Genome-wide association study of knee pain identifies associations with GDF5 and COL27A1 in UK Biobank

Weihua Meng, Mark J. Adams, Colin N. A. Palmer, Michelle Agee, Babak Alipanahi, Robert K. Bell, Katarzyna Bryc, Sarah L. Elson, Pierre Fontanillas, Nicholas A. Furlotte, Barry Hicks, David A. Hinds, Karen E. Huber, Ethan M. Jewett, Yunxuan Jiang, Aaron Kleinman, Keng‐Han Lin, Nadia K. Litterman, Jennifer C. McCreight, Matthew H. McIntyre, Kimberly F. McManus, Joanna L. Mountain, Elizabeth S. Noblin, Carrie A. M. Northover, Steven J. Pitts, G. David Poznik, J. Fah Sathirapongsasuti, Janie F. Shelton, Suyash Shringarpure, Chao Tian, Joyce Y. Tung, Vladimir Vacic, Xin Wang, Catherine H. Wilson, Jingchunzi Shi, Adam Auton, Kathleen A. Ryan, Joanne M. Jordan, Braxton D. Mitchell, Rebecca D. Jackson, Michelle S. Yau, Andrew M. McIntosh, Blair H. SmithCommunications 2019 · DOI

TL;DR

This study conducted a large-scale genome-wide association study (GWAS) to identify genetic variants associated with knee pain. They discovered two significant genetic loci, one in the GDF5 gene and another near the COL27A1 gene, which were further supported by independent cohorts. This research sheds light on the genetic underpinnings of knee pain, a common and debilitating condition.

What Problem Does This Paper Address?

Knee pain is a widespread health issue, affecting a substantial portion of the population, particularly those over 50. While osteoarthritis is a common cause, knee pain can also arise from other factors and is a significant contributor to disability and healthcare costs. Despite its prevalence, the genetic basis of knee pain itself, as opposed to just knee osteoarthritis, has not been extensively studied. Previous genetic research has largely focused on osteoarthritis, leaving a gap in understanding the specific genetic factors contributing to the experience of knee pain. This paper aims to fill that gap by performing a GWAS specifically on knee pain.

Methods

The researchers employed a genome-wide association study (GWAS) approach. For the discovery phase, they utilized data from the UK Biobank, a large cohort study. They defined "knee pain" based on a specific questionnaire response indicating pain that interfered with usual activities in the past month. To validate their findings, they used three independent cohorts: 23andMe (self-reported osteoarthritis), the Osteoarthritis Initiative (OAI), and the Johnston County Osteoarthritis Project (JoCo) (radiographic knee osteoarthritis). Standard GWAS quality control procedures were applied to filter SNPs and individuals. Statistical analysis involved logistic regression for case-control association testing, adjusting for relevant covariates. They also performed gene-set and tissue expression analyses using FUMA and investigated genetic correlations with other traits using LD Hub.

Datasets

  • Discovery Cohort: UK Biobank (N=171,516 subjects with defined knee pain cases and controls).
  • Supporting Cohorts:
    • 23andMe (self-reported osteoarthritis).
  • Osteoarthritis Initiative (OAI) (radiographic knee osteoarthritis).
  • Johnston County Osteoarthritis Project (JoCo) (radiographic knee osteoarthritis).

The UK Biobank dataset used a specific questionnaire for knee pain, differing from osteoarthritis-focused phenotypes in the supporting cohorts.

Key Insights and Ideas

  • Knee Pain as a Distinct Phenotype: The study highlights the importance of treating knee pain as a distinct phenotype, rather than solely as a symptom of knee osteoarthritis, given that not all knee pain is osteoarthritis-related and vice-versa.
  • GDF5 and COL27A1 Link to Knee Pain: The identification of GDF5 and COL27A1 as genes associated with knee pain provides crucial insights into the biological pathways involved. GDF5 is already known for its role in skeletal development and osteoarthritis, reinforcing its relevance. COL27A1, while less studied in this context, is implicated in cartilage and bone development, suggesting its potential role in knee joint integrity and pain.
  • Genetic Correlation with Socioeconomic Factors: The finding of negative genetic correlations between knee pain and educational attainment (years of schooling, college completion) and age of first baby is intriguing. This suggests that genetic predispositions to knee pain might be intertwined with lifestyle and occupational factors that are often correlated with socioeconomic status.

Key Results

  • The GWAS in the UK Biobank identified two loci reaching genome-wide significance for knee pain:
  • rs143384 in the GDF5 gene (P = 1.32 × 10<sup>-12</sup>).
  • rs2808772 near the COL27A1 gene (P = 1.49 × 10<sup>-8</sup>).
  • These associations were replicated in independent cohorts, even when using osteoarthritis as a proxy phenotype. For instance, rs143384 showed a P value of 2.44 × 10<sup>-9</sup> in the 23andMe cohort and 0.01 in the combined OAI and JoCo cohorts.
  • The SNP-based heritability of knee pain was estimated at 0.08 (SE = 0.03), suggesting a moderate genetic contribution.
  • Genetic correlation analysis revealed significant negative correlations between knee pain and years of schooling (r<sub>g</sub> = -0.29), college completion (r<sub>g</sub> = -0.36), and age of having first baby (r<sub>g</sub> = -0.30).

Limitations and Future Work

  • Phenotypic Heterogeneity: A key limitation acknowledged is the use of different, though related, phenotypes across the discovery and supporting stages. The UK Biobank's knee pain definition was broad, and the supporting cohorts used osteoarthritis (self-reported or radiographic). This broadness, while necessary due to data availability, might obscure more specific genetic signals.
  • Low Heritability: The relatively low SNP-based heritability suggests that environmental and lifestyle factors play a significant role in knee pain development, and the current GWAS might not capture all genetic contributions.
  • Future Directions: The authors suggest further investigation into the relationship between knee pain and knee osteoarthritis. They also propose exploring the genetic aetiology of knee pain in relation to socioeconomic traits, which could lead to a better understanding of its complex origins and potential interventions. The identified loci provide a starting point for functional studies to elucidate the precise mechanisms by which GDF5 and COL27A1 influence knee pain.

Summary updated

Abstract

Knee pain is one of the most common musculoskeletal complaints that brings people to medical attention. Approximately 50% of individuals over the age of 50 report an experience of knee pain within the past 12 months. We sought to identify the genetic variants associated with knee pain in 171,516 subjects from the UK Biobank cohort and seek supporting evidence in cohorts from 23andMe, the Osteoarthritis Initiative, and the Johnston County Osteoarthritis Project. We identified two loci that reached genome-wide significance in the UK Biobank: rs143384, located in <i>GDF5</i> (<i>P</i> = 1.32 × 10<sup>-12</sup>), a gene previously implicated in osteoarthritis; and rs2808772, located near <i>COL27A1</i> (<i>P</i> = 1.49 × 10<sup>-8</sup>). These findings were supported in cohorts with self-reported osteoarthritis/radiographic knee osteoarthritis without pain information. In this report on genome-wide association of knee pain, we identified two loci in or near <i>GDF5</i> and <i>COL27A1</i> that are associated with knee pain.
BibTeX
@inproceedings{meng2019genome,
  author = {Weihua Meng and Mark J. Adams and Colin N. A. Palmer and Michelle Agee and Babak Alipanahi and Robert K. Bell and Katarzyna Bryc and Sarah L. Elson and Pierre Fontanillas and Nicholas A. Furlotte and Barry Hicks and David A. Hinds and Karen E. Huber and Ethan M. Jewett and Yunxuan Jiang and Aaron Kleinman and Keng‐Han Lin and Nadia K. Litterman and Jennifer C. McCreight and Matthew H. McIntyre and Kimberly F. McManus and Joanna L. Mountain and Elizabeth S. Noblin and Carrie A. M. Northover and Steven J. Pitts and G. David Poznik and J. Fah Sathirapongsasuti and Janie F. Shelton and Suyash Shringarpure and Chao Tian and Joyce Y. Tung and Vladimir Vacic and Xin Wang and Catherine H. Wilson and Jingchunzi Shi and Adam Auton and Kathleen A. Ryan and Joanne M. Jordan and Braxton D. Mitchell and Rebecca D. Jackson and Michelle S. Yau and Andrew M. McIntosh and Blair H. Smith},
  title = {{Genome-wide association study of knee pain identifies associations with GDF5 and COL27A1 in UK Biobank}},
  booktitle = {Communications},
  year = {2019},
  doi = {10.1038/s42003-019-0568-2},
}
Figure