Decoding the Cellular Heterogeneity and Malignant Progression of Human Penile Squamous Cell Carcinoma by Single-Cell RNA Sequencing

A 66,421-cell atlas of penile squamous cell carcinoma identifies SEMA3C-high tumor cells and exhausted T cells as key drivers of metastasis

Journal: Advanced Science | Published: 2026-02-04 | Type: Journal Article | PMID: 41637545 Authors: Hu Xiheng, Shi Wensheng et al. — Department of Urology, Xiangya Hospital, Central South University; Chongqing University Three Gorges Hospital Funding/COI: Chinese government grants (National Key R&D Programmes, Natural Science Foundation of Hunan Province, Chongqing municipal programs). No conflicts of interest declared.

Summary

Researchers performed single-cell RNA sequencing on tumor and adjacent normal tissue from 10 patients with penile squamous cell carcinoma (PSCC), producing a 66,421-cell transcriptomic atlas. The central finding is that a subpopulation of malignant cells with high SEMA3C expression drives epithelial-mesenchymal transition (EMT) and correlates with lymph node metastasis and tumor stage. The paper also characterizes an immunosuppressive microenvironment dominated by exhausted T cells and SPP1-high tumor-associated macrophages — a pattern the authors note is conserved across multiple squamous cell carcinoma types.

Claims

• 66,421 high-quality cells retained after QC from 9 tumor and 6 adjacent normal samples across 10 patients; 9 major cell populations identified
• SEMA3C-high malignant cells scored highest on stemness (CytoTRACE) and showed enrichment for EMT hallmark gene sets
• Monocle3 pseudotime and CytoTRACE differentiation scores correlated strongly (Pearson R = 0.73, p < 0.001), suggesting consistent trajectory inference
• T cells in tumor tissue were predominantly in an exhausted state; regulatory T cell proportion was elevated relative to adjacent normal tissue
• SPP1-high TAMs (tumor-associated macrophages) were enriched in tumor samples and associated with pro-tumorigenic signaling; CD209+ TAMs were more prevalent in normal tissue
• A novel POSTN+ pericyte subpopulation was identified specifically in PSCC, with transcriptomic signatures of angiogenesis and extracellular matrix remodeling
• Cancer-associated fibroblasts interacted with SEMA3C-high malignant cells through multiple EMT-associated signaling pathways
• TME cell states (T cells, macrophages, CAFs) showed strong transcriptomic similarity to counterparts in cutaneous, cervical, oral, laryngeal, and lung SCCs

Study Quality

This is a descriptive, hypothesis-generating scRNA-seq study — not an interventional trial. The analytical pipeline is methodologically sound: batch correction via Harmony, copy number inference (inferCNV) to distinguish malignant from non-malignant epithelial cells, and multiple independent trajectory methods cross-validated against each other. Adjacent normal tissue was included as a comparator, which is often omitted in single-tumor atlases. The use of three separate sequencing platforms (BD Rhapsody, Singleron, Seekone) is disclosed and batch-corrected, but cross-platform technical variation is an inherent confounder that Harmony mitigates rather than eliminates.

The core limitation is sample size: 10 patients is standard for exploratory scRNA-seq but far too small to make reliable claims about SEMA3C as a clinical biomarker. The paper asserts SEMA3C "reflects cancer stage and microvessel density" and can "predict lymph node metastasis and prognosis," but this is based on computational signatures and expression correlations within a tiny cohort, not a validated biomarker study with prospective follow-up or an independent replication cohort.

Red Flags

• 10-patient cohort is the hard ceiling on everything this paper can claim; all biomarker conclusions are preliminary
• Multi-platform sequencing introduces technical heterogeneity; Harmony batch correction is imperfect and may conflate biological signal with platform artifact
• No functional validation: the proposed SEMA3C-EMT mechanism and POSTN+ pericyte roles are inferred from expression data, not confirmed in cell lines, organoids, or animal models
• SEMA3C "biomarker" claim has no independent validation cohort and no clinical cutoff established
• PSCC is rare, which makes large-sample scRNA-seq studies genuinely difficult — but rarity doesn't validate conclusions drawn from small N
• Computational trajectory methods (Monocle3, CytoTRACE) infer dynamics from static snapshots; the strong R = 0.73 correlation confirms the methods agree with each other, not that the trajectory is biologically real

Strengths

• 66,421-cell atlas is substantial for a rare cancer with limited prior single-cell data
• Inclusion of matched adjacent normal tissue allows within-study comparisons
• Cross-SCC comparison adds biological context and suggests conserved TME architecture
• Rigorous QC methodology described in detail, enabling reproducibility
• Government funding with no industry COI reduces overt bias risk
• Identifies POSTN+ pericytes as a potentially novel PSCC-specific TME component worth follow-up

Verdict

This is a solid exploratory atlas of a poorly characterized cancer's tumor microenvironment. The 66,421-cell dataset from 10 patients is about as good as you can expect for PSCC, and the identification of exhausted T cells, SPP1-high macrophages, and a potentially novel pericyte subtype gives future researchers specific targets to interrogate. What this paper cannot do — and oversells — is deliver a validated clinical biomarker. SEMA3C may well prove important; right now it's a computationally interesting candidate from a small cohort. Worth reading as a map of the terrain, not as evidence for clinical action.