Single-cell and spatial transcriptomics reveal divergent microenvironments and progression signatures in early- and late-onset prostate cancer

Geriatrics Urology Immunology Basic Medical Sciences Cell Biology Biochemistry Oncology Life Sciences Medicine
single-cell sequencing spatial transcriptomics prostate cancer heterogeneity tumor microenvironment androgen response tumor progression

Background: Age-related Heterogeneity in Prostate Cancer and New Opportunities for Precision Medicine

Prostate cancer (PCA) is the second most common malignancy and the fifth leading cause of cancer-related death among men worldwide. With the global aging population and the advancement of health screening, the age of onset for prostate cancer has become more diverse, with early-onset prostate cancer (EOPC, usually defined as onset in males aged 55 years or younger) cases significantly increasing in recent years and associated with poorer prognosis. Therefore, exploring the biological differences between prostate cancers of different age onsets, especially the construction and evolution of the tumor microenvironment (TME), holds great significance for achieving age-specific precision intervention and individualized therapy.

Clinical observations suggest that, compared to late-onset prostate cancer (LOPC), EOPC not only displays significantly different clinico-pathological features but also has lower survival outcomes. Previous studies have indicated that such heterogeneity is closely related to age-dependent fluctuations in hormone (such as androgen) levels and immune activity. However, until now, systematic investigations at single-cell and spatial resolution regarding the mechanisms of EOPC and LOPC onset/progression and microenvironment remodeling have been scarce. Traditional high-throughput omics technologies are limited by their inability to dissect the complexity at cellular and spatial levels, restricting our understanding of the microenvironment and its dynamic interactions with tumor cells. This research team has addressed this gap, delivering the first systematic “single-cell plus spatial” atlas to reveal the molecular, cellular, and spatial heterogeneity between the two types of prostate cancer and their clinical significance.

Source and Author Information

This research was published in Nature Aging, Volume 5, May 2025, pages 909–928, DOI: 10.1038/s43587-025-00842-0. The paper is titled “single-cell and spatial rna sequencing identify divergent microenvironments and progression signatures in early- versus late-onset prostate cancer”.

The first authors are Yifei Cheng, Bingxin Liu, and Junyi Xin, with corresponding authors including Bin Xu, Mulong Du, Gong Cheng, and Meilin Wang. The team hails from institutions such as Southeast University and Nanjing Medical University, China. All clinical samples, models, and data analysis in this study were conducted at top Chinese medical centers.

Overall Study Design and Technical Approach

This study employed a complex, multi-step, cross-scale, and cross-validation research system. Centered on “decoding the heterogeneity of EOPC and LOPC tumor cells and microenvironments,” it utilized single-cell RNA sequencing (scRNA-seq), spatial transcriptomics (ST-seq), large-scale cohort GWAS analysis, protein-level multiplex immunohistochemistry, flow cytometry, in vivo mouse models, and multicenter clinical data review, comprehensively characterizing cell taxonomy, spatial structure, molecular signaling pathways, and key cell-cell communications in the two tumor types.

1. Study Subjects and Sample Processing

  • Prostate Cancer Patient Tissue Collection: Ten patients with aggressive prostate cancer were enrolled (4 EOPC, 6 LOPC, all treatment-naïve), yielding a total of 63,763 high-quality single cells for downstream analysis. Additional validation cohorts were set up, with 11 EOPC and 20 LOPC FFPE samples included for supplementary support.
  • Spatial Transcriptomics Samples: One tumor tissue sample each from EOPC and LOPC was selected for spatial transcriptomics on the 10x Visium platform to systematically quantify the distribution and interactions of various cell types within tumors and their microenvironments.
  • Animal Models: In situ tumor models were established in 8-month-old (young, simulating EOPC) and 16-month-old (aged, simulating LOPC) C57BL/6 mice, divided into sham and castrated (androgen deprivation therapy) groups to verify cell population dynamics and therapeutic responses.

2. Single-cell and Spatial Transcriptomics Data Analysis

  • Single-cell Data Pipeline: After batch effect correction, cell clustering, and marker-based annotation, eight major cell types were preliminarily identified—epithelial, T cells, B cells, myeloid cells, mast cells, endothelial cells, smooth muscle cells, and fibroblasts. Further integrating CNV analysis and clustering techniques, epithelial subpopulations were precisely defined as malignant, normal, and mixed types.
  • Spatial Transcriptomics: Through anchoring and deconvolution with single-cell data, spatial positioning, and clustering were performed for each cell type and subtype, allowing detailed mapping of cellular distribution and spatially dependent intercellular communication within different tumor regions.
  • Algorithms and Innovative Tools: InferCNV was used to infer copy number variation, AUCell for pathway activity evaluation, NicheNet and CellPhoneDB for in-depth analysis of cell-cell communication, and SCFEA to estimate cellular metabolic flux. Large cohort GWAS and meta-analyses were used to genetically anchor gene sets and integrate multi-omics data.

3. Multiple Validation Approaches

  • Protein Validation: Immunohistochemistry (IHC) and multiplex immunofluorescence (MIF) were used to quantify the expression and distribution of key proteins across different tumor subtypes.
  • Flow Cytometry: Quantitative machine analysis of the differences in APOE+ macrophages and inflammatory fibroblasts content in EOPC/LOPC tumor tissues.
  • Clinical Data Analysis: Large databases such as TCGA were referenced to review relationships between molecular subtyping, survival data, and therapeutic responses.

Major Findings and Data Analysis

1. Distinct Cellular Composition of EOPC vs LOPC Tumor Microenvironments

The study first showed that EOPC tumors had significant enrichment of immune cells (such as T cells and myeloid cells) and vascular smooth muscle cells, exhibiting high immune activation and metabolic reprogramming, while LOPC tumors showed pronounced increases in epithelial and fibroblast cells. Bulk CNV analysis confirmed the tumor origins and cell type identities, laying a foundation for subsequent molecular profiling.

2. Molecular Signaling Pathways: Androgen Responsiveness in EOPC, Hormonal Resistance in LOPC

Using GSEA and CancerSEA databases, the study found that malignant epithelial cells in EOPC highly activated androgen response pathways (with genes like KLK3, NDRG1, ABCC4 highly expressed), accompanied by notable hypoxia, lipid metabolism, and TNF-NF-κB signaling activation—indicating tumors in a state of energy reprogramming and immune microenvironment co-regulation. In contrast, LOPC emphasized epithelial-mesenchymal transition (EMT) and upregulation of metastasis and drug-resistance related genes (such as CST1, TFF3), matching with late-stage cancers and low-sensitivity hormonal environments.

3. AR-related Transcriptional Program (AR-meta-program, AR-MP) Defines Subtype Core

Through non-negative matrix factorization (NMF), six main transcriptional meta-programs were identified, among which the androgen response program (AR-MP) served as the subtype core. AR-MP was significantly upregulated in EOPC tumors, consistent with protein levels (KLK3, NYP, ERP60), while LOPC was dominated by aggressive EMT programs. Spatial data indicated that AR-MP-activated tumor regions in EOPC were highly infiltrated by myeloid cells (especially APOE+ macrophages), suggesting a unique molecular microenvironment “co-evolution.”

4. APOE+ Tumor-associated Macrophages Promote EOPC Progression and Induce Immune Suppression

After dissecting 13 subtypes of myeloid cells, the study found that APOE+ tumor-associated macrophages (TAMs) were extremely enriched in EOPC, exhibiting characteristics of M2 polarization, myeloid-derived suppressor cells (MDSC), and lipid metabolic reprogramming (fatty acid degradation, cholesterol metabolism, etc.). These APOE+ TAMs were spatially colocalized with AR-MP high-expressing tumor epithelial regions and influenced tumor growth and invasiveness via axes such as FN1-integrin and EREG-EGFR, while markedly suppressing T cell effector function (increased immune suppression in CD4+ T cells). High APOE+ TAM scores were associated with increased disease recurrence, progression, and poor overall survival.

5. Inflammatory Cancer-associated Fibroblasts (iCAF) Drive EMT and Drug-resistant Phenotypes in LOPC

LOPC tumors were mainly enriched for inflammatory cancer-associated fibroblasts (iCAF, defined as PDGFRA high/FAP high), which secrete large amounts of bone morphogenetic proteins (BMP4/5/7) to mediate BMPR signaling and SMAD pathways. This strongly downregulates tumor AR-MP, significantly activates EMT-related pathways, and shapes a prostate cancer phenotype characterized by dedifferentiation, drug resistance, and metastasis. Spatial analysis of iCAF revealed that they were mainly distributed at the tumor margins, enveloping AR-MP low-expressing tumor cells and greatly influencing local microenvironments and tumor behavior. Large-scale cohort validation demonstrated that iCAF enrichment was closely associated with AR-MP degradation and shorter patient survival, with high iCAF-related BMP expression marking poor prognosis.

6. Advanced Age/Smoking Factors Promote Aberrant iCAF/EMT Activation and Induce Castration-resistant Phenotype

Another innovative discovery by the authors was the direct association of smoking history and aging with iCAF distribution and BMP secretion, further promoting tumor EMT and AR-MP downregulation, providing an explanation for drug resistance (even pre-existing CRPC subtypes) in LOPC tumors. Single-cell and spatial data, combined with three major CRPC cohorts, corroborated the link between iCAF, BMP high expression, and clinical establishment of castration-resistance.

7. Multiple Validation: Comprehensive Support from Animal Models, Protein, and Flow Cytometry

The mouse in situ tumor model showed that tumors in young mice (simulating EOPC) were more dependent on androgens, and castration therapy had a much more pronounced tumor-inhibitory effect than in aged mice. After ADT, AR-MP, high lipid metabolic, and hypoxia features in young mouse tumors were significantly downregulated, iCAF infiltration increased, and overall tumor growth slowed, providing strong evidence for a tight coupling between molecular mechanisms and clinical features. Proteomics and flow cytometry further quantitatively confirmed the abundance differences of APOE+ TAM and iCAF in EOPC/LOPC. Multi-cohort and meta-analysis data robustly validated the key biological characteristics, ensuring the scientific credibility of the findings.

Study Conclusions, Significance, and Prospects

This study is the first to map the microenvironment panorama of EOPC and LOPC at single-cell and spatial resolution, comprehensively proposing two tumor evolutionary trajectories: “EOPC—APOE+ TAM-driven—high androgen response—metabolic reprogramming—immune suppression” and “LOPC—iCAF/BMP-driven—attenuated androgen signaling—EMT—pre-existing drug resistance.” This sets a theoretical and practical foundation for age-related molecular subtyping, precision therapy, and adjunctive interventions in prostate cancer.

Scientifically, this study breaks through the “black box” of prostate cancer heterogeneity and reveals deep interactions between age, immunity, stromal components, and molecular signaling pathways; in application, it points to clinical stratified treatment directions: EOPC patients may benefit from androgen deprivation, ERBB inhibition, lipid metabolism targeting, and TAM blockade, while LOPC should focus on CAF/BMP/EMT/castration-resistance blocking strategies. The study emphasizes that targeting key microenvironmental cells (APOE+ TAM, iCAF) is expected to accelerate the implementation of precision medicine. Mechanistic links between lifestyle (such as smoking), aging, and disease molecular mechanisms also provide tangible entry points for prevention and intervention.

Research Highlights and Innovations

  1. Innovative Typing Strategy for Prostate Cancer: For the first time, the study proposes an age of onset-microenvironment-dominated typing system, thoroughly integrating genetic, molecular, cellular, spatial, and clinical signals.
  2. Identification and Function of APOE+ Tumor-associated Macrophages: It clarifies that a specific macrophage subtype in EOPC tumors dominates immune suppression and metabolic reprogramming, suggesting that dual targeting of lipid metabolism and macrophages may revolutionize EOPC therapy.
  3. Inflammatory CAF and Pre-existing Castration-resistance Mechanism Revealed: Systematic modeling of the CAF–BMP–EMT axis in LOPC opens up new directions for treating drug resistance and metastasis.
  4. Deep Multi-Method Validation by Spatial Resolution/Animal Models/Large Cohorts: The research pipeline is complete and systematic, greatly enhancing scientific credibility and clinical translation potential.
  5. Close Integration of Theory and Clinical Practice: The study not only describes mechanisms but also combines in vivo mouse models, flow cytometry, proteomics, and clinical survival analysis, promoting translation to real-world medical scenarios.

Other Noteworthy Points

  • Strict Ethics Compliance: All human and mouse experimental procedures passed review by the Nanjing Medical University Ethics Committee, meeting international ethical standards.
  • Innovative Algorithms/Platform Integration: The latest international single-cell and spatial multi-modal omics technologies and innovative algorithms were used, setting a template for subsequent research into tumor microenvironment heterogeneity.
  • Investigation into How Lifestyle Factors Like Smoking and Aging Interact with Disease Molecular Mechanisms: Offers new, feasible approaches for public health prevention and early intervention.

Conclusion

With rigorous scientific methods, cutting-edge multi-disciplinary technology, and clinically oriented thinking, this study comprehensively redraws the heterogeneity landscape of age-related prostate cancer, clarifies new mechanisms of key molecular pathways and microenvironmental cells, and brings practically feasible new strategies for individualized therapy. In the future, it is expected to inspire more breakthroughs in mechanism studies, drug development, and precision interventions, speeding malignant tumor treatment into the “precision era.”