Genetic modifiers of somatic expansion and clinical phenotypes in Huntington’s disease reveal shared and tissue-specific effects

An inherited, expanded CAG repeat in HTT undergoes further somatic expansion to cause Huntington’s disease (HD).  To gain insights into this molecular mechanism, we compared genome-wide association studies of somatic expansion in blood and somatic expansion-driven HD clinical phenotypes.  Here, we show that somatic expansion is driven by a mismatch repair-related process whose genetic modification and consequences show unexpected complexity, including cell-type specificity.  The HD clinical trajectory is further modified by non-DNA repair genes that differentially influence measures of cognitive and motor dysfunction.  In addition to shared (DNA repair genes MSH3, PMS2, and FAN1) and distinct trans-modifiers, a synonymous CAG-adjacent variant in HTT dramatically hastens motor onset without increasing somatic expansion, while a cis-acting 5’-UTR variant promotes blood repeat expansion without influencing clinical HD.  Our findings are directly relevant to the therapeutic suppression of expansion in DNA repeat disorders and provide additional clues to HD pathogenic mechanisms beyond somatic expansion.

Methods
The typed genotype data from HD subjects were subject to quality control (QC) such as call rate < 95% or minor allele frequency < 1% for subsequent pre-imputation QC  using the "HRC or 1000G Imputation preparation and checking" program (https://www.chg.ox.ac.uk/~wrayner/tools/; Version 4.3.0). Then, genotypes were imputed using the TOPMed Imputation Server (https://imputation.biodatacatalyst.nhlbi.nih.gov/#!) involving MINIMAC4 and EAGLE V2.4. The TOPMed r2 was used as the reference panel (i.e., all populations). Post-imputation, SNVs with 1) imputation r2 value < 0.5,  2) call rate < 100%, 3) Hardy-Weinberg equilibrium p value < 1E-6 except for the chromosome 4:1-5,000,000 region, or 4) minor allele frequency < 0.1% were removed. These quality control filters generated approximately 19,000,000 imputed SNVs for genome-wide association study (GWAS) for clinical (age at onset, DCL4, TFC6. TFC6 in Europeans), repeat instability (blood instability based on ABI, blood instability based on MiSeq), and quantile phenotypes (bradykinesia, chorea, dystonia, oculomotor, rigidity, SDMT, stroopword, TMS) in participants. GWAS was based on a mixed effects model using the GEMMA program. Summary results for each phenotype GWAS are available at the DRYAD.