Epigenome-wide association study of DNA methylation in maternal blood leukocytes with BMI in pregnancy and gestational weight gain

OBJECTIVES

We aimed to discover CpG sites with differential DNA methylation in peripheral blood leukocytes associated with body mass index (BMI) in pregnancy and gestational weight gain (GWG) in women of European and South Asian ancestry. Furthermore, we aimed to investigate how the identified sites were associated with methylation quantitative trait loci, gene ontology, and cardiometabolic parameters.

METHODS

In the Epigenetics in pregnancy (EPIPREG) sample we quantified maternal DNA methylation in peripheral blood leukocytes in gestational week 28 with Illumina's MethylationEPIC BeadChip. In women with European (n = 303) and South Asian (n = 164) ancestry, we performed an epigenome-wide association study of BMI in gestational week 28 and GWG between gestational weeks 15 and 28 using a meta-analysis approach. Replication was performed in the Norwegian Mother, Father, and Child Cohort Study, the Study of Assisted Reproductive Technologies (MoBa-START) (n = 877, mainly European/Norwegian).

RESULTS

We identified one CpG site significantly associated with GWG (p 5.8 × 10-8) and five CpG sites associated with BMI at gestational week 28 (p from 4.0 × 10-8 to 2.1 × 10-10). Of these, we were able to replicate three in MoBa-START; cg02786370, cg19758958 and cg10472537. Two sites are located in genes previously associated with blood pressure and BMI. DNA methylation at the three replicated CpG sites were associated with levels of blood pressure, lipids and glucose in EPIPREG (p from 1.2 × 10-8 to 0.04).

CONCLUSIONS

We identified five CpG sites associated with BMI at gestational week 28, and one with GWG. Three of the sites were replicated in an independent cohort. Several genetic variants were associated with DNA methylation at cg02786379 and cg16733643 suggesting a genetic component influencing differential methylation. The identified CpG sites were associated with cardiometabolic traits.

CLINICALTRIALS

GOV REGISTRATION NO

Not applicable.

SCRT1 is a novel beta cell transcription factor with insulin regulatory properties

Here we show that scratch family transcriptional repressor 1 (SCRT1), a zinc finger transcriptional regulator, is a novel regulator of beta cell function. SCRT1 was found to be expressed in beta cells in rodent and human islets. In human islets, expression of SCRT1 correlated with insulin secretion capacity and the expression of the insulin (INS) gene. Furthermore, SCRT1 mRNA expression was lower in beta cells from T2D patients. siRNA-mediated Scrt1 silencing in INS-1832/13 cells, mouse- and human islets resulted in impaired glucose-stimulated insulin secretion and decreased expression of the insulin gene. This is most likely due to binding of SCRT1 to E-boxes of the Ins1 gene as shown with ChIP. Scrt1 silencing also reduced the expression of several key beta cell transcription factors. Moreover, Scrt1 mRNA expression was reduced by glucose and SCRT1 protein was found to translocate between the nucleus and the cytosol in a glucose-dependent fashion in INS-1832/13 cells as well as in a rodent model of T2D. SCRT1 was also regulated by a GSK3β-dependent SCRT1-serine phosphorylation. Taken together, SCRT1 is a novel beta cell transcription factor that regulates insulin secretion and is affected in T2D.

Ghrelin suppresses insulin secretion in human islets and type 2 diabetes patients have diminished islet ghrelin cell number and lower plasma ghrelin levels

It is not known how ghrelin affects insulin secretion in human islets from patients with type 2 diabetes (T2D) or whether islet ghrelin expression or circulating ghrelin levels are altered in T2D. Here we sought out to identify the effect of ghrelin on insulin secretion in human islets and the impact of T2D on circulating ghrelin levels and on islet ghrelin cells. The effect of ghrelin on insulin secretion was assessed in human T2D and non-T2D islets. Ghrelin expression was assessed with RNA-sequencing (n = 191) and immunohistochemistry (n = 21). Plasma ghrelin was measured with ELISA in 40 T2D and 40 non-T2D subjects. Ghrelin exerted a glucose-dependent insulin-suppressing effect in islets from both T2D and non-T2D donors. Compared with non-T2D donors, T2D donors had reduced ghrelin mRNA expression and 75% less islet ghrelin cells, and ghrelin mRNA expression correlated negatively with HbA1c. T2D subjects had 25% lower fasting plasma ghrelin levels than matched controls. Thus, ghrelin has direct insulin-suppressing effects in human islets and T2D patients have lower fasting ghrelin levels, likely as a result of reduced number of islet ghrelin cells. These findings support inhibition of ghrelin signaling as a potential therapeutic avenue for stimulation of insulin secretion in T2D patients.

Phenotypic and genotypic differences between Indian and Scandinavian women with gestational diabetes mellitus

OBJECTIVE

Gestational diabetes mellitus (GDM) is a transient form of diabetes characterized by impaired insulin secretion and action during pregnancy. Population-based differences in prevalence exist which could be explained by phenotypic and genetic differences. The aim of this study was to examine these differences in pregnant women from Punjab, India and Scandinavia.

METHODS

Eighty-five GDM/T2D loci in European and/or Indian populations from previous studies were assessed for association with GDM based on Swedish GDM criteria in 4018 Punjabi Indian and 507 Swedish pregnant women. Selected loci were replicated in Scandinavian cohorts, Radiel (N = 398, Finnish) and STORK/STORK-G (N = 780, Norwegian).

RESULTS

Punjabi Indian women had higher GDM prevalence, lower insulin secretion and better insulin sensitivity than Swedish women. There were significant frequency differences of GDM/T2D risk alleles between both populations. rs7178572 at HMG20A, previously associated with GDM in South Indian and European women, was replicated in North Indian women. The T2D risk SNP rs11605924 in the CRY2 gene was associated with increased GDM risk in Scandinavian but decreased GDM risk in Punjabi Indian women. No other overlap was seen between GDM loci in both populations.

CONCLUSIONS

Gestational diabetes mellitus is more common in Indian than Swedish women, which partially can be attributed to differences in insulin secretion and action. There was marked heterogeneity in the GDM phenotypes between the populations which could only partially be explained by genetic differences.

Precision medicine in type 2 diabetes

The Precision Medicine Initiative defines precision medicine as 'an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment and lifestyle for each person'. This approach will facilitate more accurate treatment and prevention strategies in contrast to a one-size-fits-all approach, in which disease treatment and prevention strategies are developed for generalized usage. Diabetes is clearly more heterogeneous than the conventional subclassification into type 1 and type 2 diabetes. Monogenic forms of diabetes like MODY and neonatal diabetes have paved the way for precision medicine in diabetes, as carriers of unique mutations require unique treatment. Diagnosis of diabetes in the past has been dependent upon measuring one metabolite, glucose. By instead including six variables in a clustering analysis, we could break down diabetes into five distinct subgroups, with better prediction of disease progression and outcome. The severe insulin-resistant diabetes (SIRD) cluster showed the highest risk of kidney disease and highest prevalence of nonalcoholic fatty liver disease, whereas patients in the insulin-deficient cluster 2 (SIDD) had the highest risk of retinopathy. In the future, this will certainly be improved and expanded by including genetic, epigenetic and other biomarker to allow better prediction of outcome and choice of more precise treatment.

Insulin secretion and action in North Indian women during pregnancy

AIM

The relative roles(s) of impaired insulin secretion vs. insulin resistance in the development of gestational diabetes mellitus depend upon multiple risk factors and diagnostic criteria. Here, we explored their relative contribution to gestational diabetes as defined by the WHO 1999 (GDM1999) and adapted WHO 2013 (GDM2013) criteria, excluding the 1-h glucose value, in a high-risk Indian population from Punjab.

METHODS

Insulin secretion (HOMA2-B) and insulin action (HOMA2-IR) were assessed in 4665 Indian women with or without gestational diabetes defined by the GDM1999 or adapted GDM2013 criteria.

RESULTS

Gestational diabetes defined using both criteria was associated with decreased insulin secretion compared with pregnant women with normal glucose tolerance. Women with gestational diabetes defined by the adapted GDM2013, but not GDM1999 criteria, were more insulin resistant than pregnant women with normal glucose tolerance, and furthermore displayed lower insulin secretion than GDM1999 women. Urban habitat, illiteracy, high age and low BMI were independently associated with reduced insulin secretion, whereas Sikh religion, increasing age and BMI, as well as a family history of diabetes were independently associated with increased insulin resistance.

CONCLUSIONS

Gestational diabetes risk factors influence insulin secretion and action in North Indian women in a differential manner. Gestational diabetes classified using the adapted GDM2013 compared with GDM1999 criteria is associated with more severe impairments of insulin secretion and action.

Synthesis and estimation of calorific value of a structured lipid-potential reduced calorie fat

The majority of reduced calorie fats and fat substitutes available today, though similar in texture and flavor to natural fats, contain fatty acids that are not usually present in edible oils and fats and thus do not fully match the chemistry and functions of natural fats. For example, such products do not provide nutritionally important essential fatty acids (EFA). In this investigation, we prepared and evaluated a reduced calorie fat, prepared entirely from natural fats, taking advantage of the fact that long-chain saturated fatty acids (LCSFA), such as behenic acid (22:0), are poorly absorbed. Mustard oil (MO) and sunflower oil (SO) were used as substrates to yield a structured lipid (SL). The product, being derived from a natural vegetable oil, would thus provide EFA, as would a native fat, a feature not provided by the low-calorie fats available in the market. Erucic acid (22:1) was isolated from MO by a lipase (EC 3.1.1.3)-catalyzed reaction. It was then hydrogenated to behenic acid, the ethyl ester of which was subsequently enzymatically transesterified with SO to yield a plastic fat containing about 30-35% behenic acid. Absorption of this fat was studied in Wistar rats. In a preliminary single oral dose experiment, rats were fed equal amounts (2 mL) of SO and the SL. Plasma triacylglycerol (TAG) levels were estimated after 1, 2, and 3 h of feeding. The significantly lower concentration of plasma TAG in the 2-h sample, observed in the SL-fed group compared to the SO-fed group (P<0.001), indicated poor absorption of the SL. In order to estimate the calorific value of the SL, we conducted a restricted diet growth experiment over 21 d on weanling Wistar male rats with SO as caloric control. Diets for the test groups were modified by adding 5, 10, and 15% SO for the control groups, and 5 and 10% SL for the experimental groups. Food consumption of the test groups was restricted to 50% of the feed containing 5% SO that had been consumed by the ad libitum group the previous day. Body weights were recorded during the experiment. Calorific value of the SL was estimated by comparing the 21 st-d mean body weight gain of the control group with that of the experimental group. Estimated calorific value of the SL was 5.36 kcal/g. Most of the behenic acid fed was excreted, as indicated by the analysis of the fatty acids of plasma and fecal total lipid. A second growth experiment on ad libitum diet was conducted over 21 d on weanling Wistar male rats to compare the absorption behavior of the SL with that of natural oil. SO (10%) was added to the diet of the control group, and SL (10%) was added to the diet of the experimental group. Feed consumption, as well as body weights, was recorded during the experiment. The growth pattern of the experimental group was identical to that of the control group during the period of study. The mean feed intake (9.8 g/d/rat for the control group vs. 9.9 g/d/rat for the experimental group) indicated good palatability of the product. In conclusion, the enzymatically synthesized SL containing EFA and natural antioxidants has nutritional properties almost identical to those of natural fats, and can be used as a reduced calorie fat.

Associated diagnoses (comorbidity) in patients with borderline personality disorder

The authors administered the Diagnostic Interview Schedule to 21 patients with borderline personality disorder. The patients met criteria for various other DSM-III diagnoses, meeting exclusion criteria in some cases, and not in other cases. Frequency distribution of each diagnosis and the diagnoses of each individual patient, are presented. Affective disorder was the most common diagnosis (85%). Of these, 62% had primary major depression, and 23% had secondary depression. Other diagnoses include bipolar disorder, dysthymia, panic, agoraphobia, alcohol and Drug abuse, somatization disorder, and many others. The authors conclude that while borderline disorder may be a sub-affective disorder, a specific diagnostic profile for this disorder that accounts for the presence of other Axis I and Axis II syndromes has yet to be delineated.

Rhabdomyolysis, renal failure and schizophrenia

Rhabdomyolysis appears to be a complication that psychiatrists need to be aware of. A case is reported here of a patient who developed rhabdomyolysis and renal failure as a result of intense activity during a psychotic episode.

EEG sleep, lithium transport, dexamethasone suppression, and monoamine oxidase activity in borderline personality disorder

Twenty-one patients who met DSM-III criteria for borderline personality disorder (BPD) and also scored at least 7 on the Diagnostic Interview for Borderlines (DIB) were assessed on four biological markers: electroencephalographic (EEG) sleep, in vitro lithium ratio, platelet monoamine oxidase (MAO), and dexamethasone suppression test (DST). REM latency averaged 58.66 (SD 14.39); platelet MAO averaged 21.74 (SD 10.33); and lithium ratio was 0.357 (SD 0.139) in the BPD patients. All of those values were significantly abnormal. Many patients had abnormalities on three or four measures. These patients in general had multiple Axis I diagnoses from the Diagnostic Interview Schedule (DIS), and these Axis I diagnoses tended to produce patient clusters. Patients with a DIS diagnosis of schizophrenia, mania, hypomania, or schizoaffective mania had elevated lithium, low MAO, and normal EEG sleep, while those patients with coexisting major depression tended to have short rapid eye movement (REM) latency, high REM density, and normal MAO and lithium ratio. Only two patients were nonsuppressors on the DST, confirming recent reports of normal DST results in personality disorders.