Impact of malnutrition on systemic immune and metabolic profiles in type 2 diabetes

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and the natural history study protocol (12-I-073) approved by Institutional Review Boards of the National Institute of Allergy and Infectious Diseases (USA) and the National Institute for Research in Tuberculosis (India) Ethical committee approval number NIRT-IEC ID: 2013001. Before getting informed consent, the health care provider explains about the risks, benefits, the nature of the procedure, reasonable alternatives, risks and benefits of alternatives, and assessment of the participants understanding of above mentioned factors.

We enrolled small sample size of 88 study participants with T2DM, of which 44 individuals were with LBMI (Low Body Mass Index) and 44 individuals were with NMBI (Normal Body Mass Index) from Kanchipuram District, Tamil Nadu, South India (Table 1). Participant recruitment flowchart was shown in S.Figure.1 Samples were obtained from all the participants. No power calculations were done to determine sample size and it was based on samples available following matching for age and sex. All the participants were screened for diabetes and nutritional indices. Participants of both gender and age between 18 and 75 years were included. Study participants having liver disease, renal disease, cardiac disease, respiratory disease or any other acute or chronic diseases, HIV, hypertension, cancer and any previous history of anthelmintic treatment or a history of helminth infections were precluded from the study. Pregnant women were also excluded from the study. Based on 2013 American Heart Association/ American College of Cardiology guidelines (LBMI< 18.5 and NBMI between 18.5 and 24.9 kg/m²), Low and normal BMI were defined [19]. In addition, undernutrition was confirmed by the presence of low serum albumin (< 3.4 g/dl) in all the LBMI individuals [11]. BMI was calculated as body weight (in kilograms) divided by body height (in meters) squared.

An ACT 5 Diff. hematology analyzer (Beckman Coulter, Brea, CA, USA) was used to quantify haematological parameters using venous EDTA blood samples. Anthropometric assessments (comprising height, weight and waist circumference) were assessed by trained personnel [20].

Overnight fasting venous blood samples were collected from the recruited individuals in EDTA -containing tubes using standardised protocol and equipment. They were separated into two samples: the first sample containing whole blood for the measurement of HbA1c and the other plasma specimen was used for lipid profile levels and renal and liver function tests. Random Blood Glucose (RBG) test was done within two hours of eating and analysed using auto analyser. The RBG normal value should be 180 mg/dl as per the American Diabetes Association. Plasma samples were used to measure other biological parameters [2, 21].

Based on American Diabetes Association criteria, Type 2 diabetes was confirmed by an HbA1c value of 6.5% or greater and a random blood glucose of > 200 mg/dl. Overnight fasting samples were used to determine all biochemical parameters with exclusion of random blood glucose [2, 21]. All diabetic participants were newly diagnosed, not on any anti-diabetic medication at the time of blood draw and without any known complications or co-morbidities. All participants were referred to the primary health care centre for diabetic treatment.

Pancreatic hormones (insulin and glucagon), adipocytokines (adiponectin, adipsin, resistin, leptin, visfatin) and the levels of Type 1 cytokines: IFNγ, TNFα, IL-2, Type 17 cytokines: IL-17A, IL-22, Type 2 cytokines: IL-4, IL-5, IL-13, regulatory cytokine: IL-10 and pro-inflammatory cytokines: G-CSF, GM-CSF, MCP-1, MIP-1β, IL-6, IL-7, IL-8, IL-12p70 and IL-1β plasma levels were estimated using Bioplex multiplex assay system (Bio-Rad, Hercules, CA). TGF-β and IL-17F were measured by conventional ELISA according to the manufacturer’s protocol (R&D Systems, Minneapolis, MN, USA). All the measured variables such as HbA1c, cytokines and hormones were illustrated as mean, SD, intra-assay coefficient of variation (CV%), inter-assay coefficient of variation (CV%) and detection limit were shown in Table S1.

Values were expressed as geometric means. Shapiro-Wilks test was used to assess normality of the data. Nonparametric Mann-Whitney U test was used to compare the pro-inflammatory and regulatory cytokines between the group LBMI and NMBI. Spearman rank correlation was performed to assess the relationship of the BMI with glycemic parameters, pancreatic hormones, adipocytokines. Appropriate power calculation was done to assess the acceptability of the statistical findings.

Univariate and Multiple logistic regression models were used to measure the association between various factors and Low Body mass Index. Odds ratio and 95% CIs were reported for each factor in the univariate and for the factors adjusting for age and gender in the final multivariate models. Principle Component Analysis (PCA) was used with a goal of identifying potentially significant patterns of pancreatic hormones, adipocytokines and cytokines which is responsible for any of the clustering/separation between the normal and low body mass index.

Study participants demographics and biochemical parameters at baseline are presented in Table 1. The median age of the LBMI group was 37 years and the age ranged from 23 to 55 years. The median age of the NBMI group was 39 years and the age ranged from 24 to 62 years. Age, sex, lipid profile or other biochemical parameters did not differ significantly between the two groups. Baseline hematology parameters are listed in Table 2. Haematological parameters did not differ significantly between two groups.

To examine the impact of malnutrition on indices of glucose control in T2DM, we measured the levels of glycated haemoglobin (HbA1c), random blood glucose (RBG) and pancreatic hormones in LBMI and NMBI individuals. As exhibited in Fig. 1a, the levels of HbA1c (Geometric Mean (GM) of 11.27% in LBMI Vs 9.2% in NBMI), RBG (GM of 279.6 mg/dl in LBMI Vs 220.4 mg/dl in NBMI), insulin (GM of 95.88 pg/ml in LBMI Vs 54.13 pg/ml in NBMI) and glucagon (GM of 232.5 pg/ml in LBMI Vs 177.5 pg/ml in NBMI) were significantly higher in LBMI in comparison with NBMI individuals. Subsequently, we examined the impact of malnutrition on adipocytokines in T2DM by measuring the plasma levels of adiponectin, adipsin, resistin, leptin and visfatin in LBMI and NBMI individuals. As exhibited in Fig. 1b, LBMI individuals showed significantly higher levels of adiponectin (GM of 214,516 pg/ml in LBMI Vs 134,223 pg/ml in NBMI) and adipsin (GM of 29,794 pg/ml Vs 15,054 pg/ml) and in contrast, significantly lower levels of leptin (GM of 1832 pg/ml Vs 2469 pg/ml) when compared to NBMI individuals. Hence, malnutrition is connected with higher levels of the HbA1c, RBG, pancreatic hormones, adiponectin and adipsin and lower levels of leptin in T2DM individuals.

Next, we wanted to study, whether malnutrition in T2DM was linked with changes in the levels of Type 1 and Type 17 cytokines. To this end, we measured the circulating levels of Type 1 (IFNγ, TNFα and IL-2)- and Type 17 (IL-17A, IL-17F and IL-22) cytokines in LBMI and NBMI individuals. As exhibited in Fig. 2a, LBMI individuals had significantly lower levels of IFNγ (GM of 300.9 pg/ml Vs 357.7 pg/ml), TNFα (GM of 259.5 pg/ml Vs 309.8 pg/ml), IL-2 (GM of 125.6 pg/ml Vs 163.3 pg/ml), IL-17A (GM of 30.1 pg/ml Vs 38.3 pg/ml) and IL-17F (GM of 120.6 pg/ml Vs 146.8 pg/ml) in comparison with NBMI individuals.

Further, we wanted to study, whether malnutrition in T2DM was linked with changes in levels of Type 2 and regulatory cytokines. To this end, we measured the circulating levels of Type 2 (IL-4, IL-5 and IL-13) and regulatory (IL-10 and TGF-β) cytokines in LBMI and NBMI individuals. As exhibited in Fig. 2b, IL-4 (GM of 32.8 pg/ml Vs 61.9 pg/ml), IL-5 (GM of 89.4 pg/ml Vs 117.6 pg/ml), IL-13 (GM of 43.8 pg/ml Vs 65.9 pg/ml), IL-10 (GM of 96.3 pg/ml Vs 108.2 pg/ml) and TGF-β (GM of 153.1 pg/ml Vs 208.7 pg/ml) levels were significantly lower in LBMI in comparison with NBMI individuals. Hence, malnutrition with T2DM is linked to lower circulating levels of Type 1, Type 17, Type 2 and regulatory cytokines.

To examine the impact of malnutrition on other pro-inflammatory cytokines in T2DM, we measured the circulating levels of pro-inflammatory cytokines (G-CSF, GM-CSF, MCP-1, MIP-1β, IL-6, IL-7, IL-8, IL-12p70 and IL-1β) in LBMI and NBMI individuals. As exhibited in Fig. 3, LBMI individuals showed significantly lower levels of G-CSF (GM of 91.1 pg/ml Vs 107.9 pg/ml), GM-CSF (GM of 126.2 pg/ml Vs 174.2 pg/ml), MCP-1 (GM of 119 pg/ml Vs 220.6 pg/ml), MIP-1β (GM of 60.9 pg/ml Vs 86.3 pg/ml), IL-6 (GM of 38.7 pg/ml Vs 72.2 pg/ml), IL-8 (GM of 74.8 pg/ml Vs 205.5 pg/ml) and IL-12p70 (GM of 95.9 pg/ml Vs 107.1 pg/ml) when compared to NBMI individuals. Hence, malnutrition with T2DM is linked to lower circulating levels of pro-inflammatory cytokines.

To examine the relationship between glycemic parameters, pancreatic hormones, adipocytokines and BMI, we assessed the association of these parameters with BMI in all the individuals in the study. As shown in Fig. 4a, Tables 3 and 4, HbA1c, RBG, insulin and glucagon exhibited a significant negative relationship with BMI, while leptin exhibited a significant positive relationship to BMI. To examine the relationship between the systemic levels of Type 1, Type 17, Type 2, regulatory and other pro-inflammatory cytokines with BMI, we assessed the association of the cytokines with BMI in all the individuals in the study. As shown in Fig. 4b, the systemic levels of IFNγ, IL-2, IL-17F, IL-4, IL-5, IL-13, IL-10, TGF-β, G-CSF, GM-CSF, MCP-1, MIP-1β, IL-6, IL-8 and IL-12p70 each exhibited a significant positive association with BMI. In addition, we have shown the Spearman rank correlation and power analysis to demonstrate the relationship of BMI with biochemical parameters and cytokines and the correlation analysis explicitly shows sufficient correlation between the parameters (Table 3). Power analysis revealed that biochemical parameters have impact on BMI in T2DM individuals (Tables 3 and 4). Thus, biochemical and immunological parameters exhibit a complex relationship to BMI in T2DM individuals.

PCA was performed to envision variation between the groups created on the complete data set. To envision the clustering pattern of pancreatic hormones, adipocytokines and cytokines between LBMI (red circle) and NBMI (blue circle) individuals, we plotted PCA with pancreatic hormones (insulin and glucagon) and adipocytokines (adiponectin, adipsin, resistin, leptin, visfatin). In addition, the levels of Type 1 cytokines: IFNγ, TNFα, IL-2, Type 17 cytokines: IL-17A, IL-17F, IL-22, Type 2 cytokines: IL-4, IL-5, IL-13, regulatory cytokines: IL-10, TGF-β and pro-inflammatory cytokines: G-CSF, GM-CSF, MCP-1, MIP-1β, IL-6, IL-7, IL-8, IL-12 and IL-1β) were plotted between the 2 groups. The factors were reduced by excluding the factors with the commonalities as low as 0.3 and then the revised PCA was presented in the article with the following factors, PCA-1 (adiponectin, adipsin, Glucagon and insulin) and PCA-2 (Glucagon, insulin and HbA1c). The above factors were used and the plot of first two components was presented with its variability. As illustrated in Fig. 5a, PCA analysis showed that pancreatic hormone and adipocytokine clusters varied between LBMI and NBMI individuals with T2DM. The score plot of the first two components revealed 32.7 and 31.9% of overall variance, respectively. As illustrated in Fig. 5b, PCA analysis of cytokines showed two different clusters between LBMI and NBMI individuals. The factors were reduced by excluding the factors with the commonalities as low as 0.3 and then the revised PCA was presented using following factors, PCA-1 (TGF-β, GM-CSF, IL-13, IL-2, IL-17F, IL-4 and IL-8) and PCA-2 (GM-CSF, IL-2, IL-17F and IL-8). The above factors were used and the plot of first two components was presented with its variability. The score plot of the first two components revealed 35.4 and 21.9% of overall variance, respectively. The findings of PCA results were similar with multivariate and univariate analysis. Thus, PCA analysis divulge the overall influence of pancreatic hormones, adipocytokines and cytokine of T2DM individuals on BMI.

In this study, multivariate regression analysis was performed to determine the impact of confounding variables on the numerous analytes determined in T2DM individuals. As presented in Table 5, the levels of biochemical parameters such as AST, Serum triglycerides, HDL, LDL, RBG, HbA1c, diabetic parameters such as insulin, glucagon, adiponectin, adipsin, resistin, leptin and visfatin; cytokines like IFNγ, TNFα, IL-2, IL-17A, IL-17F, IL-22, IL-4, IL-5, IL-13, IL-10, TGF-β and pro-inflammatory cytokines: G-CSF, GM-CSF, MCP-1, MIP-1β, IL-6, IL-7, IL-8, IL-12p70 and IL-1β and were all significantly influenced by BMI even after adjusting for the effect of the age and gender. Hence, our data corroborate that LBMI has an impact upon numerous relevant crucial parameters in T2DM individuals, including blood glucose levels, and the levels of the adipocytokines and the more conventional cytokines.