Can proteins in your blood predict and help treat breast cancer? A new large-scale study uncovers key biomarkers and connects them to existing drugs, offering hope for targeted therapies.
Study: Genetic associations of plasma proteins and breast cancer identify potential therapeutic drug candidates. Image Credit: Yomal2233 / Shutterstock
In a recent study in the journal Communication Biology, researchers employed a novel combined Mendelian randomization (MR) analysis to identify 62 plasma proteins (including 9 with robust support, 13 with medium support, and 40 with limited support) associated with breast cancer and its Luminal A or B subtypes.
Unlike previous approaches using only a single MR analysis on limited cohorts, this research used both two-sample MR (TSMR) and summary-data-based MR (SMR) on a cohort of nearly 250,000 participants.
Findings from both human models revealed 9 robust and 13 medium-confidence plasma protein genes. Robustly associated proteins include: 1. Breast cancer – ULK3, ASIP, CSK, TLR1; 2. Luminal A – ADH5, ULK3, SARS2, UBE2N; 3. Luminal B – PEX14. Six of the nine robust proteins were supported by murine phenotype data, confirming their relevance to immune and hematopoietic systems.
However, the reduction in CSK and ULK3 expression in cancer versus healthy tissue was observed in human tissue via immunohistochemistry.
Three existing drugs (TG100801, Hydrochlorothiazide, and Imatinib) were identified as genetically or biologically linked to these proteins, but their mechanisms of action and therapeutic relevance in breast cancer remain to be elucidated.
Background
Breast cancer (BC) is a serious non-communicable disease caused by uncontrolled breast cell growth. BC and its four subtypes cause the most cancer-related deaths among women globally (24.5% incidence; 15.5% mortality, 685,000 deaths).
Recent research has turned to plasma proteins for insight due to their roles in disease biomarker identification. Past studies have shown plasma protein dysregulation in diseases like IBD and CVD.
Mendelian randomization (MR) studies use gene variation to identify causal effects between exposures and outcomes. Despite progress, past BC research has struggled with small sample sizes, limited durations, and over-reliance on single MR methods.
About the Study
This study overcomes previous limitations by integrating MR methods to examine plasma protein variation, drug discovery, and biological mechanisms.
Genotyping data came from the Breast Cancer Association Consortium (247,173 samples), and plasma proteome data came from deCODE (4,907 proteins from 35,559 individuals). Inclusion criteria for pQTLs were: (i) Genome-wide significance (P < 5 × 10⁻⁸); (ii) Outside MHC region (chr6, 25.5–34.0 Mb); (iii) Distinct LD clumping (r² < 0.01, 10,000 kb window); (iv) Cis-acting pQTLs.
MR analyses treated plasma proteins as exposures and BC subtypes as outcomes. Associations were validated via colocalization to test robustness. GeneMANIA was used for interaction mapping; pathway analysis clarified biological functions.
Mouse Genome Informatics and TICI databases provided mouse knockout and immune infiltration data. Notably, while mouse models supported gene relevance, CSK and ULK3 expression changes were seen only in human tissues. Drug candidates were matched using DrugBank, DGIdb, CheMBL, and the Therapeutic Target Database.
Study Findings
Of 4,907 proteins, 1,815 met pQTL criteria and were analyzed. Sixty-two showed an association with breast cancer or its subtypes.
Colocalization revealed: - 9 robust proteins (e.g., ULK3, ASIP, CSK, TLR1, ADH5, SARS2, UBE2N, PEX14), - 13 medium, - 40 limited. These genes play roles in immunity and blood cell regulation.
Overexpression of CSK and ULK3 in MCF-7 cells inhibited proliferation and migration, confirmed by in vitro tests. High ULK3 expression was also associated with prolonged recurrence-free survival, particularly in Luminal A breast cancer. GO analysis showed enrichment in nuclear transport, blood coagulation, and nucleocytoplasmic transport.
In cellular components (CC), the endoplasmic reticulum lumen was enriched (p = 0.001). For molecular functions (MF), serine-type endopeptidase inhibitor activity was significant (p = 0.0005). Disease Ontology (DO) analysis showed enrichment in nephritis (p = 0.026), glomerulonephritis (p = 0.014), lipid metabolism disorders (p = 0.007), and cervical squamous cell carcinoma (p = 0.017).
Three drugs (TG100801, Hydrochlorothiazide, and Imatinib) were flagged for their genetic associations with the robust proteins ULK3, CSK, and ADH5, but their clinical effectiveness in breast cancer has not been established and requires further study.
Limitations include data mostly from individuals of European ancestry and protein constraints limited to those in the deCODE database. Early-phase drugs might also have been missed.
Conclusions
This large-scale study identified 62 plasma proteins associated with breast cancer, including 9 with strong evidence. It outlined potential pathways and functions contributing to tumor behavior, such as cell migration and proliferation.
A comprehensive drug database search identified TG100801, Hydrochlorothiazide, and Imatinib as potential candidates for future breast cancer therapies.
High expression of ULK3 was associated with improved recurrence-free survival, suggesting potential prognostic value in Luminal A breast cancer. These discoveries deepen our understanding of BC biology and support the development of targeted, personalized therapies.
This study underscores the value of integrating genetic, functional, and pharmacological data in drug discovery.