ALBERT

Description: Cutaneous T cell lymphoma (CTCL) is a CD4+ T cell malignancy of the skin with heterogeneous outcomes and limited treatment options. Monoclonal antibodies directed against PD-1, such as pembrolizumab, have shown impressive efficacy in multiple advanced malignancies, and are currently tested in clinical trials in patients with CTCL. Initial data indicate that about half of the patients experience treatment response, whereas the other half are non-responders. Non-responders can be further divided into patients with stable disease versus rapid progressors. It is currently unknown why some CTCL patients respond to pembrolizumab while others rapidly progress, and no predictive biomarkers are available. Single-cell analysis approaches to identify biomarkers of response, for example quantifying the expression of PD-1 on tumor cells vs. reactive immune cells, have not enabled stratification of patients. We therefore hypothesized that more complex spatial cellular interactions within the immune tumor microenvironment (iTME) of CTCL could provide insight into the mechanisms of pembrolizumab response and enable prediction. We applied CODEX (CO-Detection by indEXing) highly multiplexed tissue imaging to study the CTCL iTME in matched biopsies before and after pembrolizumab therapy in 7 responders and 7 non-responders (see the Figure). Using 54 markers simultaneously allowed discriminating malignant CD4+ tumor cells from reactive CD4+ T cells and identified 30 different cell clusters with spatial information, including an M2 macrophage cluster that was enriched in non-responders before therapy. Unexpectedly, in pembrolizumab responders compared to non-responders, PD-1 expression levels were higher in multiple clusters of tumor cells and reactive T cells. Computational spatial analysis revealed ten distinct, conserved cellular neighborhoods in the CTCL iTME that changed in composition and frequency during therapy. Interestingly, one cellular neighborhood to be presented dramatically increased after therapy only in responders. Therefore, highly multiplexed spatial analysis of the CTCL iTME allows discovering novel, predictive biomarkers of immunotherapy response and will pave the way for future studies that functionally address the identified cell types and cellular interactions. Disclosures Khodadoust: Corvus Pharmaceuticals: Research Funding. Kim:miRagen: Research Funding; Merck: Research Funding; Medivir: Honoraria, Membership on an entity's Board of Directors or advisory committees; Innate Pharma: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Soligenix: Research Funding; Forty Seven Inc: Research Funding; Neumedicine: Research Funding; Takeda: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Trillium: Research Funding; Elorac: Research Funding; Galderma: Research Funding; Corvus: Honoraria, Membership on an entity's Board of Directors or advisory committees; Portola Pharmaceuticals: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Seattle Genetics: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Kyowa Hakko Kirin: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Eisai: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Horizon: Research Funding. Nolan:Akoya Biosciences Inc.: Consultancy, Equity Ownership, Patents & Royalties.

Description: The Wilms' tumor (WT1) gene participates in leukemogenesis and is overexpressed in most types of leukemia in humans. WT1 is also detectable in many types of lung, thyroid, breast, testicular, and ovarian cancers and melanoma in humans. Initial studies evaluated whether immune responses to murine WT1 can be elicited in mice. Murine and human WT1 are similar. Thus, mouse models might lead to resolution of many of the critical issues for developing WT1 vaccines. C57/BL6 (B6) mice were injected with synthetic peptides from the natural sequence of WT1 containing motifs for binding to major histocompatibility (MHC) class II molecules. Immunization induced helper T-cell responses specific for the immunizing WT1 peptides and antibody responses specific for WT1 protein. Screening of multiple murine cancer cell lines identified 2 murine cancers, TRAMP-C and BLKSV40, that “naturally” overexpress WT1. Immunization with MHC class I binding peptides induced WT1 peptide-specific cytotoxic T-lymphocyte (CTL) that specifically lysed TRAMP-C and BLKSV40. WT1 specificity of lysis was confirmed by cold target inhibition. No toxicity was noted by histopathologic evaluation in the WT1 peptide-immunized animals. WT1 peptide immunization did not show any effect on TRAMP-C tumor growth in vivo. Immunization of B6 mice to syngeneic TRAMP-C elicited WT1-specific antibody, demonstrating that WT1 can be immunogenic in the context of cancer cells. To evaluate whether WT1 might be similarly immunogenic in humans, serum from patients with leukemia was evaluated for pre-existing antibody responses. Western blot analyses showed WT1-specific antibodies directed against the N-terminus portion of the WT1 protein in the sera of 3 of 18 patients with acute myeloid leukemia (AML).

Description: Background:PD-1 and PD-L1/PD-L2 are expressed by malignant T-cells in mycosis fungoides (MF) and Sézary syndrome (SS). PD-1 is additionally expressed by tumor-infiltrating cytotoxic T-cells and PD-L1 is expressed by macrophages and other stromal components of the tumor microenvironment in these diseases. Moreover, reports of 9p24.1/PD-L2 translocation and CTLA4-CD28 fusion events in MF/SS support a genomic basis for immune evasion. Here, we explore the clinical activity of pembrolizumab, an immune checkpoint inhibitor of the PD-1/PD-L1 axis, in MF/SS. Methods:Patients (pts) with MF/SS stages IB-IV treated with at least 1 prior systemic therapy were enrolled in this phase 2, single-arm study coordinated by the Cancer Immunotherapy Trials Network (CITN). A Simon two-stage design was applied where stage 2 is initiated if 1 of 9 pts had an objective response. An additional 15 pts were planned in stage 2. Pembrolizumab was administered at 2 mg/kg every 3 weeks and treatment was allowed up to 2 years. The primary endpoint was overall response rate (ORR) as determined by the consensus global response criteria. Secondary endpoints were safety/tolerability, time to response (TTR), duration of response (DOR) and progression-free survival (PFS). Correlative biomarker studies included immunohistochemistry (IHC) staining for PD-L1, PD-L2, and multiple immune subsets as well as serum analysis of 62 cytokines and chemokines. Phenotypic and functional profiling of malignant and non-malignant immune cells will be performed by flow cytometry and mass cytometry (CyTOF). Results: The study completed enrollment and all 24 patients received at least one dose of pembrolizumab. Median age was 67 (range 44-85); 18 were male. Patients were advanced stage with 23 patients (96%) stage IIB or higher, including 15 patients (63%) with stage IVA SS. Most pts were heavily treated with a median of 4 prior systemic therapies (range 1-11). The median follow-up time was 40 weeks (range 9-60 weeks). The objective response rate (ORR) was 38% with 1 complete response (CR) and 8 partial responses (PR). Of the responding pts, 6 pts had 90% or greater improvement in skin disease as measured by mSWAT. An additional 9 pts (38%) had stable disease (SD). The median TTR was 11 weeks (range 8-41 weeks). Responses were durable with 8 of 9 (89%) responses currently ongoing at a median of 32 weeks of duration (4-46). The median PFS has not yet been reached, and the one-year PFS was 69%. There was no significant association between response and clinical characteristics including stage, disease type (MF vs. SS), and number of prior therapies, nor with skin tissue expression of PD-1, PD-L1, PD-L2, or infiltrating CD8+ T-cells as determined by IHC. Planned additional correlatives including CyTOF profiling, gene expression profiling, T cell receptor high throughput sequencing, multiplexed ion beam imaging (MIBI), and whole exome sequencing will explore potential predictive biomarkers of response. Adverse events (AE) were consistent with those seen in prior studies of pembrolizumab with the exception of an immune-mediated skin flare reaction seen in 6 pts (2 grade 2 and 4 grade 3). Skin flares occurred exclusively in patients with SS (6/15; 40%) and were associated with lower serum levels of the cytokines IL-7 and SCF prior to pembrolizumab treatment (p=0.01 and p=0.02 respectively, n.s. by Bonferroni correction). Pts with the skin flare reaction experienced increases in serum IFN-gamma, IL-12p40, IL-15, LIF, G-CSF, and CCL4 following treatment. There were two treatment related serious adverse events (SAE), both immune related. One pt experienced grade 2 pneumonitis which resolved with systemic corticosteroids. Another patient experienced grade 3 diarrhea secondary to steroid-refractory duodenitis. Conclusions: Pembrolizumab has significant clinical activity in pts with previously treated MF/SS. Responses were durable and were not associated with any identifiable clinical or pathologic characteristics. Treatment was well tolerated with a toxicity profile consistent with prior pembrolizumab studies, though 40% of pts with SS developed a notable skin flare reaction. These findings support further study of PD-1 blockade in the treatment of MF and SS. A phase 2 trial of pembrolizumab in combination with interferon-gamma is being developed based on these results. Disclosures Porcu: Millenium: Other: investigator in a clinical trial; miRagen: Other: Investigator in a clinical trial; celgene: Other: Investigator in a clinical trial; Innate Pharma: Other: Investigator in a clinical trial. Foss:Celgene: Consultancy, Research Funding, Speakers Bureau; Eisai: Consultancy; Seattle Genetics: Consultancy, Speakers Bureau; Spectrum Pharmaceuticals: Consultancy. Moskowitz:Bristol Myers Squibb: Honoraria; Merck: Honoraria; Seattle Genetics: Honoraria, Research Funding. Sokol:Seattle Genetics: Consultancy; Spectrum: Consultancy. Yearley:Merck: Employment. Chartash:Merck: Employment. Townson:Merck: Employment. Horwitz:Spectrum: Consultancy, Research Funding; Bristol-Myers Squibb: Consultancy; Huya: Consultancy; Infinity: Consultancy, Research Funding; Kyowa Hakka Kirin: Consultancy, Research Funding; Takeda: Consultancy, Research Funding; Seattle Genetics: Consultancy, Research Funding; Celgene: Consultancy; ADCT Therapeutics: Research Funding. Kim:Seattle Genetics: Consultancy, Other: Investigator in a clinical trial; Merck: Other: Investigator in a clinical trial; Neumedicine: Consultancy; Soligenix: Consultancy; Galderma: Consultancy; Genentech: Other: Investigator in a clinical trial; Innate Pharma: Other: Investigator in a clinical trial; Kyowa Hakko Kirin: Consultancy, Honoraria, Other, Research Funding; Millenium: Consultancy, Other: Investigator in a clinical trial; Eisai: Consultancy, Other: Investigator in a clinical trial; Actelion: Consultancy, Other: Investigator in a clinical trial; Celgene: Consultancy; MiRagen: Consultancy; Horizon: Consultancy.

Description: Background: Current treatment of advanced stage Mycosis fungoides (MF) and Sézary syndrome (SS) remains unsatisfactory. Complete responses (CR) are typically

Description: The Wilms' tumor (WT1) gene participates in leukemogenesis and is overexpressed in most types of leukemia in humans. WT1 is also detectable in many types of lung, thyroid, breast, testicular, and ovarian cancers and melanoma in humans. Initial studies evaluated whether immune responses to murine WT1 can be elicited in mice. Murine and human WT1 are similar. Thus, mouse models might lead to resolution of many of the critical issues for developing WT1 vaccines. C57/BL6 (B6) mice were injected with synthetic peptides from the natural sequence of WT1 containing motifs for binding to major histocompatibility (MHC) class II molecules. Immunization induced helper T-cell responses specific for the immunizing WT1 peptides and antibody responses specific for WT1 protein. Screening of multiple murine cancer cell lines identified 2 murine cancers, TRAMP-C and BLKSV40, that “naturally” overexpress WT1. Immunization with MHC class I binding peptides induced WT1 peptide-specific cytotoxic T-lymphocyte (CTL) that specifically lysed TRAMP-C and BLKSV40. WT1 specificity of lysis was confirmed by cold target inhibition. No toxicity was noted by histopathologic evaluation in the WT1 peptide-immunized animals. WT1 peptide immunization did not show any effect on TRAMP-C tumor growth in vivo. Immunization of B6 mice to syngeneic TRAMP-C elicited WT1-specific antibody, demonstrating that WT1 can be immunogenic in the context of cancer cells. To evaluate whether WT1 might be similarly immunogenic in humans, serum from patients with leukemia was evaluated for pre-existing antibody responses. Western blot analyses showed WT1-specific antibodies directed against the N-terminus portion of the WT1 protein in the sera of 3 of 18 patients with acute myeloid leukemia (AML).