Association of proinflammatory genes expression with serum interleukin 1β and free fatty acids in metabolically healthy and unhealthy abdominally obese individuals: a case-control study

Obesity is a major universal health threat. Worldwide obesity has more than doubled since 1980 and evidence suggest that 51% of the population will be obese by 2030 [1]. Metabolic disorders are one of the main deleterious health risks of obesity [2]. However, several studies show that there are individual differences in metabolic reactions of obesity; therefore, there are different subtypes of obesity [3, 4].

A subgroup of obese individuals who do not have metabolic complications of obesity, are called metabolically healthy obesity (MHO) [5‐9]. It is estimated that 10–25% of the obese individuals are MHO [10]; however, there is no definite criteria to describe it [4]. This phenotype is mainly characterized by normal metabolic parameters such as serum lipid and glucose levels and blood pressure, despite elevated adiposity and body mass index (BMI) [9, 11]. Accumulating data indicate that inflammation is the linking point of adiposity and metabolic disorders, but there is no convincing explanation for differences found in metabolically healthy and unhealthy obese individuals [12]. While some studies claim that MHO subjects have lower inflammation levels which contribute to their favorable metabolic profile, other reports do not show normal inflammatory profile for MHO individuals [13].

It is unclear whether inflammation starts in obesity state; however, the activity of some pathways initiating with Toll-Like Receptors (TLRs) is established. Toll-like receptor2 (TLR2), as one of the members of these metabolic sensors, has received more attention [14]. It is a pattern recognition immune receptor which recruits numerous adaptor proteins of the myeloid differentiation factor 88 (MyD88)-dependent pathway, causing subsequent inflammatory responses via activation of NFĸB [15, 16]. TLRs mediate NFĸB activation, as a principal transcriptional factor, which initiate inflammatory cascade and controls the expression of various inflammatory cytokines [16]. Animal studies have revealed that deficiency of MyD88 can reduce expression of proinflammatory genes and cytokines such as IL-1β and tumor necrosis factor-α (TNF-α) [17, 18]. IL-1β, as a pro-inflammatory cytokine, has been shown to have important effects on fat mass, fat metabolism, and body mass and induce insulin resistance [19].

It is elucidated that TLR2 dysregulation during obesity translates a metabolic challenge into an inflammatory response and contributes to obesity-associated metabolic diseases [14]. However, little is known about its activity in different obesity subtypes; moreover, their relationship with metabolic abnormalities is yet to be clarified. Abdominal obesity is more harmful than any other type of obesity with higher total body fat [20]. Researchers have revealed that abdominal (visceral) fat excess is highly concerned with metabolic diseases [21, 22]. Also, increased adipose tissue mass because of free fatty acids (FFAs) released from enlarged and hypertrophied adipocytes is a crucial in the progression of inflammation. Hence, investigation of expression levels of TLR2 with its main adaptor protein and their association with inflammatory factors seems to be demanded.

Studies on persons with diabetes and metabolic syndrome (MetS) have reported upregulated levels of TLR2 and MyD88 genes [23‐25]. In contrast, a study revealed no significant differences in gene expression levels of TLR2 in peripheral blood mononuclear cells (PBMCs) of MHO and metabolically unhealthy obesity (MUO) individuals [26]. In addition, another research indicated that expression of genes involved in inflammation had a similar alteration pattern in MHO and MUO persons [27]. There is a paucity of data about TLRs activity in MHO and MUO persons and published results are conflicting [26‐28]. It is not verified that observed expression patterns are result from adiposity or related to individuals metabolic status. For this reason, our hypothesis was that changes in serum inflammatory factors and expression levels of the related genes are independent of metabolic abnormities and are only related to the adiposity. Therefore, we considered high waist circumference (WC) as a marker of adiposity and investigated TLR2, MyD88, and NFĸB gene expression levels and serum inflammatory factors in metabolically healthy abdominally obese (MHAO) and metabolically unhealthy abdominally obese (MUAO) individuals.

The mean age of the participants in the case and control group was 35.14 ± 0.97 and 35.94 ± 1.13 years, respectively. Males comprised 51% of the subjects. Age, PAL, marital status, job, and level of education were not significantly different, when comparing MHAO with MUAO (Table 1). No statistically significant differences were found either in anthropometric measures or dietary intake between the two groups (Table 2). However, metabolic parameters including FBS (p = 0.012), TG (p < 0.001), and HDL-C (p = 0.005) were significantly different between the case and control group (Table 3). The stages of the study as a flowchart is shown in Fig. 1. Case group had higher FBS (30.6% vs. 5.9%), TG (87.2% vs. 25.8%), (systolic blood pressure (18.4% vs. 3.2%), and diastolic pressure (26.3% vs. 9.7%) than control group, Also decreased levels of HDL-C was more observed in cases than controls (77.8% vs. 44.1%). (p < 0.001 for all) (Fig. 2). Furthermore, the risk of MUAO was significantly increased with higher levels of serum FBS (OR = 7.04, 95%CI: 1.42–34.69) and TG (OR = 30.55, 95%CI: 7.48–60.67) and significantly decreased with higher HDL-C (OR = 0.22, 95% CI: 0.08–0.63) (Table 3).

There were no significant difference in the median of serum IL-1β in cases and controls (756 pg/mL, 710.5 pg/mL, respectively; p = 0.638). FFAs also did not show significant differences between the case and control group (1292 nmol/L,1326 nmol/L, respectively; p = 0.778) (Table 3). However, a significant correlation was found between log IL-1β and log FFAs in controls (r = 0.763; p < 0.001) and in cases (r = 0.760; p < 0.001) (Fig. 4). mRNA expression levels of TLR2, MyD88, and NFĸB were elevated in MUAO compared with MHAO, but it was not statistically significant (Fig. 3). The correlation of log NFĸB expression with log TLR2 (r = 0.747; p < 0.001) as well as log MyD88 (r = 0.747; p = < 0.001) was significant. Significant correlations were also noticed between log TLR2 and log MyD88 expression levels (r = 0.417; p < 0.001) only in the case group (Fig. 4).

This study for the first time focused on TLR2, MyD88, and NFĸB genes expression levels and serum levels of IL-1β and FFAs as biological indicators of the activation of these genes in MHAO vs. MUAO individuals. We found a strong association of NFĸB with TLR2 and MyD88 as well as between TLR2 and MyD88 expression levels in the PBMCs. However, the expression levels of the study genes as well as IL-1β and FFAs were similar between MUAO and MHAO groups.

As the two groups were matched on abdominal fat, it looks that gene expression levels of TLR2, MyD88 and NFĸB are highly related to abdominal obesity than to healthy or unhealthy metabolic state. There is limited data on examining the role of TLRs in metabolically healthy or unhealthy obesity and their correlations with metabolic state of obesity [26, 27] . A study showed higher mRNA expression levels of TLRs in monocytes of patients with MetS than controls after adjusting for WC [23]. In the study of Ahmad’s et al. on obese participants, the expression of TLRs and MyD88 were increased in subcutaneous adipose tissue. Migration of inflammatory PBMCs from the peripheral compartment toward the adipose tissue could be the cause of these results [29]. We used PBMCs that mostly contain monocytes and lymphocytes PBMCs are convenient for gene expression studies because they can be simply collected inadequate quantities and compared with biopsy of other tissues is less invasive [30, 31].

In line with our findings, Telle-Hansen et al. in a study on two obese groups reported no differences in the TLR2 and TLR4 gene expression in PBMCs between MHAO and MUAO groups [26]. It is notable that the significant difference was not observed even after comparison with normal-weight controls; however, in their study, the expression level of downstream adaptor proteins was not assessed. Moreover, in another cross-sectional study, the gene expression level of TLR4 and TNF-α did not differ significantly between MHAO and MUAO groups [27]. Some studies indicated that gene expression and production of IL-6 and TNFα are elevated in general and abdominal obesity [32‐34]. However, such finding was not found in the present study due to similar WC in both groups. Previous studies claim that TLRs mostly signal through the adaptor protein MyD88 via activation of NFĸB [25, 35]. In our study, the expression levels of TLR2 and MyD88 were significantly correlated with NFĸB, though their expression levels were not significantly different between the two groups. Dasu et al. in a study on patients with T2DM reported that increased expression level of TLR2 and TLR4 results in raised inflammation, mediated by NFĸB [24]. However, in the case-control study of Devaraj et al. on T1DM, TLR2 and TLR4 were significantly correlated with NFĸB expression levels [25].

In another study metabolically healthy obese woman revealed lower amounts of visceral fat and a more favorable inflammatory profile in compared to the metabolically unhealthy women [11]. However, no difference was found for IL-1β with various metabolic health criteria, even after multivariate analysis and restricting the analysis to the obese (BMI ≥ 30 kg/m²). Another research among 58 obese postmenopausal women showed more visceral adipose tissue in women with MetS, but there were no differences in levels of inflammatory markers compared to women with no MetS [36]. These studies were carried out on obese older persons and aging may affect metabolic alterations; since the higher visceral fat region is accompanied with unhealthy metabolic phenotype in obese elderly.

Previous studies have shown that MHO definitions can influence the association between inflammatory biomarkers and MHO, based on the criteria used. We used the MHO definition of Meigs et al. in which the differences between MHO and non-MHO subjects are well-distinguished [6]. In the present study, serum levels of IL1β were not significantly different between MHAO and MUAO. Our data are in agreement with the observations from Marques-Vidal et al., who found no differences in IL-1β level of MUO subjects compared to MUO [28]. When definition was based on Meigs et al., adjustment for abdominal obesity or percent body fat did not alter the non-significant differences [6]. Moreover, their results did not show a consistent association between metabolically healthy status and IL-1β level. Based on the MHO definition used, higher, similar or lower IL-1β levels were found between MHOs and MUOs [28]. However, Jialal et al. showed that persons with MetS had significantly higher levels of inflammatory cytokines (IL-1β, IL-8, and IL-6) than control subjects without MetS [23]. In their study, it was claimed that MetS is a proinflammatory state, independent of adiposity. Conversely, ABC study indicated that visceral obesity is constantly related to higher levels of C-reactive protein (CRP) and IL-6 [37]. In another study on both subcutaneous fat and visceral fat of obese persons, unfavorable lipid profile was correlated with visceral adipocyte size, since visceral adipocyte size is directly related to a visceral fat area [38]. This may be the cause of strong association between MetS and visceral obesity than subcutaneous fat. In the present research, we failed to measure visceral fat, though WC can reflect visceral adiposity [39].

Serum FFAs result from the lipolysis of adipocytes. They are implicated in the pathogenesis of obesity-related metabolic states like IR, T2DM, and CVD [40]. Succurro et al. reported increased levels of plasma total FFAs in MUO compared with MHO persons [41]. Elevated serum levels of FFAs in obese individuals are usually correlated with increased amount of adipose tissue which may justify our finding regarding the similarity of serum levels of FFAs between MHAO and MUAO subjects. Boden et al. showed that FFAs link with metabolic diseases through increased generation of deleterious proinflammatory cytokines [42]. This study confirms our results which showed that FFAs are positively correlated with IL-1β levels. In fact, any augment in serum level of FFAs in the abdominally obese persons, irrespective of their metabolic aberrations, can lead to a significant increase in serum levels of IL-1β.

In this study, dietary intake was compared between the two obese groups; therefore, no significant difference was found. However, when examining the relationship of dietary parameters (intake and composition) with inflammatory markers, weak but significant associations were observed between TLR2, dietary cholesterol, and carbohydrate percent. Nevertheless, adjusting for dietary cholesterol or carbohydrate percent could not modulate the relationship between inflammatory markers and metabolic parameters. In the present study, neither FFAs nor IL1β was significantly correlated with TLR2, MyD88, and NFĸB gene expression levels. In the study of Jialal et al., FFAs were not correlated with TLR2 as well, suggesting that other factors may be involved [23].

In general, what makes our study distinct from the majority of former ones is that our groups were matched on WC to clarify the effect of abdominal obesity in contrast to the metabolic state of persons. Most of the previous studies have been carried out on patients with MetS vs. those without the syndrome or even normal weight healthy controls as well as metabolically unhealthy vs. metabolically healthy persons, without considering WC status of participants [6, 9, 11, 43]. The present study had some limitations like smaller sample size. However, any difference between the study groups was detectable due to enough power. Although the measurement of protein concentrations along with their gene expressions is more helpful, we could not do so due to budget deficit. The major strength of the current work was that the two study groups were matched based on WC.

We found that WC may play a significant role as a mediator in the relation between proinflamatory genes expression levels and serum metabolic parameters. Likewise, serum levels of IL-1β and FFAs appear to be more related to abdominal obesity than to the metabolic state.