PD-L1 and nectin-4 expression and genomic characterization of bladder cancer with divergent differentiation

Abstract

Background: Bladder cancer with divergent differentiation (BCDD) comprises a heterogenous group of tumors with a poor prognosis, and differential expression of nectin-4 and programmed death ligand-1 (PD-L1) has been reported in BCDD. Importantly, nectin-4 expression in bladder cancer is associated with response to enfortumab vedotin, and PD-L1 expression is associated with responses to immune checkpoint inhibitors (ICIs).

Methods: The authors conducted a retrospective review identifying 117 patients with advanced or metastatic BCDD who were treated at Winship Cancer Institute from 2011 to 2021. They performed immunohistochemistry staining for nectin-4 and PD-L1 expression by histologic subtype as well as genomic analysis of these patients, including RNA sequencing, whole-exome sequencing, and fusion detection analysis as well as a subgroup genomic analysis of patients with BCDD who received ICIs.

Results: The results indicated that nectin-4 expression was highest in the groups who had the squamous and plasmacytoid subtypes, whereas the group that had the sarcomatoid subtype (70.8%) had the highest proportion of PD-L1-positive patients. Genomic analysis yielded several key findings, including a 50% RB1 mutation rate in patients who had small cell BCDD, targetable PIK3CA mutations across multiple subtypes of BCDD, and significantly higher expression of TEC in responders to ICIs.

Conclusions: In this study, the authors identified clinically relevant data on nectin-4 and PD-L1 expression in patients with rare bladder tumors. They also identified several novel findings in the genomic analysis that highlight the role of precision medicine in this population of patients. Larger, prospective studies are needed to validate these hypothesis-generating data.

Keywords: RNA sequencing (RNAseq); antibody–drug conjugate; biomarkers, bladder cancer; nectin‐4; programmed death ligand‐1 (PD‐L1); whole‐exome sequencing.

Figure 1:

Distribution of Nectin-4 H-score (Panel A) and positive Nectin-4 (H-score > 0) and PD-L1 (CPS score ≥ 10) expression (Panel B) by histologic subtype

Figure 1:

Distribution of Nectin-4 H-score (Panel A) and positive Nectin-4 (H-score > 0) and PD-L1 (CPS score ≥ 10) expression (Panel B) by histologic subtype

Figure 2:

Principal component analysis (PCA) plot for RNA from all samples plotted along principle component 1 (PC1) and principle component 2 (PC2). PC1 accounts for 31% of the total variance between samples, while PC2 accounts for 14% of the variance. Samples are colored by their histology, and ellipses calculate the probable regions for most samples of a given histology. The mixed histologies have too few samples to calculate ellipses.

Figure 3:

Volcano plots showing differential gene expression between different cancers. Comparisons are as follows: A) plasmacytoid vs all others, B) sarcomatoid vs all others, C) small cell vs all others, D) squamous vs all others, E) adenocarcinoma vs all others. Log2 fold-change is plotted on the x-axis with the negative log of the p-value is on the y-axis. Each point is a single gene. Colors correspond to significance based on arbitrarily defined threshold cutoffs of adjusted p-value < 0.05 and at magnitude of fold-change > 1.3. Genes that meet both criteria are red. Genes that meet the p-value criteria but not the fold-change cutoff are blue. Genes that meet the fold-change cutoff but not the p-value criteria are green. Genes that fall below both thresholds are grey. Genes with a positive fold-change (the right side of the volcano) have higher expression in the first group in the comparison, while genes with a negative fold-change are more greatly expressed in the ‘all others’ group. Labels identify some requested genes and some genes identified in TCGA bladder cancer literature.

Figure 3:

Volcano plots showing differential gene expression between different cancers. Comparisons are as follows: A) plasmacytoid vs all others, B) sarcomatoid vs all others, C) small cell vs all others, D) squamous vs all others, E) adenocarcinoma vs all others. Log2 fold-change is plotted on the x-axis with the negative log of the p-value is on the y-axis. Each point is a single gene. Colors correspond to significance based on arbitrarily defined threshold cutoffs of adjusted p-value < 0.05 and at magnitude of fold-change > 1.3. Genes that meet both criteria are red. Genes that meet the p-value criteria but not the fold-change cutoff are blue. Genes that meet the fold-change cutoff but not the p-value criteria are green. Genes that fall below both thresholds are grey. Genes with a positive fold-change (the right side of the volcano) have higher expression in the first group in the comparison, while genes with a negative fold-change are more greatly expressed in the ‘all others’ group. Labels identify some requested genes and some genes identified in TCGA bladder cancer literature.

Figure 3:

Volcano plots showing differential gene expression between different cancers. Comparisons are as follows: A) plasmacytoid vs all others, B) sarcomatoid vs all others, C) small cell vs all others, D) squamous vs all others, E) adenocarcinoma vs all others. Log2 fold-change is plotted on the x-axis with the negative log of the p-value is on the y-axis. Each point is a single gene. Colors correspond to significance based on arbitrarily defined threshold cutoffs of adjusted p-value < 0.05 and at magnitude of fold-change > 1.3. Genes that meet both criteria are red. Genes that meet the p-value criteria but not the fold-change cutoff are blue. Genes that meet the fold-change cutoff but not the p-value criteria are green. Genes that fall below both thresholds are grey. Genes with a positive fold-change (the right side of the volcano) have higher expression in the first group in the comparison, while genes with a negative fold-change are more greatly expressed in the ‘all others’ group. Labels identify some requested genes and some genes identified in TCGA bladder cancer literature.

Figure 3:

Volcano plots showing differential gene expression between different cancers. Comparisons are as follows: A) plasmacytoid vs all others, B) sarcomatoid vs all others, C) small cell vs all others, D) squamous vs all others, E) adenocarcinoma vs all others. Log2 fold-change is plotted on the x-axis with the negative log of the p-value is on the y-axis. Each point is a single gene. Colors correspond to significance based on arbitrarily defined threshold cutoffs of adjusted p-value < 0.05 and at magnitude of fold-change > 1.3. Genes that meet both criteria are red. Genes that meet the p-value criteria but not the fold-change cutoff are blue. Genes that meet the fold-change cutoff but not the p-value criteria are green. Genes that fall below both thresholds are grey. Genes with a positive fold-change (the right side of the volcano) have higher expression in the first group in the comparison, while genes with a negative fold-change are more greatly expressed in the ‘all others’ group. Labels identify some requested genes and some genes identified in TCGA bladder cancer literature.

Figure 3:

Volcano plots showing differential gene expression between different cancers. Comparisons are as follows: A) plasmacytoid vs all others, B) sarcomatoid vs all others, C) small cell vs all others, D) squamous vs all others, E) adenocarcinoma vs all others. Log2 fold-change is plotted on the x-axis with the negative log of the p-value is on the y-axis. Each point is a single gene. Colors correspond to significance based on arbitrarily defined threshold cutoffs of adjusted p-value < 0.05 and at magnitude of fold-change > 1.3. Genes that meet both criteria are red. Genes that meet the p-value criteria but not the fold-change cutoff are blue. Genes that meet the fold-change cutoff but not the p-value criteria are green. Genes that fall below both thresholds are grey. Genes with a positive fold-change (the right side of the volcano) have higher expression in the first group in the comparison, while genes with a negative fold-change are more greatly expressed in the ‘all others’ group. Labels identify some requested genes and some genes identified in TCGA bladder cancer literature.

Figure 4:

Study cohort heatmaps of genes identified in previous TCGA bladder cancer literature^, (Panel A) and other selected genes (Panel B). Each cell represents the expression of a single gene in a single sample. Labels above the cells show histologic subtype, labels on the right are gene names, and labels below are individual sample names. The color of a cell corresponds to the library-size normalized, log2-scaled transcript count for that gene/sample relative to the cohort median for that gene. For plotting clarity to show smaller differences, expression values are capped at a log2 fold-difference of 2, so some values with the strongest colors may exceed that value.

Figure 4:

Study cohort heatmaps of genes identified in previous TCGA bladder cancer literature^, (Panel A) and other selected genes (Panel B). Each cell represents the expression of a single gene in a single sample. Labels above the cells show histologic subtype, labels on the right are gene names, and labels below are individual sample names. The color of a cell corresponds to the library-size normalized, log2-scaled transcript count for that gene/sample relative to the cohort median for that gene. For plotting clarity to show smaller differences, expression values are capped at a log2 fold-difference of 2, so some values with the strongest colors may exceed that value.