Proteome-wide association study and functional validation identify novel protein markers for pancreatic ductal adenocarcinoma

Jingjing Zhu, Ke Wu, Shuai Liu, Alexandra Masca, Hua Zhong, Tai Yang, Dalia Ghoneim, Praveen Surendran, Tanxin Liu, Qizhi Yao, Tao Liu, Sarah Fahle, Adam S. Butterworth, Md Ashad Alam, Jaydutt V. Vadgama, Youping Deng, Hong‐Wen Deng, Chong Wu, Yong Wu, Lang WuGigaScience 2024 · DOI

TL;DR

This study leverages a novel proteome-wide association study (PWAS) approach, using genetic instruments to predict blood protein levels, to identify novel protein biomarkers for pancreatic ductal adenocarcinoma (PDAC). By functionally validating two of these novel proteins, GOLM1 and B4GALT1, the researchers demonstrate their critical roles in PDAC progression, opening avenues for improved early detection and therapeutic strategies.

What Problem Does This Paper Address?

Pancreatic ductal adenocarcinoma (PDAC) is a devastating cancer with a dismal prognosis, largely due to a lack of reliable biomarkers for early detection and effective treatment. Existing blood protein biomarkers often suffer from issues with replicability and may be confounded by various factors. This study aims to overcome these limitations by identifying novel, causally linked protein markers for PDAC risk using a more robust genetic approach.

Methods

The core of this study is a Proteome-Wide Association Study (PWAS). This approach uses genetic variants as instrumental variables to predict blood protein levels. This helps to infer causality and reduce confounding compared to traditional observational studies.

Here's a breakdown of their methodology:

  • Protein Genetic Prediction Model Development: They used plasma proteome and genotype data from the INTERVAL study. They built models to predict protein levels using genetic variants, incorporating both cis- (nearby) and trans- (distant) acting variants to enhance statistical power. Various prediction methods (BLUP, elastic net, LASSO, top1) were employed, and models were validated externally.
  • PWAS: The validated prediction models were then applied to summary statistics from a large genome-wide association study (GWAS) of PDAC risk. This allowed them to assess the association between genetically predicted protein levels and PDAC risk.
  • Robustness Analyses: To ensure the reliability of their findings, they performed several robustness checks, including using different prediction modeling methods, pruning correlated SNPs, and examining various P-value cutoffs for trans SNPs.
  • Somatic Variant Analysis: They investigated whether genes encoding associated proteins harbor deleterious somatic mutations in PDAC patients using TCGA data.
  • Pathway and Interaction Analysis: Ingenuity Pathway Analysis (IPA) was used to identify enriched pathways and biological functions associated with the identified genes. A protein-protein interaction (PPI) network was also constructed using the STRING database.
  • Drug Repurposing: They explored potential drug repurposing opportunities by identifying existing drugs that target the identified proteins using OpenTargets and DrugBank databases, further assessing potential interactions with molecular docking.
  • In Vitro Functional Validation: Two novel protein-encoding genes, GOLM1 and B4GALT1, were selected for experimental validation. This involved using RNA interference (siRNA) to silence these genes in human pancreatic cancer cell lines (PANC-1 and SU.86.86) and assessing their impact on cell proliferation, migration, and invasion through MTS assays, wound scratch assays, and Transwell invasion assays.

Datasets

  • INTERVAL Study: This dataset provided plasma proteome and genotype data from healthy European subjects, used for developing and validating protein genetic prediction models. It included approximately 830,000 genetic variants and measurements for 3,620 plasma proteins.
  • PanScan and PanC4 Consortia: These datasets provided GWAS summary statistics for PDAC risk, including data from 8,275 PDAC cases and 6,723 controls of European ancestry.
  • The Cancer Genome Atlas (TCGA): Used for analyzing somatic-level genetic mutations in genes encoding associated proteins in PDAC patients.

Key Insights and Ideas

  • Genetics as a Tool for Causal Inference in Proteomics: The study highlights the power of using genetic variants as instrumental variables (a form of Mendelian randomization) to infer causal relationships between blood protein levels and disease risk, overcoming limitations of traditional observational studies.
  • Novel Protein Biomarker Candidates: By moving beyond existing biomarkers and employing a large-scale PWAS, the study identified 16 novel proteins associated with PDAC risk, significantly expanding the potential pool of diagnostic and prognostic markers.
  • Functional Validation is Crucial: The study emphasizes that identifying associations is only the first step. The functional validation of GOLM1 and B4GALT1 provides strong evidence for their direct involvement in PDAC progression, making them particularly promising targets.
  • Drug Repurposing Potential: The identification of existing drugs targeting these novel proteins opens exciting possibilities for repurposing them to treat PDAC, potentially accelerating the development of new therapies.

Key Results

  • The study successfully developed and validated prediction models for 1,389 proteins.
  • A PWAS identified 40 proteins significantly associated with PDAC risk (FDR P < 0.05).
  • Of these, 16 proteins are novel, meaning they have not been previously reported in association with PDAC risk.
  • Functional validation showed that silencing of GOLM1 and B4GALT1 significantly inhibited cell proliferation, migration, and invasion in PDAC cell lines.
  • IPA analysis revealed that genes encoding the identified proteins are enriched in pathways such as IL-15 Production and Heparan Sulfate Biosynthesis.
  • The study identified potential drug repurposing opportunities for several proteins, with MET being a notable target for multiple existing drugs.

Limitations and Future Work

  • Causality Inference: While the PWAS approach strengthens causal inference, pleiotropic effects of genetic variants and potential unmeasured confounding cannot be entirely ruled out. Future functional studies are needed to definitively establish causality.
  • Biomarker Utility: The study focused on genetically predicted protein levels. Further research is required to assess the predictive role of these proteins in pre-diagnostic blood samples for actual PDAC risk assessment.
  • Predictor Selection: The current model development relied on statistically associated SNPs. Future work could explore additional predictors to improve model construction and evaluate more proteins.
  • Covariate Adjustment: The study acknowledges the inability to adjust for covariates like smoking and BMI in the INTERVAL dataset, which could influence protein levels.
  • Ancestry: The study primarily focused on European populations. Future research should validate these findings in diverse ancestries.
  • Drug Repurposing: While potential drug targets were identified, further preclinical and clinical studies are necessary to evaluate their efficacy in treating PDAC.

Summary updated

Abstract

PDAC is a notoriously difficult-to-treat malignancy, and our limited understanding of causal protein markers hampers progress in developing effective early detection strategies and treatments. Our study identifies novel causal proteins using genetic instruments and subsequently functionally validates selected novel proteins. This dual approach enhances our understanding of PDAC etiology and potentially opens new avenues for therapeutic interventions.
BibTeX
@inproceedings{zhu2024proteome,
  author = {Jingjing Zhu and Ke Wu and Shuai Liu and Alexandra Masca and Hua Zhong and Tai Yang and Dalia Ghoneim and Praveen Surendran and Tanxin Liu and Qizhi Yao and Tao Liu and Sarah Fahle and Adam S. Butterworth and Md Ashad Alam and Jaydutt V. Vadgama and Youping Deng and Hong‐Wen Deng and Chong Wu and Yong Wu and Lang Wu},
  title = {{Proteome-wide association study and functional validation identify novel protein markers for pancreatic ductal adenocarcinoma}},
  booktitle = {GigaScience},
  year = {2024},
  doi = {10.1093/gigascience/giae012},
}
Figure