Supplemental Table 4.1: Penn Phase 2 Cohort composition Supplemental Table 4.2: Penn Phase 2 Disease Control Diagnosis List

Supplemental Table 1: Plasma Pool Composition

Supplemental Table 8.1: PIGR values by sex Supplemental Table 8.2: ANPEP values by sex Supplemental Table 8.3: PIGR values by diabetes mellitus status Supplemental Table 8.4: ANPEP values by diabetes mellitus status Supplemental Table 8.5: Spearman correlation analysis of age and ANPEP or PIGR values

PURPOSE:Pancreatic ductal adenocarcinoma (PDAC) is typically detected too late for useful therapeutic interventions; hence, we sought blood biomarkers to detect early-stage disease. EXPERIMENTAL DESIGN:Using mass spectrometry and ELISA on plasma pools from the University of Pennsylvania (Penn) and the Mayo Clinic (Mayo), we identified aminopeptidase N (ANPEP) and polymeric immunoglobulin receptor (PIGR) as increased in early-stage (stage I/II) PDAC plasma compared with controls. We tested ANPEP and PIGR, along with prior data for thrombospondin 2 (THBS2) and carbohydrate antigen 19-9 (CA19-9), in retrospective phase II studies using separate cohorts of PDAC plasmas at different stages versus healthy or nonmalignant disease controls (DC) from Penn (n = 135) and Mayo (n = 537). RESULTS:Comparing healthy controls with stage I/II PDAC, we obtained area under the receiver-operating characteristic curves (AUC) of 0.78 [95% confidence interval (CI), 0.68-0.86]/0.80 (95% CI, 0.74-0.85; ANPEP) and 0.81 (95% CI, 0.70-0.88)/0.86 (95% CI, 0.82-0.90; PIGR) for the Penn/Mayo phase II studies, respectively. In multivariable models, CA19-9/THBS2/ANPEP, CA19-9/THBS2/PIGR, and CA19-9/THBS2/ANPEP/PIGR elicited AUC of 0.94 to 0.96 for Penn and 0.97 for Mayo. Notably, the four-marker panel elicited AUC of 0.87 for the Mayo stage I/II versus DC and 0.91 for stages I to IV versus DC. At a specificity of 95%, a plasma biomarker panel composed of CA19-9 (≥35 U/mL), THBS2 (≥42 ng/mL), ANPEP (≥2,995 ng/mL), and PIGR (≥1,800 ng/mL) yielded a sensitivity of 91.9% for PDAC stages I to IV and 87.5% for PDAC stage I/II. CONCLUSIONS:Adding ANPEP and PIGR to a plasma biomarker panel of CA19-9 and THBS2 enhances the detection of early-stage PDAC when comparing cancer versus healthy or nonmalignant DC. Given the concordance of our data in two retrospective phase II studies, assessments in prediagnostic cases are warranted.

Supplemental Table 6.1: Univariate Biomarker Analysis of Penn Phase 2 Cohort – Disease Control Reference Supplemental Table 6.2: Two-variable Biomarker Analysis of Penn Phase 2 Cohort – Disease Control Reference Supplemental Table 6.3: Multivariable Analysis of Penn Phase 2 Cohort – Disease Control Reference

Supplemental Figure 1. Representative Mayo plasma pool ELISA data, presented as interpolated concentration (error bars = SD) of two replicates for (a) ANPEP and (b) PIGR

Supplemental Table 7: Biomarker concentration cutoff points based on percentiles of distribution in Mayo Phase 2 control plasma samples (n = 140).

TMT-labeled MS Significant Protein Output from Wistar

Supplemental Table 2: LFQ MS Significant Protein Output from WUSTL

Supplemental Table 5: Demographic and clinical characteristics of patients whose samples were used in Mayo Phase 2

Supplemental Figure 2. Box-and-whisker plots with data points in jittered overlay show the mRNA expression levels of PIGR and ANPEP in tumor samples from across the TCGA cohort “pan-cancer analysis of whole genomes” (ICGC/TCGA, Nature 2020)

BACKGROUND:Limited understanding of the biology predisposing certain human papillomavirus-related (HPV+) oropharyngeal squamous cell carcinomas (OPSCCs) to relapse impedes therapeutic personalization. We aimed to identify molecular traits that distinguish recurrence-prone tumors. METHODS:Fifty HPV+ OPSCCs that later recurred (cases) and 50 nonrecurrent controls matched for stage, therapy, and smoking history were RNA-sequenced. Groups were compared by gene set enrichment analysis, and select differences were validated by immunohistochemistry. Features discriminating groups were scored in each tumor using gene set variation analysis, and scores were evaluated for recurrence prediction ability. RESULTS:Cases downregulated pathways linked to antitumor immunity (FDR-adjusted P < .05) and contained fewer tumor-infiltrating lymphocytes (P < .001), including cytotoxic T-cells (P = .005). Cases also upregulated pathways related to cell division and other aspects of tumor progression. Upregulated and downregulated pathways were respectively used to define a tumor progression score (TPS) and immune suppression score (ISS) for each tumor. Correlation between TPS and ISS (r = .603, P < .001) was potentially explained by observed upregulation of DNA repair pathways in cases, which might enhance their progression directly and by limiting cytosolic DNA-induced inflammation. Accordingly, cases contained fewer double-strand breaks based on staining for phospho-RPA32 (P = .006) and γ-H2AX (P = .005) and downregulated the cytosolic DNA sensing pathway. A combined score derived from TPS and ISS optimized recurrence prediction and stratified survival in a manner generalizable to 3 external cohorts. CONCLUSIONS:We describe a potential link in HPV+ OPSCCs between reduced DNA damage and other tumor-intrinsic and immune-related contributors to recurrence risk, opening opportunities to detect and target this high-risk biology.

Therapeutic synergy of PARP inhibition (PARPi) and CTLA4 immune checkpoint inhibition (ICI) in BRCA1 mutated (BRCAm) ovarian cancer (OC) models was dependent on increased interferon gamma (IFNg) production in the tumor microenvironment (TME). Expression of interferon-response genes in tumor samples is also linked with CTLA4 ICI efficacy in melanoma. In an assessment of interferon-response gene expression in tumor samples from women with BRCAm ovarian cancer enrolled in a phase I/II clinical trial of PARPi and CTLA4 ICI, IFIT1B was significantly associated with both treatment benefit (AUC=0.842), and progression-free survival (PFS) (n=30, p=0.029). In this cohort, IFIT1B expression was also strongly correlated with other IFIT gene family members, and positively correlated with CXCL10 expression among subjects with clinical benefit from treatment. In an analysis of TCGA data, IFIT1B expression was also significantly associated with PFS among BRCAm patients (n=16, p=0.0306), but not in the larger ovarian cancer cohort (n=299, p=0.1953). IFIT gene family expression is induced by type I and type II interferons, as well as cytosolic nucleic acid sensing. CXCL10 expression is associated with anti-tumor immunity and is prognostic in ovarian cancer. To examine the association between these candidate biomarkers and treatment outcomes, we exposed murine OC cells to PARPi, IFNg, or the combination in vitro, and compared gene expression by qPCR and production of CXCL10 and IFN beta by ELISA. Treatment with PARPi significantly increased expression of IFIT family members (Ifit1, Ifit2, Ifit3) compared with untreated cells, and additionally increased expression in combination with IFNg compared with IFNg exposure alone. PARPi-induced IFIT gene expression was independent of BRCA status and was observed in both PARPi-sensitive and -resistant BRCA1m models. PARPi also significantly increased expression of CXCL10 by qPCR, and PARPi in combination with IFNg increased levels of secreted CXCL10 by BRCA1m cells compared with IFNg alone. Although expression of IFNb was similar across all treatment conditions, high IFN beta secretion was observed in BRCA1m cells treated with PAPRi in combination with IFNg. These results indicate that PARPi can modulate tumor cell-intrinsic responses to IFNg independent of BRCAm status, including changes in secreted chemokines and cytokines associated with T cell recruitment and immune activation in the tumor microenvironment. With evidence that both IFN signaling and CXCL10 in the TME can impact treatment efficacy and disease outcomes, these data may inform the design of novel combinatorial regimens for ovarian cancer. Daniel Falcon, Alyson Yeckes, Ania Klas, Andrea Corbet, Chelsea Goff, Sandra Orsulic, Rita Serda, Phyllis Gimotty, Sarah Adams. PARP inhibition induces the expression of interferon-response genes in both BRCAm and BRCAwt high grade serous ovarian cancer cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 6163.

Abstract: Little is known about the impact of recent advances in acute myeloid leukemia (AML) treatment on racial/ethnic disparities in survival outcomes. We performed a retrospective cohort study of patients with newly diagnosed AML using data from a nationwide electronic health record–derived deidentified database. Patients were categorized based on their diagnosis date relative to venetoclax approval, as pre–novel therapy era (Pre era; 2014-2018; n = 2998) or post–novel therapy era (Post era; 2019-2022; n = 2098). Patients in the Post era were older and had more comorbidities than Pre era. Non-Hispanic Black (NHB) and Hispanic patients were younger and more likely to have lower socioeconomic status than non-Hispanic White (NHW) patients, with no differences in the distributions of key disease features. After accounting for age and comorbidity, overall survival (OS) was higher in patients in Post era than Pre era (adjusted hazard ratio [aHR], 0.90; 95% confidence interval [CI], 0.83-0.96). In Pre era, NHB had a 22% higher hazard of death than NHW (aHR, 1.22; 95% CI, 1.04-1.43), whereas worse OS was not observed for NHB in Post era (aHR, 0.86; 95% CI, 0.69-1.08; predicted 2-year survival, 45.3% vs 39.9%). Utilization of novel therapeutics in frontline therapy did not differ by race/ethnicity. Among patients receiving venetoclax-based induction, particularly those without TP53, RAS, or FLT3-ITD mutations, results suggested higher OS for NHB than NHW patients (aHR, 0.67; 95% CI, 0.45-1.01). Additional studies are needed to elucidate factors contributing to these observed survival differences and to inform strategies to optimize outcomes for all patients with AML.

9502 Background: Neoadjuvant immune checkpoint therapy has shown improvement in event-free survival outcomes in patients with resectable clinical stage III and IV melanoma. Whether there is benefit to neoadjuvant immune therapy in patients with clinical stage IIB/C melanoma is unknown. Methods: In a single arm multicenter investigator-initiated phase 2 trial, patients with clinical stage IIB/C melanoma received a single dose of neoadjuvant pembrolizumab (200 mg intravenously) 3 weeks prior to wide excision and sentinel lymph node (SLN) biopsy followed by 1 year adjuvant pembrolizumab every 3 weeks or until unacceptable toxicity or disease progression. Primary endpoint was SLN positivity rate. A sample size of 63 patients had 80% power detect a 50% difference when compared to a predetermined historical SLN positivity rate in treatment naïve patients (25% Stage IIB and 40% Stage IIC) weighted by proportion of clinical tumor stage in eligible study patients. Secondary endpoint included recurrence-free survival. Safety outcomes, including overall toxicity and immune related adverse events, were also assessed. Results: Of 63 evaluable patients (33 IIB; 30 IIC at initial biopsy), the SLN metastasis rate in the neoadjuvant study group was 27%. 28 patients (44%) had residual primary tumor after single dose pembrolizumab; 4 patients had their primary tumors upstaged to IIC. Compared to a SLN metastasis rate in a historical treatment- naïve cohort based on tumor staging at wide excision (33.1%), there was a 18% reduction in SLN positivity rate in the neoadjuvant group, although this was not statistically significant (p = 0.302). In a subgroup analysis, stage IIC patients in the neoadjuvant study group had a SLN metastasis rate of 16.7% versus 40% (p = 0.009) based on initial biopsy and 23.5% versus 40% (p = 0.0499) based on primary tumor staging at wide excision. With median follow-up of 20.4 months, the 2-year recurrence free-survival in the study group was 84% with median time to recurrence (n = 10) of 9.9 months. Overall treatment-related grade 3/4 adverse events were 14 (22%) with 9 (14%) immune-related adverse events; there was no delay in definite surgery secondary to neoadjuvant treatment. Conclusions: Rate of SLN metastasis among patients with clinical stage IIB/C melanoma undergoing neoadjuvant pembrolizumab did not differ significantly compared to expected historical rates in treatment-naïve patients; however, in a secondary subgroup analysis among patients with clinical stage IIC disease, a decrease in SLN positivity rate was noted. Neoadjuvant therapy in clinical stage IIB/C was safe and feasible, with no significant delay in surgery or new or unexpected toxicities noted in these patients. Translational studies are under way, including flow cytometric and transcriptional studies, that may reveal immunologic determinants of efficacy versus resistance. Clinical trial information: NCT03757689 .

Small extracellular vesicles (sEVs) are nanosized vesicles. Death receptor 5 (DR5) mediates extrinsic apoptosis. We engineer DR5 agonistic single-chain variable fragment (scFv) expression on the surface of sEVs derived from natural killer cells. PDGFR transmembrane domain delivers DR5-scFvs to the surface of sEVs. DR5-scFv sEVs rapidly induce apoptosis of different types of DR5 + cancer cells, myeloid-derived suppressor cells (MDSCs), and cancer-associated fibroblasts (CAFs). DR5-scFv sEVs migrate specifically to DR5 + tumors in vitro and in vivo. Systemic delivery of DR5-scFv sEVs significantly inhibits the growth of DR5 + melanoma, liver cancer, and breast cancer and prolongs mouse life span without significant toxicity. DR5-scFv sEVs are significantly more efficacious than DR5 antibodies in vivo. In organotypic patient-derived melanoma slice cultures, DR5-scFv sEVs effectively inhibit melanoma cells and MDSCs and activate CD8 + T cells. Our studies demonstrate that DR5-scFv sEVs can inhibit tumor growth by targeting tumor cells and immunosuppressive stromal cells in the TME.

Supplemental Figure S1: LAI remodels the lipidome of melanoma cells. Supplemental Figure S2: LAI induces UGCG, and UGCG inhibition augments DC661 cytotoxicity, and this effect is not seen by blocking cholesterol synthesis and key earlier steps of autophagy. Supplemental Figure S3: LAI increases the formation of GMM in cancer cells. Supplemental Figure S4: UGCG inhibition synergistically augments LAI induced cytotoxicity and abrogates GMM formation without elevating ceramide levels. Supplemental Figure S5: LAI induces UGCG-associated GMM formation in PDX cells, a combination of DC661 and eliglustat impairs tumor growth in a therapy-resistant PDX tumor model.