Investigating causal relationships between major depression and chronic pain using UK general-population datasets with whole-genome genotyping

Johnston, Keira Jacqueline Ann (2021) Investigating causal relationships between major depression and chronic pain using UK general-population datasets with whole-genome genotyping. PhD thesis, University of Glasgow.

Abstract

Chronic pain, considered here to be pain lasting 3 months or longer, imparts significant socioeconomic and public health burden around the globe. Chronic pain is associated with a wide range of conditions, illnesses, or injuries, and is categorised and investigated in many ways. Treatment and management of chronic pain is complicated by this heterogeneity, and by lack of full understanding of factors (including genetic) that influence vulnerability to developing chronic pain and biological mechanisms of chronic pain development. Major depression is commonly comorbid with chronic pain, and results of studies into potential causal direction between the two conditions are mixed. Due to symptom overlap and common comorbidity, it may be that cases of chronic pain are misclassified as major depression and vice versa. Understanding genetic factors that contribute to chronic pain vulnerability and development has the potential to improve treatment of both conditions, in addition to allowing for investigation of potential causal relationships and clinical heterogeneity.

Recently, the International Association for the Study of Pain released an updated definition of chronic pain and advocated for the study of chronic pain as a disease entity. Studying the genetics of chronic pain through genome wide association study of broad chronic pain traits, in line with this updated pain definition, may present a more tractable way to uncover common genetic variation associated with vulnerability to and mechanisms of development of chronic pain. This mode of study can also provide genome wide association study summary statistics for use in analyses that aim to investigate causality, genetic correlation and pleiotropy, and clinical heterogeneity in chronic pain and major depression.

The overall aim of this PhD project is therefore to explore causal relationships between chronic pain and MDD in large UK general-population cohorts with whole-genome genotyping data using a wide range of statistical genetic methods.

Data were obtained from two large UK cohorts with whole-genome genotyping. One, UK Biobank, is a cohort of 0.5 million participants recruited in middle age (40-79) with information on an extensive list of physical, behavioural and health related traits. Generation Scotland is a smaller (N ~ 22,000) Scottish cohort of participants recruited mainly through general practitioners in a family-based manner, again with information of physical, health, and behavioural traits. Summary statistic data were also obtained from a 23andMe-Pfizer genome wide association study of chronic pain grade.

As part of this PhD the largest genome wide association study of any chronic pain trait to date was carried out in UK Biobank. Validation of the trait (multisite chronic pain) was carried out through polygenic risk score analysis in Generation Scotland, examining the relationship between this novel chronic pain trait and chronic pain grade. Genetic correlation analyses were used to explore the genetic overlap of multisite chronic pain and a range of traits of interest, including other chronic pain phenotypes such as chronic widespread pain and chronic pain grade, in addition to major depression. Gene-level analyses were carried out to investigate genes of interest associated with chronic pain and potentially relevant to mechanisms of chronic pain development. BUHMBOX analyses were performed to test for clinical heterogeneity in chronic pain with respect to major depression and vice versa in UK Biobank. Conditional false discovery rate analyses using 23andMe-Pfizer data were also used to explore pleiotropy in chronic pain grade and major depression and to highlight pleiotropic loci of interest. Mendelian randomisation analyses, including recent mendelian randomisation methods explicitly designed to account for extensive horizontal pleiotropy, were carried out to assess potential causal relationships between major depression and chronic pain grade, and between major depression and multisite chronic pain.

Results indicated multisite chronic pain was a polygenic, moderately heritable trait. Associated genes of interest implicated a strong central nervous system component, in addition to immune related genes. Conditional false discovery rate analysis highlighted loci of interest mapped to LRFN5, a gene involved in neuroinflammation, and that were associated with regulation of gene expression at this locus. Polygenic risk scoring analysis showed multisite chronic pain to be significantly associated with both chronic pain grade and chronic widespread pain, in addition to a multisite chronic pain-like trait in Generation Scotland, validating multisite chronic pain as a trait and indicating strong genetic overlap between widespread and non-widespread pain. Genetic correlation analysis showed significant genetic overlap between multisite chronic pain and mental health traits, markedly major depressive disorder, and depressive symptoms, but a lower degree of genetic correlation with conditions associated with significant chronic pain such as rheumatoid arthritis, and no significant genetic correlation with inflammatory bowel diseases. BUHMBOX analyses showed no evidence of clinical heterogeneity in chronic pain with respect to major depression in UK Biobank or vice versa. Mendelian randomisation analyses showed no causal relationship between chronic pain grade and major depressive disorder, but a significant causal effect of multisite chronic pain on major depressive disorder.

In conclusion, I have shown that broad chronic pain traits such as multisite chronic pain present a powerful and tractable way to study mechanisms of, and factors contributing to vulnerability to, chronic pain development. Output from well-powered genome wide association studies can also be used to validate phenotypes, explore genetic overlap with traits of interest, and conduct causal analyses.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Subjects: Q Science > QH Natural history > QH426 Genetics; Q Science > QP Physiology; R Medicine > R Medicine (General)
• Colleges/Schools: College of Medical Veterinary and Life Sciences
• Funder's Name: Medical Research Council (MRC)
• Supervisor's Name: Nicholl, Dr. Barbara I., Smith, Professor Daniel, McIntosh, Professor Andrew M. and Bailey, Dr. Mark E.S.
• Date of Award: 2021
• Depositing User: Theses Team
• Unique ID: glathesis:2021-82546
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 02 Nov 2021 16:03
• Last Modified: 02 Nov 2021 16:10
• Thesis DOI: 10.5525/gla.thesis.82546
• URI: https://theses.gla.ac.uk/id/eprint/82546

Related URLs

• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI
• Article DOI