Schenk, Erin L.

Abstract Aberrant activation of GLI transcription factors has been implicated in the pathogenesis of different tumor types including pancreatic ductal adenocarcinoma. However, the mechanistic link with established drivers of this disease remains in part elusive. In this study, using a new genetically engineered mouse model overexpressing constitutively active mouse form of GLI2 and a combination of genome-wide assays, we provide evidence of a novel mechanism underlying the interplay between KRAS, a major driver of pancreatic ductal adenocarcinoma development, and GLI2 to control oncogenic gene expression. These mice, also expressing KrasG12D, show significantly reduced median survival rate and accelerated tumorigenesis compared with the KrasG12D only expressing mice. Analysis of the mechanism using RNA sequencing demonstrate higher levels of GLI2 targets, particularly tumor growth–promoting genes, including Ccnd1, N-Myc, and Bcl2, in KrasG12D mutant cells. Furthermore, chromatin immunoprecipitation sequencing studies showed that in these cells KrasG12D increases the levels of trimethylation of lysine 4 of the histone 3 (H3K4me3) at the promoter of GLI2 targets without affecting significantly the levels of other major active chromatin marks. Importantly, Gli2 knockdown reduces H3K4me3 enrichment and gene expression induced by mutant Kras. In summary, we demonstrate that Gli2 plays a significant role in pancreatic carcinogenesis by acting as a downstream effector of KrasG12D to control gene expression.

Patients with early-stage lung adenocarcinoma have a high risk of recurrent or metastatic disease despite undergoing curative intent therapy. We hypothesized that increased CD14+ cells within the tumor microenvironment (TME) could stratify patient outcomes. Immunohistochemistry for CD14 was performed on 189 specimens from patients with lung adenocarcinoma who underwent curative intent surgery. Outcomes and associations with clinical and pathologic variables were determined. In vitro studies utilized a coculture system to model the lung cancer TME containing CD14+ cells. Patients with high levels of TME CD14+ cells experienced a median overall survival of 5.5 years compared with 8.3 and 10.7 years for those with moderate or low CD14 levels, respectively (p < 0.001). Increased CD14+ cell tumor infiltration was associated with a higher stage at diagnosis and more positive lymph nodes at the time of surgery. This prognostic capacity remained even for patients with early-stage disease. Using an in vitro model system, we found that CD14+ cells reduced chemotherapy-induced cancer cell death. These data suggest that CD14+ cells are a biomarker for poor prognosis in early-stage lung adenocarcinoma and may promote tumor survival. CD14+ cell integration into the lung cancer TME can occur early in the disease and may be a promising new therapeutic avenue.

2648 Background: To characterize the safety and tolerability of PY314, an immunosuppressive macrophage depleting antibody, as a single agent and in combination with pembrolizumab in subjects with advanced refractory solid tumors including subject’s refractory to checkpoint inhibitors if approved for that indication. Methods: Two were evaluated in subjects with advanced solid tumors, single agent PY314 and PY314 in combination with 200 mg of pembrolizumab using a 3+3 dose escalation study design. Dosing was intravenous and administered once every 3-weeks, a defined cycle. Disease assessment by RECIST 1.1 was performed every 6 weeks. Each stratum included 4 dose levels of PY314 (1, 3, 10, and 20 mg/kg). Pharmacokinetics were evaluated at specified time points. Archival tumor tissue was analyzed for TREM2 expression by immunohistochemistry. Based on preclinical evaluation of TREM2 expression, HR+ HER2- and triple negative breast cancer, colorectal cancer, renal cell cancer, non-small cell lung cancer and gynecologic cancers were studied. Results: 28 subjects (median age 60 years [range 26-76], 22 females and 6 ) with an ECOG PS <2 were enrolled and all, but one was (1 subject withdrew consent after dosing). 15 subjects were treated with single agent PY314 and 13 were treated with the combination. No infusion-related reactions, dose limiting toxicities, suspected unexpected serious adverse reactions or high-grade treatment related adverse events (TRAEs) that resulted in treatment discontinuance was seen. 12 subjects experienced at least one TRAE, and in all but one subject, these were low grade. One subject experienced a treatment-related immune system disorder. serious adverse events, all unrelated to treatment. TREM2 expression in archival tumor ranged from 0.0-20%. PY314 pharmacokinetics were linear, dose proportional, unaffected by concomitant pembrolizumab and with a half-life of 8-9 days. Best radiographic response was stable disease seen in 11 subjects (39.3%) ranging in duration from 9-42 weeks. 6 subjects with stable disease have progressed and 5 remain on treatment. Conclusions: PY314 was well tolerated and has an excellent safety profile both as a single agent and in combination with pembrolizumab. A recommended dose for expansion was derived and enrollment in five prespecified cancers is ongoing. Clinical trial information: 04691375.

e21205 Background: Use of maintenance pemetrexed with pembrolizumab is the standard of care among patients with metastatic nonsquamous NSCLC without EGFR or ALK alterations treated with the KEYNOTE-189 regimen. Whether the addition of pembrolizumab to pemetrexed maintenance alters the risk of renal toxicity is not well characterized. Methods: A single center retrospective study was performed. The frequency of acute kidney injury as well as the rates of discontinuation due to renal injury was assessed. Acute renal injury was defined as ≥ 0.3 increase in serum creatinine (sCr) above the upper limit of normal or a rise in sCr ≥ 1.5 times baseline per KDIGO criteria. A Fisher exact test was conducted to compare the rate of renal injury between the two groups. Logistic regression adjusting for performance status, prior lines of treatment, and number of maintenance cycles was performed. Results: We identified 114 patients who received either maintenance pemetrexed or pemetrexed + pembrolizumab. The median number of cycles for the maintenance pemetrexed and pemetrexed + pembrolizumab groups was 5 and 7 cycles respectively. Of these, 41% (47/114) patients developed acute renal injury during their treatment course. Renal injury was seen in 14.3% (5/35) patients who received single agent pemetrexed maintenance and 25% (3/12) who received maintenance pemetrexed + pembrolizumab with no significant difference in the rates of renal injury between both arms (p = 0.403). Among patients who developed acute renal injury, 9% (4/47) permanently discontinued maintenance due to nephrotoxicity. All patients who permanently discontinued therapy received maintenance pemetrexed alone. When adjusting for covariates, ECOG performance status and number of prior lines of therapy did not increase the risk of renal toxicity. Logistic regression analysis identified that the rate of renal injury was significantly associated with the number of maintenance cycles received (p-value = 0.017, OR = 1.14, 95% CI 1.03 - 1.29). The odds of developing renal injury were 1.14 times higher with each additional cycle of maintenance therapy received. Conclusions: Renal injury is common among patients treated with patients receiving maintenance therapy. The addition of pembrolizumab to maintenance pemetrexed did not significantly increase the rate of renal injury. The risk of renal injury appears to correlate with the total number of maintenance cycles received suggesting a cumulative risk of nephrotoxicity over time.

9124 Background: Continuing a 1st generation EGFR TKI with chemotherapy upon TKI progression was not shown to be beneficial in the IMPRESS trial. However, the validity of this approach with osimertinib remains under explored. We attempted to characterize the efficacy of continuing osimertinib with chemotherapy in the post-progression setting. Methods: A single-center retrospective review of patients with metastatic EGFR mutant NSCLC who had progressed on osimertinib was performed. Clinical characteristics and treatment outcomes were noted. Progression free survival (PFS), duration of treatment (DOT), overall survival (OS) and rates of intracranial progression were captured. ANOVA or a Fisher exact test were used to identify associations between cohort characteristics and treatment outcomes. Differences in PFS, DOT and OS were assessed using a log-rank test. A Cox proportional hazard model was used to adjust for potential confounders. Results: 73 patients with EGFR mutant NSCLC with post-osimertinib treatment outcomes were identified. Cohort characteristics are summarized in Table. Median duration of follow up was 41 months. Upon progression, osimertinib was discontinued in 34 patients (Cohort A) and continued with next line of therapy in 39 patients (Cohort B). Survival analyses were adjusted for prior lines of therapy, use of platinum doublet chemotherapy, and use of immune checkpoint inhibitors in the post-progression setting. After adjusting for covariates, continuing osimertinib post-progression was associated with an improved PFS (7 vs 4 months; HR 0.58; 95% CI 0.34 – 1.00; p = 0.003) and DOT (7 vs 4 months; HR 0.52; 95% CI 0.31 – 0.87; p = 0.006). There was no difference in OS between Group A and B (52 vs 41 months; HR 0.73; 95% CI 0.43 – 1.24; p = 0.234). Rates of intracranial progression were similar between Group A and B (28% vs 23%; p = 0.649). Conclusions: After adjusting for covariates, continuing osimertinib with chemotherapy in the post-progression setting was associated with a significant difference in PFS and DOT, but with no differences in OS. Continuing osimertinib does not appear to influence the rate of subsequent intracranial progression. Prospective studies are needed to identify the optimal practice pattern.[Table: see text]

9090 Background: The clinicopathologic characteristics and optimal treatment strategies for non-sensitizing EGFR ( ns- EGFR) and HER2 activating mutations in NSCLC remain unclear. Methods: Single-center retrospective study of patients seen at University of Colorado from 2008 – 2018 with stage IV NSCLC was performed. Clinicopathologic features and treatment outcomes of patients with ns-EGFR (Exon 18, Exon 20, L861Q) and HER2 mutations were collected. Best response to TKI was determined (RECIST v1.1). PFS was calculated using Kaplan-Meier method. Results: Among 359 patients, we identified 49 ns-EGFR (36 Exon 20, 10 Exon 18, 3 L861Q) and 28 HER2 mutations (27 Exon 20, 1 gene amplification) detected via NGS (65/77), real-time PCR (9/77), FISH (1/77) and undocumented (2/77). PDL1 > 50% was seen in 44% ns- EGFR and 57% HER2. Adenocarcinoma was the most common histology (97%). Most patients were female (62%), never smokers (63%), and presented with metastatic disease (stage: I 5%, II 4%, III 6%, IV 85%). HER2+ NSCLC demonstrated a tropism for lung metastases (64%) that was significant when compared to EGFR Exon 19, EGFR L858R, ALK, ROS1, and KRAS cohorts (p < 0.001). No differences were found when other metastatic sites were compared. Among evaluable patients, response rates to TKI therapy is shown. Aggregate median PFS on TKI for ns-EGFR and HER2+ NSCLC was 6 months compared to EGFR Exon 19 (15 months; p < 0.01; HR 0.4; CI 0.24 – 0.67) and EGFR L858R (22 months; p < 0.01; HR 0.27 and 0.8; CI 0.14 – 0.54). Aggregate median OS for ns-EGFR and HER2+ NSCLC was 28 months with no differences when compared to EGFR Exon 19 and L858R subgroups. Conclusions: HER2+ NSCLC appears to have a predisposition for lung metastases. Higher DCR was observed with newer generation TKIs, but novel targeted therapeutic approaches are needed as overall outcomes remain poor. [Table: see text]

PURPOSE: There is a concern that influenza vaccination could increase the incidence of immune-related adverse events (irAEs) in patients with cancer receiving immune checkpoint inhibitors. The aim of our study was to determine the safety of influenza vaccination in this patient population. PATIENTS AND METHODS: We retrospectively identified patients who received at least 1 dose of pembrolizumab during any influenza season from September 2014 to August 2017 and reviewed medical records for irAEs. The primary endpoint was the incidence of irAEs. We used multivariable logistic regression and cumulative incidence curve with competing risks for comparison. RESULTS: Among 162 patients with cancer included in this study, 70 patients (43.2%) received at least 1 influenza vaccination. The vaccinated group was significantly older ( P = .002) and received more cycles of pembrolizumab ( P = .006). The incidence of any grade irAEs in the vaccinated group trended toward being lower (25.7% v 40.2%; P = .07) compared with the nonvaccinated group. Influenza vaccination was independently associated with fewer irAEs, with an odds ratio of 0.4 (95% CI, 0.2 to 0.9; P = .03) in multivariable analyses. The vaccinated group was less likely to have irAEs compared with the nonvaccinated group (24.7% v 34.4% at 12 months; P = .05), with death as a competing risk. The median irAE-free duration in the vaccinated group was longer than the nonvaccinated group (not reached v 28 months; P = .037). CONCLUSION: Influenza vaccination in patients with cancer receiving immune checkpoint inhibitor therapy was not associated with increased irAEs. This supports the safety of influenza vaccination in this patient population.

e14597 Background: Routine vaccination reduces preventable illness and infection in cancer patients even while undergoing therapy. Increasingly, patients with advanced malignancies receive check point inhibitors (CPI) and, with tumor response, can remain on therapy for extended periods of time. These agents promote tumor clearance by impeding a normal route of immune regulation which can result in autoimmunity. To the best of our knowledge, clinical outcomes for patients on CPI therapy who receive routine vaccinations have not been studied. Methods: The medical records of patients who received CPI therapy at Mayo Clinic in Rochester, Minnesota from March 25, 2011 to August 25, 2015 were reviewed for type and date of immunizations, CPI duration, adverse events (AE) at least possibly related to therapy, and survival from start of CPI. Patients who received > 1 dose of pembrolizumab or nivolumab were included in analysis. Patients were excluded if no follow up was done at our institution. Results: One hundred and eight patients were included in the analysis. Most patients received CPI therapy for metastatic melanoma (n = 71) or non-small cell lung cancer (n = 23). A total of 53 routine vaccinations were administered to 30 patients while on CPI therapy. Eighteen patients received a single vaccination. Annual influenza vaccination was most frequently administered (n = 38) followed by pneumococcal vaccines (n = 9). AEs were reported in 17/30 vaccinated patients and in 22/78 non-vaccinated patients (p = 0.004). In the vaccinated cohort, 11 patients experienced AEs after immunization and was not significantly different from the non-vaccinated patients (p = 0.265). For both cohorts, thyroid, rash, and pneumonitis were the most common AEs. Grade 3 AEs occurred with similar frequency between the 2 groups. Patients in the vaccinated cohort received more cycles of therapy (median 20.5 vs 6 cycles, p < 0.001). Median survival of patients who were not vaccinated was 503 days and median survival was not reached in the vaccinated group (p = 0.005). Conclusions: Routine vaccination of patients receiving CPI therapy did not significantly increase number or severity of adverse events. Routine vaccination did not reduce patient benefit from CPIs.

Abstract Our previous work has shown that in multiple hematologic and solid malignancies, an increased presence of CD14+HLA-DRlo/neg cells in the peripheral blood are associated with systemic immune suppression and poor prognosis. We have also demonstrated that glioblastoma and non-Hodgkin lymphoma tumor cells can convert healthy donor CD14+ cells into CD14+HLA-DRlo/neg. Thus, a critical step toward the generation of these immune suppressive cells is the recruitment of CD14+ monocytes to the tumor. In this study we investigated the mechanism for monocyte recruitment by lung cancer cell lines, the effect these lines have on the monocytes, and the effect the monocytes have on the tumor. Monocyte recruitment by H23, a human lung adenocarcinoma cell line, was studied using freshly isolated CD14+ cells and measured in real time with live cell imaging. Monocytes from 10 of 10 unique healthy donors robustly migrated towards H23 supernatant in a dose dependent fashion. H23 produces CCL2 at levels of 9160 ± 1615 pg/mL measured by ELISA in addition to elaborating the monocyte recruitment factors CXCL1 and IL-8 seen by cytokine array. A 2 day co-culture of H23 with monocytes from 17 of 18 unique healthy volunteers resulted in downregulated monocyte cell surface expression of HLA-DR by 49.7% ± 22.1%, a phenotype known to mediate both local and systemic immune suppression. Most importantly, monocytes had a positive effect on H23 survival under stress conditions. H23 continuously exposed to cisplatin in culture had improved survival when co-cultured with monocytes for 2 days prior to cisplatin exposure. After 6 days of low serum growth conditions, H23 survival and proliferation was bolstered by the presence of monocytes in culture resulting in a 2 fold greater number of viable H23 cells compared to H23 cultured in low serum without monocytes. These experiments have been repeated and showed that 3 of 4 additional human lung cancer cell lines demonstrated improved survival in low serum growth conditions when co-cultured with monocytes. In conclusion, we have demonstrated the active recruitment of monocytes by lung cancer cell lines, the transformation of the monocytes to an immune suppressive phenotype and the monocyte cancer cross-talk that results in improved tumor survival when exposed to the likely in vivo conditions of a nutrient poor environment or chemotherapy. Taken together these results suggest that the recruitment of monocytes may confer additional survival benefit in the tumor microenvironment beyond local immune suppression and angiogenesis. Citation Format: Erin Schenk, Allan Dietz. CD14+ monocyte recruitment, transformation and support of tumor survival in a lung cancer model. [abstract]. In: Proceedings of the CRI-CIMT-EATI-AACR Inaugural International Cancer Immunotherapy Conference: Translating Science into Survival; September 16-19, 2015; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2016;4(1 Suppl):Abstract nr A123.

Abstract Lymphocyte differentiation from naive CD4+ T cells into mature Th1, Th2, Th17, or T regulatory cell (Treg) phenotypes has been considered end stage in character. In this study, we demonstrate that dendritic cells (DCs) activated with a novel immune modulator B7-DC XAb (DCXAb) can reprogram Tregs into T effector cells. Down-regulation of FoxP3 expression after either in vitro or in vivo Treg-DCXAb interaction is Ag-specific, IL-6-dependent, and results in the functional reprogramming of the mature T cell phenotype. The reprogrammed Tregs cease to express IL-10 and TGFβ, fail to suppress T cell responses, and gain the ability to produce IFN-γ, IL-17, and TNF-α. The ability of IL-6+ DCXAb and the inability of IL-6−/− DCXAb vaccines to protect animals from lethal melanoma suggest that exogenously modulated DC can reprogram host Tregs. In support of this hypothesis and as a test for Ag specificity, transfer of DCXAb into RIP-OVA mice causes a break in immune tolerance, inducing diabetes. Conversely, adoptive transfer of reprogrammed Tregs but not similarly treated CD25− T cells into naive RIP-OVA mice is also sufficient to cause autoimmune diabetes. Yet, treatment of normal mice with B7-DC XAb fails to elicit generalized autoimmunity. The finding that mature Tregs can be reprogrammed into competent effector cells provides new insights into the plasticity of T cell lineage, underscores the importance of DC-T cell interaction in balancing immunity with tolerance, points to Tregs as a reservoir of autoimmune effectors, and defines a new approach for breaking tolerance to self Ags as a strategy for cancer immunotherapy.