Georgia Katsoula, Ana Luiza Arruda, Mauro Tutino, Ene Reimann, Norbert Bittner, Peter Kreitmaier, Karan M Shah, Diane Swift, Lorraine Southam, Siim Suutre, Galadriel Lucía Velázquez Silva, Kaspar Tootsi, Aare Märtson, Reedik Mägi, J Mark Wilkinson, Eleftheria Zeggini
5 min
Osteoarthritis is a complex, whole-joint disease with a significant genetic component, yet the specific effector genes through which risk variants operate remain largely unknown. This study aimed to bridge the gap between genome-wide association study (GWAS) signals and biological function by creating a high-resolution map of transcriptional regulation in primary joint tissues affected by osteoarthritis.
The researchers performed molecular profiling on 320 knee osteoarthritis patients, analyzing four primary tissues: macroscopically intact (low-grade) cartilage, degenerated (high-grade) cartilage, synovium, and fat pad. They generated cis- and trans-expression quantitative trait loci (eQTL) maps to identify how genetic variants influence gene expression. By integrating these maps with the largest available osteoarthritis GWAS data and functional genomic annotations—such as Hi-C chromatin interaction maps and transcription factor binding motifs—the team performed colocalization and Mendelian randomization analyses to prioritize causal effector genes.
The study identified over 10,000 unique eQTL-associated genes, with 61% not previously reported in osteoarthritis studies. By comparing low- and high-grade cartilage, the authors uncovered stage-specific genetic regulation, suggesting that regulatory mechanisms shift as the disease progresses. The integration of these data with GWAS signals allowed the researchers to prioritize 45 high-confidence effector genes. These genes are involved in critical pathways such as TGF-beta signaling, chondrocyte proliferation, and extracellular matrix remodeling. Notably, the study provides directional evidence that osteoarthritis risk alleles often amplify pro-chondrogenic and hypertrophic drivers while weakening regulatory feedback inhibitors.
The findings offer actionable insights for drug repurposing. By matching the genetically inferred direction of effect with existing pharmacological data, the authors identified potential therapeutic candidates. Specifically, they highlight inhibitors targeting LGALS3 (galectin-3) and SMAD7 as promising opportunities for osteoarthritis treatment, as these drugs align with the identified genetic regulatory mechanisms.
Osteoarthritis, a whole-joint degenerative disorder, is a major public health burden that affects nearly 600 million individuals worldwide, with no disease-modifying treatment. Molecular profiling of relevant tissues is crucial for understanding the biology of disease development. Here, we generate a comprehensive map of cis- and trans- transcriptional regulation in disease-relevant primary tissues from knee osteoarthritis patients: macroscopically intact (low-grade, n = 261) and degenerated (high-grade, n = 212) cartilage, synovium (n = 277), and fat pad (n = 92). We identify 10,166 unique expression quantitative trait loci (eQTL)-associated genes, 61.1% of which have not been reported in previous osteoarthritis eQTL studies, and uncover cartilage grade-specific genetic regulation. Using the largest osteoarthritis genome-wide association study to date, we find colocalization evidence with 136 genes and prioritize 45 high-confidence effector genes. At colocalizing loci, osteoarthritis risk alleles are associated with increased expression of genes involved in chondrogenic and hypertrophic signaling and decreased expression of genes encoding regulatory and modulatory components of these pathways. By integrating the eQTL maps with functional data, we delineate regulatory architectures for osteoarthritis risk variants, including promoter-enhancer loops and transcription factor binding effects. Finally, we provide directional evidence highlighting drugs targeting LGALS3 and SMAD7 as repurposing opportunities for osteoarthritis treatment.
Sam: [reflective] And they point to specific candidates for repurposing — SMAD7 and LGALS3 come up. How much weight can we put on those right now?
Alex: [measured, honest] These are still observational inferences. Mendelian randomization provides strong directional evidence — it's not just correlation — but functional validation in human models is the missing step before you'd want to commit serious drug development resources. The 45-gene list is a prioritization, not a confirmation.
Sam: [quiet conviction] Still, compressing from roughly a thousand risk variants down to 45 prioritized effectors is a meaningful reduction in the search space. It shifts the conversation from cataloguing associations to understanding the regulatory logic. [[RP_SECTION:limitations-and-future-directions|Limitations and future directions]]
Alex: [analytical] Right. And the limitations are worth naming clearly. The cohort is European ancestry, which is a genuine constraint — genetic architecture varies across populations, so these regulatory maps may not capture the full diversity of risk variants globally. These findings are a foundation, not a universal blueprint.
Sam: [nodding] The cross-sectional design is the other pressure point. These samples come from replacement surgeries, so you're looking at end-stage tissue. How confident can we be that the regulatory shifts are driving disease rather than just reflecting it?
Alex: [measured] That's the core interpretive challenge. You can't fully disentangle cause from consequence in that design. Some of what looks like dysregulation could be adaptive responses to damage that's already occurred. That's precisely why single-cell resolution is the logical next step — if you can show these effects are concentrated in specific chondrocyte states, like pro-inflammatory subpopulations, rather than diffuse across all cell types, the causal story becomes considerably more credible.
Sam: [sitting back] So the field is moving from static risk loci toward the dynamic regulatory landscape of the joint. More mechanistically grounded, and more actionable.
Alex: [warm, professional] That's the trajectory. The 45-gene list is most useful as a prioritization framework for functional follow-up — identifying which candidates survive validation in human models is where the clinical relevance gets established. What this study does is make that follow-up tractable. Without the eQTL integration across these four tissues, you're still sorting through a thousand candidates with no principled basis for ranking them.
Sam: [quiet conviction] A narrower, better-justified target list is exactly what translational work needs. Thanks for walking through the architecture of this one. That's it for this look at the regulatory genetics of osteoarthritis — thanks for listening to ResearchPod.