Single-cell transcriptome-wide Mendelian randomization and colocalization analyses reveal immune-cell-specific mechanisms and actionable drug targets in prostate cancer
Curation statements for this article:-
Curated by eLife
eLife Assessment
This study presents a useful compendium of triangulated single-cell eQTLs, Mendelian randomisation and colocalization of genetic signals in prostate cancer datasets. Biological interpretation in the context of the aging prostate gland, the tumour microenvironment and immune cell specificity is incomplete, so this study is a starting point for further study, and would require validation of the resulting putative causal genes.
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (eLife)
Abstract
Prostate cancer remains a major health burden with limited success in immune-targeted therapies. To identify immune-cell-specific therapeutic targets, we integrated single-cell cis-eQTL data across 14 immune cell types, bulk eQTLs, GWAS summary statistics from PRACTICAL and FinnGen, and single-cell RNA-seq data from prostate tumors. Using Mendelian randomization and Bayesian colocalization, we prioritized 80 causal eGenes with shared genetic signals, especially in CD4 and CD8 T cells. Functional analyses revealed enrichment in immune-related pathways such as antigen processing and cytokine signaling. Meta-analysis validated 52 robust eGenes across cohorts. Single-cell transcriptomics confirmed cell-type-specific expression of key genes including HLA-DQA2 , TXN , and COX6B1 within the tumor microenvironment. Drug repurposing analysis identified potential therapeutic targets such as IGF1R and FAAH , with known drug interactions mapped via DrugBank and STRING. Our integrative framework highlights immune-cell-specific genetic drivers and actionable targets in prostate cancer, offering a high-resolution resource for precision immunotherapy development.
Article activity feed
-
eLife Assessment
This study presents a useful compendium of triangulated single-cell eQTLs, Mendelian randomisation and colocalization of genetic signals in prostate cancer datasets. Biological interpretation in the context of the aging prostate gland, the tumour microenvironment and immune cell specificity is incomplete, so this study is a starting point for further study, and would require validation of the resulting putative causal genes.
-
Reviewer #1 (Public review):
Summary:
Using Mendelian randomisation on available GWAS data, the investigators identified eGenes associated with prostate cancer and applied the data to define relevant immune cell types involved. Additional analysis was performed to explore potential candidate targets and agents from licensed medicines.
This is an interesting approach as the investigators have expertise in other research fields, applied here to prostate cancers. The use of three different datasets is significant, and the approach to further analyse implicated eGenes in drug target analysis is relevant and timely.
A particular strength is taking putative genes from Mendelian randomisation analysis to target and potential drug agents.
Some aspects of the study would need to be clarified to enable interpretation of the findings in the …
Reviewer #1 (Public review):
Summary:
Using Mendelian randomisation on available GWAS data, the investigators identified eGenes associated with prostate cancer and applied the data to define relevant immune cell types involved. Additional analysis was performed to explore potential candidate targets and agents from licensed medicines.
This is an interesting approach as the investigators have expertise in other research fields, applied here to prostate cancers. The use of three different datasets is significant, and the approach to further analyse implicated eGenes in drug target analysis is relevant and timely.
A particular strength is taking putative genes from Mendelian randomisation analysis to target and potential drug agents.
Some aspects of the study would need to be clarified to enable interpretation of the findings in the context of the prostate gland and prostate cancers: expanding the descriptions of the supporting Supplementary Data and Tables, explanations of the analysis for the general reader, and clarification of the selection of eGenes (Figure 5).
-
Reviewer #2 (Public review):
Summary:
This study integrates bulk and single-cell transcriptomic-derived eQTLs from two separate consortia (PRACTICAL and Finngen) to identify immune-cell-specific therapeutic targets in prostate cancer. Mendelian randomization and Bayesian colocalization have been used to produce druggable eGene modules through STRING and DrugBank.
This is an interesting study that is attempting to address risk-associated, immune-specific transcriptomic repertoires in prostate cancer. It is knitting together concepts of drug repurposing and prostate cancer immunogenicity. This is an entirely computational study, which would benefit from some wet lab experimental validation.
It is very tricky to attribute cell-type-specific responses, especially when the majority of genes involved represent cytoskeletal or stress …
Reviewer #2 (Public review):
Summary:
This study integrates bulk and single-cell transcriptomic-derived eQTLs from two separate consortia (PRACTICAL and Finngen) to identify immune-cell-specific therapeutic targets in prostate cancer. Mendelian randomization and Bayesian colocalization have been used to produce druggable eGene modules through STRING and DrugBank.
This is an interesting study that is attempting to address risk-associated, immune-specific transcriptomic repertoires in prostate cancer. It is knitting together concepts of drug repurposing and prostate cancer immunogenicity. This is an entirely computational study, which would benefit from some wet lab experimental validation.
It is very tricky to attribute cell-type-specific responses, especially when the majority of genes involved represent cytoskeletal or stress responses, which are ubiquitous throughout the prostate microenvironment. This point is relevant for the drug repurposing section: if these drugs are targeting immune cell-specific repertoires, what would the response be of the entire environment? It would be useful to contextualize the validity of each proposed therapy in a specific prostate cancer context and the involvement of AR antagonism or radiotherapy.
Strengths and limitations of this study:
Strengths:
This is a scientifically interesting and potentially impactful study, particularly in its attempt to integrate immune-cell-specific transcriptomics, causal inference, and drug repurposing in prostate cancer. The methodology is well described, and the data (albeit limited) are well analyzed.
Limitations:
The central weakness is the overstatement of the conclusions regarding immune-cell-specific causality, without sufficiently contextualizing the biological meaning of the findings.
Highlighted genes, such as LMNA, XBP1, histone-related genes, and stress-response markers, are ubiquitous regulators involved in fundamental cellular processes, including ageing, unfolded protein response (UPR), integrated stress response (ISR), chromatin remodeling, proliferation, and metabolism. It is unclear whether these signatures truly represent immune mechanisms, or instead reflect broader inflammatory and age-associated biology expected within an ageing glandular organ such as the prostate.
Immune cell identity alone may not be sufficient to infer biological relevance because immune state characterization (e.g., exhausted versus functional T cells, or distinct macrophage/myeloid phenotypes) is largely absent from the current analysis. The assertion that specific immune populations are correlated with prostate cancer susceptibility is probably an overstatement unless the nature of these cells can also be characterized.
The interpretation of "causal variants" is not always specified, i.e., what phenotype is being associated: prostate cancer susceptibility, recurrence, progression, or treatment response (e.g. is there direct causality from immune-cell variants to prostate cancer?).
Overall, there is a need for stronger biological and translational contextualization: how do the identified pathways relate to ageing-associated inflammation, PIN, microbiome-driven inflammatory changes, and stress-response biology in the prostate gland? While the manuscript identifies network hubs and enriched pathways, it often stops short of explaining what these modules biologically represent or how they may influence prostate cancer development, progression, treatment resistance, or immune evasion.
There are additional publicly available spatial transcriptomic or single-cell datasets which could be used to validate whether the purported immune-cell-specific genes are genuinely enriched in immune populations adjacent to tumour cells. In the drug repurposing analyses, the current study does not explicitly handle prostate cancer subtypes such as HSPC, CRPC, NEPC, or DNPC and co-treatment with androgen receptor antagonism or radiotherapy.
-
