Immunological correlates of response and immune-mediated toxicity in checkpoint inhibitor (ICI)-treated metastatic urothelial carcinoma (mUC) patients (pts).

Abstract

526 Background: Myeloid derived suppressors cells (MDSC) are immune cells that create an immunosuppressive microenvironment. Increased expression of MDSC subsets is associated with worse overall survival in ICI-treated mUC pts, but their role in immune-related adverse events (irAE) is unknown. Immune profiles associated with irAE are also unknown. We investigated associations of MDSC and –omics profiles with response and irAE in ICI-treated mUC pts. Methods: Baseline (B) and on-treatment (Tx) blood samples were collected from ICI-treated mUC pts. MDSC were measured in fresh unfractionated whole blood (WB) and in peripheral blood mononuclear cells (PBMC). MDSC were identified by flow cytometry in WB, defined as LinloCD33+/HLADR-, and subclassified as polymorphonuclear (PMN)-MDSC (CD15+/CD14-), monocytic (M)-MDSC (CD15-/CD14+), and uncommitted (UC)-MDSC (CD15-/CD14-). MDSC populations were presented as % of live nucleated blood cells and as absolute numbers from WB. irAE severity was graded by CTCAE v5. In a subcohort of 17 pts, proteomics and transcriptomics were analyzed via Olink and Bulk RNAseq, respectively. Wilcoxon rank sum test compared MDSC and –omics among response and irAE groups. Kruskal-Wallis test compared –omics results between irAE responders (irAE-R), irAE non-responders (irAE-NR), and no irAE/non-responders (noAE-NR). Results: 41 ICI-treated mUC pts (25 anti-PD-L1, 16 anti-PD-1) had at least 1 MDSC sample: 28 pts at B, 30 pts at Tx, and 17 pts at both B and Tx. Primary UC sites were bladder (78%) and upper tract (22%); 73% male; median age 72 (range, 28-82); 85% had KPS > 80%; 51% had visceral metastasis. ICI was first and second-line therapy in 37% and 63% of pts, respectively. 13 pts were responders (R); 26 pts were non-responders (NR); 2 pts were not evaluable. 22 pts developed irAE. Median time to irAE was 84 days (range, 21-145); 10 pts required steroids; 3 required ICI discontinuation. UC-MDSC was predominant in WB and PMN-MDSC in PBMC in both B and Tx. Between B and Tx, WB UC-MDSC and PB UC-MDSC increased in R (n = 13; p = 0.04), but decreased in NR (n = 26; p = 0.02). In the subcohort of 17 pts, 11 had irAE (7 irAE-R; 4 irAE-NR), 6 had noAE-NR. Proteomic analysis showed increased expression of CXCL12 in noAE-NR pts (p = 0.006) and increased expression of IL-8 (p = 0.016), IL-18 (p = 0.012), and IL-18R1 (p = 0.016) in all irAE pts. At the transcriptome level, upregulation of IFN-γ was associated with response, whereas upregulation of both IFN-γ and IFN-α differentiated irAE-R from irAE-NR. Conclusions: In ICI-treated mUC pts, WB & PB UC-MDSC increased in R and decreased in NR between B and Tx. Increased expression of pro-inflammatory chemokines was observed in irAE pts, independent of response. A distinct inflammatory pathway was observed in irAE-R. Prospective investigation of blood-based biomarkers of response and irAE development is warranted.