Whole genome sequencing identifies structural variants contributing to hematologic traits in the NHLBI TOPMed program

Abstract

Genome-wide association studies have identified thousands of single nucleotide variants and small indels that contribute to variation in hematologic traits. While structural variants are known to cause rare blood or hematopoietic disorders, the genome-wide contribution of structural variants to quantitative blood cell trait variation is unknown. Here we utilized whole genome sequencing data in ancestrally diverse participants of the NHLBI Trans Omics for Precision Medicine program (N = 50,675) to detect structural variants associated with hematologic traits. Using single variant tests, we assessed the association of common and rare structural variants with red cell-, white cell-, and platelet-related quantitative traits and observed 21 independent signals (12 common and 9 rare) reaching genome-wide significance. The majority of these associations (N = 18) replicated in independent datasets. In genome-editing experiments, we provide evidence that a deletion associated with lower monocyte counts leads to disruption of an S1PR3 monocyte enhancer and decreased S1PR3 expression.

Fig. 1. A structural variant at human 9q22.1 associated with decreased peripheral monocyte count.

All p-values are derived from two-sided t-tests and are not adjusted for multiple comparisons. A Genome-wide association -log10(p-values) for 9q22.1 variants associated with peripheral monocyte counts. The purple diamond represents the trait-associated deletion (9:88923551-88924152); large circles represent other SVs; and small circles represent single nucleotide variants (SNVs) or indels. Color indicates the linkage disequilibrium (LD) calculated in the analysis sample set between the trait-associated deletion and individual SVs and SNVs. B Distribution of accessible chromatin (by DNase I sequencing) and histone modifications (H3K27ac, H3K4me1 and H3K4me3) in primary CD14 + monocytes across indicated genomic regions from ENCODE. C Virtual 4C plot of long-range chromatin interactions anchored at the trait-associated, 9q22.1 deletion (9:88923551-88924152, upper panel) and the S1PR3 promoter region (9:91605763-91606263, lower panel), shown as a grey bar, in macrophages. Yellow line highlights the S1PR3 promoter region (upper panel) trait-associated, 9q22.1 deletion (lower panel). The observed and expected chromatin contact frequencies (or counts) are represented by the black and red lines, respectively. The left Y axis displays the range of chromatin contact frequency. The statistical significance (–log10(P-value)) of each long-range chromatin interaction is represented by the blue line, with its range listed in the right Y axis. The cell line or tissue specific FDR threshold (5%) is shown as a purple horizontal dashed line, and the more stringent Bonferroni threshold (P = 0.05) is shown as a maroon horizontal dashed line. D Long-range chromatin interaction between the trait-associated, 9q22.1 deletion and S1PR3 promoter calculated in 12 different cell types. MSC (mesendoderm), NPC (neural progenitor cell), HC (hippocampus), H1 (human embryonic stem cells), LV (left ventricle), PA (pancreas), SX (spleen), DLPFC (dorsolateral prefrontal cortex), LG (lung), GM (lymphoblast), Mac (macrophages), Mon (monocytes). The circle size represents the magnitude of the -log10 p-value while the color indicates S1PR3 mRNA level. TPM: transcripts per million.

Fig. 2. Genome and epigenome editing implicates S1PR3 in the 9q22.1 monocyte association.

All p-values are derived from two-sided t-tests (unless otherwise indicated) and are not adjusted for multiple comparisons. A eQTL results between the 9q22.1 SV and genes within 1-Mb window in monocytes using data from from MESA, including n = 77 AA and n = 92 Hispanic/Latino participants. AFHI: African American and Hispanic/Latino. B Violin plot (with minima, maxima, median, and inter-quartile range) demonstrating the correlation between the 9q22.1 SV genotype and expression of S1PR3, in n = 169 from MESA. C Expression of genes within a 2 Mb window in THP-1 cells expressing dCas9-KRAB after transduction with an sgRNAs targeting the SV (orange) as compared to a neutral locus control sgRNA (blue). Relative mRNA level of each gene was represented by mean ± standard deviation (SD). N = 3 biological replicates, where each replicate is a unique cellular transduction by sgRNA cassette. **P = 0.009. Location of sgRNAs designed for CRISPRi are indicated in Fig. S8C. D–F S1PR3 gene editing impaired monocyte differentiation in vitro. D Editing efficiency in HSPCs following 3xNLS-SpCas9:sgRNA electroporation with the indicated sgRNA. Gene edits were measured after 4 days of electroporation (N = 4 biological replicates). Location of S1PR3 coding sequence targeting sgRNAs are indicated above. E Representative flow cytometry indicating CD13 + CD14 + cell populations from the neutral locus and S1PR3 targeting group after 12-day differentiation. F CD13 + CD14 + percentage in the S1PR3 targeting group and the neutral locus targeting group. N = 4 replicates where each replicate is a unique Cas9:sgRNA electroporation experiment. Mean ± SD, with Student’s two-sided t-test.***P < 0.001, ****P < 0.0001. G–J Human CD34 + HSPCs from three healthy donors were edited by Cas9 RNP electroporation (EP) targeting a neutral locus and S1PR3 coding sequence infused into NBSGW mice 24 h after electroporation. After 12 weeks, engrafted bone marrow was characterized by immunophenotyping. G Indels determined by Sanger sequencing before transplantation. (H––J) Quantification of different human cell types between the neutral locus and S1PR3 targeting group. Human chimerism, hCD45 + ; Monocytes, hCD45 + CD33 + SSClowCD14 + ; neutrophil, hCD45 + CD33 + SSChighCD16 + . N = 3 independent biological replicates, each replicate indicates one mouse. Mean ± SD, 2-sided Mann-Whitney test. *P = 0.025, **P = 0.004.