A variant in the PRDM16 locus has been correlated with QRS duration in an electrocardiogram genome-wide association study, and the deletion of PRDM16 has been implicated as a causal factor of the dilated cardiomyopathy that is linked to 1p36 deletion syndrome. We aimed to determine how a null mutation of Prdm16 affects cardiac function and study the underlying mechanism of the resulting phenotype in an appropriate mouse model. We used cardiac-specific Prdm16 conditional knockout mice to examine cardiac function by electrocardiography. QRS duration and QTc interval increased significantly in cardiac-specific Prdm16 knockout animals compared with wild-type mice. Further, we assessed cardiomyopathy-associated features by trichrome staining, densitometry, and hydroxyproline assay. Prdm16-null hearts showed greater fibrosis and cardiomyocyte hypertrophy. By quantitative real-time PCR, Prdm16-null hearts upregulated extracellular matrix-related genes ( Ctgf, Timp1) and α-smooth muscle actin ( Acta2), a myofibroblast marker. Moreover, TGF-β signaling was activated in Prdm16-null hearts, as evidenced by increased Tgfb1–3 transcript levels and phosphorylated Smad2. However, the inhibition of TGF-β receptor did not reverse the aberrations in conduction in cardiac-specific Prdm16 knockout mice. To determine the underlying mechanisms, we performed RNA-seq using mouse left ventricular tissue. By functional analysis, Prdm16-null hearts experienced dysregulated expression of ion channel genes, including Kcne1, Scn5a, Cacna1h, and Cacna2d2. Mice with Prdm16-null hearts develop abnormalities in cardiac conduction and cardiomyopathy-associated phenotypes, including fibrosis and cellular hypertrophy. Further, the RNA-seq findings suggest that impairments in ion homeostasis (Ca2+, K+, and Na+) may at least partially underlie the abnormal conduction in cardiac-specific Prdm16 knockout mice. NEW & NOTEWORTHY This is the first study that describes aberrant cardiac function and cardiomyopathy-associated phenotypes in an appropriate murine genetic model with cardiomyocyte-specific Prdm16-null mutation. It is noteworthy that the correlation of PRDM16 with QRS duration is replicated in a murine animal model and the potential underlying mechanism may be the impairment of ion homeostasis.

Hypertension or hypotension prevails as a comorbidity in patients with heart failure (HF). Although blood pressure (BP) is an important factor in managing the mortality of HF, the molecular mechanisms of changes in BP have not been clearly understood in cases of HF. We and others have demonstrated that a loss in PRDM16 causes hypertrophic cardiomyopathy, leading to HF. We aimed to determine whether BP is altered in mice that experience cardiac loss of Prdm16 and identify the underlying mechanism of BP-associated changes. BP decreased significantly only in female mice with a cardiac-null mutation of Prdm16 compared with controls, by an invasive protocol under anesthesia and by telemetric method during conscious, unrestrained status. Mice with a cardiac loss of Prdm16 had higher heart-to-body weight ratios and upregulated atrial natriuretic peptide, suggesting cardiac hypertrophy. Plasma aldosterone-to-renin activity ratios and plasma sodium levels decreased in Prdm16-deficient mice versus control. By RNA-seq and in subsequent functional analyses, Prdm16-null hearts were enriched in factors that regulate BP, including Adra1a, Nos1, Nppa, and Nppb. The inhibition of nitric oxide synthase 1 (NOS1) reverted the decrease in BP in cardiac-specific Prdm16 knockout mice. Mice with cardiac deficiency of Prdm16 present with hypotension and cardiac hypertrophy. Further, our findings suggest that the increased expression of NOS1 causes hypotension in mice with a cardiac-null mutation of Prdm16. These results provide novel insights into the molecular mechanisms of hypotension in subjects with HF and contribute to our understanding of how hypotension develops in patients with HF.

Multiple environmental factors could interact with a single genetic factor to affect disease phenotypes. We used Struct-LMM to identify genetic variants that interacted with environmental factors related to body mass index (BMI) using data from the Korea Association Resource. The following factors were investigated: alcohol consumption, education, physical activity metabolic equivalent of task (PAMET), income, total calorie intake, protein intake, carbohydrate intake, and smoking status. Initial analysis identified 7 potential single nucleotide polymorphisms (SNPs) that interacted with the environmental factors (P value < 5.00 × 10−6). Of the 8 environmental factors, PAMET score was excluded for further analysis since it had an average Bayes Factor (BF) value < 1 (BF = 0.88). Interaction analysis using 7 environmental factors identified 11 SNPs (P value < 5.00 × 10−6). Of these, rs2391331 had the most significant interaction (P value = 7.27 × 10−9) and was located within the intron of EFNB2 (Chr 13). In addition, the gene-based genome-wide association study verified EFNB2 gene significantly interacting with 7 environmental factors (P value = 5.03 × 10−10). BF analysis indicated that most environmental factors, except carbohydrate intake, contributed to the interaction of rs2391331 on BMI. Although the replication of the results in other cohorts is warranted, these findings proved the usefulness of Struct-LMM to identify the gene–environment interaction affecting disease.

No abstract

Gene-environment (G×E) interaction could partially explain missing heritability in traits; however, the magnitudes of G×E interaction effects remain unclear. Here, we estimate the heritability of G×E interaction for body mass index (BMI) by subjecting genome-wide interaction study data of 331,282 participants in the UK Biobank to linkage disequilibrium score regression (LDSC) and linkage disequilibrium adjusted kinships–software for estimating SNP heritability from summary statistics (LDAK-SumHer) analyses. Among 14 obesity-related lifestyle factors, MET score, pack years of smoking, and alcohol intake frequency significantly interact with genetic factors in both analyses, accounting for the partial variance of BMI. The G×E interaction heritability (%) and standard error of these factors by LDSC and LDAK-SumHer are as follows: MET score, 0.45% (0.12) and 0.65% (0.24); pack years of smoking, 0.52% (0.13) and 0.93% (0.26); and alcohol intake frequency, 0.32% (0.10) and 0.80% (0.17), respectively. Moreover, these three factors are partially validated for their interactions with genetic factors in other obesity-related traits, including waist circumference, hip circumference, waist-to-hip ratio adjusted with BMI, and body fat percentage. Our results suggest that G×E interaction may partly explain the missing heritability in BMI, and two G×E interaction loci identified could help in understanding the genetic architecture of obesity.