Spatially-resolved analyses of muscle invasive bladder cancer microenvironment unveil a distinct fibroblast cluster associated with prognosis.

BACKGROUND

Muscle-invasive bladder cancer (MIBC) is a prevalent cancer characterized by molecular and clinical heterogeneity. Assessing the spatial heterogeneity of the MIBC microenvironment is crucial to understand its clinical significance.

METHODS

In this study, we used imaging mass cytometry (IMC) to assess the spatial heterogeneity of MIBC microenvironment across 185 regions of interest in 40 tissue samples. We focused on three primary parameters: tumor (T), leading-edge (L), and nontumor (N). Cell gating was performed using the Cytobank platform. We calculated the Euclidean distances between cells to determine cellular interactions and performed single-cell RNA sequencing (scRNA-seq) to explore the molecular characteristics and mechanisms underlying specific fibroblast (FB) clusters. scRNA-seq combined with spatial transcriptomics (ST) facilitated the identification of ligand-receptor (L-R) pairs that mediate interactions between specific FB clusters and endothelial cells. Machine learning algorithms were used to construct a prognostic gene signature.

RESULTS

The microenvironments in the N, L, and T regions of MIBC exhibited spatial heterogeneity and regional diversity in their components. A distinct FB cluster located in the L region-identified as S3-is strongly associated with poor prognosis. IMC analyses demonstrated a close spatial association between S3 and endothelial cells, with S3-positive tumors exhibiting increased blood vessel density and altered vascular morphology. The expression of vascular endothelial growth factor receptor and active vascular sprouting were significant in S3-positive tumors. scRNA-seq and ST analyses indicated that the genes upregulated in S3 were associated with angiogenesis. NOTCH1-JAG2 signaling pathway was identified as a significant L-R pair specific to S3 and endothelial cell interactions. Further analysis indicated that YAP1 was a potential regulator of S3. Machine learning algorithms and Gene Set Variation Analysis were used to establish an S3-related gene signature that was associated with the poor prognosis of tumors including MIBC, mesothelioma, glioblastoma multiforme, lower-grade glioma, stomach adenocarcinoma, uveal melanoma, kidney renal clear cell carcinoma, kidney renal papillary cell carcinoma, and lung squamous cell carcinoma.

CONCLUSIONS

We assessed the spatial landscape of the MIBC microenvironment and revealed a specific FB cluster with prognostic potential. These findings offer novel insights into the spatial heterogeneity of the MIBC microenvironment and highlight its clinical significance.