Liraglutide and not lifestyle intervention reduces soluble CD163 after comparable weight loss in obese participants with prediabetes or type 2 diabetes mellitus

The current study was a post hoc analysis using stored serum and plasma samples from a randomized, controlled, parallel-arm study designed to assess the effects of an equal degree of weight loss, achieved by either lifestyle changes or liraglutide, on cardiometabolic variables in obese subjects with impaired glucose tolerance (IGT) and/or impaired fasting glucose (IFG) or early T2DM. The protocol and patient characteristics have been previously described [13]. In brief, 62 obese patients with prediabetes (IFG and IGT) or early T2DM were enrolled at the Obesity and Diabetes Clinics of Chieti University Hospital. In addition, 13 subjects, without obesity, diabetes mellitus or prediabetes and not on pharmacological treatment, were enrolled as controls. The patients were randomized 1:1 to receive liraglutide or lifestyle counseling. Study medication was supplied to the research pharmacy by Novo Nordisk as liraglutide 6.0 mg/mL in 3-mL prefilled pen injectors. Liraglutide treatment was administered daily by subcutaneous injection at bedtime with an initial dose of 0.6 mg/day (first week) and titrated over a 3-week period to doses of 1.2 mg daily (second week) and 1.8 mg daily (third week) based on the clinical response and side effects. The nonattainment of the 1.8 mg dose level did not constitute a withdrawal criterion. The participants in the liraglutide arm were encouraged to continue with their existing dietary and exercise habits in addition to liraglutide. The second arm consisted of an intensive lifestyle intervention. The lifestyle-arm participants received two initial days of education on the “Mediterranean diet” and food label education. The aim was a diet with an average of 30% lipids, minimum 15% protein, and maximum 10% of simple sugars, with focus on fibre rich foods, nuts, legumes, and fish rich in omega-3 fatty acids. In addition, they were recommended to consume 10 g of dark chocolate daily, as well as to reduce the intake of salt. Regarding physical activity they were recommended three hours of physical activity per week and two of the three hours were scheduled exercise with the intervention team. Both groups had regular check-ins with the intervention team, but the lifestyle group had in addition first weekly, then biweekly, and lastly monthly consultations with the team to discuss how they were doing with the lifestyle program, and to help keep the motivation up until the goal of 7% weight loss was achieved. Participants in both groups continued with their assigned treatment until they lost 7% of their initial body weight (calculated on the basis of body weight at baseline visit at the time of randomization). Six patients did not achieve this amount of weight loss within 15 months after randomization and were excluded, in addition to 16 participants that dropped out, leaving n = 20 in each treatment arm.

All study visits and procedures took place at the Clinical Research Center within Department of Medicine and Aging, Center for Advanced Studies and Technology (CAST), University of Chieti, Italy. Each patient provided written informed consent to participate, and the Protocol was approved by the Ethics Committee of the University of Chieti, and the Regional Ethical Committee in South-Eastern Norway approved the import in Norway of blood for laboratory assessments.

Venous blood samples (EDTA platelet-poor plasma) were collected at inclusion in the study and after the achievement of the 7% weight loss goal. At both visits an oral glucose tolerance test (OGTT) was performed and blood samples were taken before (T0), and 60, 90 and 120 min after a 75 g glucose load. The β cell secretion during an OGTT was estimated by applying a minimal model of glucose-induced insulin secretion to the glucose and C-peptide curves of each subject, as previously described in detail [14]. In addition, we evaluated another OGTT-based measure of B-cell function: the insulin secretion-sensitivity index-2 (ISSI-2) (defined as the ratio of the area-under-the-insulin-curve to the area-under-the-glucose curve, multiplied by the Matsuda index) [15]. All samples were frozen at − 80 °C for subsequent biochemical measurements.

Plasma levels of immune cell markers sCD163, sCD14, MPO, NGAL, sTIM-3 and sCD25 were measured by DuoSet enzyme-linked immunosorbent assays from R&D Systems (Stillwater, MN). The inflammatory marker LIGHT/tumor necrosis factor super family member 14 (TNFSF14) was measured with a Quantikine enzyme-linked immunosorbent assays from R&D Systems (Stillwater, MN). All were analyzed in a 384-format using a combination of a SELMA pipetting robot (Analytik Jena AG, Jena, Germany) and a BioTek dispenser/washer (BioTek Instruments, Winooski, VT). Absorption was read at 450 nm by using an EIA plate reader (BioTek Instruments) with wavelength correction set to 540 nm. Samples from all patients and controls were run on the same 384-well plate. Calculated limit of the detection (3*SD + OD of 0-standard) was 0.45 ng/mL (sCD163), 0.23 ng/mL (sCD14), 0.086 ng/mL (MPO), 0.035 ng/mL (NGAL), 0.0093 ng/mL (sTIM-3), 23 pg/mL (sCD25) and 10.4 pg/mL (LIGHT), respectively. The inter- and intra-assay coefficients of variation were < 10%.

In a study with 20 patients in each treatment arm (liraglutide vs. lifestyle), we were able to detect, by the end of the treatment period, a genuine difference in the mean response between the experimental and control arms, equivalent to one standard deviation of a not pre-specified continuous outcome. This was achieved with a power of 0.9 and a Type I error probability of 0.05. Comparisons of variables between groups (liraglutide versus lifestyle versus controls) and between arms (liraglutide versus lifestyle advice) were performed by χ2 test or Mann–Whitney U test. Spearman rank correlation test was used to assess relationships among continuous variables. All tests were two-tailed. All calculations were carried out using SPSS (SPSS, Chicago, IL, USA).

Clinical and biochemical baseline characteristics of the study subjects have been previously presented [12, 13, 16] and relevant variables are shown in Table 1. Patients randomized to liraglutide treatment and lifestyle intervention were similar on most parameters, except for triglycerides (TG), waist circumference and visceral adipose tissue (VAT) being higher in the liraglutide arm. Compared to healthy controls, both patient groups had a lower age, higher BMI, lower CRP higher total- and LDL cholesterol, and lower HDL cholesterol. In the liraglutide group, 10 (50%) subjects had IFG or IGT and in the lifestyle group 13 (65%) subjects had IFG or IGT.

At baseline, plasma levels of the macrophage marker sCD163, but not sCD14, was higher in patients as compared to healthy subjects (Table 2). The neutrophil marker MPO was higher in the lifestyle group, compared to the liraglutide group, but when compared to controls they were not statistically different (Table 2). There were no significant differences in plasma levels of the neutrophil marker NGAL, or of the T-cell activation markers, sTIM-3 and sCD25, when comparing cases and controls at baseline (Table 2).

In the study group as a whole, baseline levels of sCD163 correlated positively with several metabolic parameters such as BMI (rho = 0.432, p = 0.006), C-peptide (rho = 0.410, p = 0.009), insulin (rho = 0.340, p = 0.034), total cholesterol (rho = 0.358, p = 0.025), leptin (rho = 0.481, p = 0.002) and HOMA-IR (rho = 0.389, p = 0.014), and negatively with Matsuda index (rho = − 0.337, p = 0.042). Baseline levels of sCD163 also correlated positively with other markers related to inflammation, i.e., C-reactive protein (CRP, rho = 0.337, p = 0.042), IL-10 (rho = 0.559, p = 0.002), and total leukocyte counts (rho = 0.493, p = 0.001) as well as with non-alcoholic fatty liver disease (NAFLD) prior to intervention (rho = 0.356, p = 0.026, Additional file 1: Table S1).

Baseline MPO showed a negative correlation with waist-to-hip ratio (WHR, rho = − 0.345, p = 0.031) and beta-index (rho = − 0.356, p = 0.024, Additional file 1: Table S2).

All study participants except one attained the 1.8 mg dose level through-out the study period. The amount of weight loss was prespecified by the protocol to 7% of initial body weight and did not differ between the groups. Median time to predefined weight loss was 4.8 months and did not differ between the two treatment arms [13]. Concomitant therapy was unchanged during the follow-up. At the end of the intervention period (i.e. after achievement of the weight loss target) both groups experienced a reduction in several metabolic parameters, including BMI, HbA1c, fasting plasma insulin, disposition index, as well as CRP as a reliable marker of systemic inflammation, with the decrease in VAT and beta-index being more pronounced in the liraglutide arm (Table 3) [13]. The liraglutide arm showed decreased systolic blood pressure and total cholesterol, and improved beta cell function, as assessed by beta-index, and glucose tolerance, indicated by reduced fasting glucose and 1- and 2-h post load glucose levels (Table 3). On the other hand, the lifestyle intervention group showed a significant decrease in 2-h post load insulin levels, which was not significant in the liraglutide group (Table 3). There were, however, no associations between changes in the inflammatory markers (sCD163 and MPO) and changes in any of metabolic parameters (Additional file 1: Table S4).

At the end of the intervention period, we observed a significant reduction in sCD163 levels in the liraglutide arm (∆ = 87, SD = 115, p = 0.001), but not in the lifestyle arm (∆ = 24, SD = 85), with a significant between-group difference in ∆sCD163 also when adjusted for basal VAT and basal triglycerides values (p = 0.026) (Fig. 1A). In contrast, levels of sCD14 as an additional marker of monocyte/macrophage activation, were not affected by intervention in any of the two arms (Fig. 1B). Weight loss induced a significant decrease in MPO levels in the group receiving liraglutide (p = 0.048) (Fig. 1C), and not in the lifestyle intervention group, but the difference in decreases between arms was not statistically significant. For the other neutrophil markers (i.e., NGAL) and the T-cell markers (i.e., sTIM-3 and sCD25) no significant between arm difference was observed during the intervention. (Fig. 1D–F).

A 75 g oral glucose load given before the intervention, did not show any change in levels of sCD163 over a period of 120 min (p = 0.835) and there were, as expected, no differences in response between the treatment groups at baseline (p = 0.860) (Fig. 2A). After intervention, however, we observed a reduction in the levels of sCD163 in the whole study population, after the glucose load over time (p = 0.001). This effect was more pronounced in patients randomized to liraglutide compared to the lifestyle arm, although the difference between arms was not statistically significant (p = 0.081, Fig. 2B). Interestingly, and in concordance with the baseline correlations between plasma glucose and sCD163, during the OGTT post intervention, the sCD163 AUC after glucose-load was significantly and directly related to the glucose AUC only in the lifestyle arm (data not shown), suggesting that liraglutide directly affects sCD163, independently of blood glucose rise. No change in any of the other variables studied was observed during OGTT neither before nor after weight loss.

We have previously shown that the inflammatory cytokine LIGHT/TNFSF14 is increased in patients with T2DM and can induce beta cell death and impair insulin secretion [17]. Monocyte/macrophages are important cellular sources of LIGHT and based on the regulation of sCD163 in the liraglutide arm, we therefore analyzed LIGHT levels in the study group. At baseline, LIGHT was significantly increased in the patient cohort, as compared to controls (Fig. 3A), and indeed, LIGHT showed a positive correlation with sCD163 (rho = 0.417, p = 0.001) in the study group as a whole. Further, LIGHT correlated with total leukocyte counts (rho = 0.395, p = 0.017) and with ISSI-2 at baseline (rho = − 0.321, p = 0.044) (Additional file 1: Table S3). After intervention, LIGHT was reduced in the liraglutide group (p = 0.003), but not in the lifestyle group, the difference between treatment arms where however not statistically significant (Fig. 3B, C).