Increased Expression of the Innate Immune Receptor TLR10 in Obesity and Type-2 Diabetes: Association with ROS-Mediated Oxidative Stress

The pattern recognition receptors, such as TLR play a pivotal role in the recognition of pathogen- and danger-associated molecular patterns that leads to the induction and orchestration of innate immunity during the early stages of infection and also triggers the adaptive immunity or sometimes immunopathology at the later stages of infections. TLR signaling activates a cascade of adaptor proteins, protein kinases and effector transcription factors such as NF-κB and interferon regulatory factors (IRFs) to result in the expression of type-I interferons, tumor necrosis factor (TNF)-α, and interleukin (IL)-6 as determinants of the balance between beneficial host innate immune responses and immunopathology [1]. In mammals, 13 TLRs have been identified which are known to interact with structurally different ligands through the distinct extracellular domains whereas in humans, 10 different TLRs and their cognate ligands have so far been identified. TLR10 ligand(s) still remain(s) unknown [2]. The human TLR10 displays all the structural features that are characteristic of other members of human TLR family such as multiple leucine-rich repeats and Toll/IL-1 receptor (TIR) domain; TLR10 shares the highest nucleotide and amino acid sequence homology with TLR1 (50%) and TLR6 (49%) [3]. Upon activation, TLR10 forms a homo-or a hetero-dimer with TLR1 or TLR2 and transmits intracellular signal by recruiting myeloid differentiation factor (MyD)-88 to the activated receptor complex [4]. Genetic polymorphisms in the human TLR10 have been linked to various diseases such as asthma, bladder and nasopharyngeal carcinomas, and Crohn’s disease which indicates a functional role of TLR10 in these inflammatory diseases [5-8] as well as in the innate immune response to influenza virus infection [1].

In obesity and/or T2D, low-grade chronic inflammation, called metabolic inflammation, plays a key role in insulin resistance and development of metabolic syndrome [9]. Certain TLRs are known to be involved in metabolic inflammation, especially, the TLR4 has emerged as a metabolic sensor of lipopolysaccharide (LPS) as well as saturated free fatty acids (sFFAs) that are abundantly found in obese/T2D individuals [10]. Modulations in the expression of TLR1, TLR2, TLR4, TLR5, TLR6, TLR7, TLR8, and TLR9 have been reported in obesity and T2D [11-14]. Importantly, metabolic disease associated changes in the adipose tissue expression of TLR10 have remained unclear. In this study, we assessed the changes in the adipose tissue expression of TLR10 in individuals with or without T2D as well as the mechanism of TLR10 induction in monocytic cells. Herein, we present the data showing that oxidative stress leads to increased TLR10 expression which may have significance as an inflammatory marker in metabolic diseases such as obesity and T2D.

A total of 26 individuals, 13 non-diabetic (aged 26-71 years) and 13 individuals with T2D (aged 41-62 years), were recruited in the study through outpatient clinics of Dasman Diabetes Institute, Kuwait. Those of age <18yrs or with serious lung, kidney, liver, or cardiovascular disease, hematologic or immune disorders, pregnancy, type-1 diabetes, or malignancy were excluded from the study. Based on BMI, the participants were classified as lean (19 to 25 kg/m²), overweight (>25 to 30 kg/m²), and obese (>30 to 39 kg/m²). Co-morbid conditions included hypertension (two non-diabetic and six T2D patients) and hyperlipidemia (one non-diabetic and four T2D patients). The characteristics of the study participants are summarized in Table 1. All individuals gave a prior written informed consent and the study was approved by ethics committee of Dasman Diabetes Institute, Kuwait.

Anthropometric and physical parameters included body weight, height, waist circumference and systolic/diastolic blood pressure. Height and weight were measured using calibrated portable electronic weighing scales and portable inflexible height measuring bars with barefoot participants wearing light indoor clothing and waist circumference at the highest point of the iliac crest and the mid-axillary line was measured using constant tension tape at the end of a normal expiration with arms relaxed at sides. Whole body composition including body fat percentage, soft lean mass and total body water were measured using IOI353 Body Composition Analyzer (Jawon Medical, South Korea). Blood pressure was measured by using Omron HEM-907XL digital automatic sphygmomanometer (Omron Healthcare Inc. IL, USA). An average of 3 blood pressure readings, 5-10 min rest between each, was obtained. BMI was calculated using standard formula: body weight (Kg)/height (m²). Peripheral blood was collected by phlebotomist from the above-mentioned study participants following overnight (minimum 10hr) fasting and the samples were analyzed for fasting plasma glucose, glycated hemoglobin (HbA1c), fasting serum insulin, and serum lipid profile. Glucose and lipid profiles were measured by Siemens dimension RXL chemistry analyzer (Diamond Diagnostics, Holliston, MA, USA). HbA1c was measured by using Variant^TM device (BioRad, Hercules, CA, USA). Blood was centrifuged at 1200×g for 10 min and plasma was collected, aliquoted and stored at -80°C until use. Plasma triglycerides were measured by using commercial kit (Intra-assay CV% = 0.93; Inter-assay CV% = 3.05) (Chema Diagnostica, Monsano, Italy). All assays were carried out following manufacturers’ instructions.

Human adipose tissue samples (∼0.5g) were collected via abdominal subcutaneous fat pad biopsy lateral to the umbilicus using standard surgical method as previously described [15]. Briefly, the periumbilical area was sterilized by alcohol swabbing and then locally anesthetized using 2% lidocaine (2mL). Through a small superficial skin incision (0.5cm), the fat tissue sample was collected and further incised into smaller pieces, rinsed in cold phosphate buffered saline (PBS), fixed in 4% paraformaldehyde for 24hr and embedded in paraffin for further use. At the same time, freshly collected adipose tissue samples (∼50-100 mg) were preserved in RNAlater and stored at -80°C until use.

For in vitro studies, peripheral blood (30mL, EDTA tubes) was collected at phlebotomy unit in our institute for the purpose of monocyte purification as required following consent of participating adult healthy donors from laboratories support staff. The blood samples were diluted 1: 2 with sterile physiological saline (pH 7.1) and PBMC were isolated using Ficoll-Hypaque density gradient method as described [16]. Primary monocytes were purified by negative selection and following the manufacturer’s instructions (Human monocyte isolation kit II, MACS Miltenyi Biotec GmbH, Germany). Briefly, 10⁷ cells were resuspended in 30µL of PBS (pH 7.2) containing 0.5% bovine serum albumin (BSA) and 2mM EDTA, and 10µL of FcR blocking reagent. After mixing, 10µL of Biotin-Ab cocktail (anti-CD3, -CD7, -CD16, -CD19, -CD56, -CD123, and Glycophorin A antibodies) was added to label and remove T/B lymphocytes, NK cells, dendritic cells, and basophils. Following 10 min incubation on wet ice, 30µL of sample buffer and 20µL of anti-biotin antibodies conjugated MicroBeads were added, mixed and incubated on ice for 15 min. After washing, cell suspension (500µL) was applied to MACS separator column and effluent containing enriched monocytes was collected. The purified monocyte sample was fluorescently labeled with FITC-conjugated anti-CD14 and PE-conjugated anti-biotin antibodies. The purity was determined by flow cytometry which was found to be >90% for each enrichment (data not shown). For treatments, THP-1 cells, PBMC, or primary monocytes (1×10⁶ cells per well in 12-well plates) cultured in RPMI-1640 medium supplemented with 2mM L-glutamine, 10% fetal bovine serum (FBS), 100U/mL penicillin, and 100µg/mL streptomycin were initially incubated for 2hr for conditioning, then treated with H₂O₂ (9.8M, Cat. No. 822287.1000, Merck, USA) at a concentration of 20mM for 7hr, unless otherwise stated, in a humidified incubator with 5% CO₂. Cell viability was determined by trypan blue dye exclusion test and only the highly viable cells (>85% viability) were selected for use. In assays using inhibitors of signaling pathways, ROS, or anti-oxidants, THP-1 cells (1×10⁶ cells per mL per well in 12-well plates) were incubated first with selected inhibitors as recommended by manufacturers, then the cells were incubated with H₂O₂ as described above and cell pellets were lysed in RLT buffer (350µL) and stored at -80°C until use.

Total RNA was purified from adipose tissue samples using RNeasy kit and following the manufacturer’s instructions (Qiagen, Valencia, CA; USA) as described elsewhere [15]. The quantity of isolated RNA was determined by using Epoch^TM Spectrophotometer System (BioTek, Winooski, USA) and quality was assessed by formaldehyde-agarose gel electrophoresis. RNA samples (1µg each) were reverse transcribed into cDNA by using random hexamer primers and TaqMan reverse transcription reagents (High Capacity cDNA Reverse Transcription Kit; Applied Biosystems, CA, USA). For real-time RT-PCR, cDNA (50ng) was amplified using TaqMan® Gene Expression MasterMix (Applied Biosystems, CA, USA) and target gene-specific 20× TaqMan Gene Expression Assays (Applied Biosystems, CA, USA) containing forward and reverse primers (Table 2) and a target-specific TaqMan® minor groove binder (MGB) probe labeled with 6-fluorescein amidite (FAM) dye at 5’ end and non-fluorescent quencher (NFQ)-MGB at 3’ end of the probe, with 40 cycles of PCR amplification using 7500 Fast Real-Time PCR System (Applied Biosystems, CA, USA). Each cycle comprised of denaturation (95°C for 15 sec), annealing/extension (60°C for 1 min) following uracil DNA glycosylase activation (50°C for 2 min), and AmpliTaq Gold enzyme activation (95°C for 10 min). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression was used as internal control to normalize the differences in individual samples as compared with control (lean adipose tissue). The relative mRNA expression of the target gene was calculated by using 2^-ΔΔCt method and expressed as fold change (mean±SEM) over average GAPDH expression taken as 1. For determining TLR10 gene expression in THP-1 monocytic cells, PBMC or primary monocytes, total RNA was extracted using RNeasy Mini Kit and following the manufacturer’s instructions (Cat. No. 74106, Qiagen, USA). RNA samples (1µg each) were reverse transcribed into cDNA using random hexamer primers, TaqMan reverse transcription reagents and following the manufacturer’s instructions (High Capacity cDNA Reverse Transcription Kit; Cat. No. 4368814; Qiagen, USA). For real-time RT-PCR, 2µL of cDNA template was added to master mix containing TaqMan master mix (10µL, Cat. No. 4369016, Applied Biosystems), TLR10 primer (1µL, Hs_01935337), GAPDH primer (1µL, Cat. No. 4310884E, Applied Biosystems, USA) and nuclease-free water (6µL). Amplification was performed as follows: 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min (ABI 7500 thermal cycler system). The relative TLR10 mRNA expression was calculated as described above

Paraffin-embedded adipose tissue sections (4µm) were processed for immunohistochemistry as described [15]. The slides were incubated at room temperature overnight with primary antibody (1: 600 diluted rabbit anti-human TLR10 polyclonal Ab, Abcam® ab115598). After washing with PBS-Tween, slides were incubated for 1hr with secondary antibody (1: 200 diluted goat anti-rabbit conjugated with Alex Fluor 594®, Abcam® ab150088) and color was developed using 3, 3ʹ-diaminobenzidine (DAB) chromogenic substrate. Specimens were washed, counterstained, dehydrated, cleared, and mounted as described elsewhere [14]. For data analysis, digital photomicrographs of adipose tissue sections (20×; Olympus BX51 Microscope, Japan) were used to quantify staining in three different regions. The regions were delineated using ImageScope software (Aperio Vista, CA, USA). Aperio-positive pixel count algorithm version 9 was used to quantify the staining intensity. The number of positive pixels was normalized to the number of total pixels (positive and negative), color and intensity thresholds were set to detect the immunostaining as positive and the background as negative pixels. All samples were then analyzed using the same parameters. The resulting color markup of analysis was confirmed for each slide.

Formalin-fixed and paraffin-embedded sections of subcutaneous adipose tissue (8µm thick) were processed for immunofluorescent labeling using similar protocol as described earlier for IHC. Following antigen retrieval and blocking, samples were incubated overnight at room temperature with primary antibody (1: 400 diluted rabbit anti-human TLR10 Ab, abcam® ab115598). After washing thrice with PBS-Tween, slides were incubated for 1hr with secondary antibody (1: 400 diluted goat anti-rabbit Ab conjugated with Alexa Fluor® 594 (orange color); Abcam® ab150088) and washed thrice. Samples were counterstained with 4’,6-diamidino-2-phenylindole (DAPI, blue color) (Vectashield, Vector Laboratories, H1500) and mounted. Confocal images were obtained using inverted Zeiss LSM710 spectral confocal microscope (Carl Zeiss, Gottingen, Germany) and EC Plan-Neofluar 40×/1.30 oil DIC M27 objective lens. After sample excitation with 543nm HeNe laser and 405nm line of an argon ion laser, optimized emission detection bandwidths were configured by using Zeiss Zen 2010 control software. For co-localization staining, formalin-fixed paraffin-embedded adipose tissue samples were processed for immunofluorescent staining using similar protocol as described before. Samples were incubated for 2hr with 1: 400 diluted mouse anti-human adiponectin antibody (Abcam® ab22554), washed thrice and incubated overnight at room temperature with 1: 400 diluted rabbit anti-human TLR10 primary antibody (Abcam® ab115598). After three washes, samples were incubated for 1hr with 1: 400 diluted Alexa Fluor 647-cojugated goat anti-mouse secondary antibody (Abcam® ab150115, red color). Samples were washed thrice and incubated for 1hr with 1: 400 diluted Alexa Fluor 488-conjugated goat anti-rabbit secondary antibody (Abcam® ab150077, green color). After washing, samples were counterstained with DAPI (Vectashield, Vector Laboratories, H1500, blue color) and mounted. Confocal images were obtained using inverted Zeiss LSM710 spectral confocal microscope (Carl Zeiss, Gottingen, Germany) and EC Plan-Neofluar 40×/1.30 oil DIC M27 objective lens. After sample excitation with 643nm HeNe laser and 405nm line of an argon ion laser, optimized emission detection bandwidths were configured using Zeiss Zen 2010 software.

For immunofluorescent staining of THP-1 cells, 1×10⁶ cells were washed 3× with PBS-Tween and coated on slides by cytospin (500rpm for 3 min). After fixation (4% formaldehyde) and 3 washes in cold PBS, cells were permeabilized with 0.1% triton X-100 and washed thrice. Specimens were blocked by using 1% BSA for 1hr and incubated overnight at room temperature with 1: 200 diluted rabbit anti-human TLR10 primary antibody (Abcam® ab115598). Cells were washed 3× with PBS-Tween and incubated with 1: 200 diluted Alexa Fluor®594-conjugated goat anti-rabbit secondary antibody (Abcam® ab150088; orange color) and incubated at room temperature for 1hr. After three PBS washes, cells were counterstained and mounted as described earlier. For PBMC, 1×106 cells were washed, coated, fixed, permeabilized, and blocked as previously described. Later, cells were stained with 1: 200 diluted rabbit anti-human TLR10 primary antibody (Abcam® ab115598) and 1: 400 diluted Alexa Fluor®488-conjugated goat anti-rabbit secondary antibody (Abcam® ab150077; green color). Cells were washed and counterstained as stated before. Confocal images were obtained using Plan-Apochromat 63×/1.40 oil DIC M27 objective lens (Zeiss LSM710 AxioObserver inverted confocal microscope, Carl Zeiss, Gottingen, Germany). Sample excitation and imaging data analysis were performed as described above.

THP-1 monocytic cells in triplicate wells of 12-well plates (1×10⁶ cells/mL per well) cultured in RPMI-1640 complete medium were treated with H₂O₂ (10mM) and incubated for 7h in a humidified incubator (5% CO₂) and mock-treated cells served as experimental control. Harvested cells were incubated for 30 min with cell lysis buffer containing Tris 62.5 mM (pH 7.5), 1% Triton X-100, and 10% glycerol. The lysates were clarified by centrifugation at 14000rpm for 10 min and supernatants were collected. Protein concentration was measured using Quickstart Bradford Dye Reagent, 1× Protein Assay kit (Bio-Rad Laboratories, Inc., CA., USA). Samples (20 µg each) were mixed with loading buffer, heated at 95°C for 5 min and resolved by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Cellular proteins were transferred to Immuno-Blot polyvinylidene fluoride (PVDF) membrane (Bio-Rad Laboratories, USA) by electro blotting. The membranes were blocked with 5% non-fat milk in PBS for 1h, followed by overnight incubation with 1: 1000 diluted primary antibodies (p-c-Jun, pERK1/2, p-p38, p-NF-κB, and β-actin). All primary antibodies were purchased from Cell Signaling (Cell Signaling Technology, Inc., USA). Blots were washed four times with Tris-buffered saline (TBS) and incubated for 2h with horse radish peroxidase (HRP)-conjugated secondary antibody (Promega, Madison, WI., USA). Immunoreactive bands were developed using Amersham ECLPlus Western Blotting Detection System (GE Health Care, Buckinghamshire, UK) and visualized by Molecular Imager® VersaDoc^TM MP Imaging Systems (Bio-Rad Laboratories, Hercules, CA., USA). Band intensities were measured using VersaDoc^TM QuantityOne software (Bio-Rad Laboratories, Hercules, CA, USA) and β-actin served as loading control. The ratio of phosphorylated protein over loading control protein was calculated to normalize the differences in samples loading to gel. Phosphorylation fold difference in treated sample was calculated by comparing with phosphorylation level of control sample taken as 1.

For determining TLR10 protein expression, THP-1 cells or PBMC were cultured at a cell density of 1×10⁶ cells per mL per well in 12-well plates, treated with H₂O₂ (10mM concentration) in duplicate wells and incubated overnight at 37°C while control wells were mock-treated with the vehicle only. Cells were pelleted (400×g for 10 min at 4°C), resuspended in PBS containing 2% BSA and 0.2% sodium azide and washed twice. To 30µL of cell suspension in staining tubes, 5µL of PE-conjugated anti-human TLR10 antibody (anti-CD290 mAb, Cat. No. 12-2909-42, eBioscience Thermo Fisher Scientific, USA) was added while 5µL of PE-conjugated mouse IgG1 K-isotype antibody was added to control or mock tubes and incubated on wet ice in dark for 30 min. After three washes, cells were resuspended in 500µL of wash/staining buffer and flow cytometry for TLR10 expression was performed using FACSCanto II^TM flow cytometer (BD Biosciences, USA) and data were analyzed using BD FACSDIVA software (version 6.1.3).

The intracellular SOD activity against ROS induction following H₂O₂ treatment was measured using commercial SOD assay kit and following the manufacturer’s instructions (Cat. No. 706002, Cayman Chemical Co, Michigan, USA). Briefly, THP-1 cells were cultured in RPMI-1640 complete medium (1×10⁶ cells per mL per well) in 12-well plates. Experimental wells in duplicate were treated with 20mM H₂O₂ and incubated at 37°C in 5% CO₂ for 0 min, 15 min, 30 min, 1hr, 2hr, 4hr, and 7hr while controls were treated only with vehicle. Samples were collected and cell pellets were sonicated in cold 20mM HEPES buffer (Cat. No. H0887, Sigma Aldrich; pH7.2) containing 1mM EGTA (Cat. No. E4378-100G, Sigma Aldrich, USA), 210mM mannitol (Cat. No. M4125-100G, Sigma Aldrich), and 70mM sucrose (Cat. No. 50389-500G, Sigma Aldrich). Sonicated samples were centrifuged at 1500×g for 5 min at 4°C and supernatants were collected. For SOD assay, 200µL of diluted radical detector and 10µL of sample were added to designated sample wells. Similarly, 200µL of diluted radical detector and 10µL of standard were added to designated standard wells. A total of 7 standards were established in duplicate representing 0, 0.025, 0.05, 0.10, 0.15, 0.20, and 0.25 U/mL SOD activity. The reaction was initiated by adding 20µL of diluted xanthine oxidase to all wells, mixed on a plate shaker for a few seconds and then incubated at 37°C for 20 min. Absorbance was measured at 450nm (BioTek spectrophotometer, Elx808) and the average absorbance for each sample was calculated. Linearized rates (LR) of all standards were calculated as follows: LR of standard A = Absorbance of standard A/Absorbance of standard A; LR of standard B = Absorbance of standard A/Absorbance of standard B; and LR of standard C = Absorbance of standard A/Absorbance of standard C, etc. SOD activity was calculated as described below:

SOD activity (U/mL) = [(Sample LR-Y intercept/slope)×0.23/0.01]×Sample dilution factor

The data were expressed as mean±SEM values. Group means of TLR10 gene expression data were compared using Student’s t-test and the linear dependence between two variables was assessed by Pearson’s correlation coefficient ‘r’ values. Mann-Whitney t-test was used to compare means regarding lean, overweight and obese adipose tissue samples. GraphPad Prism software (version 6.05; San Diego, CA, USA) was used for statistical analysis and graphical representation of data. All P-values ≤0.05 were considered statistically significant.

We sought to determine the expression of TLR10 in the adipose tissue samples from lean, overweight and obese individuals, with or without T2D. The data show that TLR10 gene expression in the fat tissue was significantly elevated in obese individuals without (P=0.026; Fig. 1A) or with T2D (P=0.006; Fig. 1B) as compared with respective lean controls. Notably, this increase in the gene expression of TLR10 was found to be associated positively with clinical marker of corpulence BMI both in non-diabetic (r=0.81 P=0.001; Fig. 1C) and diabetic individuals (r=0.78 P=0.002; Fig. 1D). We next wanted to know if the upregulated TLR10 gene expression in obesity/T2D was also manifested at the protein level. To this end, our data show that the adipose tissue TLR10 protein expression in obesity, as determined by IHC, was remarkably elevated in both non-diabetic (Fig. 2A) and T2D individuals (Fig. 2B) as compared with respective controls. This finding was further validated by confocal microscopy which also confirms the elevated TLR10 protein expression in obese adipose tissue samples from non-diabetic (Fig. 3A) as well as T2D individuals (Fig. 3B). The IHC data representing TLR10 protein expression were further quantified as arbitrary units (AU) using Aperio software which indicated that compared with lean controls, TLR10 protein expression was significantly higher in non-diabetic (P=0.001; Fig. 3C) and diabetic (P<0.001; Fig. 3D) obese individuals. Overall, we found a strong positive correlation between gene and protein expression of TLR10 (r=0.90 P<0.0001; Fig. 3E). In addition, colocalization staining depicts the nuclei (DAPI), adiponectin (Alexa Fluor 647), TLR10 (Alexa Fluor 488), and merged confocal microscopy images of non-diabetic (Fig. 4) and T2D individuals (Fig. 5).

To see whether the increased TLR10 gene expression in the adipose tissue in obesity/T2D was associated with other inflammatory markers and TLRs’ expression (as many TLRs are known to have immunometabolic functions), we determined the gene expression of TNF-α, IL-1β, IL-6, IL-10, IL-12A, IL-18, IL-23A, CCL2, CCR2, CCL5, CCR5, CXCL10, TLR2, TLR3, TLR4, TLR7, TLR8, TLR9, TLR10, NFAT5, CLEC7A/Dectin-1, and FGL2 in the adipose tissue samples. As summarized in Table 3, TLR10 gene expression correlated positively (P<0.05) with that of TNF- α, IL-12A, IL-23A, CCR2, CCL5, CCR5, CXCL10, TLR2, and NFAT5 in non-diabetic individuals whereas it correlated with IL-1β, IL-23A, and CCL-5 in T2D individuals. TLR10 expression was found to be associated with IL-23A and CCL-5 in both groups.

Obesity is known to induce immunometabolic changes in the adipose tissue and the early changes observed are the tissue hypoxia and oxidative stress from lack of sufficient blood supply to adipose tissue during obesity. TLR10 protein, as revealed by IHC, was expressed predominantly on immune cells that were present in the interstitial spaces around adipocytes, also called crown-like structures (CLS). We, therefore, asked if the oxidative stress was involved in induction or upregulation of TLR10 expression in immune cells. To this effect, using an in vitro experimental model in which various immune cells were treated with H₂O₂ to induce ROS-mediated oxidative stress, we show that TLR10 gene expression is significantly upregulated in H₂O₂-treated THP-1 human monocytic cells (Fig. 6A), PBMC (Fig. 6B), and primary human monocytes (Fig. 6C) compared to mock-treated controls (P<0.001). TLR10 gene expression data were further validated by confocal microscopy and flow cytometry. The confocal images show more intense staining for TLR10 protein expression following H₂O₂-treatment of THP-1 monocytic cells (Fig. 7A) and PBMC (Fig. 7B). Notably, higher H₂O₂ concentration (20mM; right hand panel) induced stronger TLR10 expression than did lower concentration (10mM; middle panel). Similarly, flow cytometry also confirmed that H₂O₂ treatment resulted in enhanced TLR10 protein expression on THP-1 monocytic cells (Fig. 7C) and PBMC (Fig. 7D) compared to vehicle-treated controls (P<0.001).

We further wanted to know if H₂O₂ treatment of human monocytic cells generated ROS as an underlying mechanism of oxidative stress. In this regard, we measured SOD activity in these cells at 15 min, 30 min, 1hr, 2hr, 4hr, and 7hr post-treatment as a counter mechanism to intracellular ROS generation in response to H₂O₂ exposure. These data show a significant increase in SOD activity in a time-dependent manner in cells treated with H₂O₂ compared to mock-treated controls (Fig. 8).

We speculated that if the expression of TLR10 in monocytic cells was induced by ROS-mediated oxidative stress, it should be suppressed by treatment with antioxidants or ROS scavengers. In this regard, our data show that TLR10 gene expression was significantly abrogated in THP-1 cells (P<0.001) when treated with N-acetyl cysteine (NAC), apocyanin, curcumin, Nordihydroguaiaretic acid (NDGA), resveratrol, and trolox before exposure to H₂O₂ compared to controls that were treated with H₂O₂ only (Fig. 9).

We next wanted to know which signaling pathways were involved in TLR10 upregulation in cells challenged with ROS induction by H₂O₂ treatment. To this effect, we found that H₂O₂ treatment of THP-1 cells led to phosphorylation of transcriptional factors such as c-Jun, NF-κB, MAPK, and extracellular signal-regulated kinases (ERK)-1/2 (Fig. 10A). These data suggest that NF-κB and MAPK/AP-1 signaling pathways might be involved in ROS-mediated upregulation of TLR10 in THP-1 cells. As expected, TLR10 expression was significantly suppressed when cells were preincubated with inhibitors of NF-κB (Bay 11-7085) and MAPK/AP-1 (PD98059, SB203580, SP600125, U0126 & XMD-8-92) pathways compared to mock-treated controls (all P<0.05). However, no significant suppression of TLR10 was observed in cells treated with Triptolide which is a bioactive diterpene triepoxide that acts as an inhibitor of NF-κB mediated signaling (Fig. 10B).

Redox environment in the ER regulates ROS levels. We asked if ROS-mediated TLR10 upregulation in human monocytic cells involved the ER stress response. To this end, our data show that H₂O₂ treatment of THP-1 cells promoted expression of the ER stress responsive transcription factors or markers such as endoplasmic reticulum to nucleus signaling (ERN)-1 (P<0.001), activating transcription factor 6 (ATF)-6 (P=0.002), L-xylulose reductase (LXR)-1 (P=0.002), and C/EBP homologous protein (CHOP) (P=0.001) compared to the expression in mock-treated controls (Fig. 11).

Chronically elevated systemic levels of sFFAs have been linked to insulin resistance. Increased palmitate levels were associated with mitochondrial dysfunction, apoptosis and insulin resistance in L6 skeletal muscle cells [17]. We, therefore, asked if THP-1 cell co-treatment with palmitate and H₂O₂ could have a synergistic effect in upregulating the TLR10 expression in these cells. To this end, our data show that palmitate and H₂O₂ treatment leads to synergistic induction of TLR10 expression in monocytic cells compared with the controls treated with either palmitate or H₂O₂ (P<0.001) whereas, exposure to endotoxin (LPS) and H₂O₂ also led to increased TLR10 expression, albeit to a lesser extent than palmitate and H₂O₂ (P<0.01). On the other hand, cells treated with an unsaturated fatty acid oleate or satiety hormone leptin did not show a synergistic upregulation of TLR10 in concert with H₂O₂-induced oxidative stress (Fig. 12).

We next asked if the ROS induction was related to expression of inflammatory markers. To this end, as expected, THP-1 cell treatment with H₂O₂ led to upregulated expression of inflammatory cytokines/chemokine as compared to controls. Moreover, a synergistic increase in inflammatory markers was observed when THP-1 cells were co-treated with palmitate and H₂O₂. The data show the upregulated expression of CCL2/MCP-1 (Fig. 13A), TNF-α (Fig. 13B), IL-6 (Fig. 13C), IFN-γ (Fig. 13D), and immunoregulatory cytokine IL-10 (Fig. 13E). Notably, monocytic cells with enhanced inflammatory cytokine gene expression also displayed elevated TLR10 expression (Fig. 13F).

This study reports, for the first time to our knowledge, that the expression of TLR10 in the adipose tissue is directly modulated by advancement of obesity in humans, with or without T2D. TLR10 is a surface innate immune receptor and its modulation in obesity/T2D and the underlying mechanism(s) have remained unknown. In this regard, first, our data show that the expression of TLR10 is significantly upregulated at both gene and protein levels in the adipose tissue samples from obese individuals, with or without T2D, as compared to lean counterparts; and also that these changes were directly associated with BMI. These data indicate that obesity is a positive modulator of TLR10 expression in the adipose tissue. In agreement with this argument, it was shown that the visceral adipose tissue expression of several TLRs (TLR1, TLR4, TLR5, TLR8, TLR9, & TLR12) and downstream signaling molecules (MyD88, IRFs, NF-κB & STAT-1) was higher in diet-induced obese mice than leptin-deficient ob/ob mice [18]. Also in humans, increased expression of TLR2, TLR3, TLR4, TLR5, TLR7, and TLR8 was reported in obesity/T2D [14, 19, 20]. Among all TLRs, little is known about ligand(s) and function of TLR10. Obesity is associated with low-grade chronic inflammation in the adipose tissue. In obesity, activation of the innate immune system and elevated adipose tissue expression of TLRs contribute to a systemic acute phase response with persistent inflammation and oxidative stress [21]. Adipose tissue is a triglyceride storage and endocrine organ that plays a vital role in the energy homeostasis. This tissue is composed of adipocytes and other cells such as pre-adipocytes, fibroblasts, endothelial cells, and immune cells which secrete proteins such as cytokines, chemokines, adipokines, and hormones [22]. In patho-physiological conditions, bioactive adipokines (leptin, adiponectin, visfatin, apelin, resistin, and plasminogen activator inhibitor type 1) and proinflammatory cytokines (TNF-α, IL-1β, and IL-6) secreted by the activated monocytes/macrophages play a role to generate free radicals, promote ROS synthesis, and induce oxidative stress [23]. ROS enhances expression of the proinflammatory cytokines which allows oxidative stress cycle to continue during progression of obesity and T2D. Our confocal microscopy and IHC data represent the enhanced TLR10 expression in immune cells, most likely macrophages that form CLS around adipocytes. This observation prompted us to ask if the oxidative stress from ROS generation could induce TLR10 expression in monocytic cells. The in vitro data obtained from real-time RT-PCR, confocal microscopy and flow cytometry show that ROS generation by H₂O₂ treatment triggered the TLR10 expression in immune cells such as THP-1 cells, PBMC, and primary human monocytes. These findings are in agreement, at least in part, with previous study showing increased TLR10 expression in H₂O₂-treated THP-1 cells [24]. In the present study, we also show that ROS generation by H₂O₂ treatment leads to the upregulated TLR10 expression in PBMC as well as primary human monocytes. Treatment of cells with H₂O₂ for examining the effect of ROS on TLR10 expression seems to be relevant since H₂O₂ is the most stable and long-lasting component of ROS [25]; and it also has the properties of an inter- and intra-cellular messenger [26]. To see whether the intracellular ROS generated by H₂O₂ treatment was able to modulate the anti-oxidant activity within monocytic cells, we treated THP-1 cells with H₂O₂ for different time intervals and measured SOD activity. These data indicate that SOD activity in monocytic cells increased as a function of H₂O₂ exposure time. SOD is the major component of intrinsic anti-oxidant defense mechanism for the dismutation of highly toxic superoxide radicals. During the oxidative stress, this defense mechanism is activated to minimize lipid peroxidation through the anti-oxidant enzymes such as glutathione peroxidase and SOD [27]. In principle, SOD activity is expected to remain active until the necessary co-factors are available i.e. up until last stages of cell survival in the event of persistent oxidative stress. The overproduction of ROS incapacitates the intrinsic antioxidant mechanism in cells which eventually results in oxidative stress. As expected, the ROS-mediated TLR10 expression was abrogated when monocytic cells were treated with ROS scavengers or antioxidants before H₂O₂ treatment. This clearly indicates that by counteracting ROS and alleviating the oxidative stress, TLR10 upregulation can be largely abrogated in monocytic cells. Thus, the data from our and previous study support that oxidative stress can lead to TLR10 upregulation in monocytic cells [24].

Next, we show that the ROS induction in monocytic cells leads to phosphorylation of transcription factors e.g. c-Jun (AP-1), NF-κB, p38 MAPK, and ERK-1/2 which implies that associated signaling pathways are involved in the ROS-mediated TLR10 upregulation in human monocytic cells. As expected, THP-1 cell pretreatment with inhibitors of IκB_α (Bay 11-7085), p38 MAPK (SB203580), MAPKK (PD98059), JNK/MAPK (Sp600125), MEK-1/2 or AP-1 (U0126), and ERK5/BMK1 (XMD-8-92) significantly suppressed TLR10 expression in these cells compared to mock-treated controls whereas, treatment with Triptolide, the NF-κB/NF-AT inhibitor, did not gain level of statistical significance in causing TLR10 suppression. Our findings are supported, at least in part, by various studies showing activation of NF-κB, NF-AT, HIF-1, and AP-1 pathways after ROS induction by hypoxia. These biofactors, especially the NF-κB and AP-1, are the key transcription factors that regulate expression of genes that are involved in dynamic pathobiological processes such as inflammation, lymphoid differentiation, cellular stress, apoptosis, embryonic development, and oncogenesis [28, 29]. Although, NF-κB and AP-1 transcription factors are regulated by different mechanisms, several studies have shown that they can be activated simultaneously by same stimuli [30-32]. In agreement with our findings, a previous study showed that oxygen radicals or oxidative stress triggered the activation of both NF-κB and AP-1 transcription factors [33, 34]. Activation of the JNK by cellular stress or inflammation involves the nuclear translocation of NF-κB whereas, the expression of many genes under these conditions requires the simultaneous activation of NF-κB and AP-1, suggesting that these transcription factors may work cooperatively or even modulate the activity of each other [35]. This leads to the possibility that the cellular response to stimuli such as oxidative stress may rather be induced cooperatively by a network of simultaneously-activated transcription factors. Our data showing ROS-linked activation of p38 MAPK, ERK-1/2, and JNK are also corroborated by several studies [36-38].

We further show that ROS-associated TLR10 upregulation in human monocytic cells involves the ER stress response as inferred from the elevated expression of ER stress markers e.g. ERN1, ATF6, CHOP, and LXR1 in H₂O₂-treated cells compared to controls. Although, both oxidative stress and ER stress are found to coexist in many pathologic conditions, it is unclear whether and how these two stresses interact. In eukaryotic cells, the production of ROS as a byproduct during oxidative protein folding, mitochondrial respiration and detoxification processes has been linked to ER stress and unfolded protein response (UPR) [39]. Thus, ROS play a critical role in many cellular processes and can be produced in cytosol and organelles including mitochondria and ER. The UPR is a network of adaptive signaling pathways that ensure efficient protein folding. During morbid states, altered redox homeostasis in the ER can lead to ER stress from dysregulated disulfide bond formation and breakage. In ER and mitochondria, accumulation of ROS causes oxidative stress. Hence, the ER stress and oxidative stress are the intertwined biological processes that coexist and induce each other to influence the outcome in health or disease. Persistent oxidative stress and protein misfolding are known to play key roles in the pathogenesis of major human diseases such as diabetes, atherosclerosis, inflammation, and neurodegenerative disorders.

The emerging evidence suggests that FFAs are the common link among obesity, insulin resistance, and T2D. Plasma FFA levels are elevated in most obese individuals and increased FFAs were found to be associated with inhibition of insulin-stimulated glucose uptake in skeletal muscle [40]. LPS derived from gut microflora and sFFAs released by lipolysis signal through the TLR4/MD-2 receptor complex on macrophages and activate the NF-κB pathway to stimulate the production of proinflammatory cytokines and chemokines in the adipose tissue. While the ligands and roles of TLR2 and TLR4 in metabolic inflammation and insulin resistance are well elucidated, little is known about TLR10. Through shRNA-mediated knockdown of TLR10 in THP-1 cells, it was shown that TLR10 responded to FSL-1, LPS, and flagellin [41]. LPS and palmitate play a role in metabolic inflammation. We, therefore, asked if: (i) exposure of monocytic cells to LPS or palmitate could induce TLR10 expression; and (ii) presence of oxidative stress could promote TLR10 expression by the synergy with LPS or palmitate? In this regard, our data show that THP-1 cell exposure to an endotoxin (LPS), a saturated free fatty acid (palmitate), an unsaturated fatty acid (oleate), or the obesity-associated hormone that regulates energy balance (leptin) did not lead to significant changes in TLR10 expression. However, TLR10 expression was significantly elevated when cells were co-treated with palmitate-H₂O₂ or LPS-H₂O₂, indicating that the oxidative stress acted synergistically with palmitate or LPS to upregulate the expression of TLR10 in these cells. Palmitate-H₂O₂ co-treatment was a more potent inducer of TLR10 in monocytic cells. It is noteworthy that the inflammatory and vascular cells present in the adipose tissue produce H₂O₂ which induces oxidative stress by ROS generation and hypoxia. Thus, the elevated expression of TLR10 in the adipose tissue may be due, in part, to oxidative stress from hypoxia which is a well-known early pathologic change observed in obesity.

Of note, we also found that the synergistic interaction between palmitate and oxidative stress promoted the gene expression of several inflammatory proteins such as CCL2/MCP-1, TNF-α, IL-6, IFN-γ, and immunoregulatory cytokine IL-10. In agreement with these findings, at least in part, another study showed that palmitate and LPS acted synergistically to enhance the production of IL-1β and TNF-α in primary macrophages [42]. The key roles of proinflammatory cytokines e.g. TNF-α, IL-1β, CCL2, IL-6, and IFN-γ in the adipose tissue inflammation and insulin resistance have been well documented [43-47]. Besides, IL-10 hyporesponsiveness due to hyperglycemia-induced changes in the IL-10R and downstream signaling proteins was linked to metabolic inflammation in T2D [48]. Our data showing that palmitate and oxidative stress interact synergistically to trigger expression of the signature inflammatory cytokines and an immunoregulatory cytokine IL-10 are important since the balance between inflammatory and anti-inflammatory cytokines may play a critical role during hypoxia for regulating the switch from a normal state of homeostasis (anti-inflammatory milieu) to a morbid state of metabolic inflammation (proinflammatory milieu). The link between ROS and inflammatory mediators is obvious. On one hand, previous studies show induction of ROS by proinflammatory cytokines (TNF-α, IL-1β, and IFN-γ) or endotoxin (LPS) [49-51]. On the other hand, our data show induction of proinflammatory cytokines (TNF-α, IL-6, and IFN-γ) by H₂O₂-induced ROS. Collectively, these findings support the oxidative stress-driven model of metabolic inflammation in obesity in which the adipose tissue hypoxia may lead to enhanced production of inflammatory adipokines to promote ROS accumulation and exacerbate the oxidative stress. Chemokines, such as MCP-1, may directly influence the migration and colonization of monocytes/macrophages in the adipose tissue, a factor that will further amplify the local inflammatory responses. Our finding that co-treatment with palmitate and H₂O₂ promotes the IL-10 expression in THP-1 cells is interesting and needs further investigation. It also remains to be seen how the oxidative stress stimuli will interact with ROS in various cell types. In any case, more studies will be needed to validate and extrapolate our findings of oxidative stress-associated TLR10 upregulation and its implication for metabolic inflammation.

Taken together, our data show the increased TLR10 gene/protein expression in the adipose tissue in obesity/T2D which relates directly with BMI. The data also show that ROS-mediated oxidative stress could induce the expression of TLR10 in monocytic cells through the mechanism that involves NF-κB and MAPK signaling pathways and the ER stress response. This TLR10 induction is further promoted by the synergy between palmitate and oxidative stress i.e. a process that also promotes the expression of proinflammatory cytokines which may have implications for metabolic inflammation.

T2D (Type-2 diabetes); TLR (Toll-like receptors); PBMC (Peripheral blood mononuclear cells); H₂O₂ (Hydrogen peroxide); ROS (Reactive oxygen species); SOD (Superoxide dismutase); BMI (Body mass index); NF-κB (Nuclear factor kappa B); MAPK (Mitogen-activated protein kinases); ER (Endoplasmic reticulum); IRFs (Interferon regulatory factors); TNF-α (Tumor necrosis factor); IL (Interleukin); TIR (Toll/IL-1 receptor); MyD88 (Myeloid differentiation factor 88); LPS (Lipopolysaccharide); sFFAs (Saturated free fatty acids); HbA1c (Hemoglobin A1c); PBS (Phosphate buffered saline); BSA (Bovine serum albumin); FBS (Fetal bovine serum); RT-PCR (Reverse transcription polymerase chain reaction); MGB (Minor groove binder); FAM (6-fluorescein amidite); NFQ (Non-fluorescent quencher); GAPDH (Glyceraldehyde 3-phosphate dehydrogenase); ICH (Immunohistochemistry); DAB (3, 3ʹ-diaminobenzidine); DAPI (4’,6-diamidino-2-phenylindole); LR (Linearized rate); AU (Arbitrary units); CLS (Crown-like structures); NAC (N-acetyl cysteine); NDGA (Nordihydroguaiaretic acid); ERK (Extracellular signal-regulated kinases); ERN1 (Endoplasmic reticulum to nucleus signaling-1); ATF6 (Activating transcription factor-6); LXR1 (L-xylulose reductase-1); CHOP (C/EBP homologous protein).

This study was supported by funds (Grant #: RA-2013-002) from Kuwait Foundation for the Advancement of Sciences (KFAS).

The authors declare that there are no conflicts of interest involved.

S. Sindhu and N. Akhter contributed equally to this work.