Central obesity and sub-clinical inflammation increase metabolic risk, this study examined the intracellular inflammatory pathways in adipose tissue (AT) that contribute to this risk.
Central obesity and sub-clinical inflammation increase metabolic risk, this study examined the intracellular inflammatory pathways in adipose tissue (AT) that contribute to this risk.
This study therefore addressed the influence of NFκB and JNK activation in human abdominal subcutaneous (AbdSc) and omental (Om) AT, the effect of adiposity, T2DM status and the role of TNFα in vitro, using molecular biology techniques.
Our data showed NFκB activity is increased in Om AT versus AbdSc AT (P<0.01), which was reversed with respect to depot specific activation of JNK (P<0.01). However, T2DM status appeared to preferentially activate NFκB (P<0.001) over JNK. Furthermore, in vitro studies showed recombinant human (rh) TNFα treated AbdSc adipocytes increased NFκB activity over time (2-48 h, P<0.05) whilst JNK activity reduced (2 h, 4 h, P<0.05); inhibitor studies supported a preferential role for NFκB as a modulator of TNFα secretion.
These studies suggest distinct changes in NFκB and JNK activation, dependent upon AT depot, adiposity and T2DM status, with in vitro use of rh TNFα leading to activation of NFκB. Consequently NFκB appears to play a central role in inflammatory mediated metabolic disease over JNK, highlighting NFκB as a potential key target for therapeutic intervention.
Several studies have identified increasing central adiposity as a critical site for metabolic dysfunction and inflammatory response and, as such, adipose tissue (AT) is considered to integrate metabolic and immune functionality . Our previous studies have demonstrated that isolated human abdominal subcutaneous adipocytes (AbdSc) possess many of the key proteins in the inflammatory pathway, including nuclear factor kappa B (NFκB) [2, 3]; to date, most of our current understanding as to the function of NFκB in AT is derived through some key murine studies [4, 5]. Primarily, these studies highlight that the NFκB activator, I-kappaB kinaseβ (IKK-β), increases adipokine production in AT and links inflammation with the onset of diabetes; secondly, that inflammatory responses are markedly attenuated in the IKK-β knock-out (KO) mouse  whilst further studies implicate IKK-gamma (IKKγ) as essential for IKK activity and classical activation of the NFκB pathway .
Concurrently, the mitogen activated protein C-Jun N-terminal kinase (JNK) is also implicated as a central mediator for insulin action and inflammation in AT . JNK represents a family of serine/threonine protein kinases, consisting of three distinct JNK genes JNK 1-3 which exist in mammals. JNK1 and JNK2 are expressed across many tissues including mouse AT, liver, muscle, and macrophages, whilst JNK3 is mainly expressed in the central nervous system (CNS) [8, 9]. Genetic and biochemical studies have shown that, within in vitro systems, JNK activation results in serine phosphorylation of the insulin receptor substrate-1 (IRS-1), thus impairing the insulin signaling cascade [8, 10]. TNFα activates the JNK signaling pathway, as noted in both in vivo and in vitro studies, whilst TNFα deficient mice exhibit restored insulin signaling capacity [11, 12]. Therefore, this suggests that JNK may mediate the regulation of multiple intracellular events, including the expression of the TNFα gene through AP-1 and the insulin signaling pathway [11, 12]. JNK activity is also elevated in AT, liver and muscle tissues of obese subjects , further implicating the JNK pathway as a link between obesity, insulin action, and inflammation.
To date, the role of JNK1 in insulin resistance (IR) has been exclusively defined through murine studies, whilst the role of JNK2 is unclear, in part due to the regulatory crosstalk between the two isoforms . However, KO studies in mice indicate that JNK1 activity is an important mediator of IRS-1 dysregulation, which appears unrelated to JNK2 function . As such, JNK may influence the IR state, as both JNK1 and TNFα can disrupt insulin signaling through differential phosphorylation of the IRS [8, 13].
Whilst a role for JNK and NFκB in mouse models has been described, and its complexity and potential overlapping functionality is ascertained, the functional role and significance in human AT has not been addressed. Therefore, the aims of this study were to examine (i) the inflammatory intracellular signaling pathways involving NFκB and JNK as potential mechanisms of adipocytokine activation leading to inflammation in human AbdSc and omental (Om) AT (ii) the relative JNK and NFκB activity in fat depots and the effect of adiposity, as well as T2DM status (iii) finally, we addressed the acute and chronic in vitro effects of TNFα on NFκB and JNK activity, in human AbdSc adipocytes.
Human abdominal AT was collected from patients (age: 44.7 (mean ± SD) ± 9.3 years; BMI: 27.9 (mean ± SD) ± 7.3 kg/m2; (34 women, 7 men), fasted for 8 h) undergoing elective surgery with informed consent, obtained in accordance with Local Research Ethics Committee (LREC) guidelines and approval. AbdSc AT was obtained from T2DM patients by needle biopsy (age: 59.1±8.2 years; BMI: 35.2±9.2 kg/m2, n=8; 6 women and 2 men). Human sternum muscle (Medical Solutions, Peterborough, UK) and white blood cells were also collected following LREC approval. All tissue samples were flash frozen and/or utilized for in vitro studies, as detailed below. In total, 41 human, non-T2DM (ND) abdominal AT samples were analyzed, which were sub-divided into: AbdSc (n=23) and Om (n=18). Subjects providing fat samples were not on endocrine therapy or receiving any anti-hypertensive therapy.
RNA was extracted from AT using RNeasy Lipid Tissue Mini Kit (Quiagen, Crawley, UK) and reverse transcribed (RT) using AMV reverse transcriptase (Promega, Southampton, UK) according to manufacturers' instructions . Quantitative RT-PCR was carried out as previously described  assessing JNK1, JNK2 and CD45 sequences [11, 12, 15].
Protein concentrations were determined using the Bio-Rad Detergent Compatible (DC) protein assay kit (Biorad, Hertfordshire, UK) and quantified by Nanophotometer (Geneflow, Fradley, UK). Western blot analysis was performed using a method previously described . Protein expression in AT and/or AbdSc adipocytes utilized a rabbit polyclonal anti-JNK 1 and 2 phosphospecific antibody (286 μg/mL for both antibodies) to assess JNK expression (Biosource Nivelles, Belgium), as well as NFκB, IκBα, IκBα-phosphorylated form (IκBα-P) (1-2μg/mL, TCS Cellworks, Buckingham, UK), IKKβ and IKKγ (2 μg/mL, Abcam, Cambridge, UK), using mouse monoclonal antibodies. Equal protein loading was confirmed by densitometry using the α-tubulin antibody (2.04 μg/mL, Abcam, Cambridge, UK). A chemiluminescent detection system ECL/ECL+ (GE Healthcare, Little Chalfont, UK) enabled visualization after exposure to X-ray film.
AbdSc AT was digested with collagenase (2 mg/mL, Worthington Biochemical, Reading, USA), as previously described . Following isolation, mature AbdSc adipocytes were cultured in phenol red-free DMEM: F12 medium containing glucose (5 mM/L), penicillin (100 units/mL) and streptomycin (100 μg/mL), ready for treatment. A compacted aliquot of AbdSc mature adipocytes (200 μl; 5×105 per mL) were treated with recombinant human (rh) TNFα (Sigma, Poole, UK). Acute (2 h, 4 h) and chronic (48 h) effects of rh TNFα treatment (10, 50, and 100 ng/mL; Sigma, Poole, UK) on JNK and NFκB activity were assessed. AbdSc adipose cells were also treated with rh TNFα treatment (50 ng/mL) alone or in combination with the TNFα antagonist, WP9QY (1μM; Calbiochem, Nottingham, UK) for 2 h and 4 h. Adipocytes maintained in untreated media acted as controls. For inhibitor studies, AbdSc adipocytes were treated with NF-κB inhibitor (SN50, CalBiochem, Nottingham, UK) (50 μg/ml; 24 h) or a JNK inhibitor (SP600125, A.G. Scientific, Inc., San Diego, CA) (10 μM/mL); conditions were based on previous data [2, 3]. Viability of adipocytes was assessed as previously described . In brief, trypan blue (Sigma, Poole, UK) is used as part of a dye exclusion methodology for viable cell counting. The live viable cells do not take up the dye whereas nonviable cells take up the dye. Following treatment, the conditioned media and adipocytes were separated by centrifugation (360×g for 2 min). Conditioned media were removed, aliquoted and stored at −80°C. Protein from the adipocytes was extracted .
Conditioned media from the NFκB and JNK inhibitor studies were assessed using the human IL-6, TNFα and resistin ELISAs from R&D Systems (Abingdon, UK), intraassay variability <10%.
Protein extracted from human AbdSc AT, Om AT and AbdSc adipocytes were assayed using an ELISA-based colorimetric kit (Invitrogen, Paisley, UK); intraassay variability <10%.
AT and AbdSc adipocyte whole cell extract was obtained according to manufacturer's instructions (matched paired AT: Om and AbdSc lean: n=8, BMI: 22.3±2.1 kg/m2, Age: 43.0±6.6 years; Om and AbdSc Overweight and Obese: n=8, BMI: 31.4±3.6 kg/m2, age: 50.1±8.5 years; all female subjects). NFκB activation was expressed as percentages based on NFκB activity/total NFκB in AT. NFκB activity was assessed with Trans-AM NFκB p65 transcription factor assay kit (Active Motif, Rixensart, Belgium; detection limit: <40 ng of whole cell extract), as previously described [14, 16].
AbdSc adipocytes, human sternum, and mononuclear blood cells were incubated with a human JNK polyclonal primary antibody (1 μg/mL, Biosource UK, Nivelles, Belgium). All the human tissue slides were developed using peroxidase substrate kit (Vector Laboratories Ltd, Peterbrough, UK) and all slides were counter-stained with Mayer's hematoxylin.
For protein assessment, statistical analysis was undertaken using ANOVA for comparison of AT depots with post hoc analysis (Bonferroni). Statistical analysis was undertaken using a paired student t-test (ANOVA, SPSS 17.0, Woking, UK) for comparison of AT depots from T2DM versus ND subjects whilst matched for age, BMI and gender. Gene expression data were analyzed using an unpaired t-test (SPSS 17.0, Woking, UK). The threshold for significance was P<0.05. Data in the text and figures are presented as mean ± standard error of mean (SEM), unless otherwise stated.
NFκB, IKKβ, IKKγ, IκBα and phosphorylated IκBα (IκBα-P) protein expression were determined in Om AT and AbdSc depots. The level of NFκB protein expression was similar in both the Om AT and AbdSc depots from the obese subjects but significantly higher in the Om AT from the lean subjects in comparison with the lean AbdSc (Figure 1A). Protein expression of both IκBα and IκBα-P were significantly increased in Om compared with AbdSc AT (P<0.01) (Figure 1B), as were IKKβ and IKKγ protein expression (P<0.001, P<0.01, respectively, Figure 1C and 1D).
NFκB activity was altered by AT depot and adiposity (Figure 1E). In obese Om AT, NFκB activity (measured as NFκB activity/total NFκB expressed as a percentage) was increased in comparison with lean or obese AbdSc AT as well as with lean Om AT (P<0.05; Figure 1E). AbdSc AT NFκB activity was unaltered by adiposity (Figure 1E).
The effect of adiposity on NFκB, IκBα, IKKβ and IKKγ protein expression was also analyzed in both depots. Firstly, there was a significant reduction in NFκB protein expression with increasing adiposity (Figure 1A). Furthermore, IκBα-P expression was increased in both depots in the obese group compared with the lean subjects (Figure 1B); IKKβ was noted to increase in obesity in Om AT but no effect was observed in AbdSc AT (Figure 1C). However, IKKγ protein expression, alone, remained unaffected by increasing adiposity (Figure 1D). Om AT NFκB activity was higher in the obese subjects compared with lean Om AT (P<0.05). Additionally, total JNK protein expression, phosphorylated JNK1 and JNK2 (P-JNK1 and P-JNK2), appeared unaffected by increasing adiposity (data not shown).
Total JNK (JNK1&2 combined) expression was measured by the relative fold increase in expression of total JNK in AT depots. Om AT expressed the highest level of total JNK compared with the AbdSc AT showing a 1.51 fold increase when compared with AbdSc AT depot (P<0.05) (Figure 2A).
The relative fold increase in expression of P-JNK1 and P-JNK2 across the various AT depots was calculated in relation to AbdSc AT, which was taken as 1. AbdSc AT expressed the highest levels of P-JNK compared with Om AT (AbdSc versus Om JNK1: 1.73 fold increase, P<0.01; AbdSc Vs Om JNK2: 1.52 fold increase, P<0.01) (Figure 2B).
JNK1 and JNK2 gene expression were assessed in Om and AbdSc AT (ΔCt mean±SEM, Figure 2C) AT. There was no noted depot specific effect of JNK1 or JNK2 expression (Om AT lean: JNK1 ΔCt 11.06±0.16; JNK2: ΔCt 12.04±0.18; AbdSc AT lean JNK1 ΔCt 11.29±0.15, JNK2 ΔCt 11.81±0.18). Assessment of the effect of adiposity noted no differences in Om JNK expression (Om AT Obese: JNK 1 ΔCt 11.17±0.21; JNK2: ΔCt 12.13±0.18) whilst in AbdSc AT JNK1 expression between lean and obese subjects was significantly altered (AbdSc obese: JNK1 ΔCt 11.59±0.15, P<0.01) but did not reach significance for JNK2 expression (AbdSc obese: JNK2 ΔCt 12.05±0.2; P=NS).
CD45 represents a marker of macrophages and may represent an important site for JNK expression, thereby in order to ascertain that the effect of JNK expression was not due to macrophage content we assessed CD45 expression in the depots. However there was no significant difference in expression observed between AT depots analyzed (Om CD45 lean ΔCt 20.44±0.27; obese ΔCt 19.79±0.31; AbdSc ΔCt 20.04±0.52).
AbdSc AT expressed the highest ratio of P-JNK to total JNK expression (JNK activity/total JNK expression in AT; Figure 2D) in comparison to Om AT. Furthermore, the difference in relative activity was significant when comparing Om versus AbdSc AT (P<0.05, Figure 2D). We further examined the effect of adiposity in AbdSc and Om AT. No statistically significant difference was observed with increasing adiposity (Figure 2D).
Assessment of JNK activity in AbdSc AT from T2DM subjects compared with AbdSc AT from nondiabetic (ND) subjects showed a significant reduction due to T2DM status (JNK1: P<0.001, JNK2: P<0.001; Figure 3A). Assessment of NFκB activity determined that activity significantly increased in AbdSc AT from T2DM subjects compared with ND (P<0.001; Figure 3A).
Further analysis of P-JNK 1 and P-JNK 2 expression in AbdSc AT from T2DM compared with AbdSc AT from ND subjects showed, again, a significant reduction in JNK1 and JNK2 expression in T2DM (P<0.01, P<0.001 respectively, Figure 3B).
NFκB activity was altered by rh TNFα both over time and at a range of concentrations (10, 50 and 100 ng/mL rh TNFα; n=6) in the in vitro treated AbdSc adipocytes.
An increase in NFκB activity was detected at 2 h and 4 h post-treatment with rh TNFα 50 ng/mL concentration compared with control untreated cells, whilst at a concentration of 10 ng/mL rh TNFα, only a significant increase 4 h post-treatment was observed (P<0.01; Figure 4). JNK1 appeared to reduce % expression compared with control at 2 h (rh TNFα: 10 ng; P<0.05, P<0.01, respectively) and 4 h treatment (rh TNFα: 50 ng; P<0.01, P<0.01, respectively). With chronic treatment (48 h) NFκB activity remained elevated (P<0.05; Figure 4).
AbdSc adipocytes treated with rh TNFα led to a significant reduction in the P-JNK1 at both time points (P<0.05, P<0.01, n=4, Figure 5). Concurrently, AbdSc adipocytes treated with TNFα antagonist (AGT, 1 ng) and rh TNFα (50 ng/mL) increased JNK1 protein expression at 2 h and 4 h (P<0.01, Figure 5) whilst AGT alone led to a reduction in JNK1 protein expression (P<0.01, Figure 5).
AbdSc adipocytes treated with rh TNFα only altered P-JNK2 protein expression at 4 h 100 ng/mL treatment (Figure 5; 4 h: TNFα 100 ng/mL: 0.97±0.71 ODU*) whilst rh TNFα in combination with AGT led to an increase in P-JNK2 protein expression at 4 h (4 h TNFα 50 ng/mL/TNFα AGT 1 ng: 10.23±3.01 ODU↑*), whilst again AGT alone led to a reduction in JNK2 protein expression at 4 h (4 h TNFα AGT 1 ng: 2.12±0.62 ODU↓*; n=4, Figure 5).
AbdSc adipocytes treated with an NFκB inhibitor significantly reduced IL-6, TNFα and resistin, whilst no change was noted in JNK inhibitor treated AbdSc adipocytes (Table 1).
|Adipokine secretion||Control (untreated cells), mean ± SE||NFκB inhibitor, mean ± SE||JNK inhibitor, mean ± SE|
|TNFα (pg/ml)||184 ± 38||102 ± 31*||111 ± 21|
|IL-6 (pg/ml)||2912 ± 321||1781 ± 211**||2563 ± 311|
|Resistin (pg/ml)||124 ± 32||65 ± 24*||96 ± 17|
Positive expression of JNK was identified in human AbdSc adipocytes (Supplemental Figure 1).
Previous studies have implicated both NFκB and JNK inflammatory pathways as potential therapeutic molecular targets for the treatment of obesity mediated insulin resistance and T2DM [7, 8, 10, 17]. This particular study addressed two fundamental questions:  is one inflammatory pathway predominantly activated in AT from either obese individuals or people with T2DM, contributing to metabolic risk;  does one pathway appear to be preferentially activated in vitro in response to proinflammatory mediators, such as rh TNFα which is elevated in obesity and T2DM subjects [2, 20, 21].
From our ex vivo findings, AbdSc and Om AT from ND subjects appear to have distinct inflammatory mechanisms. Increased IκBα, IKKβ and IKKγ protein expression were apparent in Om AT compared with the AbdSc AT depot, which corresponded with increased NFκB activity in Om AT compared with AbdSc AT. With increasing adiposity, both IκBα and IKKβ were upregulated in AbdSc and Om AT, whilst NFκB protein expression decreased. This may potentially arise due to the rapid turnover of NFκB mediated by increased IκBα-P expression, which translocates NFκB to the nucleus. This rapid turnover of NFκB was supported by our NFκB activity data, which determined increased NFκB activity in Om AT taken from obese patients compared with lean subjects.
Our studies further addressed the activity of NFκB in AbdSc AT from T2DM subjects compared with age and BMI matched ND subjects. Interestingly, NFκB activity increased substantially in AbdSc fat taken from the T2DM subjects; this could arise due to the known systemic proinflammatory milieu in such subjects, leading to a continuing vicious cycle of inflammation from AbdSc AT. Such a finding develops our understanding of NFκB in T2DM, with several studies highlighting the importance of this inflammatory switch in other tissues, such as the heart, liver and peripheral mononuclear blood cells [22, 23], as well as the relevance of therapeutic targeting of NFκB activity [19, 24].
Concurrent to NFκB analysis, our studies investigated JNK1 and 2. Positive immuno-staining for JNK expression in AT was noted; mRNA and protein expression studies affirmed the presence of JNK1 and 2 in human ND AT, which appeared unrelated to macrophage presence.
Assessment of total JNK protein denoted increased expression in Om AT compared with the AbdSc AT. Additionally, both P-JNK1 and P-JNK2 forms were also increased in Om AT. In order to examine the association of total JNK and P-JNK activity more accurately, we evaluated the relative activity of JNK as a ratio (expressed as a percentage of P-JNK versus total JNK) within a comparable quantity of AT from each depot; a similar equation was utilized to assess NFκB activity. Of note was that JNK1 and JNK2 % activity were substantially less in the Om AT compared with AbdSc AT, in contrast to NFκB activity observed in the respective depots. Furthermore active JNK expression appeared unaffected by increasing adiposity, whilst reduced in AT taken from T2DM subjects - in contrast to NFκB activity, as well as previous murine studies . However, such findings may imply that NFκB is the more dominant pathway and is preferentially activated over JNK pathways or that JNK activation may occur under certain proinflammatory mediated circumstances, such as T2DM. Further to this, studies by Sourris et al. affirm JNK activation in AbdSc AT as a determinant of insulin resistance .
Both obesity and T2DM are associated with increased circulating TNFα levels. JNK has been shown to be activated by insulin and TNFα in animal models and differentiated human cultured preadipocytes [7, 11, 13, 27]. We therefore explored the direct action of rh TNFα on the P-JNK and NFκB activity. Specifically, we examined the effect of rh TNFα on P-JNK and NFκB activity in mature AbdSc adipocytes as a stimulus for the inflammatory pathways. Our own previous in vitro studies have shown that human AbdSc adipocytes respond to various influences, such as lipopolysaccharide, insulin, resistin and TNFα [2, 3, 28], with rh TNFα shown to stimulate other adipokines, such as angiotensin II .
In contrast to previous murine studies [7, 27, 30], our findings demonstrated that rh TNFα appeared to reduce JNK activation, which was reversed with the use of a TNFα antagonist (AGT). Additionally, rh TNFα was observed to mediate NFκB activation at 2, 4, and 48 h post-treatment. These findings suggest that, within human AbdSc adipocytes, TNFα may possess a dual influence on intracellular signaling, leading to a reduction in JNK activity, predominantly influencing JNK1 expression, whilst activating NFκB more effectively. As such, a reduction in JNK1 expression may occur through cross-talk between the JNK and IKK pathways, where both IKKα and IKKβ can phosphorylate the same site of IRS1 as JNK, whilst JNK activation may also be inhibited by NFκB target genes [31, 32]. Furthermore NFκB and JNK inhibitors were utilized to examine the influence on TNFα release. NFκB blocker treated cells reduced TNFα secretion and other proinflammatory cytokines. In contrast, AbdSc adipocytes treated with a JNK inhibitor did not affect any adipokine secretion, a similar finding being noted in cultured human monocytes . Taken together, these data suggest that in AbdSc AT, in T2DM subjects, TNFα may preferentially activate NFκB, which mediates a reduction in JNK activity, therefore NFκB acts as the dominant intracellular inflammatory pathway [4, 24]. However, JNK may still mediate an inflammatory response where NFκB target genes are less dominant [26, 27, 33].
In conclusion, these studies demonstrate distinct depot specific differences in AbdSc and Om AT with respect to expression and activity of NFκB and JNK. Furthermore these studies illustrate that JNK phophoshorylation does not appear to be central in AbdSc AT, for mediating inflammatory responses in the pathogenesis of obesity mediated T2DM, as previously identified in rodent models. As such, these studies, in conjunction with previous data, further strengthen the role for NFκB as a potential key target for therapeutic drug intervention in the management and prevention of obesity and T2DM in humans.
We acknowledge the British Heart Foundation for Dr. Alison Harte's Intermediate fellowship, research support from Birmingham Science City and The Rowlands Trust. We thank the Thai Government for Warunee Kumsaiyai PhD Sponsorship and the Saudi Arabian Government for Dara Al-Disi's PhD sponsorship. We also thank the surgeons and theatre staff at UHCW Hospital, Coventry, and Jane Starcynski for immunohistochemical support.