1. Stratified Medicine & Immunobiology Prof Ben Willcox b.willcox@bham.ac.uk Director, Cancer Immunology and Immunotherapy Centre www.ciic.org.uk b.willcox@bham.ac.uk

2. Cancer Immunotherapy “Tumour immunology has long had a bright future” “For those mice in the audience, it’s good news...” “Immunotherapy earns its spot in the ranks of cancer therapy” “…a tipping point in the fight against cancer”

3. My research Cancer Immunotherapy in the news

4. Type, Density and Location of Immune Cells within Human Colorectal Tumours Predict Clinical Outcome Galon et al, Science 313, pp 1960-1 (2006) Examined TIL levels in colorectal cancer cohorts Assessed TIL levels, types, density, location in tumours (eg Core of tumour (CT) vs Invasive Margin (IM) TIL profile a better predictor of patient survival than current histopathological methods (eg UICC-TNM staging). Therefore adaptive immune response has strong influence on behaviour of human tumours In situ analysis of TILs could be a useful prognostic tool in CRC treatment and possibly other tumours

5. Features of the tumour microenvironment or stroma Frances R. Balkwill, Melania Capasso and Thorsten Hagemann; Journal of Cell Science 2012 (125, pp. 5591-5596

6. The tumour microenvironment: a core feature of cancer Hallmarks of Cancer: The Next Generation Douglas Hanahan, Robert A. Weinberg. Cell Vol 144, Issue 5, p646–674, 4 March 2011

7. B cell tumour ALL/CLL Engineering “smarter” immune cells – CAR therapy Emily Whitehead - “The girl that lived” “CD20” Target antigen killing “CAR” T cells “designer receptor” (Chimeric Antigen Receptor = CAR) Experimental white blood cell treatment shows ‘remarkable’ promise in leukaemia Relapsed, refractory ALL patients treated Complete remission in 90% of patients Very effective tumour clearance “On target” side effects, can be severe Long-lasting responses…potential cures What are safe targets on other tumours ? CRUK website blog, October 2014

8. Tumour “Removing the brakes” on the anti-cancer immune response – checkpoint blockade Unleash “pacified” T cells > Emphasis on understanding the immune profile of patients tumours - potential fundraising campaign Why do/don’t some patients respond ? Which patient groups to apply new therapies in ? Can you predict and avoid harmful side-effects ? Why do only some patients/tumours respond ? Patients with metastatic cancers Responses long-lasting in some patients Effects in range of tumours

9. The potential for cure Chemotherapy vs standard treatment Immunotherapy : durable responses in some patients Melanoma, anti-CTLA-4 Melanoma, anti-CTLA-4 +/- anti-PD-1 Anti-CTLA-4 Control NSCLC Meta-analysis, JCO, 2008

10. Essentials of T cell antigen recognition Cell-cell contact is essential T cells are exquisitely sensitive to peptide antigens presented on the surface of APC by MHC molecules Petide antigens are typically 8-10 amino acids (class I MHC), and up to ~ 25 amino acids long (class II MHC) T cell recognition is highly regulated Two signals required for T cell priming, and activation Signal 1 involves T cell receptor (TCR) recognising antigen in context of MHC Signal 2 involves recognition of “costimulatory” ligands by receptors on the T cell

11. Checkpoints regulate T cell antigen recognition CTLA-4/CD80/86 and PD1/PDL1/L2 have emerged as important checkpoint controls

12. Why have we evolved checkpoint controls ? Inhibitory checkpoints are critical to dampen down T cell responses ctla-4 gene KO results in massive autoimmunity and death at young age Mutations and/or polymorphisms in the CTLA-4 gene have been associated with a wide range of autoimmune diseases, including insulin- dependent diabetes, celiac disease, systemic lupus erythematosus, multiple sclerosis, primary biliary cirrhosis and other autoimmune diseases. pd-1 gene KO live for > 1 year before presenting SLE-like symptoms Implications : will blocking these checkpoints be safe in humans ?

13. Checkpoint blockade – CTLA-4 After TCR/pMHC recogntion, CD28/CD80/86 provides second signal, leading to upregulation of CTLA-4 CTLA-4 has higher affinity for CD80/86, interrupting activation signal, downregulating T cell function and inhibiting excessive T cell expansion Anti-CTLA-4 antibodies bind to CTLA-4, block interaction with CD28, which is free to interact with CD80/86 prolonging T cell activation and amplifying T cell mediated immunity vs tumors

14. Checkpoint blockade – PD1/PDL1 T cell activation by DCs PD1 pathway blockade shifts balance of signals delivered by the DC from suppressive to activating Antigen-specific T cell effector function In the tumour microenvironment, tumour-specific T cells potentially recognise tumour cells but can then be inactivated by expression of PD-L1 or PD-L2 on the tumour cell >>>>tolerance/anergy Blockade rescues T cell function in the periphery, leading to full effector function and tumour killing

15. Checkpoint blockade: predictions of the model CB target expression on tumour cells or tumour-specific T cells is important Efficacious CB might result in autoimmune responses Blockade of multiple checkpoint inhibitors might be beneficial to elicit increased anti- tumour immunity CB requires antigen to be present on tumour cells

16. Checkpoint blockade of PD-1 in NSCLC N Engl J Med 2015;372:2018-28.

17. (L) MRI lung scan, 51 year old patient, active tumour progression despite chemo (red arrows = metastases). (R) excellent response, metastases massively reduced. Checkpoint blockade of anti-PD-1 in NSCLC Large, international, Phase I, Keynote 001 trial Advanced NSCLC, either untreated or treated Assessed PD-L1 expression as potential predictive biomarker Results: - tolerable side effect profile - durable responses in some patients Ipilumumab (anti-CTLA-4) Adverse reactions in up to 1/3 rd of patients “immune-related serious adverse effects” or irSAEs, up to grade 3 or 4 Ranging from dermatitis to server chronic colitis or acute hepatitis Cf Ipiliumumab (anti-CTLA-4 in melanoma:up to 1/3 rd of patients “immune-related serious adverse effects” or irSAEs, up to grade 3 or 4; ranging from dermatitis to chronic/acute hepatitis

18. Blockade of PD-1 in NSCLC: PDL1 as a potential biomarker Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients Roy S. Herbst, et al. Nature 515, November 2014 Clinical and correlative biomarker results from a phase 1 clinical trial in patients with different solid tumours are presented; the findings indicate that PD-L1 expression on tumour-infiltrating immune cells is associated with clinical response to MPDL3280A (anti-PD-L1). PD-1 blockade induces responses by inhibiting adaptive immune resistance Paul C. Tumeh, et al. Nature 515, November 2014 The dynamics of T-cell responses are investigated in tumour tissue from patients with advanced melanoma who were treated with a PD-1-blocking monoclonal antibody, revealing that clinical efficacy of the treatment correlates with increased frequencies of pre-existing CD8 + T cells and PD-1 and PD-L1 expression. Evidence pre-treatment PD-L1 expression correlates with response to anti-PD-1 therapy Many PD-L1+ tumours don’t respond, some PD-L1-negative tumours do respond ! PD-L1 expression on TILs could be relevant

19. Checkpoint blockade: unleashing responses to mutated tumour antigens Predicting immunogenic tumour mutations by combining mass spectrometry and exome sequencing Mahesh Yadav et al. Nature 515, November 2014 A combination of genome-wide exome and transcriptome analysis, mass spectrometry and computational structural modelling are used here to identify immunogenic neo-antigens in two mouse tumour cancer cell lines; mice vaccinated with predicted immunogenic peptides yielded therapeutically useful cytotoxic T-lymphocyte responses. Checkpoint blockade cancer immunotherapy targets tumour-specific mutant antigens Matthew M. Gubin et al. Nature 515, November 2014 A carcinogen-induced mouse tumour model is used here to show that mutant tumour-specific antigens are targets for CD8 + T-cell responses, mediating tumour regression after checkpoint blockade immunotherapy, and that these antigens can be used effectively in therapeutic vaccines; this advance potentially opens the door to personalized cancer vaccines. Intracellular proteins > class I MHC pathway Presentation of “mutated self”

20. Understanding the immunogenetics of tumours Mutational burden in human tumours varies Does tumour immunogenicity relate to mutational burden ? The prevalence of somatic mutations across human cancer types. Signature of mutational processes in human cancer; Alexandrov et al, Nature 500, 415-21, (2013)

21. CIRC expression in CRC stratifies the patients into 4 groups Group AGroup BGroup C Group D Cluster matrix with mutation/pathological data Group A has relatively strong CIRC expression (A) Groups B and especially D have relatively low CIRC expression

22. Anti-PD-1 unleashes responses to mutations in NSCLC Science Vol 348, pp 124-8 (2015)

23. Anti-PD-1 unleashes responses to mutations in NSCLC Higher nonsynonymous mutation burden in tumours correlates with improved objective response to Pembrolizumab, durable clinical benefit, and progression-free survival (more important than tumour expression of PD-L1) Efficacy correlated with molecular smoking signature, higher neoantigen burden, and DNA repair pathway mutations

24. Anti-PD-1 unleashes responses to mutations in NSCLC

25. Anti-PD-1 compared to SOC chemotherapy International, prospective, randomised, phase 3 trial

26. Anti-PD-1 compared to chemotherapy Advanced squamous-cell NSCLC with disease progression after first-line chemotherapy Nivolumab (anti-PD-1 vs docetaxel; BMS drug) monotherapy compared to docetaxel OS, RR, PFS significantly better with nivolumab vs docetaxel Efficacy was regardless of PD-L1 expression level Risk of death 41% lower with nivolumab than docetaxel (P<0.001)

27. Checkpoint blockade and lung cancer - challenges Why do only some patients respond ? Relatively low level of mutations Presence of specific immunosuppressive pathways in some patients Involvement of multiple CB receptors Absence of (relevant) antigen presentation pathways New checkpoint blockade therapies being developed – how best to combine? How to rationally combine chemotherapy with checkpoint blockade approaches ? New checkpoint blockade drugs very expensive – how to develop best biomarkers ? Smoking status Genetics of tumour – eg non-synonymous mutation rate How to cope with low mutation tumours ? New therapeutic approaches Vaccination ? Antigen delivery ?

28. Strategic advantages to Birmingham lung immunobiology workstreams Computational biology efforts Birmingham Centre for Computational Biology CRC as a template Bioinformatics as a key hypothesis driving step 8 Key role for Birmingham in SMP2/ Matrix Molecular Pathology expertise Philippe Taniere Multiplexed IHC platform; InForm and Definiens Tissue Studio software (automated scoring) ATF-based so GCLP compliant Aligned RNA-based analyses of IHC sections Sample availability

29. Survival rates for children with cancer are excellent Acknowledgements Prof Gary Middleton Dr Philippe Taniere Mr Neeraj Lal Mr Andrew Beggs CIIC grouping