Theses Doctoral

Utilizing functional genomics approaches to characterize risk genes in alopecia areata

Erjavec, Stephanie O'Toole

Understanding the genetic architecture of complex disorders is important for identifying disease mechanisms and potential molecular targets for therapeutic interventions. Genetic diseases are broadly classified as either Mendelian (monogenic) diseases or complex (polygenic) diseases. Common, polygenic disorders result from inheritance of multiple common variants with low penetrance. In contrast, monogenic, Mendelian disorders are caused by rare variants with high penetrance at a single genetic locus. However, an increasing number of studies support a role for rare variants of moderate effect size in complex diseases. As a result, genetic approaches previously utilized for discovering rare variants in Mendelian diseases, such as next generation sequencing, can now effectively be applied to complex diseases to define the contribution of both rare and common variants to the genetic burden of polygenic traits and diseases.

Alopecia Areata (AA) is a complex autoimmune disease characterized by non-scarring hair loss that is due to a combination of both enviornmental and genetic factors. Our previous Genome-wide Association Study (GWAS) identified at least 14 genetic regions contributing to AA disease susceptibility. Although useful in identifying disease-associated loci and surrounding linkage disequilibrium (LD) blocks, GWAS is not sufficiently granular to 1) elucidate causal (association-driving) variants; and 2) discover rare risk variants. This level of resolution can only be achieved by deep sequencing followed by functional validation of variants.

The goal of this thesis was to address these challenges in AA using two genetic approaches that have not been previously utilized in the context of this disease. In Chapter 2, I performed a hypothesis-driven analysis of common variants in a GWAS-associated locus using targeted genomic sequencing. In Chapter 3, I utilized whole exome sequencing (WES) in an unbiased approach to assess rare variant contribution in AA disease risk. To conduct these analyses simultaneously, we designed a whole exome sequencing (WES) chip that also included custom capture of 24 Mb of genomic regions covering the 14 genetic loci previously identified using GWAS. I applied these two sequencing approaches in a large AA patient cohort and identified potentially causal variants in several genes. To interrogate the consequences of these variants, I performed functional analyses to determine the effects of disease-causing variants on the target organ of AA attack, the hair follicle (HF).

In Chapter 2, I report on the use of targeted genomic sequencing to interrogate the coding and non-coding regions surrounding a previously implicated GWAS locus. This approach provided fine mapping of coding and non-coding common variants in regions that may be contributing to disease risk. In this thesis, I focused on the effect of genetic perturbations on the end-organ HF, and consequentially prioritized my functional analysis using two criteria: 1) genes expressed in the (HF) and; 2) GWAS regions that were not previously implicated in other autoimmune diseases. One of the regions that satisfied these criteria harbored the Syntaxin 17 (STX17) gene, which encodes a SNARE protein involved in autophagy and mitochondrial fission. Targeted genomic sequencing of the STX17 region in 849 AA cases identified 35 non-coding and 0 coding variants in high LD with the GWAS SNP. Thirty-three variants were significantly enriched in cases compared to controls, and the remaining two were nominally significant. Thirty-two of the significantly associated AA variants were confirmed to be AA skin eQTLs that downregulated expression of STX17 in affected scalp skin of AA patients. Downstream analyses incorporated in silico and functional cell assays that uncovered a novel autophagy-independent role for STX17 in melanocyte biology. I discovered that a reduction of STX17 expression was associated with an accumulation of a melanocyte-specific antigen and increased immunogenicity, as seen by CD8+ T cell infiltrates in the skin of AA patients with low levels of STX17 expression. I used a targeted sequencing approach to successfully identify candidate causal variants driving the GWAS association at the STX17 locus, and propose a novel mechanism underlying STX17-dependent melanocyte perturbation and AA disease.

In the second section of this thesis, we used the WES feature of the chip to assess the genetic contribution of rare variation in AA, in a genome-wide and unbiased manner. WES data and gene-level burden analyses of 18,653 genes in 849 AA patients was compared to 15,640 controls to identify rare variants associated with AA. Unexpectedly, this analysis identified one gene, encoding a hair-specific keratin, Keratin 82 (KRT82) that harbored significantly more rare damaging mutations in AA cases compared to controls (p=2.68E-06). Eleven rare damaging mutations were found in 51 AA patients in the heterozygous state (6.01%) compared to 2.58% controls. These variants resided in evolutionary conserved amino acid residues, and nine out of the eleven mutations were located in established disease-causing domains in keratin proteins. I determined that KRT82 expression was absent or largely reduced in AA hair follicles, including the bulb region, the site of AA immune attack. Moreover, AA patients with damaging variants and reduced KRT82 expression had increased perifollicular CD8+ T cell infiltrates in comparison to control HFs with intact KRT82 expression remaining. I proposed that damaging mutations in the coding regions of KRT82 resulted in loss of functional protein, thereby weakening the protective HF cuticle and predisposing the HF to immune attack.

In summary, I used two genetic approaches (targeted genomic sequencing and WES) to identify common (Chapter 2) and rare variants (Chapter 3) with novel contributions to the complex genetic architecture of AA. I focused my functional studies on genes expressed in the target HF, with the goal of defining the role of unidentified, variant-mediated end-organ disruption in the predisposition of AA patient HFs to aberrant autoimmune attack. Up to now, most efforts in AA mechanistic studies have focused on the aberrant immune response. The work in my thesis uncovered novel roles for perturbations in the HF itself as a participating factor in AA disease risk.


  • thumnail for Erjavec_columbia_0054D_16250.pdf Erjavec_columbia_0054D_16250.pdf application/pdf 6.43 MB Download File

More About This Work

Academic Units
Genetics and Development
Thesis Advisors
Christiano, Angela M.
Ph.D., Columbia University
Published Here
October 20, 2020