Supplementary Figure 4. Predicted HLA alleles (A) and their associated immunogenic epitopes (B) in patients treated with mRNA-4157 monotherapy.

Supplementary Figure 8. Study design for Part A and Part D.

Supplementary Figure 2. Characterization of mRNA-4157 neoantigen pool-specific T cells post-expansion at pre-treatment and post-treatment for patients who received mRNA-4157 monotherapy (A–C) or combination therapy (D–K). (L) Alluvial plot illustrating neoantigen reactivity.

Supplementary Table 2. Safety summary for patients who received a dose of mRNA-4157.

Supplementary Figure 3. Patient responses to individual neoantigens at pre-treatment and post-treatment for patients who received mRNA-4157 monotherapy (A–C) or combination therapy (D–J) using ELISpot assay.

Supplementary Figure 6. mRNA-4157 in combination with pembrolizumab activates an adaptive immune response.

Supplementary Table 4. Cellular markers used for ex vivo immune phenotyping.

Supplementary Table 3. Cellular markers used for post-expansion intracellular staining and identification of CD4 and CD8 responses.

Supplementary Figure 7. Distribution of T cell subsets pre-treatment and post-treatment for high (Patients 7 and 6) and low (Patients 13 and 14) immune responders.

Supplementary Table 1. Representativeness of study participants.

Abstract mRNA-4157 (V940) is an individualized neoantigen therapy targeting up to 34 patient-specific tumor neoantigens to induce T-cell responses and potentiate antitumor activity. We report mechanistic insights into the immunogenicity of mRNA-4157 via characterization of T-cell responses to neoantigens from the first-in-human, phase 1, KEYNOTE-603 study (NCT03313778) in patients with resected non–small cell lung cancer (Part A: 1-mg mRNA-4157, n = 4) or resected cutaneous melanoma (Part D: 1-mg mRNA-4157 + 200-mg pembrolizumab, n = 12). Safety, tolerability, and immunogenicity were assessed. All patients experienced ≥1 treatment-emergent adverse event; there were no grade 4/5 adverse events or dose-limiting toxicities. mRNA-4157 alone induced consistent de novo and strengthened preexisting T-cell responses to targeted neoantigens. Following combination therapy, sustained mRNA-4157-induced neoantigen-specific T-cell responses and expansion of cytotoxic CD8 and CD4 T cells were observed. These findings show the potential of a novel mRNA individualized neoantigen therapy approach in oncology. Significance: The safety and immunogenicity results from this phase 1 study of mRNA-4157 as adjuvant monotherapy or combination therapy with pembrolizumab show generation of de novo and enhancement of existing neoantigen-specific T-cell responses and provide mechanistic proof of concept to support further development of mRNA-4157 for patients with resected solid tumors. See related commentary by Berraondo et al., p. 2021

Background The identification of cancer-specific T cell receptor (TCR) sequences is paramount to the advancement of cancer immunotherapies. Recent studies and clinical trials have shown that monoclonal T cell therapy is prone to immune evasion of cancer cells by loss of HLA heterozygosity and low antigen heterogeneity. Cocktail T cell therapy which comprises of TCRs corresponding to multiple HLAs and antigens has been proposed to improve the efficacy of adoptive cell transfer therapy. In addition to CD8+ cytotoxic T cells, neoantigen-specific CD4+ T cells, while identified as important for immunotherapy-induced anti-tumor responses, remain a largely untapped therapeutic resources due to the challenging nature of identification and isolation. Hence, a rapid and high-throughput discovery of both CD8+ and CD4+ TCRs against multiples Class I and II HLAs and cancer antigens is an urgent need. We engineered peptide-bound major histocompatibility complex (pMHC) proteins as capture agents for cancer-specific T cells. The design of these single-chain-trimers (SCTs) enables high-throughput multiplexing for identification and isolation of cancer-targeting CD4+ and CD8+ T cells from multiple patients against large panels of cancer antigens. We applied the technology to identify CD8+ and CD4+ TCRs against oncogenic proteins E6 and E7 from HPV-16, which is the leading cause of cervical cancer. Methods A panel of 200+ Class I SCTs and 100+ Class II SCTs were designed and expressed in a high-throughput platform. PBMCs from precancerous HPV-16+ patients with cervical lesions were collected and enriched with CD8+ and CD4+ T cells. A large pool of 200+ Class I SCT tetramer pool with barcode as antigen identifier was used to capture cancer-specific CD8+ T cells. A computational analysis pipeline was established to pair TCR α and β. HLA-matching cognate antigen was assigned to each TCR pair after UMI count correction and noise removal. The antigen-specific TCRs are subsequently sequenced, validated for functionality, and analyzed for therapeutic applications. Results We identified 43 CD8+ TCR pairs against E6 and E7 oncoproteins from HPV-16 and they are in progress for pre-clinical validation. Conclusions The SCT platform enables rapid identification of cancer-specific CD+ and CD4+ T cells and allows detailed characterization of anti-tumor T cells for which alternative solutions are extremely limited. We applied the technology to PBMCs extracted from HPV-16 related precancerous patients in a clinical trial and discovered cancer-specific TCRs. In summary, the application of the SCT technology is of high value to the fundamental and clinical immune-oncology studies.

Background Recent advances in the field of cancer immunotherapy have identified CD8 + T cell responses against tumor-specific mutations as a key driver of tumor regression and overall survival. ADXS-NEO is a personalized Listeria monocytogenes ( Lm )-based immunotherapy designed to target a patient’s mutation-derived tumor-specific neoantigens. The objective of this study is to demonstrate the feasibility of using the ADXS-NEO platform to target tumor-specific point mutations and control tumor growth by generating neoantigen-specific T cell responses using a pre-clinical mouse tumor model. Methods Whole-exome sequencing of the MC38 mouse tumor cell line identified 2870 unique non-synonymous mutations. The netMHCcons algorithm was used to predict 137 potential neoantigens. We validated 20 immunogenic neoantigens either by peptide immunization followed by ELISPOT or by the presence of CD8 + T cells recognizing the neoantigen peptide following checkpoint inhibitor treatment. Two ADXS-NEO vectors were constructed; Lm20, targeting 20 validated immunogenic neoantigens, and Lm19, targeting most of the non-validated NSMs. Results Both Lm19 & Lm20 significantly slowed tumor growth in C57BL/6 mice compared to control. An accumulation of ADXS-NEO-specific TILs was observed in tumor bearing mice treated with either Lm19 or Lm20. Examination of the tumor microenvironment in Lm19 or Lm20 treated mice revealed a decrease in the frequency and absolute number of Tregs, TAMs, MDSCs, and PD1 high exhausted CD8 + T cells as well as an increase in the frequency and absolute number of effector CD8 + T cells, relative to control. Conclusion ADXS-NEO is a potent immunotherapy capable of driving immune responses against tumor-specific mutations and leading to tumor control in mice.

Abstract Introduction: Neoantigens derived from tumor-specific mutations have been shown to drive tumor specific CD8+ T cell responses leading to tumor regression and extending overall survival. Frameshift mutations are estimated to generate up to nine times more neoantigens per mutation compared to in-frame mutations. However, it is not clear if vaccination against frameshift mutations induces neoantigen-specific CD8+ T cell responses that result in control of tumor growth. ADXS-NEO is a personalized Listeria monocytogenes (Lm)-based immunotherapy designed to target mutation-derived tumor-specific neoantigens. Advaxis' Lm-based immunotherapies consist of live attenuated bacterial vectors that are bioengineered to secrete an antigen-adjuvant fusion protein consisting of a truncated non-hemolytic fragment of listeriolysin O, which has adjuvant properties, and tumor-specific antigens. Here, we demonstrate the feasibility of using the ADXS-NEO platform to target tumor-specific frameshift mutations in order to generate neoantigen-specific T cells that control tumor growth. Results: Whole-exome sequencing of the CT26 and MC38 mouse tumor cell lines identified 30 and 31 unique frameshift mutations respectively. Individual frameshift mutations ranged in size from 12 to as many as 150 amino acids (aa). Lm vectors targeting the two longest frameshift mutations were constructed for each tumor model. The therapeutic efficacy of Lm vectors expressing either a single 57 aa (Lm-57) or a single 150 aa (Lm-150) MC38 frameshift mutation were evaluated in C57BL/6J mice. Both Lm vectors generated multiple unique frameshift-specific TILs and slowed tumor growth. Furthermore, we evaluated the tumor microenvironment following Lm-57 or Lm-150 treatment and observed a decrease in the frequency and absolute number of Tregs, TAMs, and MDSCs and an increase in the frequency and absolute number of total cytotoxic granzyme A+ effector CD8+ T cells. Similarly, Lm vectors expressing either a 64 aa (Lm-64) or a 93 aa (Lm-93) CT26 frameshift mutation were evaluated in the CT26 tumor model. Both Lm-64 and Lm-93 significantly controlled tumor growth. Additionally, an influx of neoantigen-specific TILs and a significant decrease in the frequency of intratumoral Tregs was observed. Conclusion: ADXS-NEO induced potent immune responses against tumor-specific frameshift mutations and controlled tumor growth. Advaxis' Lm platform is able to target frameshift mutations ≥150 aa and generate multiple neoantigen-specific T cells per frameshift. ADXS-NEO controls tumor growth via multiple mechanisms, including the generation of tumor-specific cytotoxic TILs, by secreting tumor-derived neoantigens directly into dendritic cells and by attenuating the suppressive tumor microenvironment. Citation Format: Brandon Coder, Daniel O. Villarreal, Susan Armington, Elena Filippova, Andrew L'Huillier, Dipti Kelkar, Xiaoming Ju, Cristina Mottershead, David Balli, Kim Ramos, Hyewon Phee, Jim Johnston, Robert Petit, Michael Princiotta. Targeting frameshift mutations with a Listeria monocytogenes immunotherapy drives neoantigen-specific antitumor immunity in the MC38 and CT26 mouse tumor models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-148.

Abstract Introduction: Recent advances in the field of cancer immunotherapy have identified CD8+ T cell responses against tumor-specific neoantigens as a key driver of tumor regression and prolonged survival. ADXS-NEO is a personalized Listeria monocytogenes (Lm)-based immunotherapy designed to generate immune responses against mutation-derived tumor-specific neoantigens. Advaxis' Lm-based immunotherapies consist of live highly-attenuated bacterial vectors that are bioengineered to secrete a fusion protein consisting of a truncated non-hemolytic fragment of listeriolysin O, which has adjuvant properties, and tumor-specific neoantigens that harbor nonsynonymous point mutations (NSMs). The objective of this study is to demonstrate the feasibility of using the ADXS-NEO platform to target tumor-specific point mutations to generate neoantigen-specific T cells and control tumor growth. Results: Whole-exome sequencing of the MC38 mouse tumor cell line identified 2870 unique NSMs. Among these, the IC50 of 138 NSMs were predicted to be less than 500 nM by the netMHCcons algorithm. We evaluated the immunogenicity of 37 NSMs, and found that 12 immunogenic NSMs elicited a CD8+ T cell response following peptide immunization. Moreover, we identified 10 additional immunogenic NSMs in MC38-bearing mice treated with a check point inhibitor. Altogether, we identified 22 immunogenic and 23 non-immunogenic NSMs. Two ADXS-NEO vectors were constructed, Lm-19 & Lm-20, targeting 19 non-immunogenic and 20 immunogenic NSMs respectively. The ability of Lm-19 and Lm-20 to control MC38 tumor growth was evaluated in C57BL/6J mice. We found that both Lm-19 & Lm-20 led to an accumulation of neoantigen-specific CD8+ TILs and significantly slowed tumor growth. Moreover, both Lm-19 and Lm-20 decreased the frequency and absolute number of intratumoral Tregs, TAMs, and MDSCs and increased the frequency and absolute number of effector CD8+ T cells. Interestingly, expression of PD-L1 was decreased in TAMs and MDSCs and the frequency and total number of granzyme A+ CD8+ effector T cells was increased. Furthermore, the proportion of phenotypically exhausted PD-1hiLAG3+ TILs was decreased. Together, these data suggest the tumor microenvironment in mice receiving Lm-19 and Lm-20 becomes more cytotoxic and less suppressive. Conclusion: ADXS-NEO is a potent immunotherapy capable of driving immune responses against tumor-specific mutations and leading to tumor control. The effectiveness of the Lm platform is demonstrated by the generation of neoantigen-specific T cells to peptide sequences that were identified as “non-immunogenic” using a conventional peptide-adjuvant immunization. This study is a clear demonstration that T cell mediated anti-tumor responses can be generated by targeting tumor-derived NSMs with the ADXS-NEO Listeria monocytogenes vector. Citation Format: Brandon Coder, Hyewon Phee, Cristina Mottershead, Dipti Kelkar, Elena Filippova, Xiaoming Ju, Bryan Vander Lugt, Olga Pryshchep, Justin Lesch, Xian Liu, Jason DeVoss, Keegan Cooke, Claret Liu, Jinghui Zhan, Petia Mitchell, Kim Ramos, Daniel O. Villarreal, Jim Johnston, Robert Petit, Michael Princiotta. Neoantigens that fail to elicit measurable T cell responses following peptide immunization can control tumor growth when delivered using a Listeria-based immunotherapy platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-150.

Abstract Introduction: Virtually all tumors contain somatic mutations that can result in novel antigenic sequences that may be targeted by the host cellular immune response. Some of these mutations occur in preferential regions of specific genes commonly referred to as hotspot mutations. Hotspot mutations are commonly shared by cancer patients both within and across multiple tumor types. These hotspot mutations often confer loss or gain of function contributing to oncogenesis, which makes them promising therapeutic targets. One such mutation commonly found in several human tumor types is an aspartic acid substitution for glycine at position 12 (G12D) in KRAS. This same mutation occurs in the CT26 murine colorectal tumor model. To determine if expression of the KRAS G12D sequence in a bacterial immunotherapy vector can control tumor growth in the CT26 murine model, the Advaxis Listeria monocytogenes (Lm)-based platform was engineered to express a 21-amino acid KRAS sequence peptide containing the G12D mutation (Lm-Hot KRAS_G12D). In addition, we evaluated control of tumor growth using an ADXS-HOT construct (ADXS-503) that expresses multiple shared human hotspot and tumor-associated antigens, including the G12D KRAS. The ADXS-HOT clinical program is comprised of several Lm-based immunotherapies designed to target multiple shared hotspot and tumor-associated antigens commonly found in specific cancer types. In this study, we demonstrate control of tumor growth in a mouse model by targeting a commonly shared hotspot mutation using an Lm-based immunotherapy. Results: We show that the Lm-HOT KRAS_G12D therapy significantly delayed tumor growth and improved long-term survival in the murine CT26 colon carcinoma model. This response was associated with an increase in the frequency of tumor infiltrating antigen-specific CD8 T cells and γδ T cells within the tumor microenvironment and a decrease in the frequency of intratumoral regulatory T cells (Tregs). Furthermore, tumor-specific CD8 T cells displayed lower expression of exhaustion markers as well as increased functionality upon restimulation. Interestingly, our proprietary ADXS-503 (a clinical ADXS-HOT construct) which includes KRAS G12D as one of its multiple targets, was also capable of significantly suppressing tumor growth in the CT26 tumor model. Conclusion: These results suggest that our ADXS-HOT platform is a promising approach to target shared hotspot mutations. That ADXS Lm constructs targeting a single hotspot mutation can significantly control tumor growth whether it is in a single or multi-target construct. These data describe an exciting translatable discovery with the potential for broad utility across multiple tumor types and patients who share common hotspot mutations. Citation Format: Daniel Villarreal, Brandon Coder, Susan Armington, Andrew L'Huillier, Cristina Mottershead, Elena Filippova, Nithya Thambi, Kim Ramos, David Balli, Robert Petit, Michael Princiotta. Targeting shared hotspot cancer mutations with a Listeria monocytogenes immunotherapy induce potent anti-tumor immunity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr LB-149.

### P189 Rational combinations of intratumoral T cell and myeloid agonists mobilize abscopal responses in prostate cancer #### Casey Ager1, Matthew Reilley2, Courtney Nicholas1, Todd Bartkowiak1, Ashvin Jaiswal1, Michael Curran1 ##### 1Department of Immunology, University of Texas MD Anderson