Characterizing aging-related genetic and physiological determinants of spinal curvature

Abstract

Background: Increased spinal curvature is one of the most recognizable aging traits in the human population. However, despite high prevalence, the etiology of this condition remains poorly understood.

Methods: To gain better insight into the physiological, biochemical, and genetic risk factors involved, we developed a novel machine learning method to automatically derive thoracic kyphosis and lumbar lordosis angles from dual-energy X-ray absorptiometry (DXA) scans in the UK Biobank Imaging cohort. We carry out genome-wide association and epidemiological association studies to identify genetic and physiological risk factors for both traits.

Results: In 41,212 participants, we find that on average males and females gain 2.42° in kyphotic and 1.48° in lordotic angle per decade of life. Increased spinal curvature shows a strong association with decreased muscle mass and bone mineral density. Adiposity demonstrates opposing associations, with decreased kyphosis and increased lordosis. Using Mendelian randomization, we show that genes fundamental to the maintenance of musculoskeletal function (COL11A1, PTHLH, ETFA, TWIST1) and cellular homeostasis such as RNA transcription and DNA repair (RAD9A, MMS22L, HIF1A, RAB28) are likely involved in increased spinal curvature.

Conclusions: Our findings reveal a complex interplay between genetics, musculoskeletal health, and age-related changes in spinal curvature, suggesting potential drivers of this universal aging trait.

Fig. 1. Machine learning model evaluation.

a Example of manually annotated kyphotic (T5-T12) and lordotic (L1-L4) Cobb angles in lateral UK Biobank lateral DXA scan. b-c Correlations of Cobb angles as assessed by average annotator and machine learning derived kyphotic angles and lordotic angles in training and testing datasets.

Fig. 2. Distribution of kyphosis and lordosis.

a Kyphotic and lordotic angles as a function of age, stratified by sex. Spinal curvature increases with age in both sexes with approximately the same slope, but kyphotic and lordotic angles are greater in females. Confidence intervals represent SEs. b Prevalence of hyperkyphosis and hyperlordosis increases by age category. c Mean age increases jointly in higher kyphotic and lordotic angle categories.

Fig. 3. Spinal curvature associations with pre-existing conditions and biochemical measures.

Angle as a function of a diseases (a) or biomarker values (b). Beta is a degree change in kyphosis and lordosis angle with the presence of a prior condition or per standard deviation increase in biochemical measure. All analyses were adjusted by age at visit, sex, BMI, smoking status, and Townsend deprivation index.

Fig. 4. Physiology associations.

a Association of musculoskeletal traits with spine curvature. Analyses were adjusted by age at visit, sex, BMI, smoking status, and Townsend deprivation index, except analyses examining BMI, height, total fat mass, and visceral adipose tissue were not adjusted for BMI. Beta is a degree change in angle per standard deviation increase in physiological measure. b Importance of cross-validated gradient boosting algorithm (GBM)-selected variables for predicting spine curvature.

Fig. 5. Genetics of spine curvature.

a MTAG GWAS manhattan plot with fine-mapped lead SNPs annotated. b GWAS catalog phenotypic correlations with kyphosis and lordosis angle. In green are significant associations where pFDR < 0.05. c Comparison of genetic and phenotypic correlations. Displayed in the upper left triangle are genetic correlations for selected musculoskeletal traits. Displayed in the lower right triangle are phenotypic correlations of selected musculoskeletal traits. Both traits show similar patterns of genetic and phenotypic correlations, with greater kyphosis and lordosis angles anti-correlated with other traits. Significant correlations are marked with asterisks. d Genetic summary figure of GWAS, colocalization (eQTL and sQTL), and MR associations by lead SNP and gene.

Fig. 6. Multi-system physiology effect on spine curvature as a function of age.

Loss of bone mineral density, decreased muscle mass, and decreased lung capacity are associated with increased kyphosis and lordosis angles. Adiposity has opposing effects on the thoracic and lumbar spine: decreasing kyphosis and increasing lordosis. Aging is a common risk factor for declines in physiological function and presentation of comorbid conditions such as osteoporosis, type 2 diabetes, and esophageal diseases. Both traits are strongly genetically correlated and share common risk alleles such as PTHLH → BMD, COL11A1 → connective tissue, and ETFA → muscle mass, as well as a number of genes involved in DNA repair and cellular homeostasis processes (RAD9A, MMS22L, HIF1A, RAB28).