Whole-exome and whole-genome sequencing of 1064 individuals with type 1 diabetes reveals novel genes for diabetic kidney disease.

AIMS/HYPOTHESIS: Diabetic kidney disease (DKD) is a severe diabetic complication that affects one third of individuals with type 1 diabetes. Although several genes and common variants have been shown to be associated with DKD, much of the predicted inheritance remains unexplained. Here, we performed next-generation sequencing to assess whether low-frequency variants, extending to a minor allele frequency (MAF) ≤10% (single or aggregated) contribute to the missing heritability in DKD. METHODS: We performed whole-exome sequencing (WES) of 498 individuals and whole-genome sequencing (WGS) of 599 individuals with type 1 diabetes. After quality control, next-generation sequencing data were available for a total of 1064 individuals, of whom 541 had developed either severe albuminuria or end-stage kidney disease, and 523 had retained normal albumin excretion despite a long duration of type 1 diabetes. Single-variant and gene-aggregate tests for protein-altering variants (PAV) and protein-truncating variants (PTV) were performed separately for WES and WGS data and combined in a meta-analysis. We also performed genome-wide aggregate analyses on genomic windows (sliding window), promoters and enhancers using the WGS dataset. RESULTS: In the single-variant meta-analysis, no variant reached genome-wide significance, but a suggestively associated common THAP7 rs369250 variant (p=1.50 × 10-5, MAF=49%) was replicated in the FinnGen general population genome-wide association study (GWAS) data for chronic kidney disease and DKD phenotypes. The gene-aggregate meta-analysis provided suggestive evidence (p<4.0 × 10-4) at four genes for DKD, of which NAT16 (MAFPAV≤10%) and LTA (also known as TNFß, MAFPAV≤5%) are replicated in the FinnGen general population GWAS data. The LTA rs2229092 C allele was associated with significantly lower TNFR1, TNFR2 and TNFR3 serum levels in a subset of FinnDiane participants. Of the intergenic regions suggestively associated with DKD, the enhancer on chromosome 18q12.3 (p=3.94 × 10-5, MAFvariants≤5%) showed interaction with the METTL4 gene; the lead variant was replicated, and predicted to alter binding of the MafB transcription factor. CONCLUSIONS/INTERPRETATION: Our sequencing-based meta-analysis revealed multiple genes, variants and regulatory regions that were suggestively associated with DKD. However, as no variant or gene reached genome-wide significance, further studies are needed to validate the findings.

Protein glycation products associate with progression of kidney disease and incident cardiovascular events in individuals with type 1 diabetes.

BACKGROUND: Despite improved glycemic treatment, the impact of glycation on pathological consequences may persist and contribute to adverse clinical outcomes in diabetes. In the present study we investigated the association between serum protein glycation products and progression of kidney disease as well as incident major adverse cardiovascular events (MACE) in type 1 diabetes. METHODS: Fructosamine, advanced glycation end products (AGEs), and methylglyoxal-modified hydro-imidazolone (MG-H1) were measured from baseline serum samples in the FinnDiane study (n = 575). Kidney disease progression was defined as steep eGFR decline (> 3 mL/min/1.73 m2/year) or progression of albuminuria (from lower to higher stage of albuminuria). MACE was defined as acute myocardial infarction, coronary revascularization, cerebrovascular event (stroke), and cardiovascular death. RESULTS: Fructosamine was independently associated with steep eGFR decline (OR 2.15 [95% CI 1.16-4.01], p = 0.016) in the fully adjusted model (age, sex, baseline eGFR). AGEs were associated with steep eGFR decline (OR 1.58 per 1 unit of SD [95% CI 1.07-2.32], p = 0.02), progression to end-stage kidney disease (ESKD) (HR 2.09 per 1 unit of SD [95% CI 1.43-3.05], p < 0.001), and pooled progression (to any stage of albuminuria) (HR 2.72 per 1 unit of SD [95% CI 2.04-3.62], p < 0.001). AGEs (HR 1.57 per 1 unit of SD [95% CI 1.23-2.00], p < 0.001) and MG-H1 (HR 4.99 [95% CI 0.98-25.55], p = 0.054) were associated with incident MACE. MG-H1 was also associated with pooled progression (HR 4.19 [95% CI 1.11-15.89], p = 0.035). Most AGEs and MG-H1 associations were no more significant after adjusting for baseline eGFR. CONCLUSIONS: Overall, these findings suggest that protein glycation products are an important risk factor for target organ damage in type 1 diabetes. The data provide further support to investigate a potential causal role of serum protein glycation in the progression of diabetes complications.

Modeling binding specificities of transcription factor pairs with random forests.

BACKGROUND: Transcription factors (TFs) bind regulatory DNA regions with sequence specificity, form complexes and regulate gene expression. In cooperative TF-TF binding, two transcription factors bind onto a shared DNA binding site as a pair. Previous work has demonstrated pairwise TF-TF-DNA interactions with position weight matrices (PWMs), which may however not sufficiently take into account the complexity and flexibility of pairwise binding. RESULTS: We propose two random forest (RF) methods for joint TF-TF binding site prediction: ComBind and JointRF. We train models with previously published large-scale CAP-SELEX DNA libraries, which comprise DNA sequences enriched for binding of a selected TF pair. JointRF builds a random forest with sub-sequences selected from CAP-SELEX DNA reads with previously proposed pairwise PWM. JointRF outperforms (area under receiver operating characteristics curve, AUROC, 0.75) the current state-of-the-art method i.e. orientation and spacing specific pairwise PWMs (AUROC 0.59). Thus, JointRF may be utilized to improve prediction accuracy for pre-determined binding preferences. However, pairwise TF binding is currently considered flexible; a pair may bind DNA with different orientations and amounts of dinucleotide gaps or overlap between the two motifs. Thus, we developed ComBind, which utilizes random forests by considering simultaneously multiple orientations and spacings of the two factors. Our approach outperforms (AUROC 0.78) PWMs, as well as JointRF (p<0.00195). ComBind provides an approach for predicting TF-TF binding sites without prior knowledge on pairwise binding preferences. However, more research is needed to assess ComBind eligibility for practical applications. CONCLUSIONS: Random forest is well suited for modeling pairwise TF-TF-DNA binding specificities, and ComBind provides an improvement to pairwise binding site prediction accuracy.

Whole-genome sequencing identifies variants in ANK1, LRRN1, HAS1, and other genes and regulatory regions for stroke in type 1 diabetes.

Individuals with type 1 diabetes (T1D) carry a markedly increased risk of stroke, with distinct clinical and neuroimaging characteristics as compared to those without diabetes. Using whole-exome or whole-genome sequencing of 1,051 individuals with T1D, we aimed to find rare and low-frequency genomic variants associated with stroke in T1D. We analysed the genome comprehensively with single-variant analyses, gene aggregate analyses, and aggregate analyses on genomic windows, enhancers and promoters. In addition, we attempted replication in T1D using a genome-wide association study (N = 3,945) and direct genotyping (N = 3,263), and in the general population from the large-scale population-wide FinnGen project and UK Biobank summary statistics. We identified a rare missense variant on SREBF1 exome-wide significantly associated with stroke (rs114001633, p.Pro227Leu, p-value = 7.30 × 10-8), which replicated for hemorrhagic stroke in T1D. Using gene aggregate analysis, we identified exome-wide significant genes: ANK1 and LRRN1 displayed replication evidence in T1D, and LRRN1, HAS1 and UACA in the general population (UK Biobank). Furthermore, we performed sliding-window analyses and identified 14 genome-wide significant windows for stroke on 4q33-34.1, of which two replicated in T1D, and a suggestive genomic window on LINC01500, which replicated in T1D. Finally, we identified a suggestively stroke-associated TRPM2-AS promoter (p-value = 5.78 × 10-6) with borderline significant replication in T1D, which we validated with an in vitro cell-based assay. Due to the rarity of the identified genetic variants, future replication of the genomic regions represented here is required with sequencing of individuals with T1D. Nevertheless, we here report the first genome-wide analysis on stroke in individuals with diabetes.

Genetic Risk Score Enhances Coronary Artery Disease Risk Prediction in Individuals With Type 1 Diabetes.

OBJECTIVE: Individuals with type 1 diabetes are at a high lifetime risk of coronary artery disease (CAD), calling for early interventions. This study explores the use of a genetic risk score (GRS) for CAD risk prediction, compares it to established clinical markers, and investigates its performance according to the age and pharmacological treatment. RESEARCH DESIGN AND METHODS: This study in 3,295 individuals with type 1 diabetes from the Finnish Diabetic Nephropathy Study (467 incident CAD, 14.8 years follow-up) used three risk scores: a GRS, a validated clinical score, and their combined score. Hazard ratios (HR) were calculated with Cox regression, and model performances were compared with the Harrell C-index (C-index). RESULTS: A HR of 6.7 for CAD was observed between the highest and the lowest 5th percentile of the GRS (P = 1.8 × 10-6). The performance of GRS (C-index = 0.562) was similar to HbA1c (C-index = 0.563, P = 0.96 for difference), HDL (C-index = 0.571, P = 0.6), and total cholesterol (C-index = 0.594, P = 0.1). The GRS was not correlated with the clinical score (r = -0.013, P = 0.5). The combined score outperformed the clinical score (C-index = 0.813 vs. C-index = 0.820, P = 0.003). The GRS performed better in individuals below the median age (38.6 years) compared with those above (C-index = 0.637 vs. C-index = 0.546). CONCLUSIONS: A GRS identified individuals at high risk of CAD and worked better in younger individuals. GRS was also an independent risk factor for CAD, with a predictive power comparable to that of HbA1c and HDL and total cholesterol, and when incorporated into a clinical model, modestly improved the predictions. The GRS promises early risk stratification in clinical practice by enhancing the prediction of CAD.

Whole-exome sequencing identifies novel protein-altering variants associated with serum apolipoprotein and lipid concentrations.

BACKGROUND: Dyslipidemia is a major risk factor for cardiovascular disease, and diabetes impacts the lipid metabolism through multiple pathways. In addition to the standard lipid measurements, apolipoprotein concentrations provide added awareness of the burden of circulating lipoproteins. While common genetic variants modestly affect the serum lipid concentrations, rare genetic mutations can cause monogenic forms of hypercholesterolemia and other genetic disorders of lipid metabolism. We aimed to identify low-frequency protein-altering variants (PAVs) affecting lipoprotein and lipid traits. METHODS: We analyzed whole-exome (WES) and whole-genome sequencing (WGS) data of 481 and 474 individuals with type 1 diabetes, respectively. The phenotypic data consisted of 79 serum lipid and apolipoprotein phenotypes obtained with clinical laboratory measurements and nuclear magnetic resonance spectroscopy. RESULTS: The single-variant analysis identified an association between the LIPC p.Thr405Met (rs113298164) and serum apolipoprotein A1 concentrations (p=7.8×10-8). The burden of PAVs was significantly associated with lipid phenotypes in LIPC, RBM47, TRMT5, GTF3C5, MARCHF10, and RYR3 (p<2.9×10-6). The RBM47 gene is required for apolipoprotein B post-translational modifications, and in our data, the association between RBM47 and apolipoprotein C-III concentrations was due to a rare 21 base pair p.Ala496-Ala502 deletion; in replication, the burden of rare deleterious variants in RBM47 was associated with lower triglyceride concentrations in WES of >170,000 individuals from multiple ancestries (p=0.0013). Two PAVs in GTF3C5 were highly enriched in the Finnish population and associated with cardiovascular phenotypes in the general population. In the previously known APOB gene, we identified novel associations at two protein-truncating variants resulting in lower serum non-HDL cholesterol (p=4.8×10-4), apolipoprotein B (p=5.6×10-4), and LDL cholesterol (p=9.5×10-4) concentrations. CONCLUSIONS: We identified lipid and apolipoprotein-associated variants in the previously known LIPC and APOB genes, as well as PAVs in GTF3C5 associated with LDLC, and in RBM47 associated with apolipoprotein C-III concentrations, implicated as an independent CVD risk factor. Identification of rare loss-of-function variants has previously revealed genes that can be targeted to prevent CVD, such as the LDL cholesterol-lowering loss-of-function variants in the PCSK9 gene. Thus, this study suggests novel putative therapeutic targets for the prevention of CVD.

Genetic factors affect the susceptibility to bacterial infections in diabetes.

Diabetes increases the risk of bacterial infections. We investigated whether common genetic variants associate with infection susceptibility in Finnish diabetic individuals. We performed genome-wide association studies and pathway analysis for bacterial infection frequency in Finnish adult diabetic individuals (FinnDiane Study; N = 5092, Diabetes Registry Vaasa; N = 4247) using national register data on antibiotic prescription purchases. Replication analyses were performed in a Swedish diabetic population (ANDIS; N = 9602) and in a Finnish non-diabetic population (FinnGen; N = 159,166). Genome-wide data indicated moderate but significant narrow-sense heritability for infection susceptibility (h2 = 16%, P = 0.02). Variants on chromosome 2 were associated with reduced infection susceptibility (rs62192851, P = 2.23 × 10-7). Homozygotic carriers of the rs62192851 effect allele (N = 44) had a 37% lower median annual antibiotic purchase rate, compared to homozygotic carriers of the reference allele (N = 4231): 0.38 [IQR 0.22-0.90] and 0.60 [0.30-1.20] respectively, P = 0.01). Variants rs6727834 and rs10188087, in linkage disequilibrium with rs62192851, replicated in the FinnGen-cohort (P < 0.05), but no variants replicated in the ANDIS-cohort. Pathway analysis suggested the IRAK1 mediated NF-κB activation through IKK complex recruitment-pathway to be a mediator of the phenotype. Common genetic variants on chromosome 2 may associate with reduced risk of bacterial infections in Finnish individuals with diabetes.

Genome-wide association study on coronary artery disease in type 1 diabetes suggests beta-defensin 127 as a risk locus.

AIMS: Diabetes is a known risk factor for coronary artery disease (CAD). There is accumulating evidence that CAD pathogenesis differs for individuals with type 1 diabetes (T1D). However, the genetic background has not been extensively studied. We aimed to discover genetic loci increasing CAD susceptibility, especially in T1D, to examine the function of these discoveries and to study the role of the known risk loci in T1D. METHODS AND RESULTS: We performed the largest genome-wide association study to date for CAD in T1D, comprising 4869 individuals with T1D (cases/controls: 941/3928). Two loci reached genome-wide significance, rs1970112 in CDKN2B-AS1 [odds ratio (OR) = 1.32, P = 1.50 × 10-8], and rs6055069 on DEFB127 promoter (OR = 4.17, P = 2.35 × 10-9), with consistent results in survival analysis. The CDKN2B-AS1 variant replicated (P = 0.04) when adjusted for diabetic kidney disease in three additional T1D cohorts (cases/controls: 434/3123). Furthermore, we explored the function of the lead discoveries with a cardio-phenome-wide analysis. Among the eight suggestive loci (P < 1 × 10-6), rs70962766 near B3GNT2 associated with central blood pressure, rs1344228 near CNTNAP5 with intima media thickness, and rs2112481 on GRAMD2B promoter with serum leucocyte concentration. Finally, we calculated genetic risk scores for individuals with T1D with the known susceptibility loci. General population risk variants were modestly but significantly associated with CAD also in T1D (P = 4.21 × 10-7). CONCLUSION: While general population CAD risk loci had limited effect on the risk in T1D, for the first time, variants at the CDKN2B-AS1 locus were robustly associated with CAD in individuals with T1D. The novel finding on ß-defensin DEFB127 promoter provides a link between diabetes, infection susceptibility, and CAD, although pending on future confirmation.