Multiscale genetic architecture of donor-recipient differences reveals intronic LIMS1 mismatches associated with kidney transplant survival
Zeguo Sun, Zhongyang Zhang, Khadija Banu, Ian W. Gibson, Robert B. Colvin, Zhengzi Yi, Weijia Zhang, Bony De Kumar, Anand Reghuvaran, John Pell, Thomas D. Manes, Arjang Djamali, Lorenzo Gallon, Philip J. O’Connell, John Cijiang He, Jordan S. Pober, Peter S. Heeger, Madhav C. Menon
Zeguo Sun, Zhongyang Zhang, Khadija Banu, Ian W. Gibson, Robert B. Colvin, Zhengzi Yi, Weijia Zhang, Bony De Kumar, Anand Reghuvaran, John Pell, Thomas D. Manes, Arjang Djamali, Lorenzo Gallon, Philip J. O’Connell, John Cijiang He, Jordan S. Pober, Peter S. Heeger, Madhav C. Menon
View: Text | PDF
Research Article Genetics

Multiscale genetic architecture of donor-recipient differences reveals intronic LIMS1 mismatches associated with kidney transplant survival

Abstract

Donor-recipient (D-R) mismatches outside of human leukocyte antigens (HLAs) contribute to kidney allograft loss, but the mechanisms remain unclear, specifically for intronic mismatches. We quantified non-HLA mismatches at variant-, gene-, and genome-wide scales from single nucleotide polymorphism (SNP) data of D-Rs from 2 well-phenotyped transplant cohorts: Genomics of Chronic Allograft Rejection (GoCAR; n = 385) and Clinical Trials in Organ Transplantation-01/17 (CTOT-01/17; n = 146). Unbiased gene-level screening in GoCAR uncovered the LIMS1 locus as the top-ranked gene where D-R mismatches associated with death-censored graft loss (DCGL). A previously unreported, intronic, LIMS1 haplotype of 30 SNPs independently associated with DCGL in both cohorts. Haplotype mismatches showed a dosage effect, and minor-allele introduction to major-allele-carrying recipients showed greater hazard of DCGL. The LIMS1 haplotype and the previously reported LIMS1 SNP rs893403 are expression quantitative trait loci (eQTL) in immune cells for GCC2 (not LIMS1), which encodes a protein involved in mannose-6-phosphase receptor (M6PR) recycling. Peripheral blood and T cell transcriptome analyses associated the GCC2 gene and LIMS1 SNPs with the TGF-β1/SMAD pathway, suggesting a regulatory effect. In vitro GCC2 modulation impacted M6PR-dependent regulation of active TGF-β1 and downstream signaling in T cells. Together, our data link LIMS1 locus D-R mismatches to DCGL via GCC2 eQTLs that modulate TGF-β1–dependent effects on T cells.

Authors

Zeguo Sun, Zhongyang Zhang, Khadija Banu, Ian W. Gibson, Robert B. Colvin, Zhengzi Yi, Weijia Zhang, Bony De Kumar, Anand Reghuvaran, John Pell, Thomas D. Manes, Arjang Djamali, Lorenzo Gallon, Philip J. O’Connell, John Cijiang He, Jordan S. Pober, Peter S. Heeger, Madhav C. Menon

×

Figure 4

Variant-level mismatches at the LIMS1 locus were associated with death-censored graft loss.

Options: View larger image (or click on image) Download as PowerPoint
Variant-level mismatches at the LIMS1 locus were associated with death-c...
(A) Kaplan-Meier plot shows that the SNP rs893403 mismatch was associated with DCGL. P value was derived from log-rank test in comparison of mismatch versus non-mismatch group. (B) Forest plot for the association of the rs893403 mismatch with DCGL in all and European-to-European (E-to-E) D-R pairs, with univariate and multivariable models adjusted by genome-wide mismatch, HLA mismatch score, donor status, and induction therapy. (C) Venn diagram shows the number of top candidate SNPs (other than rs893403) associated with DCGL (nominal P ≤ 0.05) within the LIMS1 gene region in Cox regression analyses adjusted by rs893403 mismatch or within the rs893403 non-risk stratum for all and E-to-E D-R pairs. (D) LocusZoom plot shows the association P values (on –log10 scale) of variants with high frequency of any mismatch (≥5% D-R pairs) within the LIMS1 locus in adjusted analysis in all D-R pairs. The linkage disequilibrium metric R2 was calculated for SNPs surrounding 1 of the top candidate SNPs, rs200106875, within the LIMS1 gene region.