Increased somatic mosaicism in autosomal and X chromosomes for suicide death

Abstract

Mosaic chromosomal alterations (mCAs) are classified as mosaic deletions (loss), copy-neutral loss of heterozygosity (CN-LOH), and duplications (gain), attracting special attention as biological aging-related acquired genetic alterations. While these mCAs have been linked with aging and various diseases, no study has investigated their association with suicide risk which is associated with abnormal biological aging. Here, we examined the association between suicide deaths and mCAs, including mosaic loss of the X (mLOX) and Y chromosomes, by leveraging blood-derived single nucleotide polymorphism-array data. The first (410 suicide decedents and 88,870 controls) and the second (363 suicide decedents and 88,870 controls) cohorts were analyzed and integrated using meta-analyses (773 suicide decedents and 177,740 controls). Total mCAs in autosomal chromosomes were significantly increased in suicide (p = 1.28 × 10^-6, odds ratio [OR] = 1.78), mostly driven by loss (p = 4.05 × 10^-9, OR = 2.70) and gain (p = 1.08 × 10^-3, OR = 2.23). mLOX were significantly increased in female suicide (p = 2.66 × 10^-21, OR = 4.00). The directions of effects of all mCAs in autosomal and sex chromosomes on suicide were the same in the first and second sets. Subgroup analyses suggest that our findings were mostly driven by suicide itself, and not confounded by comorbid psychiatric disorders or physical diseases, smoking status, sample location, or postmortem sample status. In conclusion, we provide the first evidence for aberrant mCAs in somatic autosomal and X chromosomes in suicide, which may contribute to an improved understanding of the genomic pathophysiology underlying suicide.

Fig. 1. Frequency of detectable mosaicism stratified by age.

The x- and y-axes represent the age category and percentage occurrence of each detectable mosaicism, respectively. Error bars indicate standard error. Blue lines denote the control group, and magenta lines denote the suicide decedent group. The p-values for the first and second sets were calculated using logistic regression analyses adjusted for sex (only for autosomal mosaicisms), age, age squared, and array. p-values for the meta-analysis were calculated using a fixed-effect model. CN-LOH copy-neutral loss of heterozygosity, mCA mosaic chromosomal alteration, mLOX mosaic loss of the X chromosome, mLOY mosaic loss of the Y chromosome.

Fig. 2. Forest plots of the odds ratios of suicide in detectable mosaicisms.

Averages of the odds ratios and 95% confidence intervals of suicide in (A) all autosomal mCAs, (B) loss, (C) CN-LOH, (D) gain, (E) mLOX, and (F) mLOY are shown. p-values were calculated by meta-analyses using a fixed-effect model. CN-LOH copy-neutral loss of heterozygosity, mCA mosaic chromosomal alteration, mLOX mosaic loss of the X chromosome, mLOY mosaic loss of the Y chromosome.

Fig. 3. Forest plots of the odds ratios of suicide in detectable mosaicisms after excluding suicides with comorbid psychiatric disorders.

Averages of the odds ratios and 95% confidence intervals of suicide in (A) all autosomal mCAs, (B) loss, (C) CN-LOH, (D) gain, (E) mLOX, and (F) mLOY are shown. p-values were calculated by meta-analyses using a fixed-effect model. CN-LOH copy-neutral loss of heterozygosity, mCA mosaic chromosomal alteration, mLOX mosaic loss of the X chromosome, mLOY mosaic loss of the Y chromosome.