Platelets impact the responsiveness of immune checkpoint blockade therapy in solid tumors.

e15023 Background: A growing body of evidence has linked platelets (plts) with immune suppression in the tumor microenvironment (TME). Our group has demonstrated, plts are the dominant source for TGF-β in TME and pharmacologic inhibition of plt function enhances multiple forms of immunotherapy (Sc Immunol 2017 PMID 28763790; J Immunol 2019 PMID 31358658; Sci Transl Med 2020 PMID 31915300). To further delineate this relationship, we examined the roles of plts on T cell exhaustion in a preclinical model and on response to immune checkpoint inhibitors (ICI) in a large cohort of patients (pts) with stage 4 cancers. Methods: The mouse MC-38 colon adenocarcinoma model was used in age-matched female C57Bl/6J mice. Antiplatelet therapy (APT) consisting of aspirin and clopidogrel (both at 30 mg/kg daily) was delivered p.o. and anti-PD-1 antibody (100 mg/mouse every 3 days) was administered i.p. starting on days 5 and 9 respectively. TME analysis via multispectral histology or flow cytometry was performed. Retrospective analysis was carried out on 826 pts who received ICI from 2011-2017 at the Ohio State University. Baseline plt count was collected within 7 days before initiating ICI. Repeat plt count was obtained prior to initiation of cycle 2. Normal plt counts were defined as 150,000 – 450,000/µl blood, and thrombocytosis as plt ≥450,000/µl blood. Kaplan Meier and log-rank analysis were performed to estimate median survival and determine the association with plt count. Results: In pre-clinical models, pretreatment with APT (1) synergized with PD-1 blockade to enhance T cell infiltration into MC-38 tumors resulting in immediate tumor reduction, and (2) decreased tumor infiltrating CD8 + T cell TOX expression, a transcription factor associated with T cell exhaustion (Nat Immunol 2019 PMID: 31427776). Among pts receiving ICIs, 46 (5.6%) pts with thrombocytosis had a significantly reduced median OS vs pts with normal plt counts: 6.0 (95 CI: 1.5—10.6) months (mos) vs 11.6 (95 CI: 9.7—13.4) mos (p = 0.002). Fluctuations ≥50,000 plt/mL in either direction between cycle 1 and 2 were associated with a significant reduction in median OS: 8.3 (95 CI: 6.4—10.1) mos vs 13.6 (95 CI: 11.2-16.1) mos (p < 0.001). Conclusions: There is a strong association between plts and failure of ICI in both the preclinical and clinical settings, likely via modifying the amount of active CD8 + T cells infiltrating into tumors. These findings merit further study to delineate the underlying mechanism for plt-mediated immune suppression and strategies to overcome it.