Tertiary lymphoid structure and neutrophil-lymphocyte ratio coordinately predict outcome of pembrolizumab

Abstract

Emerging evidence suggests that the presence of tertiary lymphoid structures (TLS) and neutrophil-lymphocyte ratio (NLR) in peripheral blood is associated with the treatment response to checkpoint inhibitors (CPIs), whereas there is limited knowledge regarding whether these factors reciprocally impact the treatment outcomes of CPIs in metastatic urothelial carcinoma (mUC). Herein, we investigated treatment outcomes of platinum-refractory mUC patients (50 cases with whole-exome and transcriptome sequencing) treated with pembrolizumab. The pathological review identified 24% of cases of TLS in the specimens. There was no significant difference in the NLR between the TLS- and TLS+ groups (p = 0.153). In the lower NLR group, both overall survival and progression-free survival were significantly longer in patients with TLS than in those without TLS, whereas the favorable outcomes associated with TLS were not observed in patients in the higher NLR group. We explored transcriptomic differences in UC with TLS. The TLS was comparably observed between luminal (20%) and basal (25%) tumor subtypes (p = 0.736). Exploring putative immune-checkpoint genes revealed that ICOSLG (B7-H2) was significantly increased in tumors with lower NLR. KRT expression levels exhibited higher basal cell markers (KRT5 and KRT17) in the higher NLR group and lower differentiated cell markers (KRT8 and KRT18) in patients with TLS. In conclusion, the improved outcomes of pembrolizumab treatment in mUC are restricted to patients with lower NLR. Our findings begin to elucidate a distinct molecular pattern for the presence of TLS according to the NLR in peripheral blood.

Keywords: neutrophil-lymphocyte ratio; pembrolizumab; tertiary lymphoid structure; urothelial carcinoma.

FIGURE 1

(A) Representative image of urothelial carcinoma with tertiary lymphoid structures (TLS) along the tumor. The presence of TLS was confirmed by spotting aggregated CD20⁺ cells, within which BCL6⁺ germinal center‐like cells are present. (B) Neutrophil–lymphocyte ratio (NLR) according to the presence of the TLS. The difference was assessed by the Mann–Whitney U test. (C) Kaplan–Meier curves on overall survival (OS) and progression‐free survival (PFS) in higher (≥5.00) and lower (<5.00) NLR groups. Log‐rank test was utilized to examine the difference in survival. (D) OS and PFS for patients with or without the TLS. Log‐rank test was utilized to examine the difference in survival.

FIGURE 2

(A) Kaplan–Meier curves on OS and PFS for patients with or without the TLS. Patients were stratified according to the median NLR. Log‐rank test was utilized to examine the difference in survival. (B) Best objective response (BOR) for the pembrolizumab treatment according to the TLS presence. Fisher's exact test was examined to assess the distribution differences.

FIGURE 3

(A) Summary of the total cohort (n = 100) with the best objective response for pembrolizumab treatment. Fifty cases underwent next‐generation sequencing (NGS). (B) Oncoprint for 50 patients treated with pembrolizumab. (C) Gene Set Enrichment Analysis (GSEA) plotting between patients with/without the TLS in all human MSigDB collections (Hallmark, C1‐8: 23684 gene sets) by false discovery rate q‐value (FDR‐q) and normalized enrichment score (NES). (D) GSEA of “HINATA_NFKB_IMMU_INF” (left panel), and “EACTOME_CD22_MEDIATED_BCR_REGULATION” (right panel) that were top two upregulated pathways in cases with TLS. (E) Best objective response (BOR) for the pembrolizumab treatment according to the UNC molecular subtypes. Fisher's exact test was examined to assess the distribution differences. (F) Pie charts of the presence of TLS in luminal and basal molecular subtypes. Fisher's exact test was examined to assess the distribution differences. (G) Heatmap of RNA expression level of immune‐checkpoint genes and KRT differentiated genes in 50 patients treated with pembrolizumab. The right panels exhibit the p‐value for the difference in the expression level of each gene according to the presence of TLS and NLR. The difference was assessed using the Mann–Whitney U test. (H) RNA expression level according to the presence of TLS and NLR. The difference was assessed by the Mann–Whitney U test.