Faculty Bibliography

The tumor microenvironment (TME) in pancreatic ductal adenocarcinoma (PDAC) is a complex ecosystem that drives tumor progression; however, in-depth single cell characterization of the PDAC TME and its role in response to therapy is lacking. Here, we perform single-cell RNA sequencing on freshly collected human PDAC samples either before or after chemotherapy. Overall, we find a heterogeneous mixture of basal and classical cancer cell subtypes, along with distinct cancer-associated fibroblast and macrophage subpopulations. Strikingly, classical and basal-like cancer cells exhibit similar transcriptional responses to chemotherapy and do not demonstrate a shift towards a basal-like transcriptional program among treated samples. We observe decreased ligand-receptor interactions in treated samples, particularly between TIGIT on CD8 + T cells and its receptor on cancer cells, and identify TIGIT as the major inhibitory checkpoint molecule of CD8 + T cells. Our results suggest that chemotherapy profoundly impacts the PDAC TME and may promote resistance to immunotherapy.

Hepato-pancreatico-biliary (HPB) malignancies are difficult-to-treat and continue to to have a high mortality and significant therapeutic resistance to standard therapies. Immune oncology (IO) therapies have demonstrated efficacy in several solid malignancies when combined with chemotherapy, whereas response rates in pancreatic ductal adenocarcinoma (PDA) are poor. While promising in hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA), there remains an unmet need to fully leverage IO therapies to treat HPB tumors. We therefore defined T cell subsets in the tumor microenvironment of HPB patients utilizing a novel, multiparameter flow cytometry and bioinformatics analysis. Our findings quantify the T cell phenotypic states in relation to checkpoint receptor expression. We demonstrate the presence of CD103⁺ tissue resident memory T cells (T_RM), CCR7⁺ central memory T cells, and CD57⁺ terminally differentiated effector cells across all HPB cancers, while the anti-tumor function was dampened by expression of multiple co-inhibitory checkpoint receptors. Terminally exhausted T cells lacking co-stimulatory receptors were more prevalent in PDA, whereas partially exhausted T cells expressing both co-inhibitory and co-stimulatory receptors were most prevalent in HCC, especially in early stage. HCC patients had significantly higher T_RM with a phenotype that could confer restored activation in response to immune checkpoint therapies. Further, we found a lack of robust alteration in T cell activation state or checkpoint expression in response to chemotherapy in PDA patients. These results support that HCC patients might benefit most from combined checkpoint therapies, whereas efforts other than cytotoxic chemotherapy will likely be necessary to increase overall T cell activation in CCA and PDA for future clinical development.

Background Solid tumors comprise >90% of cancers. Nonsmall cell lung cancer (NSCLC), metastatic colorectal cancer (CRC), and pancreatic cancer are the leading causes of cancerrelated mortality (5-year overall survival: 26%, 15%, and 11%, respectively).1 Chimeric antigen receptor (CAR) T-cell therapy has demonstrated clinical efficacy in hematologic malignancies.2,3 However, translating engineered T-cell therapies to solid tumors has proven to be challenging due to a lack of tumor-specific targets that can discriminate cancer cells from normal cells. Previous studies using carcinoembryonic antigen (CEA) T-cell receptors and mesothelin (MSLN) CARs resulted in dose-limiting on-target, off-tumor toxicities.4,5 To create a therapeutic safety window, Tmod CAR T-cell therapy utilizes dual-signaling receptors to create a robust logic gate capable of killing tumor cells, while leaving healthy cells intact.6,7 The 2 receptors in Tmod CAR T-cell therapy comprise an activator that recognizes an antigen on the surface of tumor cells that may also be present on normal cells, such as CEA and MSLN, and a blocker that recognizes a second surface antigen from an allele lost only in tumor cells (figure 1).8,9 Human leukocyte antigen (HLA) loss of heterozygosity (LOH) offers a definitive tumor versus normal discriminator target for CAR T-cell therapy.10 The frequency of HLA LOH among advanced NSCLC, CRC, and pancreatic cancers in the Tempus real-world dataset is 16.3% with a range of 15.6%- 23.1%.11 LOH can be reliably detected using the Tempus xTOnco next-generation sequencing (NGS) assay.12,13 Different activator/blocker combinations can be engineered with the Tmod platform technology and may be applied to T cells and natural killer cells in autologous and allogeneic settings. BASECAMP-1 is a currently enrolling observational study with key objectives: 1) To identify patients with somatic HLA LOH eligible for Tmod CAR T-cell therapy, and 2) Subsequent apheresis and manufacturing feasibility for the future EVEREST CEA or MSLN Tmod CAR T-cell studies. Methods BASECAMP-1 (NCT04981119) patient eligibility has 2 parts (figure 2): 1) Patients will be initially screened to identify germline HLA-A*02 heterozygosity by central NGS. If HLA-A*02 heterozygosity is confirmed, primary archival tumor tissue will be analyzed for somatic mutations by xTOnco NGS testing; 2) If the tumor demonstrates HLAA* 02:01 LOH and the patient is eligible after screening, the patient will undergo apheresis. Banked T cells will be available for the autologous EVEREST Tmod CAR T-cell therapy interventional study to reduce waiting time at relapse. (Figure Presented)

Background Chimeric antigen receptor (CAR) T-cell therapy has shown clinical efficacy in hematologic cancers, but success is limited in solid tumors due to a lack of tumor-specific targets that distinguish cancer from normal cells and an immunosuppressive tumor microenvironment.1 Integrating synthetic biology and comprehensive molecular profiling of tumors may provide active and tolerable approaches to CAR T-cell therapy in patients with solid tumors. Human leukocyte antigen (HLA) loss of heterozygosity (LOH) in tumors offers a definitive tumor vs normal discriminator target for CAR T-cell therapy.2 The Tmod platform3,4 is a modular logic-gated CAR T system comprising different versions including a carcinoembryonic antigen (CEA)- or mesothelin (MSLN)-targeting CAR activator and a separate blocker receptor targeting HLA-A*02 or other HLA alleles to protect normal cells. Compared with existing immunohistochemistry (IHC) tests, Tempus xT-Onco is a standard-of-care next-generation sequencing (NGS) assay5 that detects somatic alterations including HLA LOH and generates whole transcriptome RNA data (eg, CEA or MSLN expression) and a tumor immune infiltration profile, which can effectively identify patients appropriate for Tmod CAR T-cell therapy. Methods HLA LOH in solid tumors was assessed with paired germline and somatic DNA sequencing. Common driver mutations, microsatellite instability status, and tumor mutational burden were examined in HLA-A LOH or HLA-A intact cohorts. Tumor expression of CEA and MSLN was evaluated via RNA sequencing and compared with immunohistochemistry (IHC) results. Results A total of 21,053 tumor samples in the Tempus database were compared with their matched-normal samples. HLA-A LOH was detected in 16% of 10,867 advanced solid tumors (table 1) and similar LOH frequencies were observed among common HLA-A alleles. Clinical factors and molecular biomarkers were similar between HLA-A LOH and HLA-A intact cohorts. High CEA expression was seen in IHC-positive patients. Conclusions The frequency of HLA-A LOH in solid tumors in the Tempus database is similar to that reported in the Cancer Genome Atlas.6 Tempus xT-Onco reliably detects HLA LOH and quantifies CEA and MSLN expression. Based on these data, patients with solid tumors are now being prospectively screened for HLA LOH using xT-Onco in an ongoing tissue banking study (BASECAMP-1, NCT04981119), preparing for future interventional protocols

Background Nearly all colorectal and most pancreatic and lung cancers express carcinoembryonic antigen (CEA). However, due to its expression in normal gut epithelial cells, CEAtargeted therapies have resulted in on-target, off-tumor toxicity. To overcome this, we have developed TmodTM, a logicgated T-cell therapy platform. Tmod constructs are composed of an activating CAR or T-cell receptor that targets a tumor antigen and an inhibitory receptor recognizing an antigen expressed on normal healthy tissues, but not on tumor cells due to loss of heterozygosity (LOH).1,2 A2B530 is a CEAdirected Tmod construct utilizing an LIR-1-based inhibitory receptor (blocker) targeting human leukocyte antigen A*02 (HLA-A*02). Methods To generate CEA Tmod, T cells from HLA-A*02(+) donors were transduced with a single lentivirus to express i) the CAR, ii) the blocker, and iii) an shRNA targeting b2M. Cytotoxicity was measured by culturing CEA(+) target cell line pairs (A*02[-] and A*02[+]), expressing either GFP or RFP, with engineered T cells and quantifying live target cells over time. In vivo activity was examined using NSG mice subcutaneously implanted with normal (CEA[+]A*02[+]) and tumor cells (CEA[+]A*02[-]), in the right and left flanks. Mice were treated intravenously with CEA Tmod cells or control T cells. Results Control CEA CAR T cells killed CEA(+) target cell lines in vitro irrespective of HLA-A*02 expression. In contrast, CEA Tmod cells selectively killed tumor cells (CEA[+]A*02[-]) while sparing normal cells (CEA[+]A*02[+]). In mixed target cell cultures, CEA Tmod cells killed only the A*02(-) target cells, whereas the CEA CAR T cells killed both the A*02(- ) and A*02(+) cell lines. Further, CEA Tmod cells exhibited bidirectional control between the activated and blocked states. While mice treated with control CEA CAR T cells experienced a reduction in volume and bioluminescence of both normal and tumor grafts, CEA Tmod cells specifically cleared A*02(-) tumors in mice (table 1). Finally, although expansion of Tmod cells in peripheral blood trended lower than CAR and TCR controls, anti-tumor activity was comparable in these groups. Conclusions A2B530 is an autologous CEA Tmod cell product that exploits common LOH at the HLA locus in cancer cells, enabling these engineered T cells to discriminate between normal and tumor cells. BASECAMP-1 (NCT04981119), an observational study identifying patients with somatic HLA LOH, is recruiting. Eligible patients with metastatic colorectal, pancreatic, or non-small cell lung cancer will be apheresed for a future A2B530 EVEREST-1 interventional study

Background Mesothelin (MSLN) is expressed on a variety of solid tumors, including mesothelioma and ovarian, uterine, gastric, pancreatic, and lung cancers.1 However, efforts to target MSLN using cellular therapies have been hampered by severe on-target, off-tumor toxicities associated with damage to normal tissues expressing MSLN.2 To avoid these toxicities, we have developed a logic-gated engineered cell therapy, TmodTM, which is composed of two chimeric antigen receptors (CARs): an activator that targets a tumor-associated antigen and an inhibitory receptor (blocker) gated by an antigen expressed on normal tissue but lost in tumor cells due to loss of heterozygosity (LOH). A2B694 is an MSLN-specific Tmod construct combining a third-generation MSLN CAR with an LIR-1-based inhibitory receptor specific for human leukocyte antigen A*02 (HLA-A*02). Methods Lentivirus encoding i) the CAR, ii) the blocker, and iii) an shRNA targeting b2M was used to transduce T cells from HLA-A*02 donors and generate MSLN Tmod cells. In vitro cytotoxicity measurements were performed using fluorescence-based imaging and luciferase readouts. In vivo assessments were performed in NSG mice subcutaneously implanted with normal cells (MSLN[+]A*02[+]), or tumor cells (MSLN[+]A*02[-]), in the left and right flanks, respectively. Following engraftment, mice were randomized and treated intravenously with MSLN Tmod cells or controls. Grafts were measured via caliper. Results MSLN Tmod cells preferentially killed tumor cells (MSLN[+]A*02[-]) over normal cells (MSLN[+]A*02[+]) in vitro, unlike clinically active comparator M5 CAR T cells, which indiscriminately killed both target cell types (figure 1A). Soluble MSLN, tested across a 0-2 mg/mL range, did not impact MSLN Tmod function. Additionally, in mixed cell cultures where T cells and tumor and normal cells were simultaneously cultured (1:1:1 ratio), MSLN Tmod cells selectively killed tumor targets while sparing normal cells. Further, MSLN Tmod cells cycled between activated and blocked states in vitro when repeatedly challenged with tumor or normal target cells. Finally, while MSLN CAR T cells killed both normal and tumor grafts in vivo, MSLN Tmod cells selectively killed tumor grafts while sparing normal grafts (figure 1B, C). Conclusions A2B694 is an autologous MSLN Tmod cell product that leverages LOH at the HLA locus in cancer cells, providing a mechanism to discriminate between normal and tumor cells. BASECAMP-1 (NCT04981119), an observational study that will identify patients with somatic HLA LOH, is currently recruiting. Eligible patients with metastatic colorectal, pancreatic, or non-small cell lung cancer will be apheresed for a future A2B694 interventional study (EVEREST-2)

Background Previously reported single institution data on family history of pancreatic adenocarcinoma (PDAC) showed that most individuals with a germline pathogenic or likely pathogenic variant (PV/LPV) in a PDAC susceptibility gene who were diagnosed with PDAC would not have met current recommendations for PDAC surveillance established by the National Comprehensive Cancer Network, the American College of Gastroenterology, or International Cancer of the Pancreas Screening Consortium. These recommendations rely on the assumption that PV/LPV carriers with family history of PDAC are at greater risk for developing PDAC as compared to carriers without a family history. This study is a multi-site collaboration to validate the previous findings. Methods Individuals with PDAC who had a germline PV/LPV in ATM, BRCA1, BRCA2, EPCAM, MLH1, MSH2, MSH6, PALB2, or PMS2 were assessed for family history of PDAC in first- (FDR) or second-degree relatives (SDR). A comparison group of individuals with PDAC who had no germline PV/LPV identified through multigene panel testing was also assessed. Chi-square and t-tests were used to determine statistical significance. Results Nine institutions compiled a cohort of 196 individuals with PDAC who had a germline PV/LPV in one of the aforementioned genes. See Table 1 for demographics. Fifty (25.5%) had an FDR and/ or SDR affected by PDAC and 146 (74.5%) had no family history of PDAC. The cohort was significantly more likely to have a PDACaffected FDR or SDR than individuals with PDAC who had no germline PV/LPV (p = 0.004). Significance was also reached for affected FDR alone (p = 0.003), but not for affected SDR alone (p = 0.344). See Table 2. Conclusions This multi-site study confirms that most individuals with PDAC and a PV/LPV in ATM, BRCA1, BRCA2, EPCAM, MLH1, MSH2, MSH6, PALB2, or PMS2 would not meet current pancreatic cancer surveillance recommendations because they do not have family history of PDAC. Family history, particularly an affected FDR, enriches the cohort but alone is insufficient in identifying the majority of high-risk individuals who are at risk for developing PDAC. (Table Presented)

Pancreatic ductal adenocarcinoma (PDAC) is one of the most aggressive malignancies, with a dismal survival rate. Screening the general population for early detection of PDAC is not recommended, but because early detection improves survival, high-risk individuals, defined as those meeting criteria based on a family history of PDAC and/or the presence of known pathogenic germline variant genes with PDAC risk, are recommended to undergo screening with MRI and/or endoscopic ultrasound at regular intervals. The Pancreatic Cancer Early Detection (PRECEDE) Consortium was formed in 2018 and is composed of gastroenterologists, geneticists, pancreatic surgeons, radiologists, statisticians, and researchers from 40 sites in North America, Europe, and Asia. The overarching goal of the PRECEDE Consortium is to facilitate earlier diagnosis of PDAC for high-risk individuals to increase survival of the disease. A standardized MRI protocol and reporting template are needed to enhance the quality of screening examinations, improve consistency of clinical management, and facilitate multiinstitutional research. We present a consensus statement to standardize MRI screening and reporting for individuals with elevated risk of pancreatic cancer.

The stem-cell-like behavior of cancer cells plays a central role in tumor heterogeneity and invasion and correlates closely with drug resistance and unfavorable clinical outcomes. However, the molecular underpinnings of cancer cell stemness remain incompletely defined. Here, we show that SNHG1, a long non-coding RNA that is over-expressed in ~95% of human muscle-invasive bladder cancers (MIBCs), induces stem-cell-like sphere formation and the invasion of cultured bladder cancer cells by upregulating Rho GTPase, Rac1. We further show that SNHG1 binds to DNA methylation transferase 3A protein (DNMT3A), and tethers DNMT3A to the promoter of miR-129-2, thus hyper-methylating and repressing miR-129-2-5p transcription. The reduced binding of miR-129-2 to the 3'-UTR of Rac1 mRNA leads to the stabilization of Rac1 mRNA and increased levels of Rac1 protein, which then stimulates MIBC cell sphere formation and invasion. Analysis of the Human Protein Atlas shows that a high expression of Rac1 is strongly associated with poor survival in patients with MIBC. Our data strongly suggest that the SNHG1/DNMT3A/miR-129-2-5p/Rac1 effector pathway drives stem-cell-like and invasive behaviors in MIBC, a deadly form of bladder cancer. Targeting this pathway, alone or in combination with platinum-based therapy, may reduce chemoresistance and improve longer-term outcomes in MIBC patients.