Heterogeneous associations between interleukin-6 receptor variants and phenotypes across ancestries and implications for therapy

Xuan Wang, Molei Liu, Isabelle-Emmanuella Nogues, Chien‐Jen Chen, Xin Xiong, Clara-Lea Bonzel, Harrison G. Zhang, Chuan Hong, Xia Yin, Kumar Dahal, Lauren Costa, Jing Cui, Sumitra Muralidhar, Jennifer Moser, Jennifer E. Deen, Philip S. Tsao, J. Michael Gaziano, Elizabeth R. Hauser, Amy M. Kilbourne, Shiuh‐Wen Luoh, Michael E. Matheny, Dave Oslin, Lori Churby, Stacey B. Whitbourne, Jessica V. Brewer, Shahpoor Shayan, Luis E. Selva, Saiju Pyarajan, Kelly Cho, Scott L. DuVall, Mary T. Brophy, Brady Stephens, Todd Connor, Themistocles L. Assimes, Adriana M. Hung, Henry R. Kranzler, Samuel M. Aguayo, Sunil K. Ahuja, Kathrina Alexander, X. Michelle Androulakis, Prakash Balasubramanian, Zuhair K. Ballas, Jean C. Beckham, Sujata Bhushan, Edward J. Boyko, David Cohen, Louis J. Dell’Italia, L. Christine Faulk, Joseph Fayad, Daryl Fujii, Saib Gappy, F. A. Gesek, Jennifer Greco, Michael Godschalk, Todd Gress, Samir Gupta, Salvador Gutierrez, John B. Harley, Kimberly Hammer, Mark B. Hamner, Robin A. Hurley, Pran Iruvanti, Frank J. Jacono, Darshana Jhala, Scott Kinlay, Jon Klein, Michael Landry, Peter S. Liang, Suthat Liangpunsakul, Jack Lichy, Charles S. Mahan, Ronnie Marrache, Stephen Mastorides, Elisabeth Mates, Kristin Mattocks, Paul Meyer, Jonathan P. Moorman, Timothy R. Morgan, Maureen Murdoch, James E. Norton, Olaoluwa Okusaga, Kris Ann Oursler, Ana Palacio, Samuel Poon, Emily Potter, Michael Rauchman, Richard J. Servatius, Satish Sharma, River Smith, Peruvemba Sriram, Patrick Strollo, Neeraj Tandon, Gerardo VillarealScientific 2024 · DOI

Abstract

The Phenome-Wide Association Study (PheWAS) is increasingly used to broadly screen for potential treatment effects, e.g., IL6R variant as a proxy for IL6R antagonists. This approach offers an opportunity to address the limited power in clinical trials to study differential treatment effects across patient subgroups. However, limited methods exist to efficiently test for differences across subgroups in the thousands of multiple comparisons generated as part of a PheWAS. In this study, we developed an approach that maximizes the power to test for heterogeneous genotype-phenotype associations and applied this approach to an IL6R PheWAS among individuals of African (AFR) and European (EUR) ancestries. We identified 29 traits with differences in IL6R variant-phenotype associations, including a lower risk of type 2 diabetes in AFR (OR 0.96) vs EUR (OR 1.0, p-value for heterogeneity = 8.5 × 10<sup>-3</sup>), and higher white blood cell count (p-value for heterogeneity = 8.5 × 10<sup>-131</sup>). These data suggest a more salutary effect of IL6R blockade for T2D among individuals of AFR vs EUR ancestry and provide data to inform ongoing clinical trials targeting IL6 for an expanding number of conditions. Moreover, the method to test for heterogeneity of associations can be applied broadly to other large-scale genotype-phenotype screens in diverse populations.

Introduction

Clinical trials remain the gold standard for medical evidence, yet they often suffer from high costs, limited duration, and a lack of ancestral diversity. This lack of diversity makes it difficult to identify heterogeneous treatment effects—instances where a drug's efficacy or safety profile varies significantly across different patient populations. To address this, researchers utilize genetic proxies, such as variants in the interleukin-6 receptor (IL6R) gene, to simulate the effects of pharmacological interventions (like IL6R antagonists) across broad populations in Phenome-Wide Association Studies (PheWAS).

However, detecting differences in these associations across ancestries is statistically difficult. Standard methods often lack the power to identify signals in smaller subgroups within large, predominantly European-centric datasets. This study introduces a novel statistical framework to overcome these hurdles and applies it to investigate the ancestry-specific effects of IL6R signaling.

Methodology

The hetFDR Framework

The researchers developed hetFDR (False Discovery Rate controlled heterogeneity testing), a three-step statistical procedure designed to maximize power in diverse cohorts:

  1. Statistic Construction: For each trait, the model calculates an overall mean effect and a heterogeneity statistic representing the difference between ancestries.
  2. Adaptive Weighting: The algorithm uses the mean effect as auxiliary information to weight the p-values of the heterogeneity test, assuming that traits with strong average effects are more likely to exhibit inter-group differences.
  3. FDR Control: A final filtering step ensures that the false discovery rate remains below 10%.

Data Sources

The primary analysis was conducted using the VA Million Veteran Program (MVP), encompassing 439,309 individuals of European (EUR) ancestry and 105,838 individuals of African (AFR) ancestry. Findings were further validated using the UK Biobank and the Mass General Brigham Biobank.

Results: Ancestry-Specific Effects of IL6R Blockade

The study identified 29 traits with significant heterogeneity between AFR and EUR populations. The most notable findings involved metabolic and hematologic traits.

Metabolic Benefits in African Ancestry

A key discovery was the "two-fer" effect for Type 2 Diabetes (T2D). In the AFR group, the IL6R variant was associated with a significantly lower risk of T2D (OR 0.96) and lower HbA1c levels. In contrast, no such protective effect was observed in the EUR group (OR 1.0). This suggests that IL6R-targeting drugs may offer unique metabolic benefits for patients of African ancestry.

Comparison of Odds Ratios for clinical phenotypes between African (AFR) and European (EUR) ancestries.

Hematologic and Infection Risk Paradox

The study revealed a massive difference in white blood cell (WBC) associations (p-heterogeneity = $8.5 \times 10^{-131}$). While the variant was associated with higher WBC counts and lower odds of neutropenia in the AFR group, it also correlated with a modest increase in the risk of serious infections (OR 1.03). This highlights a complex biological trade-off: while the variant appears to bolster certain immune markers in AFR individuals, it may simultaneously increase vulnerability to specific infections.

Comparison of effect sizes for laboratory traits between African (AFR) and European (EUR) ancestries.

Discussion and Future Directions

These findings provide critical data for ongoing clinical trials targeting IL6. The evidence suggests that IL6R blockade may be particularly effective for treating or preventing T2D in African ancestry populations. However, the increased infection risk noted in this group warrants careful monitoring.

While the MVP cohort is massive, the imbalance between EUR and AFR sample sizes remains a limitation for detecting subtle effects. Future research must expand to include Hispanic, Asian, and other global populations to fully map drug-gene interactions. Furthermore, while genetic proxies provide lifelong insights, clinical trials are necessary to confirm how these ancestral differences translate to acute pharmacological interventions.

Summary updated

Abstract

The Phenome-Wide Association Study (PheWAS) is increasingly used to broadly screen for potential treatment effects, e.g., IL6R variant as a proxy for IL6R antagonists. This approach offers an opportunity to address the limited power in clinical trials to study differential treatment effects across patient subgroups. However, limited methods exist to efficiently test for differences across subgroups in the thousands of multiple comparisons generated as part of a PheWAS. In this study, we developed an approach that maximizes the power to test for heterogeneous genotype-phenotype associations and applied this approach to an IL6R PheWAS among individuals of African (AFR) and European (EUR) ancestries. We identified 29 traits with differences in IL6R variant-phenotype associations, including a lower risk of type 2 diabetes in AFR (OR 0.96) vs EUR (OR 1.0, p-value for heterogeneity = 8.5 × 10<sup>-3</sup>), and higher white blood cell count (p-value for heterogeneity = 8.5 × 10<sup>-131</sup>). These data suggest a more salutary effect of IL6R blockade for T2D among individuals of AFR vs EUR ancestry and provide data to inform ongoing clinical trials targeting IL6 for an expanding number of conditions. Moreover, the method to test for heterogeneity of associations can be applied broadly to other large-scale genotype-phenotype screens in diverse populations.
BibTeX
@inproceedings{wang2024heterogeneous,
  author = {Xuan Wang and Molei Liu and Isabelle-Emmanuella Nogues and Chien‐Jen Chen and Xin Xiong and Clara-Lea Bonzel and Harrison G. Zhang and Chuan Hong and Xia Yin and Kumar Dahal and Lauren Costa and Jing Cui and Sumitra Muralidhar and Jennifer Moser and Jennifer E. Deen and Philip S. Tsao and Sumitra Muralidhar and J. Michael Gaziano and Elizabeth R. Hauser and Amy M. Kilbourne and Shiuh‐Wen Luoh and Michael E. Matheny and Dave Oslin and J. Michael Gaziano and Philip S. Tsao and Lori Churby and Stacey B. Whitbourne and Jessica V. Brewer and Shahpoor Shayan and Luis E. Selva and Saiju Pyarajan and Kelly Cho and Scott L. DuVall and Mary T. Brophy and J. Michael Gaziano and Philip S. Tsao and Brady Stephens and Todd Connor and Themistocles L. Assimes and Adriana M. Hung and Henry R. Kranzler and Samuel M. Aguayo and Sunil K. Ahuja and Kathrina Alexander and X. Michelle Androulakis and Prakash Balasubramanian and Zuhair K. Ballas and Jean C. Beckham and Sujata Bhushan and Edward J. Boyko and David Cohen and Louis J. Dell’Italia and L. Christine Faulk and Joseph Fayad and Daryl Fujii and Saib Gappy and F. A. Gesek and Jennifer Greco and Michael Godschalk and Todd Gress and Samir Gupta and Salvador Gutierrez and John B. Harley and Kimberly Hammer and Mark B. Hamner and Adriana M. Hung and Robin A. Hurley and Pran Iruvanti and Frank J. Jacono and Darshana Jhala and Scott Kinlay and Jon Klein and Michael Landry and Peter S. Liang and Suthat Liangpunsakul and Jack Lichy and Charles S. Mahan and Ronnie Marrache and Stephen Mastorides and Elisabeth Mates and Kristin Mattocks and Paul Meyer and Jonathan P. Moorman and Timothy R. Morgan and Maureen Murdoch and James E. Norton and Olaoluwa Okusaga and Kris Ann Oursler and Ana Palacio and Samuel Poon and Emily Potter and Michael Rauchman and Richard J. Servatius and Satish Sharma and River Smith and Peruvemba Sriram and Patrick Strollo and Neeraj Tandon and Philip S. Tsao and Gerardo Villareal},
  title = {{Heterogeneous associations between interleukin-6 receptor variants and phenotypes across ancestries and implications for therapy}},
  booktitle = {Scientific},
  year = {2024},
  doi = {10.1038/s41598-024-54063-3},
}
Figure