Prostate cancer is the second most diagnosed cancer in men worldwide, with approximately 1.4 million new cases and 375,000 deaths reported globally in 2020. While survival has improved in many countries due to better screening and treatments, it remains a major public health burden, and new approaches are urgently needed to understand its molecular biology and develop more effective therapies.
Prostate cancer progresses through three main stages: intraepithelial neoplasia (PIN), a precancerous state; hormone-sensitive prostate cancer (HSPC), which responds to treatments that reduce male sex hormones (androgens); and castration-resistant prostate cancer (CRPC), which continues to grow even when androgen levels are suppressed. CRPC represents the most deadly and treatment-resistant form of the disease.
The central driver of prostate cancer is the androgen receptor (AR), a protein that is activated by testosterone and dihydrotestosterone to control the expression of hundreds of genes essential for prostate cell growth. This knowledge has led to several approved drugs including abiraterone, enzalutamide, apalutamide, and darolutamide that target the AR pathway at different points. However, resistance to these drugs almost inevitably develops.
To understand how prostate cancer originates, progresses, and develops treatment resistance, researchers have employed a wide array of omics technologies: high-throughput tools that measure all molecules of a particular class in a biological sample simultaneously. These include genomics, epigenomics, transcriptomics, proteomics, and metabolomics, each providing a different layer of molecular insight.
Prostate cancer has a comparatively low mutation rate but a high degree of copy number alterations, where large genomic segments are duplicated or deleted. The most common early genetic alteration is a chromosomal rearrangement fusing the TMPRSS2 and ERG genes, an androgen-regulated fusion that enhances ERG expression and occurs in roughly half of prostate cancer cases. Mutations in the SPOP and FOXA1 genes are also frequent early events.
As prostate cancer progresses to metastatic hormone-sensitive disease (mHSPC), mutations accumulate in the TP53 tumor suppressor, the DNA damage repair pathway (including BRCA1, BRCA2, and ATM genes), and the PTEN/PI3K pathway. These changes are associated with worse clinical outcomes and provide targets for emerging therapies, including PARP inhibitors now approved for BRCA-mutated prostate cancers.
In metastatic castration-resistant prostate cancer (mCRPC), the most critical changes involve amplification of the AR gene itself, mutations in the AR ligand-binding domain that make the receptor resistant to AR antagonists, and loss of the tumor suppressors RB1 and TP53. The emergence of AR-V7, a splice variant of the AR that lacks the ligand-binding domain and is therefore constitutively active regardless of drug treatment, is a major mechanism of resistance.
Nearly 270 genetic risk loci identified by genome-wide association studies (GWAS) have been linked to prostate cancer susceptibility. Most of these risk variants are in non-coding regulatory regions, affecting gene transcription rather than protein-coding sequences. Ethnic differences in mutation frequencies, such as higher rates of certain PTEN and AR alterations in Black men and different FOXA1 mutation rates in Asian men, highlight the importance of diverse patient cohorts in genomic research.
Epigenomics studies heritable changes in gene expression that do not involve DNA sequence changes. In prostate cancer, altered DNA methylation patterns are widespread, with many tumor suppressor genes silenced by hypermethylation of their promoter regions. A distinct DNA methylation profile characteristic of treatment-emergent small-cell neuroendocrine prostate cancer has been identified, enabling subtype classification from tissue samples.
Histone modifications, including acetylation and methylation of specific amino acids on histone proteins, control whether genes are accessible for transcription. Specific patterns of histone marks at AR-bound enhancers, particularly H3K27 acetylation, are essential for androgen-driven gene activation. Striking differences in these patterns are observed between primary and metastatic prostate cancer, reflecting the dramatic transcriptional reprogramming that occurs during progression.
Cistromics maps where proteins like the AR physically bind across the entire genome. The AR does not bind gene promoters directly but instead contacts distal regulatory elements called enhancers. The AR cistrome changes dramatically between normal prostate, primary cancer, and CRPC, shifting gene programs from differentiation and growth suppression toward survival and proliferation. Treatment with AR antagonists reverses these binding patterns, providing insight into drug mechanisms.
Non-coding RNAs, including microRNAs and long non-coding RNAs (lncRNAs), add another regulatory layer. The lncRNA HOTAIR is upregulated by androgen deprivation therapy, stabilizes the AR, and promotes cancer growth, invasion, and metastasis. Some tumor-suppressive microRNAs are downregulated in prostate cancer and have been proposed as potential therapeutic agents, while others serve as diagnostic biomarkers in blood and urine.
Transcriptomics measures the complete set of RNA molecules in a cell, providing a snapshot of which genes are active. In prostate cancer, transcriptomic profiling has identified gene expression subtypes that predict biochemical recurrence, metastasis, and response to treatment. A 70-transcript signature predicting elevated risk of recurrence and metastasis was identified early using microarray technology, and more recently RNA sequencing has provided much higher resolution and sensitivity.
A critical finding from transcriptomics is the emergence of AR splice variants, particularly AR-V7, in castration-resistant prostate cancer. AR-V7 lacks the ligand-binding domain, making it constitutively active and insensitive to enzalutamide and abiraterone. Its expression is strongly correlated with resistance to these therapies, and it activates distinct gene programs including cell cycle genes and DNA damage response pathways that differ from those controlled by full-length AR.
Single-cell RNA sequencing has uncovered previously unrecognized cell populations in normal prostate tissue with progenitor function that may be involved in cancer initiation. In metastatic disease, single-cell studies reveal an unexpected immune microenvironment in bone metastases, with actionable immune signatures not apparent from bulk tissue analysis. This suggests that immune checkpoint therapies, which have limited efficacy in primary prostate cancer, might be more effective in specific metastatic contexts.
Transcriptome analysis of steroid synthesis genes reveals that multiple enzymes involved in intratumoral androgen biosynthesis are elevated in CRPC compared to primary tumors. This allows CRPC cells to produce their own androgens even when circulating levels are suppressed, representing a key resistance mechanism. These findings directly informed the development of abiraterone acetate, which blocks intratumoral androgen synthesis.
Proteomics measures the complete protein content of a cell or tissue, capturing the final functional output of gene expression. In prostate cancer, large-scale proteomic studies have consistently identified upregulation of proteins involved in fatty acid synthesis and lipid metabolism in both primary and metastatic tumors. Cell cycle and DNA damage response pathway proteins are the most consistently altered in advanced disease across independent studies.
Critically, the correlation between protein levels and the corresponding RNA or DNA data is often limited in prostate cancer, particularly in advanced tumors. This means that genomic or transcriptomic data alone cannot reliably predict which proteins are actually present and active in cancer cells, underlining the value of dedicated proteomics studies to provide independent molecular insight.
Post-translational protein modifications, including phosphorylation, ubiquitylation, and glycosylation, have emerged as important regulators of cancer progression. The ubiquitin ligase adaptor protein SPOP, mutated in about 11% of primary prostate cancers, controls the stability of key cancer-driving proteins including the AR, BRD4, and TRIM24. SPOP mutations prevent these proteins from being degraded, leading to their abnormal accumulation and enhanced oncogenic signaling.
Metabolomics reveals extensive metabolic rewiring in prostate cancer. Elevated de-novo lipogenesis (fat synthesis) is a hallmark of prostate cancer, and androgens directly drive upregulation of lipid synthesis genes. In CRPC, increased cholesteryl ester accumulation and altered phospholipid composition have been documented. Urine metabolomics offers a non-invasive approach to identifying diagnostic and prognostic biomarkers, with multiple studies identifying altered metabolites related to energy production, amino acid metabolism, and the TCA cycle.
Individual omics approaches each capture only a slice of the complex molecular landscape of cancer. Multi-omics integration combines data from genomics, epigenomics, transcriptomics, proteomics, and metabolomics to reveal the interconnections between different regulatory levels that no single approach can provide. Methods including iCluster, Bayesian consensus clustering, and network-based algorithms have been developed specifically for this purpose.
A key insight from multi-omics integration is that the different omics layers are only partially correlated. Genomic copy number changes do not reliably predict transcript levels, transcript levels do not reliably predict protein abundance, and protein abundance does not always predict metabolic activity. This disconnect means that each omics layer provides genuinely independent and complementary information, making integration necessary rather than redundant.
Multi-omics analysis of prostate cancer has led to a new molecular classification system, identified crosstalk between epigenome alterations and metabolic dysregulation, and uncovered novel potential drug targets not apparent from single-omics approaches. Large public repositories including The Cancer Genome Atlas, the International Cancer Genome Consortium, and the Cancer Proteome Atlas have made multi-omics data from thousands of tumor samples available to researchers worldwide.
Emerging spatial transcriptomics adds geographic context to gene expression, revealing how gene activity varies not just between patients but within different regions of the same tumor, including normal tissue, precancerous areas, and invasive cancer. This spatial resolution is critical for understanding the tumor ecosystem and the interactions between cancer cells and their surrounding microenvironment.
The convergence of multi-omics technologies with improved computational methods promises to transform prostate cancer management. As multi-omics datasets grow larger and analytical tools become more sophisticated, it will become possible to define patient subgroups at unprecedented molecular resolution, identifying which patients are most likely to benefit from specific therapies and which are at risk for rapid disease progression.
Key challenges remain, including the curse of dimensionality: the number of molecular variables measured (potentially millions) vastly exceeds the number of patients studied, making statistical analysis prone to overfitting. Harmonizing data from different omics platforms, institutions, and time points also requires careful computational approaches to avoid technical artifacts confounding biological signals.
The growing availability of liquid biopsies, which analyze circulating tumor DNA, RNA, proteins, and metabolites in blood or urine, opens the possibility of serial molecular monitoring without repeat invasive tissue biopsies. These non-invasive approaches could track how the molecular landscape of a patient's cancer evolves under treatment, enabling earlier detection of emerging drug resistance and real-time treatment adaptation.
The ultimate goal is a precision medicine framework where each prostate cancer patient's treatment is guided by their individual molecular profile across multiple omics layers, ensuring that therapy targets the specific vulnerabilities of their tumor. Combined with advances in medical imaging and artificial intelligence, the molecular characterization provided by multi-omics approaches will form the foundation for a new era of personalized prostate cancer care.