Novel insights into the pleiotropic health effects of growth differentiation factor 11 gained from genome-wide association studies in population biobanks

Abstract

Background: Growth differentiation factor 11 (GDF11) is a member of the transforming growth factor-β (TGF-β) superfamily that has gained considerable attention over the last decade for its observed ability to reverse age-related deterioration of multiple tissues, including the heart. Yet as many researchers have struggled to confirm the cardioprotective and anti-aging effects of GDF11, the topic has grown increasingly controversial, and the field has reached an impasse. We postulated that a clearer understanding of GDF11 could be gained by investigating its health effects at the population level.

Methods and results: We employed a comprehensive strategy to interrogate results from genome-wide association studies in population Biobanks. Interestingly, phenome-wide association studies (PheWAS) of GDF11 tissue-specific cis-eQTLs revealed associations with asthma, immune function, lung function, and thyroid phenotypes. Furthermore, PheWAS of GDF11 genetic variants confirmed these results, revealing similar associations with asthma, immune function, lung function, and thyroid health. To complement these findings, we mined results from transcriptome-wide association studies, which uncovered associations between predicted tissue-specific GDF11 expression and the same health effects identified from PheWAS analyses.

Conclusions: In this study, we report novel relationships between GDF11 and disease, namely asthma and hypothyroidism, in contrast to its formerly assumed role as a rejuvenating factor in basic aging and cardiovascular health. We propose that these associations are mediated through the involvement of GDF11 in inflammatory signaling pathways. Taken together, these findings provide new insights into the health effects of GDF11 at the population level and warrant future studies investigating the role of GDF11 in these specific health conditions.

Keywords: Asthma; Growth differentiation factor 11; PheWAS.

Fig. 1

Genetic variants significantly associated with GDF11 expression as detected in the GTEx project. A GDF11 cis-eQTLs (± 1 Mb of the transcription start site) separated by tissue. Specific tissues are denoted by colors. The most significant eQTL in each tissue is denoted by a black square. B Linkage disequilibrium (LD) blocks for GDF11 cis-eQTLs from the GTEx project. The LD heatmap reports pairwise LD values (R²) of the QTL variants. The corresponding eQTL normalized effect size (NES) bar chart heatmaps are located above the LD heatmap. Row labels (to the left of each chart) denote the tissue type and number of samples. The y-axis (to the right of each chart) is the -log₁₀(p-value). Tissue abbreviations: ARTAORT = Artery – Aorta; BRNCHA = Brain – Cerebellum; MSCLSK = Muscle – Skeletal; SKINNS = Skin – Not Sun Exposed (Suprapubic); WHLBLD = Whole Blood. TSS = transcription start site; TES = transcription end site

Fig. 2

Traits associated with variants and predicted tissue-specific expression of GDF11 and MSTN in Open Targets Genetics and the TWAS Hub. Traits were sorted by disease category to identify the most prevalent associations in each database. Disease categories are denoted by colors in the legend. Traits that did not fit a category were placed into the “other” category

Fig. 3

PheWAS results from Open Target Genetics reveal associations between the SNP rs1689510 and asthma, immune function, lung function, and hypothyroidism. The x-axis represents categories of disease traits, and the y-axis represents the -log₁₀ (p-value) of each association. The red significance line represents p = 1e-5. Associations that met the threshold p < 1e-14 are labeled in the figure. Results are sourced from FinnGen, UK Biobank, and GWAS Catalog

Fig. 4

Bulk tissue gene expression for GDF11 from the GTEx dataset. The x-axis represents separate tissues, and the y-axis represents expression values. Expression values are shown in TPM (transcripts per million), calculated from a gene model with isoforms collapsed to a single gene

Fig. 5

Overarching study design and results flow chart for GDF11, MSTN, HEY1, and FOXO1. Each step of the study design and corresponding databases are depicted to the left of the vertical black line. Overarching results for each gene identified from each step are depicted to the right of the vertical black line