Melanoma, o status da imunoterapia e o futuro próximo

III International Symposium on Immuno-Oncology Saturday, October 7 th 9:3 1: am Melanoma, o status da imunoterapia e o futuro próximo Antoni Ribas, M.D., Ph.D. Professor of Medicine, Surgery, Molecular and Medical Pharmacology Director, Tumor Immunology Program, Jonsson Comprehensive Cancer Center (JCCC) Director, Parker Institute for Cancer Immunotherapy (PICI) Center at UCLA University of California Los Angeles (UCLA) Chair, Melanoma Committee at SWOG

Overall Survival: Metastatic Melanoma Phase III Studies 1-year OS 25 35% 1 46% 2 47% 4 56% 6 7% 7 71% 8 71% 1 Pembrolizumab 74% 11 Dab + tram d 75% 12 Vem + cobi 73% 13 Nivo + ipi 199 21 211 212 213 214 215 216 217 15% 1 24% 2 29% 4 Ipi 3% 6 45% 7 58% Nivo 8 53% 11 Dab + tram 48% 12 Vem + cobi 55% 14 Pembrolizumab 2-year OS 64% 13 Nivo + ipi 3-year OS 22% 3 44% 15 Dab + tram 37% 16 Vem + cobi 5-year OS 18% 5 Ipi 35% 9 Nivo (ph I) Georgina V Long, MIA 1. M. Middleton Ann Oncol 27;18:1691. 2. Hodi FS et al NEJM 21;363:711. 3. Schadendorf D et al JCO 215:33:1889. 4. Robert C et al NEJM 211;364:2517. 5. Maio M et al. JCO 215; 33:1191. 6. Chapman PB et al Poster presentation SMR 215. 7. Hauschild A et al Poster presentation ESMO 214. Abstract 192. 8. Atkinson V et al Poster presentation SMR 215. 9. Hodi FS et al Oral presentation AACR 216. Abstract CT1. 1. Carlino M et al Oral presentation AACR 216. Abstract CT4. 11. Long GV Lancet Oncol 216. 12. Ascierto PA et al Lancet Oncol 216:17:1248. 13. Larkin J et al AACR 217:1558. 14. Schachter J et al Oral presentation ASCO 216. Abstract 954. 15. Flaherty K et al Oral presentation ASCO 216. Abstract 952.16. MacArthur GA et al. Poster presentation SMR 216

PFS, probability 1..8.6 High LDH High tumour volume More metastatic sites Brain metastases PD-L1 negative/lack of immune signature 5% 45-5% 43% Normal LDH Low tumour volume Fewer metastatic sites CR to treatment.4.2 4-45% 39% 3% 24%. Georgina V Long, MIA 6 12 18 24 3 36 42 48 Time from randomisation, months Dabrafenib + trametinib 1 (n = 352) Vemurafenib + cobimetinib 5 (n= 247) Nivolumab 2 (n = 21) Nivolumab 3 (n = 316) a Pembrolizumab 4 (n = 279) Nivo + ipi 3 (n = 314) 1. Robert C Oral ESMO 216. 2. Atkinson et al Poster SMR 215. 3. Larkin J et al Oral AACR 217. 4. Schachter J et al Oral ASCO 216. 5. Ascierto PA et al Lancet Onc 216.

CTLA-4 and PD-1 Checkpoint Blockade Abril-Rodriguez and Ribas, Snapshot, Cancer Cell 217

Progression-free Percentage Survival of PFS (%) Checkmate 67: Progression-Free Survival 1 9 8 7 Median PFS, mo (95% CI) HR (95% CI) vs. IPI HR (95% CI) vs. NIVO NIVO+IPI (N=314) NIVO (N=316) IPI (N=315) 11.7 (8.9 21.9).42 (.34.51).76 (.62.94) 6.9 (4.3 9.5).54 (.45.66) 2.9 (2.8 3.2) -- -- -- 6 5 4 3 5% 43% Grade 3/4 tox: 59% 21% 43% 37% 2 1 NIVO+IPI NIVO 18% IPI 3 6 9 12 15 18 21 24 27 3 33 36 Months Patients at risk: NIVO+ IPI 314 218 176 156 137 132 125 118 11 14 71 16 NIVO 316 178 151 132 12 112 17 13 97 88 62 16 IPI 315 136 77 58 46 43 35 33 3 27 16 5 Larkin et al. AACR 217, Wolchok et al. NEJM 217 28% 12% Database lock: Sept 13, 216, minimum f/u of 28 months

Overall Percentage Survival of PFS (%) Checkmate 67: Overall Survival 1 9 8 7 6 5 4 3 2 1 NIVO+IPI NIVO IPI 73% 74% 67% Median OS, mo (95% CI) HR (98% CI) vs. IPI HR (95% CI) vs. NIVO 3 6 9 12 15 18 21 24 27 3 33 36 39 Months Patients at risk: NIVO+IPI 314 292 265 247 226 221 29 2 198 192 17 49 7 NIVO 316 292 265 244 23 213 21 191 181 175 157 55 3 IPI 315 285 254 228 25 182 164 149 136 129 14 34 4 64% 59% 45% NIVO+IPI (N=314) NIVO (N=316) IPI (N=315) NR.55 (.42.72)*.88 (.69 1.12) NR (29.1 NR).63 (.48.81)* 2. (17.1 24.6) -- -- -- *P<.1 Larkin et al. AACR 217, Wolchok et al. NEJM 217 Database lock: Sept 13, 216, minimum f/u of 28 months

CTLA-4 and PD-1 Checkpoint Blockade Abril-Rodriguez and Ribas, Snapshot, Cancer Cell 217

Melanoma response to PD-1 blockade is mediated by pre-existing infiltrates of CD8s inhibited by reactively expressed PD-L1 PD1/PDL1 1 Anti-PD-1 Anti-PD-L1 Melanoma cell or tumor macrophage IFN-g Tumeh et al., Nature 214

PD-1 blockade induces responses by inhibiting adaptive immune resistance CD8 PD-1 PD-L1 Response Progression Melanoma cell or tumor macrophage Melanoma cell or tumor macrophage Interferon gamma Hypothesis formulated based on quantitative IHC analyses of 46 cases from UCLA Adapted from Tumeh et al. Nature 214 Patient #1 Patient #2

What differentiates anti-pd-1-responsive from non-responding melanomas? ORR: 33% ORR in previously untreated: 45% Pembrolizumab Keynote 1 trial. Central radiology review by RECIST v1.1 Ribas et al. JAMA 216

What differentiates anti-pd-1-responsive from non-responding melanomas? Ayers et al, JCI 217 Pembrolizumab Keynote 1 trial. Central radiology review by RECIST v1.1 Ribas et al. JAMA 216

What differentiates anti-pd-1-responsive from non-responding melanomas? Mutational load and response to anti-pd-1 in NSCLC Mutational load and response to anti-pd-1 in MSI high colon cancer Mutational load and response to anti-pd-1 in melanoma Rizvi et al, Science 215 Le et al, NEJM 215 Hugo et al, Cell 216 McGranahan et al. Science 216 Pembrolizumab Keynote 1 trial. Central radiology review by RECIST v1.1 Ribas et al. JAMA 216

What differentiates anti-pd-1-responsive from non-responding melanomas? Hugo, Zaretsky et al. Cell 216 Willy Hugo, PhD Jesse Zaretsky IPRES (Innate anti-pd-1 Resistance) signature Roger S. Lo, MD, PhD Pembrolizumab Keynote 1 trial. Central radiology review by RECIST v1.1 Ribas et al. JAMA 216

The Cancer Immunogram Tumor foreignness Mutational load Tumor sensitivity to immune effectors MHC expression IFN-g sensitivity General immune status Lymphocyte count Absence of inhibitory tumor metabolism LDH, glucose utilization Immune cell infiltration Intratumoral T cells anti-pd-1/l1 Absence of soluble inhibitors IL6->CRP/ESR Absence of Checkpoints PD-L1 Blank, Haanen, Ribas, Schumacher. Science 216

Desmoplastic melanoma: Defined by a dense collagenous fibrous tissue a) DM1 DM2 b) Baseline 2-3 months later Tumor (S1) S1 Trichrome Collagen (Trichrome) a) a) S1 S1 Trichrome Trichrome S1 S1 A rare subtype of melanoma (less than 4%) S1 A dense fibrous reaction Trichrome Trichrome DM1 CM1 A known relationship to UV light damage High NF1 mutation rate and no known actionable genes for targeted therapies. ome CM1 CM2 DM2 CM2 blue collagenous stroma, red cytoplasm and brown nucleus CASE 1 b) b) CASE CASE 1 CASE CASE 2 CASE 2 CASE CASE 3 SE 3 Baseline Baseline 2-3 months after anti-pd1 2-3 months therapy later Zeynep Eroglu Siwen Hu-Lieskovan Jesse Zaretsky (submitted)

High response rate and high mutational load in Desmoplastic melanoma 7% overall response rate 18% complete response rate n=57 (out of 154 cases Reviewed*) 1 siiic 3 M1a 2 M1b 35 M1c 4 5 6 *Retrospective Review o s e (M o n th s ) B) E s tim a te d O v e ra ll S u r v iv a l (% ) % Overall survival (OS), median not reached 1 9 8 7 6 5 4 3 2 1 Estimated 2 year OS 73% (CI 62-88). 1 2 3 4 5 6 7 Time (months) T im e S in c e F irs t d o s e (M o n th s ) Non-Desmoplastic Melanoma Desmoplastic Melanoma = Progressive Disease = Response (RECIST1.1) Zeynep Eroglu Siwen Hu-Lieskovan Jesse Zaretsky (submitted)

Response, % Primary and Acquired Resistance to PD-1 Blockade Waterfall plot of RECIST responses in 51 patients treated with pembrolizumab in the Keynote 1 trial KM of duration of response in patients treated with pembrolizumab or ipilimumab in the Keynote 6 trial Primary resistance 1 9 8 7 6 Acquired resistance 5 4 3 2 1 4 8 12 16 2 24 28 32 36 No. at risk Pembrolizumab Ipilimumab Time, months 233 22 199 175 16 136 118 78 24 46 42 32 29 25 23 19 13 3 Ribas et al. JAMA 216 Robert et al. ASCO 217

How does the cancer sense IFN-gamma and reactively expresses PD-L1? PD1/PDL1 1 Melanoma cell or tumor macrophage Interferon gamma Tumeh et al., Nature 214

Interferon gamma receptor pathway regulating reactive PD-L1 expression Melanoma cell IRF-1 PDL1 Promoter PD-L1 Adapted from Shin et al. Cancer Discovery 217 and Garcia-Diaz et al. Cell Reports 217

Primary resistance to PD-1 blockade by disabling PD-L1 adaptive expression Would be useless to try to inhibit PD-1:PD-L1 Melanoma cell IRF-1 PDL1 Promoter PD-L1 Adapted from Shin et al. Cancer Discovery 217 and Garcia-Diaz et al. Cell Reports 217

Is JAK loss associated with primary resistance to PD-1 blockade? Responders Non-Responders JAK1 homozygous all heterozygous JAK1 homozygous 1/23 melanoma cases with high-allele frequency JAK1 mutation 1/16 colorectal cases with high-allele frequency JAK1 mutation data from Le DT et al. NEJM 215 Shin et al, Cancer Discovery 217

Lesion Size (mm 2 ) 6 4 2 Case #1 2 4 Days on Therapy Lesion Size (mm 2 ) 15 1 5 Case #2 2 4 6 8 Days on Therapy Non-Synonymous Mutations Non-Synonymous Mutations Baseline Baseline CNV Relapse CNV * Baseline CNV Relapse CNV Baseline (M42) CN Total CN Total Relapse 1 2 3 >=4 CN Minor Allele (LOH) Baseline Relapse 1/2 3 4 5/6 >6 CN Minor Allele 1 (LOH) New mutation * Relapse (Tumor) Jesse Zaretsky UCLA MSTP JAK1 Q53* * Relapse (M464) JAK2 F547_splice

Interferon-gamma Sensitivity Model: Initially, benefits of PD-L1 suppression outweigh immune sensitizing effects. PD-L1 expression is of no benefit after PD-1/L1 blockade; selective pressure is flipped. The cancer has an incentive to lose INF-γ sensitivity, avoid apoptosis, enhanced antigen presentation. Acquired Resistance JAKs Melanoma cell Growth inh/apoptosis PD-L1 APM Adapted from Zaretsky et al. NEJM 216

How prevalent are JAK and B2M loss-of-function mutations? 2 2 Neither 8 IFNGR, JAK/STAT pathway mutations B2M mutations n=12 In 8 additional paired biopsies: One additional B2M LoF mutation No additional JAK1/2, IFNGR, IRF1 or STAT1/3/5 LoF mutations Jesse Zaretsky, Antoni Ribas (unpublished)

Confirmation of B2M and JAK as genetic mechanisms of resistance to anti-pd-1 using CRISPR/Cas9 knock out sublines MC38 wild-type MC38 B2M knockout MC38 JAK2 knockout tumor volume (mm3) 8 6 4 2 MC38 WT untreated MC38 WT apd-1 tumor volume (mm3) 8 6 4 2 B2M untreated B2M apd-1 tumor volume (mm3) 5 4 3 2 1 JAK2 untreated JAK2 apd-1 5 1 15 days after tumor injection 5 1 15 2 days after tumor injection 1 2 3 days after tumor injection Davis Torrejon, Gabriel Abril Rodriguez, Siwen Hu-Lieskovan (unpublished)

Defects in the IFNγ pathway induce resistance Genome-wide CRISPR mutagenesis reveals essential genes for the effector function of T cells in a target cell. Loss of Ptpn2 increases IFNγ sensing by tumour cells Functional loss of APLNR reduces efficacy of cancer Immunotherapy, which IPs with JAK1

Conclusions Inhibiting adaptive immune resistance is the mechanistic basis of the antitumor activity of PD-1 blockade therapies Combination therapies aimed at increasing T cell infiltration in tumors may improve the antitumor activity of PD-1 blockade Loss of function mutations in IFN-gamma receptor signaling or antigen presenting machinery mediate some cases of primary resistance and acquired resistance to PD-1 blockade therapy

Intervalo

III International Symposium on Immuno-Oncology Saturday, October 7 th 1:3 1:5 am Combinação em Imunoterapia: Deve ser o standard? Qual o melhor parceiro para combinar com imunoterapia? Antoni Ribas, M.D., Ph.D. Professor of Medicine, Surgery, Molecular and Medical Pharmacology Director, Tumor Immunology Program, Jonsson Comprehensive Cancer Center (JCCC) Director, Parker Institute for Cancer Immunotherapy (PICI) Center at UCLA University of California Los Angeles (UCLA) Chair, Melanoma Committee at SWOG

Intrinsic or primary resistance to immune checkpoint therapies- cellular microenvironment Cold tumors- induce TILs? Alternate immune suppressive mechanisms? Teng, Ngiow, Ribas, Smyth. Cancer Research, 215

Intrinsic or primary resistance to immune checkpoint therapies- cellular microenvironment 41% 45% 12% 2% Teng et al., Can Res, 215; Taube et al., Sci Trans Med, 212

2 out of 48 melanoma cell lines had JAK1/2 LOF mutations and did not respond to IFN-gamma by expressing PD-L1 PD-L1 baseline surface expression M395: JAK1 chr1:6539827c>t, Exon 17 D775N M368 : JAK2 chr9:555668 G>A, Exon 8 D313_Splice JAK1 5 # Mutations D775N JAK2 5 Pkinase_Tyr Pkinase_Tyr 2 4 6 8 1 1154 aa # Mutations D313_splice SH2 Pkinase_Tyr Pkinase_Tyr 2 4 6 8 1 1132 aa Daniel Shin, MD Shin et al. Cancer Discovery 217, 7, 188-21

Management of cancer in the anti-pd-1/l1 era Anti-PD-1/anti-PD-L1 Bring T cells into tumors: Generate T cells: + anti-ctla4 + immune activating antibodies or cytokines + TLR agonists or oncolytic viruses + IDO or macrophage inhibitors + targeted therapies Vaccines TCR engineered ACT CAR engineered ACT Modified from Ribas, Cancer Discovery 216

PD-1 / PD-L1 and IDO in the T cell inflamed phenotype T cell T cell mediated INFγ release triggers both PD-L1 expression IDO expression in tumor cells and in the µ-environment PD-1 INFγ INFγ - IDO STAT PD-L1 Tumor cell TCR MHC Kynurenine Tryptophan - +

Efficacy data: ECHO-22, Keynote-6, Checkmate 67 Response ECHO-22 Epacadostat + Pembro 1 PD-1 single agent Checkmate 67 2 / Keynote-6 3 Ipi + Nivo Checkmate 67 2 ORR (CR+PR) 56% 44% / 42% 58% CR 14% 16% / 13% 19% PR 41% 3% / 29% 42% SD 16% 1% / 21% 11% PD 29% 38.6 / 29% 23.6 Not evaluable 3% 7% / - 6% Median PFS 12.4 months 6.9 / 8.3 months 11.5 months PFS @ 18 months 52% 43% / - 5% Grade 3/4 toxicity 2% 21% / 17.5% 59% 1 Hamid, ESMO 217; 2 Larkin, AACR 217 and Wolchok, NEJM 217; 3 Robert, ASCO 217 (2 pembro arms pooled). Note: data not randomized head to head, comparison only for indication

Double immune checkpoint inhibition: PD-1 plus LAG-3 CTLA-4 Blockade PD-1 Blockade + LAG-3 Blockade Dendritic cell MHC B7 TCR CD28 +++ - - - B7 CTLA-4 +++ T cell T cell MHC TCR ++ + PD1 PD-L1 - - - anti-pd1 LAG3 - - - Tumor cell anti-ctla-4 anti-lag-3 MHC c II ASCO 217-Ascierto et al Nivolumab + BMS-98616 (anti-lag3) Patients refractory or resistant to anti-pd-1

Best Percent Change in Sum of Target Lesion Diameters From Metastatic Melanoma with Prior-IO Cohort LAG-3 1% a n = 22 ORR, 2% LAG-3 <1% a n = 12 ORR, 7.1% LAG-3 Unknown n = 8 ORR, 1 8 6 1 8 6 1 8 6 LAG-3 expression enriched for responses in IO-experienced patients Baseline b 4 2 2 4 6 4 2 2 4 6 4 2 2 4 6 Nearly a 3-fold increase in ORR was observed in patients with LAG-3 1% vs LAG-3 <1% (2% vs 7.1%) Overall response rate was 13% 8 8 8 1 1 1 6 PRs: 2 prior PD; 3 prior PR; 1 unk DCR, disease control rate; ORR, objective response rate. a LAG-3 expression (percent of positive cells within invasive margin, tumor, and stroma) evaluated using immunohistochemistry (IHC) assays on formalin-fixed, paraffin-embedded tumor sections. Immune cell LAG-3 expression ( 1% or <1%) determined using mouse antibody clone 17B4. b Response-evaluable patients (n = 48; all progressed on prior anti PD-1/PD-L1 therapy). Six patients had clinical progression prior to their first scan and are not included in the plot. One patient with best change from baseline >3% had an unconfirmed best response of SD. Ascierto et al. ASCO 217

Adding intralesional therapies to anti-pd-1/l1 Anti-PD-1/anti-PD-L1 IT injection oncolytic virus TLR agonist STING agonist + Anti-PD-1/anti-PD-L1 Then the benefit should only be in the patients who were unlikely to respond to anti-pd-1/l1 alone because their T cells were not in the tumor

Kaplan-Meier Percent Kaplan-Meier Percent Percentage Change from Baseline Patient 2 MASTERKEY-265: T-Vec + pembrolizumab T-VEC Intralesional Pembrolizumab Wk:1 6 3 Baseline (Week -5) Week Week 12 1 75 5 25 N = 21 Stage IIIB (N = 1) Stage IIIC (N = 6) Stage IV M1a (N = 2) Stage IV M1b (N = 4) Stage IV M1c (N = 8) 25 5 75 1 62% objective response rate 33% complete response rate 1 8 6 PFS 1 8 6 OS 4 4 2 T-VEC plus pembrolizumab (N = 21) Median (95% CI) 3 6 9 12 15 18 Study Month Number of Patients at Risk: 21 16 14 14 13 13 1 2 T-VEC plus pembrolizumab(n = 21) Median (95% CI) 3 6 9 12 15 18 Study Month Number of Patients at Risk: 21 2 2 2 2 17 6 Ribas et al. Cell 217

T-VEC increases tumor CD8 and PD-L1 in patients responding to combination with pembrolizumab Pt #1 Baseline Week 6 Week 3 Pt #2 CD8 antibody with red chromogen T-VEC T-VEC+ pembro T-VEC T-VEC+ pembro

Adding BRAF targeted therapies to anti-pd-1/l1 Anti-PD-1/anti-PD-L1 BRAFi+MEKi Wilmott et al. CCR 213 Frederick et al. CCR 213 BRAFi+MEKi + anti-pd-1/l1

Combination of BRAFi+MEKi+Anti-PD-1 NIH Director s Blog* Knocking Out Melanoma: Does This Triple Combo Have What It Takes? Posted on March 31, 215 by Dr. Francis Collins Comment on: Hu-Lieskovan S, et al. Sci Transl Med. 215;7:279ra41. Hu-Lieskovan et al. Sci Transl Med. 215 Mar 18;7(279):279ra41 *Available at: https://directorsblog.nih.gov/215/3/31/knocking-out-melanoma-does-this-triplecombo-have-what-it-takes/. Accessed on August 18, 216.

Change from Baseline, % Clinical trials combining BRAFi+MEKi+anti-PD-1/L1 dabrafenib+trametinib +durvalumab dabrafenib+trametinib +pembrolizumab vemurafenib+cobimetinib +atezolizumab 1 8 6 4 2-2 -4-6 -8-1 * * * * * 4 8 12 16 2 24 Time, months * * Ribas et al. J Clin Oncol 33, 215 (suppl, abstr 33 ASCO) * 4 8 12 16 2 24 Time, months Ribas et al. ESMO, 217 Hwu et al. Annals of Oncology 27; 216 (supp 6; abstr 119PD ESMO)