F), a natural halogen, non-radioactive isotope of fluorine. Fluorine-dense perfluorocarbon (PFC) nanoemulsions, engineered to be endocytosed by non-phagocytic cells in culture, will be used to label our NK cell preparations by simple incubation with PFC in the media, for variable times (up to one day), followed by a wash step. Alternatively, we will test PFPE, a linear perfluoropolyether polymer

described as an effective 19F MRI cell tracking reagent. Optimization parameters to be tested will include dose of PFC in media, cytoplasmic volume and uptake properties of cells, together with viability, phenotype and killing activity in vitro. PFC-labelled human NK cells will be assayed by MRI detection for three times over a week after injection in non-obese diabetic (NOD) mice. PET/CT imaging methods based on oxine-conjugated zirconium89 (89Zr) will be explored as well, for high potential sensitivity, in collaboration with Xxxx. Xxxxx Xxxx of the Nuclear Medicine Dept. at the University of Pisa. This radionuclide conjugate has been successfully used as a simple and clinically translatable cell label to track immune cells with PET/CT in mouse models. It readily permeabilizes the cellular membrane and is retained in the cells, without no changes in CD16 or CD56 expression, and cytotoxicity against K562 cells, for up to 6 days.

Milestones Task 4.7 IEO:

4.7.1 Optimized labelling of ex vivo selected and engineered NK cells with nano-emulsion formulations for in vivo MRI, with information about NK labelling efficiency, persistence and biodistribution.

4.7.2 Validation of radionuclide-based labelling and PET/CT scan as an alternative approach for

in vivo NK cell tracking.

Feasibility, Risk Assessment and Contingency plans

The Candiolo Cancer Institute – FPO IRCCS, Coordinating Centre of WP3, has been hosting for several years research activities that, altogether, can cover the entire CAR-T field, from cancer patients to preclinical models, from CAR-target validation to CAR-killer cell engineering and testing, in vitro and in vivo. The WP3 Coordinator, Xxxx Xxxxxx, has a long-standing experience in development, characterization and exploitation of patient-derived cancer models. He currently coordinates the H2020 EU project EDIReX, aimed at the development of a European Research Infrastructure focused on patient-derived cancer models. Furthermore, his experience and collaborative networks in integrative cancer genomics, functional genomics and bioinformatics ensure optimal integration of preclinical model data from WP3 partners. Feasibility for Partner- specific activities follows.

FPO. TASK 1: we expect about 20 new melanoma and sarcoma samples, with corresponding PBMCs (in 20% of cases), every year. This rate should cover the expected failure-rate of about 40% in generating tumour cultures or a conservative 60% in case of 3D models. In the case of low success rate of 3D models, we will evaluate alternative or parallel methodologies, including perfusion-based flow-dynamic systems (e.g. AIM Biotech 3D Cell Culture Chip), improving distribution of tumor-tissue-structures throughout the scaffold, availability of nutrients, oxygen, drugs and “infused” lymphocytes. TASK 2: We already have the CAR constructs necessary to conduct the planned experiments, plus expertise and positive preliminary data about transduction and functional assessment of CIK and NK-92 cells. In the unlucky case of suboptimal activity of CAR constructs, we will design multiple variants of the same CAR that might address issues of target-binding affinity, antitumor activity and in vivo persistence. TASK 3: We have long-standing experience with in vitro testing of drug- and cell-therapeutic approaches to cancer. Specifically, the antitumor activity in vitro of anti-CD44v6 CAR-CIK against selected soft tissues sarcomas was first reported by our group in a recent publication. It is possible that some assays already in place in our labs provide suboptimal information, in which case we will network with the project partners to improve the tests or acquire new ones. TASK 4: Our group has a decennial experience in the field of cancer adoptive immunotherapy, both in vitro and in vivo. We can count on ongoing funded

projects in cancer adoptive immunotherapy and CAR-engineering, that support shared experimental platforms and materials with the present research, allowing a reduction in the planned experimental costs. If unexpected toxicities should emerge from early in vivo essays, we may re-design the experiments with lower doses or reduced number of infusions. The analysis of the lowest effective E/T ratio obtained in vitro will help us to decide possible dose reductions of CAR-CIK in the following in vivo experiments. If we do not observe significant benefit by blocking PD1 in CAR- CIK or CAR/NK, we may explore, or integrate, additional checkpoint inhibitors. Anti-TIM3 or anti- LAG3, currently explored in clinical trials, would be intriguing alternatives to explore based on our preliminary analysis.

FPG

Besides the profound experience in the isolation of CSC from primary tumors and the generation of stable CSCs lines, the research group at FPG/IRCSS successfully isolated mAbs targeting GB CSCs. The group has access to laboratories equipped with facilities dedicated to automated cell culture. The flow cytometry facilities are equipped with high-throughput instrumentation. In addition, the PI's laboratory can count on high-profile collaborations that will assist in the optimal accomplishment of the proposal's goals. GB and mCRC samples will be provided by the Gemelli hospital. We are convinced that the final part of the existing working pipeline for the identification of potentially therapeutic CARs will then seamlessly integrate with WP3 to obtain valid preclinical in vivo data. To improve the chances of success, we will take advantage of profound expertise of our collaboration partner OPBG for the generation of CAR T cells. Therefore, we presume to produce positive preclinical outcomes in mouse experiments, at least with the already validated mAbs, and thus open the opportunities for further clinical development.

INT

The scientific and technical staff has the appropriate background, knowledge and skills to achieve the completion of the project; in fact, the majority of them have 8 to more than 20 years experience in the specific field of interest. The constant interactions among the participants will allow regular discussions on the progress and status of the project. The research will be conducted in the laboratories of the DRAST (Department of Applied Research and Technological Development) which have all equipment, instruments and facilities necessary for the project. The majority of reagents needed for the project are directly produced in our laboratory or purchased from commercial companies and fully available. All the in vitro experiments will be performed in triplicates and in at least 2-3 independent experiments to enable a correct statistical analysis of the results. For in vivo experiment we can take advantage of a certified animal house service with expertise in developing transgenic and xenografts mouse model. For in vivo analysis, according to INT OPBA rules, at least 5 animals/group, will be evaluated and at least 2-3 independent experiments will be performed. Parametric and non-parametric statistical approaches will be used, as appropriate. The P values of all statistical tests will be two-sided and a P-value <0.05 will be accepted as significant.

OPBG

The feasibility of the WP4 proposal is guaranteed by the significant expertise of OPBG unit in the field. In particular, we do not predict major risks of failure to fulfil the expected milestones. Minor issues could be associated to each Task. TASK1: In the generation of patient-derived cell cultures, we need take in consideration that around 10% of failure could be observed (based on our previous experience). Historically, culturing cancer cells from solid tumours has generally not been rapid or readily feasible. In addition to this challenge, patients presenting with metastatic disease often undergo a diagnostic needle biopsy rather than surgical resection, and the biopsy material may be relatively limited. OPBG has established conditions that allow for more robust and, at times, otherwise unattainable efficiency in culturing cancer cells from surgical or biopsy samples. TASK2:

We may face issues in the validation of a scale-up phase of the GMP process needed to generate CAR NK and CAR γδ T cells for the generation of an allogenic third-party bank, without the use of a cellular feeder and in the absence of fetal bovine serum. Thus, OPBG will optimize the process proceeding with the modification of the culture conditions in terms of cytokines and pleiotropic stimulations. TASK3: After the generation and characterization of CAR-effector cells, we could realize that the CAR conformation is not optimal. Thus, OPBG will perform novel cloning of the CAR constructs based on its own previous experience, to insert modification in the hinge and the transmembrane domain, as well as evaluating alternative co-stimulation. TASK4: In vivo preclinical studies will be performed by applying already established model for neuroblastoma and leukaemia. To also develop xenograft mouse models of sarcoma, glioblastoma, CRC and PDAC, OPBG will perform an in vivo study with an escalation dose of the tumor cells to establish the correct number of tumor cells to be implanted to achieve a robust animal model.

FT/SGER

We have solid expertise with leukaemia mouse models and full access to the facility of University Milano Bicocca. We have established standardized methods for non-viral gene manipulation, functional analysis and use of pre-clinical models. Reagents and protocols for the construction of the AML-CARs including the dual-CARs are already available. In particular, our group has solid and documented expertise in the generation and characterization of CIK CAR-effector cells, and in this context we will take advantage in particular of: (i) our solid expertise with the assessment of leukaemia mouse models and with all the related toxicity assays; (ii) our well established flow cytometry unit for several in vitro test of proliferation, apoptotic and immunologic markers assays;

(iii) our great expertise in molecular genetics methods including advanced platforms such us NGS. Possible risks and contingencies: (i) NGS analysis on AML PDX may be insufficient to support directly the creation of advanced mouse models, (e.g. by genetic engineering). In this contingency, we will continue the study with a wider group of AML PDX and this part will be funded from other sources, beyond the budget of the present grant; even in case of limited contribution to the efficacy against AML targeting the immune suppressive microenvironment, this will allow to obtain new knowledge in this field; (iii) GvHD occurrence: CIK cells have limited alloreactivity and the treatment foresee a single infusion, therefore the applicant expect minimal GvHD; (iv) emergence of AML- epitope loss: for anti-CD33 CAR, these issues are not expected to occur, since the extensively documented CD33 persistence upon anti-CD33 antibody-drug conjugate Mylotarg treatment; (v) treatment-related toxicity, or lack of efficacy: these issues could be addressed by developing a dual-CAR strategy, ensuring better safety and selectivity; to evaluate the treatment toxicity and to prevent its occurrence in patients, infused animals will be monitored according to in- house standard procedures. In case of lack of efficacy and persistence, we will explore multiple infusions.

IRST

IRST is a dynamic Oncology Hospital and Cancer Research Institute founded in 2007. Its research activities are strongly oriented to the full integration of preclinical, translational and clinical research. All the required expertise is present: medical oncologists, biologists, biotechnologists, bioinformatics and biostatisticians. The laboratory’s equipment includes all of the instruments necessary to perform all the WP3-related molecular, biochemical and cellular experiments.

IOV

The Immunology and Oncology Molecular Diagnostics (IDMO) Section of IOV offers several supports through its main core facilities and a >400 m2 Specific Pathogen-Free animal facility for rodents, equipped with in vivo imaging systems. The PI’s Laboratory focuses on the development of the most innovative strategies of adoptive immunotherapy and antitumor vaccination approaches. Lab coworkers have an excellent expertise in in vitro evaluation of T, NK-92 and CIK cell

phenotypic profile, cytotoxic activity, cytokine release, and transgenic T cell Receptor (TCR) and CAR transduction with viral vectors; moreover, PI and co-workers have vast skills and competences in both syngeneic and xenogeneic mouse cancer models of different histotypes, in assessment of pharmacokinetics and pharmacodynamics of new investigational drugs, and in in vivo biodistribution studies. Additionally, a close collaboration has been established with Prof. Peipp’s research group in Kiel (Germany). Prof. Peipp has great experience in developing novel approaches to enhance antibody-based immunotherapy of cancer by manipulating effector cell recruitment. Several immunoligands, bsAb and engineered antibodies have been developed by the group and assessed for their capability in triggering effector cell cytotoxicity. Intriguing results have also been achieved in the investigation of enhancement of NK cell-mediated lysis of acute lymphoblastic leukemia by the Fc-engineered anti-CD19 antibody. All these tools have been made available to the PI based on a collaborative agreement. We have paid maximal attention to advance a proposal that is realistic and has the premises to be successfully carried out and completed. The overall rationale of the project appears to be well grounded in the preliminary results reported. We have a great competence in mAb development, purification, labelling with fluorophores and radionuclides, and in vitro and in vivo use. All other techniques and procedures required in the project (CAR receptor development, CAR-T, NK-92 and CIK cell generation and functional assessment in vitro and in vivo) are well set up, including the establishment of PDX, and nuclear medicine approaches that rely on group’s competence and the recent installation of an exceptional trimodal imaging apparatus for SPECT/PET/CT scanning.

IEO.

Our group is well trained in the use of intensified HSCT programs with mobilized PBSC in the management of hematologic malignancies. The improvement of treatment options for patients with refractory disease is a major issue in our research studies since long time. In this view, we have been using the Baltimore haplo-HSCT program with PBSC as graft source for patients with high- risk haematological malignancies. More recently, we have started a new program with donor- derived NK infusion following haplo-HSCT in refractory patients, according to the procedure developed by the Seattle group and detailed in the Preliminary result section. Taken together, all these premises make the program quite feasible from the clinical point of view. Possible issues: (i) Lymphocyte labelling can be challenging due to their small cellular and cytoplasmic size that limits the number of particle droplets it can hold. In addition, lymphocytes are not naturally phagocytic. Nonetheless, PFC nanoemulsions have been specifically engineered to be endocytosed, even by non-phagocytic cells in culture. Moreover, NK cells labelled with PFPE nanoemulsion exhibited unaltered viability and phenotype, and were detectable by longitudinal MRI up to 8 days after intratumoral injection in NSG mice. Relevant to our project, 19F MRI cell detection techniques have employed in clinical trials and feasibility has been established in a first-in-human clinical study. In conclusion, while preliminary, all available data indicate that 19F MRI-labelled NK cells can be traced in vivo. In case of limited sensitivity, the use of radiolabelled NK cells will be adopted. This sensitive radioactive technique has been extensively tested in humans and represents the method of choice in a variety of clinical settings. (ii) A systematic approach to deplete the vast majority of NK inhibitory receptors has never been done before. Current protocols employ antibodies or molecules to inhibit single inhibitory receptors and our proposed approach will eliminate the heterogeneity and the adaptation of the NK cell to the inhibition of a single inhibitory receptor. Other points that may undermine the activation of NK cells when these inhibitory receptors are depleted: 1. Tolerance of NK to MHC class I-negative cells is actively maintained, and is quickly lost when tolerant NK cells are mixed with MHC class I-positive cells. Blocking of some inhibitory receptors may halt the development of a functional NK cell population; 2. NK cells that lack inhibitory receptors for self MHC class I exist but are hypo-responsive. 3. NK cells that receive stronger inhibitory signals through MHC class I recognition acquire a stronger capacity to respond to activation signals.

WP4 Novel strategies to increase activation, expansion, survival, tissue penetration, cytotoxicity and monitoring of CAR expressing cells

WP leader: Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IRST); Partecipanti: Ospedale San Xxxxxxxx (OSR), Ospedale Pediatrico Bambin Gesù (OPBG), Istituto Nazionale Tumori Regina Xxxxx (IRE), Fondazione Tettamanti/Ospedale San Xxxxxxx (FT/SGER), Centro di Riferimento Oncologico (CRO), Rete AMORe (Xxxxxxx, Oncologico Bari, CROB)

Rational

Indeed although adoptive immunotherapy with T cells engineered with CARs have achieved unprecedented results in acute lymphoblastic leukaemia and diffuse large B cell lymphoma treatment, their efficacy against solid tumours still lags behind. The reasons for this difference are poorly understood and may comprise the lack of sufficiently restricted target antigens in solid tumours, tumour inherent heterogeneity and distinct anatomic structure, the presence of a highly immunosuppressive microenvironment, the inability of infused cells to reach the neoplastic site, to proliferate in the context of the tumour microenvironment and to persist for a long time as well as the need for engineered T cells to be able to carry out effector functions in these particularly unfavourable clinical contexts.

The adverse tumour microenvironment (TME) is composed of diverse molecular and cellular elements, making CAR-T cells dysfunctional. For example, oxidative stress, nutritional shortage, acidic pH, and oxygen absence threaten the efficiency of CAR-T cells. The secretion of immunosuppressive cytokines, namely TGF-β and IL-10; suppressive immune cells, such as myeloid-derived suppressor cells (MDSCs), regulatory T cells (Tregs) and tumour-associated macrophages (TAMs) or neutrophils (TANs); and checkpoint inhibitory proteins including PD-L1 affect an efficient anti-tumour function of the CAR cells. Targeting TME by CAR T therapy represents an attractive strategy to treat solid tumours as it has been recently proven in immunocompetent murine models.

Tumour extrinsic pathways responsible for T cell exclusion are sustained by stromal cells that may limit T cell trafficking within the TME by different mechanisms, including the secretion of soluble factors and extracellular matrix components (ECM) whose density and organization influence immune cell migration. Cancer activated fibroblasts (CAFs) are the major producers of ECM and become activated as a consequence of various stimuli in the TME. The characterization of the composition of the tumour mass, of the extracellular matrix as well as of the expression of specific tumour markers involved in the conditioning of the cells in the TME, is of pivotal importance and it is believed to drive the design of the next generation of strategies to tackle refractory solid tumours. In this perspective the development of analytical procedures to describe the complexity of the tissue, to derive hints about possible therapeutic targets and at the same time to monitor and localize CAR expressing cells, will be crucial first for our understanding of their interaction with tumour and/or tumour-associated/derived components and, second, to rank and to compare constructs endowed with specific properties transferred to the effector cells. In parallel the use of standardized platforms to analyse the properties of effector cells at different stages of the process (from manufacture to storage and post-infusion) will help to define the best conditions of production and the impact of specific changes in the protocols/constructs in the view of a continuous improvement. In this contest, a particular focus will be on the visualization and quantification of cells in the tumour mass as a key aspect to develop very effective CAR cells and unleash their maximal potential in the context of a disfavoured TME.

Beyond improving the efficacy for solid tumours, also the translation of success of CAR cells towards other haematological malignancies represents a challenge, especially for those types of leukaemia characterized by an immune suppressive microenvironment, such as Acute Myeloid Leukaemia (AML) and Chronic Lymphocytic Leukaemia (CLL). In the AML context, current standard treatments for AML ensure high rates of remission, but long-term disease control remains

unsatisfactory and the overall survival rate is around 30%, highlighting the urgent need for new treatments.

Concerning AML, the first pioneering clinical trials showed transient and partial responses after CAR T-cell therapies against several targets (Xxxxx Y, NKG2D, CD33 and CD123). Overall, these evidences highlight the need to find novel strategies to boost CAR T-cell therapy against such aggressive and resistant haematological malignancies.

Multiple Myeloma (MM) as well is a disease involving malignant transformation of plasma cells that thrive in the bone marrow and accounts for more than 60.000 new cases yearly in Europe and United States. The identification of novel therapies for MM is still a high clinical priority. Adoptive cell therapy with CAR redirected T cells as well as with cells secreting interfering factors represent promising and innovative immunotherapeutic approaches. Anti-CD19 CAR-T therapy has demonstrated consistently high antitumor efficacy in haematological malignancies, but CD19 expression is not generally associated with MM. Syndecan-1 (SYC1, CD138) may represent an important target for CAR-T therapy in MM being highly expressed on malignant plasma cells with a reported key role in disease progression.

WP4 is structured around 5 main tasks:

1. Development of platform to be used across the activity of the WP4

2. Improving Efficacy

3. Modulation of immunosuppressive elements along with targeting and overcoming the inhibition exerted by TME

4. Deep characterization of the immunosystem for the retrieval of the relevant features to optimize CAR T cell approach

5. Development of novel therapeutical Synergy to optimize CAR cell approach

Preliminary results and Feasibility Scientific activities performed at OSR

X. Xxxxxx’ lab has shown that TSCM can be selectively expanded in vitro utilizing an OSR- proprietary method that utilizes total T cell stimulation by anti-CD3/anti-CD28 beads in the presence of IL-7+IL-15 (Ciceri Blood 2015). These cells are also efficiently transduced in vitro with lentiviral vectors encoding tumor-specific CARs and TCRs and have strong therapeutic activity against hematological and solid human tumor xenografts in NSG mice (Provasi Nat Med 2012, Xxxxxxx Blood 2013, Mastaglio Blood 2017, Xxxxxxx Xxx Med 2018). Bonini has also shown that spontaneous donor-derived TSCM can be detected, persist very long term and correlate with disease remission in patients undergoing hematopoietic stem cell transplantation for acute leukemia (Xxxxxxxx Sci Transl Med 2015, Xxxxxxxx Xxx Commun 2019). In addition, Xxxxxx has pioneered the concept of endogenous TCR gene disruption and complete editing of the specificity of TSCM cells, to allow the single expression of tumor specific TCRs and optimize their anti-tumor effect in vivo (Provasi Nat Med 2012, Mastaglio Blood 2017). This approach exploits now the CRISPR/Cas9 technology and is succesfully applied also for deleting the expression of any selected inhibitory immune receptor expressed by TSCM cells, in order to make them resistant to the specific immunosuppressive signals present in the tumor microenvironment (Xxxxxxxxxx, Bonini unpublished). T cell tracking (via TCR gene sequencing and vector integration mapping) in longitudinal samples harvested from patients enrolled in gene therapy studies have shown that genetically engineered TSCM cells are endowed with high expansion and persistence capacity (Xxxxxxxx STM 2015). Overall, T cell manufacturing protocols designed to enforce a TSCM phenotype to CAR-T cell products are currently implemented in clinical trials.

X. Xxxxxxxx and P. Xxxxxxxxx’x lab has contributed to the discovery of human CD1d-restricted Natural Killer T cells (Dellabona J Exp Med 1994) and to their developmental and functional

characterization. iNKT cells are strongly implicated in tumor immunosurveillance and Xxxxxxxx and Xxxxxxxxx have shown that the presence of functional iNKT cells in patients and in animal models correlates with the control of Acute Leukemia, Chronic Lymphocytic Leukemia and Prostate Cancer progression (de Lalla J Immunol 2011, Gorini Blood 2017, Cortesi Cell Rep 2018). Furthermore, this strong anti-tumor function can be achieved also by adoptively transferred iNKT cells and is mediated by their unique capacity to localize to the tumor site and selectively restrict immunosuppressive myelomonocytic cells, while supporting the pro-inflammatory ones (Gorini Blood 2017, Cortesi Cell Rep 2018). These results highlight an indirect anti-tumor function of iNKT cells that nicely complements the direct function exerted by effector T cells derived from the TSCM compartment. Furthermore, they also point to the use of iNKT cells engineered with tumor- specific CARs for the dual targeting of malignant and non-malignant infiltrating immunosuppressive myeloid cells at the same time. Published evidence (Heczey Blood 2014, Xxxx X Clin Invest 2016, Rotolo Cancer Cell 2018) show that human iNKT cells can be efficiently engineered with anti-CD19 or anti-GD2 and targeted against acute leukemia and neuroblastoma, respectively, in humanized animal models. Our preliminary proof-of-concept results confirm that human iNKT cells can be engineered with CD19-CAR and kill as efficiently as T cells CD19- expressing targets in vitro (Figure 1).

Finally, all these approaches are being assessed in a highly sophisticated pre-clinical mouse model set up at OSR by X. Xxxxxxx (see also WP5) based on the humanization of NSG immunodeficient mice that express three transgenic human cytokines Stem Cell Factor (SCF), GM-CSF and IL-3 to support rapid and complete engraftment of human hematopoietic stem cell transplantation (huNSG- SGM3) (Xxxxxxx Xxx Med 2018), which support the development of human myelomonocytic cells that reconstitute the myeloid component of the TME (Xxxxxxx Xxx Med 2018, X. Xxxxxxx unpublished results).

Dr. Xxxxxxx’x lab is also working on strategies to increase the efficacy of CAR-T cell therapy against solid malignancies. In particular, as proof-of-concept using N-glycosylation-defective pancreatic tumor cell lines, generated by knocking-out the expression of the glycosyltransferase Mgat5, they showed that hampering tumor glycosylation resulted in a dramatic increase of tumor targeting by CAR-T cells (Figure 2, X. Xxxxx ASGCT 2019, manuscript in preparation). To exploit this mechanism in order to increase the efficacy of CAR-T cells against solid tumors, X. Xxxxxxx’x lab sought to block tumor N-glycosylation with the glucose/mannose analogue 2-Deoxy-D-glucose (2DG). Similarly to glycosylation knocked-out cells, treatment with 2DG also sensitized tumor cells to recognition by CAR-T cells. Notably, 2DG alone proved to be ineffective, suggesting a synergistic effect with CAR-T cells. In line with in vitro data, mice receiving CAR-T cells highly

benefited from 2DG administration, which conversely was unable to mediate any antitumor effect alone. Interestingly, improved antitumor activity was accompanied by a decrease in the frequency of CAR-T cells co-expressing exhaustion and senescence markers. Thanks to metabolic deregulation (Warburg effect), 2DG is expected to selectively accumulate in cancer cells compared to healthy tissues, supporting the safety of the combined approach. Accordingly, the same doses of 2DG able to enhance tumor recognition by CAR-T cells failed to induce protein de-glycosylation and to increase the elimination of healthy cells, such as keratinocytes.

Figure 2. Combining CAR-T cells with 2DG increases antitumor efficacy.

A. CAR-T cells kill more efficiently glycosylation-defective (mtag5 Knock-Out and WT + tunicamycin) than glycosylation-competent (WT) T3M4 pancreatic tumor cells. B-C. De-glycosylation agent 2DG potentiates tumor cell killing by 00x0.XXX T-cells and XXX XXX-X xxxxx. X. 0XX does not increase the targeting of primary keratinocytes by 44v6 CAR-T cells. E. 2DG improves antitumor activity by 44v6 CAR-T cells in vivo. F. 2DG ameliorates the exhaustion profile of tumor infiltrating CAR-T cells in vivo.

Scientific activities performed at OPBG

Adoptive transfer of CAR-T cells has resulted into less striking effects in solid tumours as compared to B-cell lymphoid malignancies. Although active tumour-mediated immunosuppression may play a role in limiting efficacy, functional changes in T lymphocytes following their ex vivo manipulation may also account for cultured CAR-T cell reduced ability to penetrate stroma-rich solid tumours. We therefore studied the capacity of human in vitro-cultured CAR-T cells to degrade components of the extracellular matrix (ECM). In contrast to freshly isolated T lymphocytes, we found that in vitro-cultured T lymphocytes lack expression of the enzyme heparanase (HPSE) that degrades heparin-sulphate proteoglycans, which are main components of ECM. We found that HPSE mRNA is down-regulated in in vitro-expanded T cells, which may be a consequence of p53 binding to the HPSE gene promoter (Nat Med. 2015 May; 21(5): 524–529). Moreover, as it is well known, solid tumours secrete a variety of chemokines like CXCL5 and CXCR2, which compose the signal path preventing T cells from migrating to the advanced cancer. We have also demonstrated that the immunosuppressive tumour microenvironment (TME) recruits immunosuppressive cells, such as myeloid cells and fibroblasts, which constitute the fibrotic extracellular matrix surrounding the tumour and inhibiting the infiltration of T cells into the tumour.

Scientific activities performed at IRE

Bioinformatics analysis to identify potential innovative targets in TME for CAR T cell therapy.

Tumour-Immune Dysfunction and Exclusion (TIDE), a computational method to model two primary mechanisms of tumour immune evasion, which include the T cell dysfunction and prevention of T cell infiltration in tumours, has been developed (Xxxxx et al 2019). In order to identify the major factors involved in T cell exclusion we started from the CTL-signature (CTL: Cytotoxic T Lymphocytes), described in [TIDE] as the expression of 5 genes representing the CTL activity within a tumour sample (i.e. XXXXX, XXX0, XXXXX, XX0X, XX0X). We extracted from the TCGA Dataset of Lung Adenocarcinoma (TCGA-LUAD) the geometric mean of the CTL signature (CTL-signal) and partitioned the dataset in quartiles accordingly. The patients in the first quartile were defined CTL-low and the ones in the last CTL-high, in order to extract information from the “exceptional infiltrated” vs “exceptional excluded”, differently from the methods usually applied, namely finding a threshold to bipartite the dataset (Figure 3).

Figure 3. Heatmap showing relative expression levels of the CTL signature for 41 patients in TCGA-LUAD (18 CTL-low vs 23 CTL-high). Other >100 patients from the second and third quartile of the CTL distribution were excluded from visualization. Only a subset of patients of the overall dataset (>500 pts) were analysed, based on raw reads counts availability from the Broad Institute Firehose. In the final profiling, all the overall RSEM-normalized dataset will be employed. We proceeded by comparing the expression signature of CTL-excluded (low) and CTL-rich (high). The resulting differentially expressed genes were filtered according to two different strategies: 1. a wide filter in order to capture all the low expressed genes comprising the whole stroma, immune background and the overall microenvironment 2. A strict filter able to capture the candidate biomarkers for immune exclusion. The two analyses returned 5374 and 60 genes, respectively. We are currently validating in our casuistry the subset of 60 genes with a strong modulation (log2FC > 3) and an elevated level of expression in at least one group (CPM > 5), that includes the modulation of Ephrin type-A receptor 3 (EphA3) (log2FC -1.3,pval 0.02), whose expression supports the cancer’s growth and metastatic potential (Offenhäuser et al. 2018).

Identification of targets specific of pro-invasive CAF subtype. In the Unit of Tumour Immunology and Immunotherapy we have a huge collection of fibroblasts derived from both NSCLC and PDAC tumours (CAF), paired with fibroblasts distal (DF) sites as well as normal tissue. CAFs have been characterized being positive for αSMA and FAP expression and for hMENAΔv6, the mesenchymal associated isoform of the cytoskeleton regulatory protein hMENA, which spliced isoform extensively characterized in our group involved in the epithelial to mesenchymal transition in the tumours (Di Modugno et al 2012) and recently in CAF functionality. We have found that, although heterogeneous, hMENAΔv6 expression is higher in CAFs compared to DFs; moreover, our unpublished data indicate a role of hMENAΔv6 in CAF activation as shown by the ability of hMENAΔv6 to increase collagen gel contraction, MMP2 activation and Matrigel invasion of CAFs.

By mass spectrometry-based proteomics analysis (LC-MS/MS) of CAFs with high hMENAΔv6 expression and pro-invasive activity compared to CAF with low expression of hMENA and normal fibroblasts DFs (Figure 3A) we identified a signature related to high hMENAΔv6 expression in pro- invasive CAFs. The signature includes GAS6, the ligand of the AXL receptor, a druggable pathway

involved in cancer cell invasiveness and therapy resistance and EphA3 which we have also found higher in the CTL low with respect to CTL rich dataset (see previous paragraph). The LC-MS/MS data on GAS6 were validated by ELISA and qRT-PCR (Figure 4 B-D). We also demonstrated that CAFs express the GAS6 receptor AXL (Figure 3E).

Figure 4. A signature related to high hMENAΔv6 expression in pro-invasive CAFs includes the ligand of AXL receptor GAS6 and EphA3. A. Heatmap of proteins differentially expressed in the secretoma of pro-invasive CAFs highly expressing hMENAΔv6 (CAF high) with respect to normal fibroblasts (NF) and CAF expressing low hMENAΔv6 (CAF low), as evaluated by LC-MS/MS. B-C Validation of the differential secretion by XXXXX (B) and expression by qRT-PCR (C) of GAS6 in the three subgroups of fibroblasts. D. As a further validation of the data shown in A the depletion of hMENA by siRNA reduces the expression of GAS6 in CAFs as evaluated by QRT-PCR. E. Western blot analysis showing that CAFs from different patients also express the GAS6 receptor AXL.

NSCLC displaying a high stromal FN1 and expressing hMENA/ hMENAΔv6 but negative for the epithelial hMENA11a isoform show organized T and B lymphocytes at the tumour margin and excluded from the tumour nests

Based on our published data describing that the pattern of hMENA isoform expression impacts the ECM composition and NSCLC patient outcome through fibronectin/β1 integrin axis (Di Modugno et al Oncogene 2018), we have analysed by immunohistochemistry the presence and localization of T and B cells in this cohort of primary tumour tissues (Figure 5). Preliminary data indicate that T and B lymphocytes (organized in TLS) are preferentially localized at the periphery of the tumour nests (PT) in the presence of high stromal fibronectin (FN1) expression and low expression of the epithelial hMENA11a on tumour cells. Conversely, tumours expressing high level of hMENA11a and low stromal fibronectin favour the inclusion of T and B cells in the humoral area (not shown). These data suggest that hMENA is at the cross-road between ECM composition and T and B cell homing in the tumour area.

Figure 5. NSCLC primary tumours lacking the hMENA11a isoform in the presence of hMENA/hMENAΔv6 show high stromal FN1 and exclusion of organized B (CD20) and T (CD3) lymphocytes from the tumour core. Correlation between the pattern of hMENA isoform expression and stromal FN1 level with B and T lymphocytes (organized in TLS) at the periphery of the tumour (PT). Consecutive sections of a representative case is reported.

Scientific activities performed at FT/SGER

We are currently cloning different constructs to be further compared in vitro and in vivo. In particular, we are generating single XX0x.XXX, XX00.XXX and XX00.XXX together with a bispecific Dual CAR XX-0x.XXX/XX00.XXX and a variant carrying an IL3 mutant, which limits the binding of IL3 to its receptor CD123 on AML cells, in order to get a suboptimal signal 1.

Since we do not know the binding kinetics of the new scFv for CD146, we are currently cloning several XX000.XXXx carrying different spacers in order to identify the “optimum of distance” that can allow the binding of the CD146 epitope and the subsequent transduction signal upon target antigen encounter.

We have generated XX00.XXX+ T lymphocytes from 4 different samples collected from CLL patients. The polyclonal activation of T cells from all the patients analysed stimulated preferentially CD8+ cells on CD4+ cells, with high percentage of CD8+ CD45RO+ CD62L- T effector memory (TEM) cells. XX00.XXX+ T cells showed an efficient in vitro cytotoxic activity against CLL CD23+ MEC1 cells, specific cytokine release and proliferation upon MEC1-target stimulation.

Scientific activities performed at CRO

We have developed a new immunogenomic approach for the delivery of a tumour-suppressor xxXXX using engineered primary B lymphocytes (LBs) armed with an anti-CD38 antibody (daratumumab). LBs have been chosen because of their ability to synthesize xxXXX, and the ease to arm them with antibody via their unique Fc receptor (CD32) to facilitate targeting of tumour cells.

1) We have reported a system for synthesis and delivery of miRNAs in primary LBs as well as in activated CD8+ T lymphocytes, as a proof of principle for the project proposed.

Moreover, we showed that upon programming, primary LBs secrete xxXXX encapsulated in EV. Experiments show that LBs can be programmed to produce EV with a predetermined xxXXX content enriched >15 times over that of EVs released by cells under physiological conditions. Notably, EVs from programmed LBs protected NSG mice from orthotopic tumours induced by implantation of triple negative human breast cancer cells.

2) We have obtained evidence that miR-21 antagonism decreases MM growth in vivo and attenuates MM BD by reduction of RANKL/OPG ratio, which is one of the main triggers of MM BD. In addition, preliminary data show that miR-21 antagonism impairs T helper 17 (Th17) cell differentiation and function. Since Th17 cells are increased in MM osteolytic disease, miR-21 antagonism impairs TH17 mediated OCL activity.

3) With respect to targeting programmed LBs, we have obtained proof-of-principle evidence that in dynamic flux conditions, LBs incubated with anti-CD38 antibody (Daratumumab) physically interact with MM cells immobilized on coated glass. Once armed, LBs are able to go directly to the tumour microenvironment.

Recent studies show that many tumours, including Epithelial Ovarian Cancer (EOC), have thousands of alternative splicing events, undetectable in normal samples and some of these variants may be considered putative neoantigens.

To identify possible neo-antigens produced by alternative splicing in PT-resistant EOC we have generated several isogenic PT-resistant EOC cells and primary culture from naïve and PT-resistant tumours (See Table 1 and 2).

Table 1. EOC samples collected at CRO (last update at Xxx 2019). All samples have been collected upon written consent of the patient and stored at -80°C. Table reports the number of patients (Pts) and the number of corresponding samples. HGSOC=High Grade Serous OC; Endo= Endometriod; CCC=Clear Cell Carcinoma; MUC=Mucinous Carcinoma; UNK=Unknown. Number of samples already processed for DNA, RNA and protein is indicated. STPC = Short Term Primary

Table 2. Isogenic cellular models of PT-resistant (PT-Res) cells. PT-Res clones have been generated by 20 subsequent cycles of treatment with increasing doses of PT, as described in (5). Table reports the PT IC50 of 4 representative clones that will be used in this study. IC50 has been calculated by exposing the cells to increasing doses of PT for 72 hours. Endo = High Grade Endometriod EOC; CCC = Clear Cell Carcinoma; HGSOC = High Grade Serous EOC. *Likely to be HGSOC;

**Generated since widely used in EOC studies but “likely not

One common feature of PT-resistant EOC cells is their increased ability to adhere to the mesothelium. In search of the mediators of this greater adhesion to mesothelium and laminins (the ECM proteins produced by mesothelial cells) our preliminary data show that the PT-resistant cells have a greater expression of the Integrin α6 (ITGA6) that mediated their increased adhesion to laminins (Fig. 9 and not shown). Also in PT-resistant cells express one particular ITGA6 splicing variant (ITGA6B) and ITGA6 appears to be aberrantly glycosylated in these cells (Fig. 10).

Our recent data show that EOC cells treated with PT activate a survival mechanism associated with greater invasive capacity and clearance of the mesothelium. These evidences suggest an involvement of the remodelling activity of the extracellular matrix. Preliminary proteomics data obtained on conditioned media of parental and PT-resistant cells demonstrated an increased expression and secretion of the TIMP1 a metalloprotease inhibitor in all tested PT-resistant EOC cells. Interestingly our preliminary data suggest that TIMP1 expression is induced by platinum treatment via the activation of MAPK signalling pathway (Fig. 11 and not shown).

Figure 11. TIMP1 is overexpressed in and secreted by PT-Resistant EOC cells. A. qRT- PCR analyses of parental and PT-resistant (MI- res) cell clones in TOV-112D and OVSHAO cells evaluating the expression of TIMP1 mRNAs. B. Western blot analyses of TIMP1 protein expression in the conditioned medium (CM) of the cells described in A. B. Western blot analyses of TIMP1 protein expression in the CM of parental and PT-resistant (MI-res) TOV-112D cells clones treated with the

For a greater clinical translatability of this kind of researches it is necessary to build and characterize appropriate models that can better recapitulate the human pathology. For this reason we are generating PDX models from patients with sensitive platinum and PT-resistant EOC. We have established procedures to engraft PDX in NSG mice and to date 3 PDX from PT-resistant patients are under development in the lab. The study of immune-therapy also requires the availability of manageable immune-competent syngeneic models that could recapitulate the human pathology. To this aim we obtained and are now characterizing the murine ID8 cells modified for the expression of TP53, BRCA1, BRCA2 and PTEN genes by the CRISPR technology. This gene editing technology generated transplantable murine EOC cell lines that recapitulate critical mutations in human high grade EOC. Moreover published data showed that in ID8 cells TP53 knock out increases CCL2 expression and induces marked increases in immunosuppressive myeloid populations within solid tumour, making this model an attractive tool to study intratumoral immune cells penetration and persistence.

To better characterize the immune-infiltrate of mouse and human specimens we have set up and validated (Xxxxxxx et al. in preparation) a multiplex analyses evaluating 4 different antigens (i.e. CD8, CD4, CD68 and CD274) on the same slide (Fig. 12). This technology allow to precisely quantify not only the number and the percentage of immune cells in the tumour mass but also to obtain their spatial organization and precisely identify double or triple positive immune cells.

Figure 12. Validation of a multiplex immune assay. A single tissue slide from EOC patients’ specimens was sequentially immunolabeled with anti-CD4, CD8, CD68 and CD274 (PD-L1) to visualize tumor cells, using the OPAL® system (Perkin Xxxxx). Stained slides were then acquired and analyzed using an Olimpus microscope equipped with a multispectral camera and interfaced with the MANTRA system (Perkin Xxxxx). All staining and analyses procedures are set for both human and mouse tumors and immune cells.

Scientific activities performed at Rete AMORe

Rete AMORe has developed a LLMPP's Lymph2Cx assay that is a digital gene-expression (NanoString) based test for COO assignment in formalin-fixed paraffin-embedded tissue (FFPET. The test is based on the Lymph2Cx gene expression assay which profiles 15 classifier genes and 5 housekeeping genes to compute a linear predictor score (LPS) and determine the COO subtype. The Laboratory of the Haematology Unit of the Xxxxxxx Institute has a consolidated experience in the routine application, in addition to the conventional diagnostic approaches, of the LLMPP's Lymph2Cx test. The application of this test to a series of DLBCL naive, observed at the IRCCS Xxxxxxx, has enabled a more accurate molecular subtyping and prognostic evaluation of these patients. Recently, through the use of computational methods, such as "CIBERSORT" ("Cell type Identification by Estimating Relative Subsets of known RNA Transcripts"), the research group at the IRCCS of Bari has demonstrated the possibility of gaining highly sensitive quantitative and qualitative information, on the cellular composition of TME in DLBCL. This approach is based on a prognostic panel of 45 genes (30 of which pertaining to myofibroblastic, 10 associated with dendritic cells and 5 to CD4 T lymphocytes) capable of defining, on FFPE bioptic tissues and using "Nanostring" technology, the level of infiltration and the functional biology of specific microenvironmental cell populations, correlating them to the degree of response to standard R- CHOP-based immunochemotherapy.

The expertise of the Rete AMORe in the field of gene expression with the Nanostring approach is also documented by a recent preliminary study carried out at the Melanoma, Cancer Immunotherapy and Innovative Therapies Unit of the IRCCS Xxxxxxx. In this analysis, gene expression profiles were evaluated on tumour samples and liquid biopsies obtained from metastatic melanoma patients treated with anti-PD1 or anti-CTLA4 antibodies. The study was conducted exploiting the Nanostring nCounter platform and Pan Cancer IO360 and CP panels for the evaluation of gene and protein expression.

Scientific activities performed at IRST

To test CAR-expressing cells in an in vivo-like melanoma environment, an organotypic skin melanoma culture (OMC), developed in Prof Xxxx Xxxxxx’x Laboratory (Radboudumc, Nijmegen, The Netherlands), will be made available through the project (manuscript under revision). The proposed OMC is built using de-cellularized dermis, obtained by physical decellularization process, as scaffold for stromal cells (fibroblasts and keratinocytes) and tumour cells seeding. Quantitative assessments of the de-cellularized dermis confirmed the lack of cellular (nuclei) and vascular (CD31) components. Concomitantly, histochemical evaluation of dermal markers demonstrated that physical decellularization of the human dermal scaffold did not disrupt its complex extracellular matrix architecture (elastin and collagen fibres) (Figure 13). Preliminary findings have shown that by seeding melanoma cells within the reconstructed skin a growing tumour can be mimicked. Indeed, similar to organoids, these OMC reproduces the spatial distribution of cells, such as fibroblasts and keratinocytes, naturally constituting the skin. The feasibility of culturing different commercial melanoma cell lines within the present system has also been assessed. Evidence that the OMC is effectively suitable to address human host-tumour interactions was collected by studying the behaviour of a subset of circulating natural occurring human dendritic cells (DCs). Moreover, live imaging and multispectral immunohistochemistry end point analyses of fixed OMCs showed that myeloid cells interact with cells in their surrounding niche and engage in cell-to-cell interactions with melanoma cells (Figure 14). We also attested that by deconstructing these in vitro engineered micro tissues each viable population can be recovered, thus also allowing deeper flow cytometry as well as gene-expression analyses (Figure 15). Besides the illustrated use of primary naturally circulating DCs, we assessed the feasibility of culturing T lymphocytes within the OMC, and found that the percentage of live CD3+CD4+ and CD3+CD8+ cells was not affected by the culture conditions. Furthermore, the decellularized human matrix used as scaffold for the OMC allows a better representation of the extracellular matrix (ECM) components and is potentially more suitable for assessing T cell-infiltration mechanisms governed by degradation of ECM elements. Immunohistochemical analysis on FFPE sections belonging to melanoma organotypic cultures or patient-derived melanoma specimens is largely in use in our Institute. In this context, we recently optimized a protocol for Multiplexed Immunohistochemical Consecutive Staining on Single Slides (MICSSS). This approach allows for the construction of multi-parametric panels (up to 10 colours) on a single FFPE tissue slide. We are currently developing 2 large multiplex IHC panels to visualize and characterize myeloid and lymphoid cells in the context of the tumour microenvironment (TME). As an example, Figure 16 shows the feasibility of performing 4 consecutive rounds of staining without losing staining efficiency for CD4 and CD8. This approach will be made available throughout the project for assessing the immune contexture within, but not limited to, human samples or organoid-derived FFPE blocks.

Experimental plan

Task 1: Development of platform to be used across the activity of the WP4

IRST-Task 1.1 Creation of an innovative vector which integrates three transduction modalities and allows allogeneic CAR therapy and interference with CAR T cell exhaustion As first, the vector is designed to allow three independent transduction strategies by harboring sequences recognized by specific enzymes. In particular, the vector will be compatible with a classical lentiviral process, with the Sleeping Beauty transposon (SB) and by site-specific integration mediated by CRISPR / Cas9 co-transfected enzyme. Within this vector, sequences will

also be included that allow interference with TCR exposure, i.e. PEBL anti- - microglobulin expression (i.e. Cas9 mediated cleavage) to minimize both GVHD and rejection risks, respectively. Variants of this vector will also allow us to interfere with the expression of CAR-T inhibitory molecules (i.e. PD-1; LAG3). In order to test the transduction efficiency with various methods, EGFP and/or specific CAR moieties will be allocated as transgenes and analyzed by flow cytometry. T cells transduced with CARs recognizing different antigens will be tested for their ability to penetrate into the humoral mass in 3D tumor cell models and for their exhaustion status.

IRST-Task 1.1 milestones:

1.1.1 Generation of a series of vectors with different combinations of elements compatible with the different transduction systems (Months 1-12)

1.1.2 Introduction of sequences that allow interference with potentially toxic or functional exhausting signals (Months 1-12)

1.1.3 Evaluation of T cell transduction efficiency by the different strategies and of the interference with the expression/exposition of specific proteins (Months 6-12)

1.1.4 Cloning of specific scFvs and evaluation of CAR T penetration in 3D tumor cell models and quantification of CAR T functional exhaustion (Months 12-24)

IRST-Task 1.2 Generation of a technological composite platform for the functional characterization of CAR-expressing cells

In this task three technologies will be combined into a platform to investigate CAR expressing cell functions regarding their interaction with the tumour and its microenvironment, the exhaustion phenotype and CAR cells penetration ability into solid tumours. The platform will be available to all the participating partners.

1. nCounter CAR-T Characterization Panel (Nanostring platform) will be used to profile CAR- expressing cells at the transcriptional level. Besides the 780 built in probes of the panel, an enlarged set of 20 additional probes will be designed to assay specific genes of interest to the partners of the consortium. Within the aims of the current work-package this technology will be exploited for characterizing the activation status, exhaustion profile and other CAR relevant features for CAR engineered cells after manufacturing and in the contest of tumor organotypic cultures.

2. Multiplexed Immunohistochemical Consecutive Staining on Single Slides (MICSSS) approach can be performed to analyze the expression of up to 10 markers on a single FFPE slide through sequential staining cycles. The technology will be instrumental to interrogate the immune contexture and thus being used for the characterization of immunecompetent organotypic cultures. A lymphoid panel (CD45, XX0, XX0, XX0, XX00, XX00, Xxxx0, XX0, XXX0, XXX0, XX00X0, CD27) as well as a myeloid panel (CD45, XX00, XX000, XX00, XXXXX, XX00x, CD209/DC- SIGN, LAG3, PDL1, TIM3) will be designed and tested.

3. RNAscope technology allows the contemporary detection of antigens and RNA by combining IHC and RNA in situ hybridization (ISH) on FFPE samples derived from organotypic cultures or on patient’ biopsies whenever available throughout the project. We will exploit RNAScope to: i) localize CAR cells in the contest of tumor; ii) analyze the expression of tumor derived exhaustive cues; iii) characterize antigen expression and localization; iv) visualize RNAs specific for cytokines/chemokines that are traditionally difficult targets for IHC.

IRST-Task 1.2 milestones:

1.2.1 Implementation of the nCounter CAR-T Characterization Panel (Nanostring) including additional specific probes from the partners of WP4 (Months 1-6)

1.2.2 MICSSS myeloid panel (Months 3-6)

1.2.3 Application of RNAscope technology to organotypic cultures including or not CAR expressing cells (Months 1-6)

1.2.4 Collection of samples from the Consortium and running of the platform (Months 6-24)

CRO Task 1.3 Development of a digital pathology platform

OPAL technology coupled with the MANTRA system will be developed to characterize the tumour immune CAR cell infiltrate as well as tumour microenvironment. In particular, digital pathology analysis will be largely used to precisely quantify the presence of different type of immune cells infiltrating or surrounding the tumours using up to 7 different antigens on the same tissue slides. The great advantage of this approach is that it not only allow to unbiasedly define the number and percentage of immune cells in each area of the specimens but also to map the infiltrating immune cells and their interaction with the tumour cells.

CRO-Task 1.3 milestones:

1.3.1 Application of digital pathology analysis to investigate CAR cell approach optimization (Months 1-24)

IRST-Task 1.4 Use of hNIS as in vivo CAR T display system

IRST envisions to integrate the hNIS gene into its transduction platform in order to monitor and localize CAR transduced cells accurately in vivo both in preclinical models and, in perspective, in the clinical setting.

IRST-Task 1.4 milestones:

2.10.1 Generation of vectors harboring the hNIS gene (Months 12-18)

2.10.2 Analysis of efficacy and in vivo localization of specific CAR-T transduced cells in PDX mouse harboring solid tumors (Months 12-24)

Task 2 Improving Efficacy

OSR-Task 2.1 Assessing the optimization of CAR therapy efficacy in colorectal adenocarcinoma model as well as tumour hepatic metastases by the use of CEA-CAR TSCM and iNKT cell subsets

We will assess the anti-tumor activity of TSCM, iNKT and total T cells redirected against a tumor antigen by transduction with lentiviral vectors containing the CEA-CAR gene or, as control, the genes encoding a TCR specific for HLA-A2-restricted peptide SLLMWITQC from NY- ESO1/LAGE1 (Mastaglio Blood 2017), a tumour associated antigen expressed by >50% colorectal cancer. The redirected effector cell products will be investigated initially in vitro. Moreover, we will assess the anti-tumor activity of CEA-CAR TSCM, iNKT and total T cells in different pre- clinical animal settings, in which human CEA/NY-ESO1-expressing colorectal cancer cells are injected either S.C., or through mesenteric vein to generate liver metastases, into immunodeficient NSG-SGM3 or huNSG-SGM3 mice. By combining the information obtained in these settings, we aim at identifying the critical qualities of the candidate redirected-T cell products that can inform their transfer to patients.

OSR-Task 2.1 milestones:

2.1.1 Generation of CEA-CAR TSCM, iNKT and total T cells. (Months 1-11);

2.1.2 Generation of NY-ESO1-TCR TSCM, iNKT and total T cells. (Months 1-11);

2.1.3 Evaluation in vitro of CEA-CAR TSCM, iNKT and total T cell anti-tumour activity. (Months 3- 6);

2.1.4 Evaluation in vitro of NY-ESO-TCR TSCM, iNKT and total T cell anti-tumour activity. (Months 3-6);

2.1.5 Evaluation of CEA-CAR TSCM, iNKT and total T cell anti-tumour activity in humanized animal models. (Months 5-24);

2.1.6 Evaluation of NY-ESO1-TCR TSCM, iNKT and total T cell anti-tumour activity in humanized animal models. (Months 5-24).

OPBG-Task 2.2 Generation of a CAR construct in combination with the cytokine IL-15 and its high affinity receptor, to increase proliferation following antigenic stimulation.

OPBG has already cloned a specific antigen for an antigen expressed on neuroblastoma tumour cells (i.e. GD2) in a retroviral construct, but also from several adult cancers, including glioblastomas, sarcomas, melanomas and breast carcinomas. OPBG intends to design a cloning capable of also integrating the expression of the human cytokine IL15 in combination with the membrane receptor IL15R. CAR.GD2-IL15 / R will be tested in different tumour models in order to verify the increase in activation, expansion and survival following specific stimulation with the GD2 antigen.

OPBG-Task 2.2 milestones:

2.2.1 Generation of the construct carrying CAR.GD2 in combination with IL15/IL15R (1-6 months)

2.2.2 In vitro study to evaluate efficacy of CAR.GD2/IL15/IL15R cells in comparison to CAR.GD2 cells. (Months 6-12)

2.2.3 One in vivo experiment to evaluate efficacy of CAR.GD2/IL15/IL15R cells in comparison to CAR.GD2 cells (Months 10-24)

OPBG-Task 2.3 Generation of a bi-specific CAR construct able to improve CAR-T activation even in the event of loss of expression of one of the target antigens.

OPBG has already cloned a specific CAR for the CD19 antigen and another specific CAR for the CD22 antigen in two retroviral constructs. Both antigens are expressed by the leukaemia B cells of patients with ALL. Immunotherapy with genetically modified T cells with a CAR.CD19 receptor showed a high clinical efficacy in many patients with Bcp-ALL or B-cell non-Hodgkin lymphoma (B-NHL). Among the main causes of failure, the appearance of CD19-negative clones has been well documented. For this reason, the creation of a bi-specific CAR targeting both the CD19 antigen and the CD22 (antigen expressed in most cases of Bcp-ALL) could reduce the risk of recurrence associated with the loss or down-regulation of one of the two antigens.

OPBG intends to design a cloning capable of integrating a bi-specific CAR into a retroviral construct. In vitro and in vivo models will be generated to validate the use of bi-specific CARs in case of loss of the target antigen.

OPBG-Task 2.3 milestones:

2.3.1 Generation of the construct for the bispecific CAR (Months 1-6)

2.3.2 Generation of leukaemia or lymphoma lines characterized by the loss of CD19 or CD22 (Months 6-12)

2.3.3 In vitro study to evaluate the activity of the bispecific CAR-T cells (Months 12-18)

2.3.3 Beginning of the in vivo study to assess the bispecific CAR-T cells activity (Months 18-24)

OPBG-Task 2.4 Generation of a CAR construct with an innovative co-stimulation to improve cell in vivo persistence.

It has been reported that the presence of CD28 domains, as well as of the ITEMs of CD3 in the CAR molecule, rapidly induces a state of "exhaustion" of transduced T cells, characterized by a high PD-1 expression, poor persistence in vivo and insufficient antitumor efficacy. OPBG has already cloned a specific CAR for the GD2 antigen that is characterized by a cytotoxic signal transduction relative to the DAP12 and KIR molecule. The association between the

intracytoplasmic domains of KIR activators and DAP12 regulates the activation signal in cells of the innate immune system (NK and γδT). OPBG aims at studying the activity of this new CAR molecule in experimental models in vitro (2D and 3D) and in vivo (xenogeneic models of immunodeficient mice).

OPBG-Task 2.4 milestones:

2.4.1 End of the in vitro study to evaluate activity of CAR.KIR T cells (Months 1-6)

2.4.2 End of the in vivo study of the activity of CAR.KIR T cells (Months 6-18)

2.4.3 Evaluation of subsequent optimizations for CAR.KIR molecule (Months 12-24)

CRO- Task 2.5 Optimization of CAR T cell approach in MM

This task will be devoted to anti-CD138 CAR-T production by transfecting T cells of MM patients (hereafter hCD138-CAR-T). The hCD138-CAR-T production will take advantage of the close collaboration with research groups known to have a well-recognized expertise in the field.

In particular, the production of three anti-CD138 antibodies will be requested to a private company. The best scFv will be used to clone second generation CAR. Lentiviral vector will be generated to genetically modify T cells from MM patients with the CAR construct. The killing capability of hCD138-CAR-T will be evaluated both in vitro and in vivo against multiple myeloma cell lines and primary MM cells. Cytokine production and expression of molecules of immunosuppression, adhesion and migration, will be evaluated after CAR T activation. For in vivo model, a xenograft model will be established in NOD-SCID-IL2R null (NSG) mice by using MM-lines (MM-line mice) or primary-derived MM cells (PDX-MM mice). In particular, we will develop a humanized mouse model of MM (NSG) to reproduce the human BM microenvironment. After this initial set- up, we will optimize the system by the use of primary human LBs engineered to express and release in culture medium sncRNAs (anti-miR-21 and miR-30c-2 and/or miR-214 and miR-335) to target miR-21, XBP1 and SOX-4 in MM cells, respectively. The use of engineered LBs will be tested on humanized mouse model of MM (NSG) to reproduce the human BM microenvironment. Mice will be grafted with patient-derived MM cells (or with selected MM cell lines). Engineered LBs loaded or not-loaded with anti-CD38 antibody will then be implanted into bone chips of mice grafted with patient-derived MM cells. We will consider vectors "coding" for four anti-miR and miR elements or a combination of them (anti-miR-21 and miR-30c-2 and/or miR-214 and miR-335). MM tumour growth will be monitored to assess the effect of engineered LBs miR-based therapy coupled with anti CD38 LBs loading. In selected experiments, hCD138-CAR-T cells will be injected in the mouse model. After sacrifice, retrieved synthetic/human/rib bone chips will be evaluated to assess impairment of osteolytic activity by miR-based therapy and the putative synergistic effect of engineered LBs with CAR T cells.

CRO-Task 2.5 milestones:

2.5.1 Generation of the anti-CD138 CAR construct (Months 1-4);

2.5.2 hCD138-CAR-T production and characterization (Months 5-12);

2.5.3 In-vitro killing efficacy of hCD138-CAR-T (Months 9-12).

2.5.4 hCD138-CAR-T/MM cell/tumour microenvironment interaction (Months 13-20);

2.5.5 Establishment of xenograft MM mouse model (Months 13-21);

2.5.6 In vivo killing efficacy of hCD138-CAR-T (Months 18-24).

2.5.7 Generation of plasmid DNA vector (pCMV-MIR) to express sncRNAs (Months 1-13)

2.5.8 Generation of pCMV-MIR encoding both the anti-sense RNA and the miRNAs. (Months 4- 14)

2.5.9 Establish the best method to engineer primary human LBs in order to induce the synthesis and release of sncRNAs in culture medium. (Months 4-18)

2.5.10 Functional evaluation of engineer primary human LBs in MM cells (Months 7-18)

2.5.11 Development of a mouse model of MM (NSG) to reproduce the human BM microenvironment. (Months 10-24)

2.5.12 Analysis of infiltration of bone chips by LBs in mice treated with anti-CD38 antibody (Months 12-24)

2.5.13 Assess the therapeutic potential of antibody loading on engineered LBs in combination with CAR T (Months 14-24)

CRO- Task 2.6 Optimization of CAR in epithelial Ovarian Cancer (EOC) toward the validation of novel target antigens

Based on our preliminary data, we aim to identify novel variants of the integrin ITGA6 specifically expressed on PT-resistant EOC. We will study the mechanisms that regulate the different expression, splicing and glycosylation of ITGA6 in order to develop specific CAR constructs designed to target PT-resistant cells with high metastatic capacity. Interestingly, our preliminary data using a blocking Ab directed against ITGA6 (i.e. GoH3) showed that it completely prevents the adhesion of PT-resistant cells to laminins, suggesting that ITGA6 could represent a useful antigen for CAR targeting of PT-resistant EOC preventing their adhesion to mesothelial cells and therefore their spreading in the abdominal cavity.

Finally, the expression of the specific isoforms identified for CAR T targeting will be evaluated in primary EOC available in our biobank (see Table 1) by IHC and/or qRT-PCR approaches and then related to response to PT of the patients.

CRO-Task 2.6 milestones:

2.6.1 Definition of the role of ITGA6 in the generation of PT-resistant phenotype (Months 1-12)

2.6.2 Role of ITGA6 splicing variants and its glycosylation as possible antigens for CAR development. (Months 8-20)

IRST-Task 2.7 Cell–therapy based on SSTR induction in Pancreatic Neuroendocrine Neoplasia

Neuroendocrine tumours from gastro-entero-pancreatic tract (GEP-NETs) over-express somatostatin receptors (SSTRs) 2, 3 and 5. SSTR subtypes expression shows marked heterogeneity and differences in tumour sites and differentiation. Latest diagnostic and therapeutic tools, such as PRRT are radiolabelled somatostatin analogues targeting SSTRs. Although GEP – NETs show SSTRs overexpression, about 30% of patients are refractory to PRRT suggesting intratumoral heterogeneity of expression. We have identified a signature of 3 circulating miRNAs that strongly correlates with the metabolic status of pancreatic NETs (PET positivity) and, especially for one of them, with lower expression of SSTR-2 by tumour cells. These observations lead to the hypothesis that the miRNAs are directly involved in SSTR expression interference that could be restored by delivering competitive molecules (blockmirs/antagomirs) through a paracrine mechanism in Trans. To this aim we will first clone blockmirs and/or antagomirs for the microRNAs previously identified into two different DNA vectors to induce transduced B cells to express and secrete both type of molecules. We will transduce primary human B cells to express blockmirs and/or antagomirs for the microRNAs previously identified with the idea that those molecules released by the cells can affect neighbouring tumour cells properties (in particular SSTR expression). Transduced B cells will be co-cultured with NET cell lines to test their ability to interfere with SSTR regulation (overexpression is expected) and survival of NET cells.

IRST-Task 2.7 milestones:

2.7.1 Generation of DNA vectors for the different antagomir/blockmir of interest (Months 1-6)

2.7.2 Human B cell collection, transduction and arming with antibodies towards SSTR2 and their characterization in vitro (Months 7-18)

2.7.3 Binding, killing and efficacy evaluation of transduced and antibody armed B cells on co- cultured tumour cell lines (e.g. NT-3 PNETs cell lines). (Months 13-24)

FT/SGER-Task 2.8 Obtaining the bispecific dual CAR XX-0x.XXX/XX00.XXX for the treatment of acute myeloid leukaemia (LMA).

We propose to investigate a double targeting model (with signalling in trans) to improve the selectivity for CD123+ / CD33+ cells using a first generation anti-CD123 CAR (XX-0x.XXX), combined with an anti-CD33 co-stimulating receptor (CCR) without activation signalling domains. The Dual CAR XX-0x.XXX/XX00.XXX functionality will be tested by in vitro assays against AML primary cell lines and primary blasts by cytotoxic activity, cytokine production and proliferation capacity in response to the target. Following in vitro characterization, we will study the functional properties of Dual CAR CIK cells compared to CIK cells expressing single CARs (XX0X.XXX and XX00.XXX) through in vivo models of NSG mice, xenotransplanted with AML cells. At the end of the experiment, the mice will be sacrificed, the tissues collected and the remaining cells analysed for the presence of tumour cells and CIK by flow cytometry.

FT/SGER -Task 2.8 milestones:

2.8.1 Obtaining the bi-specific CAR XX-0x.XXX/XX00.XXX (Months 1-6)

2.8.2 In vitro characterization of the efficacy and safety profiles of CIK Dual CAR cells IL- 0x.XXX/XX00.XXX. (Months 1-10)

2.8.3 Characterize the efficacy and safety profiles in vivo of Dual CAR CIK cells. (Months 8-24)

Task 3 Modulation of immunosuppressive elements along with targeting and overcoming the inhibition exerted by TME

OSR-Task 3.1 Increase in the function of CEA-CAR TSCM and iNKT cells using CRISPR/Cas9 technology to eliminate the expression of immunosuppressive molecules induced in the metastatic microenvironment.

Preliminary analysis of T lymphocytes infiltrating liver colorectal cancer metastases (X. Xxxxxx, unpublished results) showed the expression of several immune suppressor molecules (including checkpoint inhibitors), defining a series of potential targets for elimination by genetic recombination with CRISPR/Cas9. In a proof-of-concept approach, the action of CEA-CAR TSCM or iNKT cells genetically manipulated to eliminate the expression of the PD-1 suppressor molecules, will be analysed in the humanized xenograft model generated by injecting colorectal carcinoma cells engineered to express PD-L1 into NSG-SGM3 or huNSG-SGM3.

OSR-Task 3.1 milestones:

3.1.1. Generation of CEA+ COLO-205 or SW620 colorectal cancer cells expressing constitutive PD-L1. (Months 6-7);

3.1.2 Production of second-generation CEA-CAR TSCM and iNKT cells by selective deletion of PD- 1 gene. (Months 6-12);

3.1.3. Evaluation of PD-1-edited CEA-CAR TSCM and iNKT cell anti-tumour activity in animal models. (Months 8-24);

IRE- Task 3.2 Identification of novel TME-derived targets suitable for CAR T cell therapy

To identify TME-derived targets we posit that tumours with low T cell infiltration express high levels of molecules involved in the physical barrier generated by the presence of specific stromal

cells. To this aim we will use different approaches including: i. bioinformatics tools; ii. An in depth characterization of CAFs isolated from tumour tissues of NSCLC patients; iii. Preclinical (3D and murine) models.

Our preliminary bioinformatics evidence will be extended taking advantage of single-cell RNA-seq. These single cell analyses will allow us to query both specific Tumour-associated or Stroma- associated signatures enabling the identification of distinctive features of tumour stroma as targets for novel therapies. Among the cellular components participating in the generation of the physical barrier hampering CAR T cell trafficking to the tumour sites, CAFs are crucial players. CAFs are heterogeneous but more genetically stable than cancer cells and consequently less prone to antigen escape and generally conserved among different tumours. Therefore therapies targeting CAF may have the potential to be effective in a broad spectrum of tumours.

We will validate the role of CAF in tumour T cell exclusion, by analysing in immunohistochemistry the expression of EphA3, Axl and TGFβ isoforms in the stromal compartment of NSCLC primary tumours already evaluated by our group for T and B cell infiltration. Once the new technology NanoString® GeoMxTM digital spatial profiling (DSP) will be available in the Consortium or thanks to our collaboration with Nanostring, we will analyse the expression and spatial distribution of T cells with respect to stromal cells overexpressing the selected molecules. We will use CAFs overexpressing EphA3, Axl or TGFβ to set-up in vitro 3D bioprint cultures of tumour and CAF embedded into pulmonary ECM. We will measure matrix stiffness by Brillouin microscopy, in collaboration with the Italian Institute of Technology in Rome, in 3D co-cultures of tumour cells with NF which we expect to express low level of hMENA, EphA3, Axl and TGFβ isoforms and CAFs highly expressing these molecules. We will then evaluate the effect of EphA3, Axl and TGFβ isoforms specific antibodies on the matrix stiffness. To validate the contribution of this CAF subtype in T cell exclusion, we will study the capability of T cells to penetrate 3D co-cultures depending on the expression of these molecules. Moreover, monoclonal antibodies against the targets described will be generated against the selected factors, by other members of the ACC network and/or outsourced to specialized companies. These novel antibodies will be used in combination with already available CAR T or their sequences employed for the design of innovative bi-specific CAR T able to target simultaneously tumour cells and the TME. The effect of these novel constructs will be tested in preclinical models available in both our Institution and in the ACC consortium.

IRE-Task 3.2 milestones:

3.2.1 Identification of molecules involved in the process of T cell exclusion from the tumour site. (Months 1-12)

3.2.2 Characterization of EphA3, Axl and TGFβ isoform expression in CAFs and correlation with CAF functionality. (Months 1-12)

3.2.3 Analysis of the presence and spatial distribution of T cells with respect to stromal cells overexpressing the selected molecules in primary tumours. Characterization of the role of EphA3, Axl or TGFβ isoforms in the preclinical model of 3D co-cultures in terms of CAF-mediated ECM composition, stiffness and T cell infiltration. (Months 12-24)

3.2.4 Generation of new monoclonal antibodies against targets of the tumour microenvironment. (Months 12-24)

OPBG-Task 3.3 Development of strategies for the optimization of CAR-T cell penetration in solid tumours.

In order to optimize the CAR T-cell approach in solid tumours, it is necessary to consider that an important component in neoplastic tissues is represented by the extracellular matrix, composed of a series of elements that physically limit the penetration of CAR-T cells. OPBG has already evaluated the expression of a series of human metalloproteinases (MMPs), which could synergize with CAR

constructs in neoplastic systems related to solid tumours. In the present task, OPBG aims at developing in vitro models (2D and 3D) to test the experimental hypothesis.

OPBG-Task 3.3 milestones:

3.3.1 Generation of constructs containing MMP sequences (Months 0-12)

3.3.2 End of the in vitro study of the CAR-T cells/ MMP activity (Months 6-18)

3.3.3 Beginning of the in vivo study of the CAR-T cells/ MMP activity (Months 18-24)

FT/SGER -Task 3.4 Generation of a bispecific CAR directed against the cells of the leukemic niche.

We propose a strategy based on the use of CIK cells engineered with bi-specific CARs, with a specificity directed against the leukemic cell and the other against mesenchymal stromal cells (MSC) of the leukemic niche. The idea will be tested in vivo using the humanized medullary niche model that our group has developed. We intend to clone the scFv sequences CD146 and CD33 within a previously described CAR construct. While the proof of principle will be obtained by using CD146 (MSC) and CD33 (LMA), we will focus on the identification of a surface target selectively expressed on the stromal cells of the malignant niche. CD146 / XX00.XXX CIK cells will be generated by nucleofection using a non-viral platform optimized by our group. We will evaluate the effector functions of CD146 / XX00.XXX CIK cells in vitro against MSCs and leukemic cell lines and LMA blasts in comparison with our previously validated XX00.XXX CIK cells. Specifically, we plan to carry out in vitro tests such as cytotoxicity assays, evaluation of cytokine production and proliferation tests. To demonstrate the in vivo efficacy of CD146 / XX00.XXX CIK cells, these will be re-infused intravenously in an immunodeficient mouse model, in which we will have previously transplanted humanized ossicles and injected leukemic cell lines and / or primary cells. Flow cytometry, immunohistochemistry and molecular analyses will be used to evaluate the specificity of this system against human CD146 + MSCs in the ossicle and their effect on leukaemia.

FT/SGER-Task 3.4 milestones:

3.4.1 Obtaining bi-specific CAR directed against MSC CD146 + cells and CD33 + leukemic cells (Months 1-6).

3.4.2 Evaluation of in vitro efficacy of CD146 / XX00.XXX CIK cells. (Months 1-10)

3.4.3 Demonstration of in vivo efficacy and leukaemia niche targeting. (Months 8-24)

CRO- Task 3.5 Evaluation of the role of metalloproteases activity in the modulation of immune infiltration in ovarian tumours.

By an unbiased proteomic approach we have identified TIMP1 (TIMP metallopeptidase inhibitor 1) a natural inhibitors of the matrix metalloproteinases (MMPs), as a molecule commonly overexpressed and actively secreted by PT-resistant cells but not by their parental counterpart. TIMP1 expression has been linked to platinum resistance and its high serum levels predict shorter overall survival of EOC patients. Based on these data, we will study the role of TIMP1 in the modulation of the penetration and persistence of immune cells in EOC tumour masses. We will assay the role of TIMP1 in vivo using the available ID8 syngeneic model. In particular, we will modify TIMP1 expression in ID8-TP53NULL cells that form tumours with higher infiltration of immunosuppressive myeloid cell populations and by using the OPAL technology coupled with the MANTRA system, we will fully characterize the tumour immune infiltrate in mice injected with ID8-TP53NULL cells modified for TIMP1 or treated with anti-TIMP1blocking Ab. Finally, thanks to the availability of a large number of plasma samples from 356 EOC patients already assayed for their immune infiltration we will verify if TIMP1 levels in the tissue or in the circulation correlates with the immune infiltration and the response to therapy.

CRO-Task 3.5 milestones:

3.5.1 Definition of the role of TIMP1 in the generation of PT-resistant phenotype (Months 1-12)

3.5.2 Role of TIMP1 in immune cells tumour infiltration in EOC. (Months 8-18)

3.5.3 Role of TIMP1 in immune cells tumour infiltration in EOC (Months 10-24)

3.5.4 Characterization of different in vivo models to study the efficacy of CARs in EOC (Months 1- 14)

3.5.5 Characterization of different in vivo models to study the efficacy of CARs in EOC (Syngeneic ID8 Cells) (Months 1-20)

3.5.6 Characterization of different in vivo models to study the efficacy of CARs in EOC (PT- resistant PDX) (Months 1-24)

Rete AMORe-Task 3.6 Identification of molecular targets expressed in the tumour microenvironment of Melanoma and DLBCL

On tumours explanted from mouse models of melanoma, available in the Network, the nCounter system technology will be used applying the panel "nCounter®PanCancer Mouse IO360", in order to analyse the pathways most involved in the tumorigenesis process: angiogenesis, remodelling of the extracellular matrix (ECM), epithelial-mesenchymal transition (EMT), metastatization, etc. The panel includes 770 genes and combines vital components involved in the complex interaction between tumour, microenvironment and anti-tumour immune response, allowing a multi-faceted characterization of disease biology and evasion mechanisms. The study will also include a retrospective analysis of tissue samples from patients with metastatic melanoma (FFPE tissues) obtained from archives of the Pathology Unit at IRCCS Xxxxxxx. The RNA extracted from the slides will be used for analysis at the Nanostring with the panel "PanCancer Human IO 360". Similarly, archival samples and those from DLBCL patients undergoing CAR-T-based treatments, as also related to the COO, will be collected and analysed through the nCounter PanCancer Immune Profiling Panel, to profile TILs from lymphoma microenvironment.

Rete AMORe-Task 3.6 milestones:

3.6.1 Collection of biological samples obtained from treated mouse models, processed according to standardized methods (SOP) (Months 1-18).

3.6.2 Collection and selection of melanoma metastases from patients operated on at the IRCCS Xxxxxxx (Months 1-8).

3.6.3 Data processing and analysis for the identification of new molecular markers of the tumour microenvironment, tumorigenesis markers and components of the extracellular matrix (Months 9- 18).

TASK 4 Deep characterization of the immunosystem for the retrieval of the relevant features to optimize CAR T cell approach

OPBG-Task 4.1 Development of post-infusion immunological monitoring strategies of CAR-T cells.

In the view of the generation of optimal CAR T cell modifications to increase activation, in vivo expansion and tissue penetration, OPBG will deeply characterize tumour microenvironment and long lasting persisting CAR T cells in infused patients. In particular, two Phase II clinical trials are ongoing in OPBG. The first one for treatment with CAR-T cells of patients with neuroblastoma (OPBG-CAR.GD2-001), whereas the second one for the treatment of B-ALL/NHL (OPBG- CAR.CD19-001). OPBG has already developed a flow-cytometry platform for immuno-monitoring of enrolled patients, designed to study the emergence of suppressive myeloid populations, regulatory T cells and expression profiles on the CAR-T cell surface of markers associated with potential exhaustion function of CAR-T cells. In this TASK, OPBG aims to correlate the outcome

of the treated patients with the immunomonitoring data that will be collected during the study. Furthermore, in order to highlight new markers, OPBG will apply single cell analysis platforms based on the association of membrane and RNAseq markers.

OPBG-Task 4.1 milestones:

4.1.1 Immune-monitoring analysis of at least 15 patients enrolled in the OPBG-CAR.GD2-001 study (Months 1-24)

4.1.2 Immune-monitoring analysis of at least 15 patients enrolled in the OPBG-CAR.19-001 study (Months 1-24)

4.1.3 Statistical evaluation of the biological factors that correlate with the clinical progress of the patients included in the study (Months 18-24)

4.1.4 Development of "single cell" platforms designed to deepen the characterization of patients treated with CAR T during their follow-up (Months 12-24)

Rete AMORe-Task 4.2 Analysis of the relationships between the Cell-of-Origin (COO), the tumour microenvironment and the response to CAR-T cell therapy (Yescarta, Kymriah) in patients with DLBCL

In this part of the study we plan to collect, in collaboration with the working groups of the Rete AMORe, and with the other IRCCSs involved in the project, fresh, frozen or FFPE samples from DLBCL patients who received registered CAR-T cells products (Yescarta; Kymriah) in a real-life setting and/or clinical studies. Tumour samples will undergo nanostring-based COO definition and results correlated with most relevant clinical outcomes, including response duration and the PR / CR conversion rate, also in the context of the C-MYC/BCL-2 gene status (FISH) and expression of immune checkpoints (PDL-1, CD47, etc.) in tumour cells. Results will be finally correlated with expansion patterns and temporal modifications of CAR-T cells, which will be studied through a combined approach of NGS, Nanostring and multiparametric flow cytometry. In this way, it will be possible to have a stringent dataset on persistence of immunologic memory cells from infused CAR-T products. In addition, the NanoString approach will be used to characterize T cells in each step of the CAR-T workflow using the nCounter Human CAR-T Characterization Panel, allowing to understand the mechanisms that influence CAR-T cell performance. 780-gene expression panel incorporates content to measure eight essential components of CAR-T cell biology including T-cell activation, metabolism, exhaustion, and receptor diversity with optional customization for measuring transgene expression with NanoString’s Protein Barcoding Service. A different set of studies will be conducted through application of a prognostic panel of 45 genes identified by the IRCCS of Bari through a "CIBERSORT" strategy. Using the same collected material it will be therefore possible to define, on pre-CAR-T treatment biopsy tissues and, using Nanostring technology, the level of infiltration and the functional architecture of specific microenvironmental cell populations, correlating them to the COO and to the mid- and long-term degree and quality of response to CAR-T therapies in patients with DLBCL.

Rete AMORe-Task 4.2 milestones:

4.2.1 Collection of biological samples obtained from patients with DLBCL candidates for CAR-T cell therapy in centres related to WP and Rete AMORe. (Months 1-12)

4.2.2 Analysis and identification of the COO of the collected samples and relative quantification of the different cellular, immune and stromal populations, infiltrating the lymph node microenvironment of patients with DLBCL that will be treated with CAR-T. (Months 6-18)

4.2.3 Definition of the main mechanisms of interaction between cellular populations of the microenvironment (selected on the basis of computational analysis) and DLBCL tumour cells, for the identification of new antigenic and therapeutic targets. (Months 12-20)

4.2.4 Processing and analysis of data obtained in the first two design phases and correlation with clinical data, in order to validate its prognostic / predictive power. (Months 20-24)

IRST- Task 4.3 Evaluation of exhaustion phenotype and tumour microenvironment-mediated immune suppression of CAR-expressing cells using in vitro preclinical 3D organotypic tumour cultures.

The exhaustion phenotype and the tumour microenvironment (TME)-mediated immune suppression of CAR- expressing cells will be analysed taking advantage of the set 3D organotypic tumour cultures. The CAR-cell related exhaustion signature will be derived from the gene expression data collected through the Nanostring platform_nCounter CAR-T Characterization Panel.

To recapitulate an immunosuppressive-like TME, myeloid cells (e.g. monocytes) will be injected in the OMC and the interaction of CAR-expressing cells with myeloid cells will be assessed by sequential IHC (MICSSS) and retrieved CAR cells analysed using the Nanostring platform to assess their exhaustion profile. Additionally, for one selected autologous CAR T cell-melanoma cell 3D culture, a proof-of-principle single cell transcriptomic analysis (10X Genomics) of tumour (CD45- negative) and CAR T cells will be performed in order to evaluate mechanisms of tumour escape from treatment. This will identify crucial elements responsible for the ineffective clearance of tumour cells by the effector cells and changes in the phenotypic heterogeneity of primary melanoma cells upon CAR-T cells therapy.

IRST-Task 4.3 milestones:

4.3.1 OMC containing myeloid cells characterized (Months 6-9)

4.3.2 Analysis of the exhausted phenotype of CAR expressing cells sorted from myeloid cell- conditioned OMC performed (at least from 3 samples) and compared to results obtained in WP3 (Task 3.b.iii) (Months 12-16)

4.3.3 Single cell expression analysis performed from one autologous CAR T cell/tumour cell assessed within primary organoid culture (Months 15-24)

4.3.4 Melanoma heterogeneity portrayed upon CAR T cell co-culture (Months 15-24)

IRST- Task 4.4 Study of T cell exhaustion and of potential mechanisms of T cell costimulation in acute myeloid leukaemia (AML)

T cells will be isolated from AML patients at diagnosis (pre-treatment, n = 10) by negative immunomagnetic selection and analysed by single-cell RNA sequencing. Dimensional reduction analysis will allow to define the cellular subtypes according to the transcriptional profile together with their frequency. The heterogeneity of the T-cell compartment will be investigated by unsupervised clustering approaches. In addition, signatures of T cell phenotype, exhaustion/pre- exhaustion, activation and persistence, obtained from gene sets available through the literature (e.g. MSigDB and nanostring CAR-T panel) will be applied to the whole population and its subtypes. This analysis will identify new potential inhibitory/costimulatory molecules and/or their combinations to be functionally validated.

IRST-Task 4.4 milestones:

4.4.1 Patient enrolment (Months 1-15)

4.4.2 Single cell expression profile of T cell subpopulations in AML patients at diagnosis (M6- M18)

4.4.3 Data analysis (Months 12-21)

4.4.4 Signature identification (Months 18-24)

Task 5 Development of novel therapeutical synergy to optimize CAR cell approaches

FT/SGER-Task 5.1 In vivo evaluation of XX00.XXX T cells derived from CLL patients in combination with lenalidomide.

We will explore the combinatorial effects of lenalidomide with patient-derived T cells redirected with a CAR specific for the CD23 molecule, a tumour-associated antigen (TAA) over-expressed by the CLL. We aim to validate in vivo the efficacy and functionality of patient-derived T cells transduced with an anti-CD23 CAR, in collaboration with the group of Xxxx. Xxxxx Xxxx (San Xxxxxxxx Hospital, Milan) and Dr. Xxxxxxx Xxxxxxxxxxx (MD Xxxxxxxx, Houston, USA). XX00.XXX T cells will be generated, functionally tested by in vitro cytotoxicity assays and transferred to Rag2

- / - γc - / - mice transplanted with the line of CLL MEC-1, a mouse model previously developed by our collaborators. Different conditions will be tested and compared, including PBS, non-transduced T cells (NT) or XX00.XXX T cells alone or in combination with lenalidomide or IL-2. Efficacy will also be evaluated in terms of improved survival of the in vivo combination of CAR-based immunotherapy with lenalidomide. Treatment with lenalidomide has been shown to block LLC- induced actin dysfunction in T-cell immunological synapses and to reduce the expression of CLL- inhibitory ligands and their T-cell receptors. To see if the addition of lenalidomide, in particular in T cells redirected by CAR, affects the expression of molecules of the supramolecular activation (SMAC) clusters, the expression levels of the main components of the immunological synapse will be tested, the expression levels of T cell inhibitors and molecules normally altered on contact with LLC cells. For this purpose, healthy T cells, NT T cells and XX00.XXX will be co-cultured for 48 hours in vitro with normal B cells or CL cells (such as the MEC-1 cell line, which exactly recapitulates the properties of primary cells) in the presence or absence of optimized doses of lenalidomide.

FT/SGER-Task 5.1 milestones:

5.1.1 Evaluation of tissue homing of XX00.XXX T cells in association with lenalidomide treatment. (Months 8-24)

5.1.2 Functional analysis by immunohistochemistry and flow cytometry. (Months 8-24)

5.1.3 In vitro characterization of immunological synapses of anti-CD23 T cells in combination with lenalidomide. (Months 1-10)

OPBG-Task 5.2 Generation of a high-throughput screening platform for the evaluation of the synergistic effect between CAR T and drug libraries.

In order to enhance the activity of CAR-T cells, OPBG has already developed a "high-throughput" screening platform of drug libraries approved by the FDA and EMA, as well as drugs in late pre- clinical development.

The platform allows evaluating the antineoplastic activity of thousands of drugs in an extremely repeatable way. In the present TASK, OPBG will aim at investigating the same platform in association with the CARs. In particular, it will be evaluated not only the anti-neoplastic activity of the drugs, but also their synergistic activity with the CAR-T cells, or if a specific drug negatively affects the proliferation and activation of the CAR-T cells themselves.

OPBG-Task 5.2 milestones:

5.2.1 Definition of the high-throughput screening platform to evaluate drug libraries in association with CAR-T cells (Months 1-12)

5.2.2 In vivo validation of pharmacological targets in association with CAR-T cells (Months 12-24)

IRE-Task 5.3 Oncolytic vectors as gateways for CAR-Ts in solid tumours Assessment of tumour stromal- and cancer-cell targeting by oncolytic viruses

This task aims at subverting the hostile TME by using oncolytic viruses as immunostimulators favouring T cell trafficking and activation in the tumour (Kellish et al. 2019). Through a collaboration with Xxxx. Xxxxxxx, Head of Laboratory of Immunovirotherapy, University of Helsinki we have access to oncolytic vectors based on different viruses such as adenovirus, vaccinia virus, herpes virus and semliki forest virus. These virus display a different tropism and can be engineered to increase their tropism against tumour or stromal cells, sparing normal cells. We will assess and compare tropism of a panel of oncolytic vectors towards different malignant tumour and stromal cells of both human and murine origin. We will evaluate the possibility to improve tropism of OVs towards stromal cells such as CAFs. This also based on recent data demonstrating that OVs can be directed towards cancer-associated fibroblasts, which overexpress fibroblast activation protein-α (FAP), leading to enhanced antitumor activity (de Sostoa et al. 2019; Xxxxxxxx et al. 2018). We will characterize the infectivity, replication, and gene expression profile of different OV constructs in primary human CAFs of NSCLC...

Assessing the combinatory strategy of oncolytic virotherapy and CAR T cell therapy in a syngeneic murine orthotopic model of head and neck cancer

We will test the hypothesis that pre-treating tumours with OVs, leading to local oncolysis, would enhance the antitumor activity of CAR-T cell therapy in a syngeneic murine orthotopic model of head and neck cancer, previously established in our Institute (AT-84/AT-84 E7; Xxxxxx et al 2015). This murine model allows the direct intratumoral administration of oncolytic vectors and CAR T cells. The efficacy of the treatment will be monitored by the Perkin Xxxxx 3D optical imaging IVIS platform available in our Institute which precisely and non-invasively detects the location of primary and metastatic tumours.

Both human papillomavirus (HPV) - and HPV+ HNSCC express epidermal growth factor receptor (EGFR), making the tumour cells amenable for XXXX.XXX T cell recognition and killing (Xxxxxxxx Xxxx et al. 2017). We will obtain CAR vectors targeting EGFR to transduce T cells from the Consortium. The functionality of XXXX.XXX T will be tested in vitro by assessment of cytotoxic activity, production of cytokines and cell proliferation assay on tumour cells expressing EGFR. Then we will evaluate the effects of the treatment with OVs and EGFR.CAR-T on tumour growth and T cells infiltration.

IRE-Task 5.3 milestones:

5.3.1 Definition and modulation of tropism of a panel of oncolytic vectors towards different malignant tumour and stromal cells. (Months 1-12)

5.3.2 Analysis of immune cell infiltration in the HNSCC murine model and development of oncolytic vectors to be used in combination with the CAR- T treatment. (Months 1-12)

5.3.3 Studies of the combinatory strategy of oncolytic virotherapy and CAR T cell therapy in the murine model (Months 12-24)

Contingency plan

Tasks

Risk

Proposed risk-mitigation measures

Task 1.1

The generation of complex vectors harboring several modules could be troublesome if specific features are not compatible (i.e. transposition sequences with LTRs or sequences for site specific recombination).

In order to overcome these potential problems it will be possible to switch to a second transposition system (PB) or to change the Cas9 cleavage site (by changing the specific guide RNA) and accordingly the flanking regions in the vector to be recombined by DNA homology repair

within the T cell genome.

Task 1.2

We may not be able to come up with a full 10- color multiplex IHC panel for each proposed

subpopulation panel.

More than one antibody will be tested for each marker for repeated round of stainings to attest the

relative efficiency throughout the whole protocol.

		The definitive panel of antibodies will be then selected based on their specificity and resistance to reiterative rounds of stripping. Alternatively, in the absence of good commercially antibodies for a given marker, RNA in situ hybridization (ISH) technology will be used to directly visualize the endogenous RNA of membrane localized proteins.
Task 1.4	The expression level of human NIS should be high enough to grant sufficient iodine/technetium uptake from transduced cells to be detected in vivo	There is already literature about the use of human NIS in CAR T cells. However, promoters endowed with increasing strength will be tested to find the best solution.
Task 2.1	Specific difficulty in generating hepatic metastasis by injecting human colorectal cancer cell lines in NSG mice	1. We will perform proof-of-concept study with sc tumors to provide an overview of the ACT strategy for solid tumors; 2. We have several colon cancer cell lines that grow well sc and will be assessed in the metastatic model; 3. Two ways of injection of cancer cells in the liver (intra-mesenteric vein; direct intra-hepatic) can be performed to increase the chance of tumor take.
Task 2.2	Not enough space in the retroviral vector for both CAR molecules and the complex IL15/IL15R.	Cloning into a lentiviral vector
Task 2.3	The bi-specific CAR construct could show suboptimal leukemia targeting.	In this case, taking advantage from the interaction with biophysics and biostatisticians, we will proceed by the optimization of the bispecific construct, and if not sufficient, we will clone a lentiviral vector to carry the two single specific CAR molecules.
Task 2.4	CAR construct with an innovative co- stimulation represented by a specific KIR sequence does not show an optimal in vitro activity.	We will design a second novel construct based on the choice of a different KIR molecule.
Task 2.4	Failure in establishing a CRISPR/Cas KO AML cell line for CD33 and/or CD123	In alternative, use of established cell lines that express or not the CD33 and/or CD123 antigens
Task 2.5	The chosen first antibody to produce the anti- CD138 CAR-T is not effective.	- other two anti-CD138 antibodies will be tested to produce the anti-CD138 CAR-T. - simultaneously employing the three anti-CD138 CAR-T could be also tested to increase the killing efficacy for MM-line cells and primary MM cells.
Task 2.6	The available blocking Ab against ITGA6 is toxic in vivo.	N-Glycosylation is specifically different between parental and PT-Resistant cells therefore we could produce new specific Abs for the different glycosylation sites and test them as blocking Ab.
Task 2.7	Single antagomirs/blockmirs can fail to restore and enhance SSTR-2 expression level if more miRNAs targeting different regions of the transcript are co-expressed in the tumor cells.	The problem can be overcome by transfecting cells with vectors endowed with 2 or 3 blockmirs/antagomirs at once.
Task 2.8	Obtainment of non-functional CARs: - Epitope accessibility and optimum of distance - scFv design spacer design	Problems in CAR functionality against the target will be overcome by an extensive CAR design approach, comparing different constructs carrying different scFv and/or spacers
Task 3.1	1.The CRISPR/Cas9 action may induce off- target mutations in T cell DNA. 2. ACT with gene-edited CAR- T or iNKT cells may still not sufficiently improve.	1. Off-target activity of targeting nucleases will be monitored by using GUIDE-seq established in OSR through the collaboration with Prof. X. Xxxxxxx’x group (Schiroli G. Cell Stem Cell. 2019). 2. Harness other genes and pathways that might be

		discovered in the other activities performed in WP4. Should this be the case, additional gene transfer/ genome editing activities will be designed.
Task 3.2	We may not be able to optimize the 3D in vitro models to validate the novel TME- derived targets suitable for CAR T cell therapy	We will take advantage of animal models also available within the consortium.
Task 3.4	Obtainment of non-functional CARs: - Epitope accessibility and optimum of distance - scFv design spacer design	Problems in CAR functionality against the target will be overcome by an extensive CAR design approach, comparing different constructs carrying different scFv and/or spacers
Task 3.4	Low expression of CD146 on primary MSCs culture	Identification of established cell lines highly expressing CD146 or engineering cell lines for the overexpression of CD146
Task 3.4	High toxicity of Dual CARs on normal HSCs within the hematopoietic niche	Optimize the suboptimal activation of the signal 1 within the trans-signaling strategy
Task 3.4	Low or absence of CD146 / XX00.XXX CIK cells within the ossicle previously implanted in vivo with AML cells	Adoption of alternative ossicle models based on the use of different artificial scaffolds
Task 3.5	It is also possible that TIMP1 serum levels taken at baseline will not correlate with tumor infiltration and/or patients’ survival.	In this case we will test the levels of plasma- TIMP1 in patients treated with chemotherapy a different time points.
Task 3.6	A small number of samples obtained from treated mouse models could cause non- significant results	The availability and sharing of samples within the Network will be essential for the advancement of research.
Task 4.1	No biomarkers association with patient's outcome among the ones tested by FACS.	We will develop novel flow panels considering the data obtained through RNAseq analysis and the patient's samples will be analyzed retrospectively.
Task 4.2	Collection of adequate numbers of samples (FFPE lymph node biopsies) from DLBCL patients candidate to CAR-T treatments (registered indication) is strongly dependent from shared criteria for eligibility and strong commitment from participating centres of the rete AMORe.	At project kick-off, a first meeting is planned to share eligibility criteria and promote commitment among centres. TC meetings will be scheduled twice a month to monitor patient accrual and address possible intercurring problems. Centres will be asked to refer, to a shared central registry, DLBCL patiente deemed eligible with biweekly frequence. Pathology Units from the AMORe centres will be engaged in all meetings and TCs.
Task 4.4	Few results from T cell characterization in AML due to heterogenity.	Data will be analyzed at interpatient and intrapatient level after normalization and harmonization. Gene signatures for data analysis will be appropriately selected.
Task 4.5	It is possible that the murine-derived ID8 cells will not represent a good study model.	In this case we will focus on human xenograft (both PT-resistant cells and PDX) to better appreciate its contribution.
Task 5.1	Low or absence of XX00.XXX T cells within hematopoietic organs	Optimization of the balance between number of XX00.XXX T cells infused and lenalidomide dose concentration
Task 5.1	Low or lack of lenalidomide combination effect on immune synapse between XX00.XXX T cells and CLL cells	Optimization through lenalidomide in vitro titration
Task 5.2	Failure of the high-throughput screening platform for the discovery of novel drugs characterized by synergistic activity with CAR-T cells	We will consider those compounds showing antagonist effects with CAR-T cells, to discover novel mechanism of CAR-T cell resistance
Task 5.3	Tropism of OV is dependent on the nature of the vector	We plan to test different OVs to achieve the optimal transduction efficiency.

WP5: Strategies to improve the safety profile of CAR T cells.

WP leader: OSR; Partcipant Institutions: OPBG, AMORe network, ICH, INT.

Rational

So far, CAR-T cells have shown a suboptimal safety profile. In fact, a significant proportion of treated patients have experienced infusional toxicities and, rarely, lethal events have occurred (June Science 2018). The CRS is the most serious adverse event, which consists of an acute systemic inflammatory reaction leading to fever, hypotension, capillary leak, coagulopathy and occasionally multiorgan failure. The management of CRS toxicity still poses a major medical challenge. Some risk factors for the development of severe CRS and/or neurotoxicity, including patient and treatment characteristics, have been identified in numerous CAR-Ts clinical trials. In addition, several groups have identified some biomarkers able to predict in advance which patients will suffer from severe CRS and/or neurotoxicity during cell therapy with CAR-Ts, facilitating early intervention strategies (Xxxxx Blood 2019). However, additional studies are needed to better understand the biology and related risk factors for CRS and/or neurotoxicity and to determine if such markers can be used prospectively to predict the risk of adverse events. The scope of WP5 is to gain further insight into the definition of the mechanistic aspects underlying CRS in CAR-T cell therapy, for the identification of both efficacious treatments and predictive biomarkers, with the ultimate goal of devising preventive strategies that can increase the safety of the approach. We know that CRS associates with marked elevations of serum cytokine levels and its onset and severity correlate with the pharmacokinetic characteristics of CAR-T cells and tumor burden, respectively (Maude NEJM 2018; Park NEJM 2018). Recent advances in animal modelling of CRS revealed that it is triggered by CAR-T cell activation and amplified by innate immune cells, e.g. monocytes and macrophages, which produce high amounts of IL-1, IL-6 and nitric oxide (Xxxxxxx Xxx Med 2018; Xxxxxxxxx Nat Med 2018). Interleukin-1 (IL-1) is a central mediator of CRS and neurotoxicity, and IL-1 blockade is protective in these conditions (Giavidris Nature Medicine; Norelli Nature Medicine). IL-1 activity is tightly regulated at multiple levels by diverse mechanisms including a receptor antagonist (IL-1Ra), a decoy receptor (IL-1R2), and a negative regulator (IL-1R8) (Garlanda Immunity 2013; Mantovani Immunity 2019, Molgora Nature 2017). The decoy receptor IL-1R2, expressed by monocytes, polarized M2 macrophages, neutrophils, and Treg cells, negatively regulates IL-1 activity by forming a complex with IL-1 and the IL-1RAcP, and exerting a dominant-negative effect. IL-1R2 is released in a soluble form and increased blood concentrations of soluble IL-1R2 have been detected in a wide range of human inflammatory disorders. IL-1R8 (also known as SIGIRR) is an atypical receptor of the IL-1R family, acting as a negative regulator of NF-kB and JNK activation by interfering with the recruitment of TIR-containing adaptor molecules, as well as mTOR kinase activity in lymphocytes, following stimulation of IL-1R family members or TLRs. IL-1R8 expression in lymphocytes is associated with their functional activation (Molgora Nature 2017). IL-1 represents a major stimulus for releasing the long pentraxin PTX3, a fluid-phase pattern recognition molecule that acts as a key component of humoral innate immunity and regulator of inflammatory responses. PTX3 is produced and released in particular by cells of myeloid origin in response to pro-inflammatory mediators, and has been developed as a novel inflammation and tissue damage biomarker (Garlanda Physiol Rev 2018). PTX3 behaves as an acute phase protein, and its plasma concentration increases more rapidly than C-reactive protein (CRP) in inflammatory conditions, reaching hundreds nanograms and even micrograms/ml during endotoxic shock, sepsis, and other inflammatory and infectious conditions.). Monocyte-macrophages have a central role as a source of these cytokines and biomarkers, as well as in the immuneregulation of the tumor microenvironment (Mantovani Nat Rev Clin Oncol 2017). The tetraspan surface molecule MS4A4A is selectively expressed by cells of the monocyte-macrophage lineage(s), and it is

upregulated in macrophages by M2 or M2-like signals, including IL-4 and dexamethasone and in monocytes in patients treated with methylprednisolone (Mattiola Nature Immunol 2019). In addition, while it has been suggested that the presence of an immunosuppressive tumor microeneviroment (TME) may negatively regulate effectiveness and persistence of CAR-T cells (Becker Cancer Immunol Immunother 2013), few data exist on the putative role of dendritic cells (DCs) subsets in such a setting. Similarly, while early data suggested that specific DCs subsets may play a role in modulating CAR-T-specific toxicity, no specific studies have been so far produced to finally clarify the issue (Titov Cell Death Dis. 2018). Recently, also endothelial activation has been reported to play a crucial role in the pathogenesis of CRS (Hay Blood 2018).

The other most commonly observed toxicity with CAR-T cell therapies is neurotoxicity, which commonly presents with delirium, headache, decreased level of consciousness and speech impairment. Similarly to CRS, neurotoxicity can worsen and lead to life-threatening complications, such as acute cerebral edema. The severity of neurotoxicity is strictly associated with the severity of CRS, suggesting a close relation between these two toxic manifestations (Gust Cancer Discovery 2017). However, while we are starting accumulating evidences on the pathogenesis of CRS, neurotoxicity still remains a poorly known event and a major need for a proper handling of CAR-T cell therapeutics.

Although their application to solid tumors is still rare, compared to haematological malignancies, the incidence of systemic, severe CRS upon CAR-T-cell infusion in patients with solid malignancies has been reported to be uncommon (Xxxxx J Clin Oncol 2015; Xxxx CCR 2015; Xxxxx Xxx Ther 2017; Xxxxx NEJM 2015; X’Xxxxxx Sci Transl Med 2017; Xxxxx J Immunother 2017). However, in most trials in solid tumours the therapeutic efficacy of CAR-T cells was also limited, possibly because of specific obstacles, such as poor T-cell trafficking to the tumour site, poor antigen recognition and an extremely immunosuppressive tumour microenvironment. Different strategies are currently under investigation to increase the efficacy of CAR-T cell therapy in solid tumours. Since the CRS was not common in haematological cancer patients who did not have a clinical response, strategies to increase antitumor efficacy might potentially impact the safety of CAR-T cell therapy also in solid malignancies. To this aim, in addition to working on the above host’s derived molecular and cellular mediator of toxicities, a way to control CAR-T cell toxicity may be that of directly modulating their effector functions in vivo. For instance, the infusion of CAR-T cells that have been modified with the inducible caspase-9 (iC9) suicide gene may prove efficacious in reducing acute toxicities, in as much as the strategy efficiently reverts GvHD symptoms in acute leukemia recipients of haploidentical donor iC9.T cells after αβ T-cell depleted haplo HSCT. Furthermore, it may be possible to interfere pharmacologically with the production of pro-inflammatory mediators by CAR-T cells, without affecting their direct anti-tumor cytoxicity.

Finally, considering the CAR-T clinical application site, EMA recommendations do not strictly define the inclusion and exclusion criteria or the most appropriate setting of use of CAR-T cells. However, based on the experience in the clinical trials and on the preliminary reports regarding the use of CAR T-cell therapy in real life, it is emerging that the application of strictly inclusion criteria can minimize the severity of adverse events (AE) and can guide the selection of patients who can be successfully treated. Noteworthy, the assessment and grading of CRS and immune effector cell‐associated neurotoxicity syndrome (ICANS) vary considerably across clinical trials and across institutions, making difficult to compare the safety of different products and hindering the ability to develop optimal strategies for the management of these toxicities. Only recentl,y the first uniform consensus for definition and grading of CRS and ICANS has been proposed for use across clinical trials, while no consensus is yet available for their management in the clinical practice.

Hence, altogether, the above considerations highlight many different unmet levels of intervention on CAR-T cell toxicity that can be defined by leveraging increased knowledge of the subtending mechanisms, combined with advanced genetic control of effector T cells and a thorough clinical monitoring of patients.

Document Metadata

Table of Contents

CONVENZIONE

Document Metadata