Prostate cancer is the second most common cancer in men globally and the fifth leading cause of male cancer death. While the genetic landscape of prostate cancer is relatively well understood, this review argues that understanding proteins, the functional products of genes, provides a fundamentally different and essential layer of insight.
There is a critical gap between genetics and biology: not every genetic mutation or change in gene expression translates into a corresponding change in protein levels. In prostate cancer, only about 10% of protein abundance variation can be explained by RNA expression levels, meaning that gene-level studies miss the majority of what is actually happening in the cancer cell.
Proteins are the molecules that actually carry out cellular functions. They are the targets of most successful cancer drugs. Understanding which proteins are present, in what quantities, how they are modified, and how they interact with each other provides the functional readout that genome sequencing alone cannot capture.
Modern mass spectrometry (MS)-based proteomics can now reliably quantify thousands of proteins simultaneously from small tissue samples, enabling comprehensive studies of prostate cancer biology that were impossible just a decade ago.
The androgen receptor (AR) is the master regulator of prostate biology and the central driver of prostate cancer. It binds androgens (male sex hormones) in the cytoplasm and then moves to the nucleus to activate genes that drive cell growth and tumor progression.
Androgen deprivation therapy (ADT) is the first-line treatment for metastatic prostate cancer and works by reducing androgen levels to starve cancer cells. However, 80% of cases eventually develop castration-resistant prostate cancer (CRPC), where cancer continues to grow despite low androgen levels. Understanding how this happens at the protein level is a major focus of current research.
The most common genetic event in prostate cancer is a chromosomal rearrangement creating the TMPRSS2-ERG fusion, occurring in approximately 50% of cases. This places the ERG transcription factor under androgen control, connecting the gene fusion to androgen-driven tumor growth and metastasis.
Other key molecular alterations include loss of the tumor suppressor PTEN (in up to 50% of advanced tumors), amplification of the oncogene c-MYC, and mutations in AR itself, all of which accumulate as cancer progresses from localized disease to lethal metastatic CRPC.
Liquid chromatography-tandem mass spectrometry (LC-MS/MS) is the core technology enabling modern proteomics. It separates proteins from a tissue sample, fragments them into peptides, and measures their masses with high precision to identify and quantify which proteins are present.
Several quantification strategies have been developed. Label-free approaches compare protein intensities directly between samples, while labeling methods like SILAC and TMT tag proteins with heavy isotopes or chemical tags to enable more precise relative quantification across multiple samples simultaneously.
Beyond measuring total protein levels, specialized sub-proteomic approaches exist: phosphoproteomics captures proteins modified by phosphorylation (a key regulatory switch in signaling), glycoproteomics studies sugar modifications, and interactomics uses proximity labeling or immunoprecipitation to capture proteins that interact with a specific target protein.
A major technical advance has been the development of protocols for formalin-fixed paraffin-embedded (FFPE) tissue, which allows proteomics on archived tumor samples that were preserved decades ago, dramatically expanding available research material.
Across multiple large-scale studies, the most consistently altered protein pathways in primary prostate cancer involve metabolism: specifically, upregulation of fatty acid synthesis, changes in the TCA cycle (Krebs cycle), glycolysis, oxidative phosphorylation, and amino acid metabolism. This metabolic reprogramming is a hallmark of how cancer cells alter their energy production to fuel rapid growth.
The second most common alteration category is cell adhesion and cytoskeletal changes, with adhesion proteins consistently downregulated in cancer tissue. Reduced cell adhesion allows cancer cells to detach from their original location and migrate, a key step toward metastasis. Integrin pathways are particularly affected.
Proteomics has given new prominence to RNA processing and protein turnover pathways, which were less visible in earlier sequencing studies. Components of the spliceosome, ribosomal biogenesis, protein folding, and the ubiquitin-proteasome degradation system are all dysregulated in prostate cancer tissue.
Interestingly, intra-patient variability in protein profiles is higher in benign tissue than in cancer, suggesting that tumor tissue converges on a more consistent molecular phenotype than the normal tissue it arose from.
As prostate cancer progresses from localized primary tumor to castration-resistant and metastatic disease, the proteome undergoes dramatic additional changes, with heterogeneity increasing at every step. Bone metastases show significantly more protein diversity than primary tumors or benign tissue.
The most prominent newly altered pathways in advanced disease involve cell cycle regulation and DNA damage repair. While proliferative changes are not strongly evident in primary tumor proteomes, they become overt in CRPC, with cyclin-dependent kinase activities enriched in metastatic disease based on phosphoproteomic analysis.
Metabolic changes initiated in primary cancer are amplified in CRPC. The TCA cycle undergoes a second wave of changes during the development of treatment resistance, with distinct metabolic states identifiable for primary cancer versus CRPC. Fatty acid oxidation is particularly altered in distal metastases.
A striking finding from longitudinal comparison is that even treatment-naive metastatic samples cluster proteomically with CRPC samples rather than with primary tumors, suggesting that aggressive molecular features required for metastasis are acquired early, before any treatment is administered.
The most important message from integrative studies combining genomics, transcriptomics, and proteomics is the poor correlation between different molecular layers. Only about 2% of proteins have their abundance meaningfully associated with gene copy numbers, despite prostate cancer being substantially driven by copy number alterations.
The deletion of the major tumor suppressor PTEN, which affects RNA expression of roughly half of all studied genes, impacts the abundance of only 2.7% of proteins. This dramatic disconnect illustrates why targeting a gene mutation does not guarantee a predictable effect on protein function.
RNA expression levels explain only about 10% of protein abundance variability in primary prostate cancer, and this correlation weakens further as the disease progresses to CRPC. Less than one-third of pathways identified as altered by RNA analysis were also identified by proteomics, and vice versa, meaning the two approaches reveal largely non-overlapping biology.
Specific mechanisms driving this disconnection include post-transcriptional regulation by microRNAs (miRNAs), post-translational protein modifications, differential protein stability, and translational control. These regulatory layers are invisible to genomic and transcriptomic approaches but are captured by proteomics.
The androgen receptor interactome has been studied in significant detail using both immunoprecipitation-based and newer proximity labeling approaches. These studies reveal that AR interacts with hundreds of proteins beyond its known transcriptional cofactors, including components involved in RNA splicing, protein synthesis, and DNA repair.
AR mutations that arise during castration resistance, such as T877A (present in the widely used LNCaP cell line), produce altered interaction profiles where the receptor gains the ability to bind and be activated by non-androgen steroids including progesterone and glucocorticoids, providing a mechanism for treatment escape.
Studies using the BioID proximity labeling method have mapped which proteins are physically near AR when it is activated by androgens versus blocked by the drug enzalutamide. Enzalutamide significantly reduces AR interactions, while DHT-activated AR recruits chromatin remodeling complexes, methyltransferases, and demethylases that regulate gene accessibility.
These interaction maps have clinical implications: identifying proteins that uniquely interact with mutant or constitutively active AR variants in CRPC could reveal new therapeutic targets, especially proteins that are required for AR function in treatment-resistant disease but not in normal cells.
A major application of large-scale proteomics is the identification of better biomarkers for patient stratification. While the current standard biomarker PSA is limited in its ability to predict disease behavior, proteomics has already identified potential panels of proteins that better distinguish low-risk from high-risk cancers and patients likely to experience biochemical recurrence.
Protein clusters identified from proteomic data define patient subtypes independent of genomic subtypes, suggesting that the same genetic cancer can have fundamentally different protein activity patterns requiring different treatment approaches. This supports the concept of proteome-guided personalized medicine.
Phosphoproteomics in particular holds promise for guiding therapy decisions, since kinase activities visible in phosphoproteomic data directly indicate which signaling pathways are active and thus which kinase inhibitor drugs might be effective in a given patient's cancer.
The field is moving toward understanding not just what proteins are present in bulk tumor samples, but how proteins behave within the tumor microenvironment, including cancer-associated fibroblasts which play important roles in cancer progression and may themselves be therapeutic targets, as suggested by findings on the collagen-crosslinking enzyme LOX2.