Combined All-Extremity High-Intensity Interval Training Regulates Immunometabolic Responses through Toll-Like Receptor 4 Adaptors and A20 Downregulation in Obese Young Females

Modern lifestyle-induced immunometabolic disorders, like insulin resistance (IR), type 2 diabetes, cancer, autoimmune diseases, and neurodegeneration, have become a global burden because they promote the premature death [1, 2]. Due to the strong association between obesity, metabolic disturbances, and mortality risk, obesity has become a causal element in the etiology of these chronic diseases [2-4]. Although obesity is a multifactorial state, chronic low-grade systemic inflammation, known as metainflammation, is a hallmark that partly links all of these disturbances through promotion of the activation and interactions of many inflammatory pathways in multiple cells [1]. During the development of obesity, pathologic expansion of adipocytes in adipose tissue (AT) induces hypoxia and adipocytes necrosis, which act as initiators of chronic low-grade systemic inflammation. In this state, more endogenous ligands are produced, which activate the immune system and impair AT homeostasis [2, 5]. These changes in AT lead to tissue-resident inflammatory/anti-inflammatory immune cell imbalance through an increase in the recruitment of circulating immune cells and a difference in the phenotype of tissue-resident macrophages (inflammatory/anti-inflammatory macrophages [M1/M2] polarization) in AT [5-8]. At this state, endogenous ligands (glucose and saturated fatty acid) activate inflammatory pathways like toll-like receptors (TLR) [9, 10], considered the key regulators of metabolic disorder progression [1]. As TLR4 becomes activated, signaling transition is conducted by activating the intracellular toll/interleukin (IL)-1 receptor [7] domain, which recruits the important myeloid differentiation factor 88 (MyD88) and TIR domain-containing adaptor-inducing IFN-β (TRIF) adaptors. Consequently, these adaptors provoke MyD88- and TRIF-dependent downstream cascades, which in turn activate the nuclear factor (NF)-κB and interferon regulatory factor 3 (IRF3) inflammatory transcriptional factor. Activation of these signaling pathways is mediated by interaction with the tumor receptor-associated factor 6 (TRAF6) adaptor, which leads to the production of many inflammatory cytokines including IL-6, IL-12, IL-1β, interferon γ (IFNγ) and tumor necrosis factor α (TNFα). These products are strong inducers of M1 polarization [5] and insulin dysfunction in obese (OB) and IR states [1], intensifying the inflammatory condition [3, 11]. Although TLR4 signaling is a finely tuned pathway in preventing the enormous immune responses by different negative regulators, like A20, it is assumed that the abnormal activation and/or interaction of these negative regulators could be partly involved in suppression of TLR4 signaling [12-14]. Accordingly, diminishment of AT inflammation and the decrease in M1 polarization due to a reduction of the inflammatory activities of circulating immune cells suggest a treatment strategy at the early stages of inflammation [5, 8].

Entering adulthood in an OB state increases the mortality risk and, on the contrary, an active lifestyle induces lower rates of TLR4 activation, inflammatory monocytes, and chronic diseases [4, 15, 16]. So, it is believed that regular exercise training could be an effective strategy in inhibiting the progress of inflammation. As a result, exercise therapy has been recommended for the treatment of obesity and the related metabolic disorders as a potent, inexpensive, nonpharmacological approach with minimum side effects [17, 18]. The immunomodulatory properties of exercise training have indicated a reduction in AT and systemic inflammation [19]. It happens following suppression of macrophage infiltration, an increase in macrophage polarization toward M2 [20, 21], and TLR2/4 downregulation in peripheral blood mononuclear cells (PBMCs) [22-25]. However, a lack of time and work-life imbalance constitute the biggest obstacles for participating in exercise training programs [26]. To overcome this phenomenon, new time-efficient exercise protocols like HIIT and all-extremity training have received more attention among researchers. These protocols improve cardiometabolic function through the reduction of TLR2 expression and IR in patients with obesity [23, 27]. In this context, a few studies which have focused on resistance training (RT) have indicated a low effect on metabolic factors in the youth [28].

In this study, we investigated how a new all-extremity combined HIIT (CHIIT) protocol can change the metainflammatory state. We assessed the TLR4 signaling cascade responses in PBMC of overweight (OW) and OB females. To the best of our knowledge, there is almost no study relating long-term exercise training to the TLR4 pathway and its adaptors and negative regulators. Knowledge of whether the induced adaptations of CHIIT consisting of both aerobic training (all-extremity cycling) and RT could modify the TLR4 pathway among sedentary young women would be practically beneficial.

This cross-sectional nonrandomized controlled study design was applied by comparing the control (CG) and exercise groups (EG). The subjects in the EG performed the exercises 4 days a week. All measures were taken first at baseline and then, after 10 weeks (38–40 sessions) of training.

All of the participants were recruited via posters posted throughout the University of Isfahan; 86 volunteers were selected for this study (Fig. 1). Seventy-one out of 86 individuals met the inclusion criteria, i.e., healthy young (18–25 years) and OW or OB (BMI = 25–35 kg/m²) patients. As for the exclusion criteria, 41 out of 71 subjects were omitted due to having a nonsedentary lifestyle (participated in any exercise program 6 months before our study), smoking, consuming any medicine, being on a diet, and having cardiovascular diseases (CVD) or kidney, hypertension, IR, and diabetes issues. In the end, a total of 30 eligible women were randomly allocated to the CG (n = 15) and the EG (n = 15); they were selected from a pool.

All full supervised training sessions (38–40) were performed at the Laboratory of the Sport Science Faculty. The training protocol consisted of a progressive all-extremity CHIIT, including 4 sets of 4-min intervals separated by 3 periods of 3-min intervals of active rest (leg cycling) (Figure 2). Each interval involved 4 exercises, i.e., 2 aerobic exercises (arm and leg simultaneous cycling; Monark Ergomedic 839 E and 831 E; Monark, Sweden) and 2 resistance exercises (1 upper body and 1 lower body exercise; Vectra Fitness On-Line 4800; USA) according to a circuit training program where an RT followed all-extremity cycling training. The RT part of the protocol, which involved the large muscle groups of the body, consisted of chest presses, lateral pulldowns, leg presses, and leg flexions, where 1 upper and 1 lower body exercise was performed in every interval. The intensity of aerobic training increased gradually according to the age-predicted (220 – age) maximum HR percentage (HRmax) and monitored using a Polar watch (F4 Electro; Oy, Kempele, Finland). The RT load was adjusted according to the one maximum repetition (1RM) percentage. The training intensity and the RT volume increased during the preparation and main phases in a gradual manner. During the preparation phase, in the first 2 weeks, the aerobic training intensity and the RT volume increased from 50 to 70% HRmax and from 40 to 50% 1RM, respectively. The training intensity and volume were reached to 90% HRmax and 70% 1RM in the remaining 8 weeks, provided the subjects were able to perform more than 12–15 repetitions of each exercise for 3 consecutive training days, at an increased load of about 5–10%. Each session began with a warm-up of 10–12 min of cycling, stretching, and kinetic movements and finished with a cool-down consisting of cycling and stretching for 5 min.

The maximal strength of every resistance workout was assessed before the blood testing day in 2 sessions. On the first day, participants were instructed to learn about the proper lifting technique and allowed to practice 2 sets of 8 repetitions. Following the instructions of the 1–6 RM testing procedure, the participants performed the test for all movements on the second day. Finally, strength exercises were evaluated based on the adjusted 1RM% equation proposed by Nutter [29].

To become adapted to the laboratory environment and facilities, monitoring the HR, synchronizing the arm and leg cycling, and reaching the target HR were all done during 2 familiarization sessions on 2 nonconsecutive days after the 1RM testing day. Each of the participants performed 2 sets of the protocol, with an increase in the rest time at a 1:1 ratio in the first session, followed by the 2 or 3 intervals of the protocol on the second day.

All of the subjects in both groups were tested twice, i.e., at baseline and in a follow-up run, with the same procedure a day before 1RM testing and ∼48 h after the last training session to avoid the acute impact of exercise on the test results.

Weight and standing height were measured in light clothing without shoes using a Seca 220-gauge stadiometer (Seca, Hamburg, Germany). BMI was calculated as weight (kg) divided by height (m) squared. Waist circumference (WC), hip circumference (HC), and the waist-to-hip ratio (WHR) were assessed to indicate the distribution of abdominal fat using an elastic measuring tape. WC and HC were measured between the lowest rib and the iliac crest and in the broadest part of the gluteal muscle, respectively, while participants stood erect [30]. The body fat percent (BF%) was calculated by measuring the abdominal, triceps, and suprailiac 3-site subcutaneous skinfold thickness in triplicate using a Lange skinfold caliper (Cambridge Scientific Industries, Cambridge, MD, USA) and the Jackson-Pollock equation for 3-site measurement in women [31].

A YMCA submaximal test was run on an electronically braked cycle ergometer (Ergomedic 839 E; Monark, Varberg, Sweden) to determine the subjects’ maximal oxygen consumption (VO_2max) [32]. Participants cycled with constant RPM during every stage, while the workload was increased in 3-min intervals. The test was automatically terminated when the HR became constant in 2 consecutive stages at 110–150 bpm. Arterial blood pressure (BP) and resting HR were measured using an automated brachial cuff after 45 min of seated rest.

To ensure that the results of this intervention were not affected by other concurring factors like daily physical activities and dietary changes, the participants were instructed not to change their routine diet or physical activities throughout the study period. Also, to monitor the daily physical activities and dietary energy intake, a pedometer-based step count (AM3; iHealth Labs Inc., USA) and a self-reported 3-day diet recall were applied to record the daily steps and food intake data during the study period on 2 weekdays and one weekend, respectively. These assessments were run at week 1, the baseline, week 4 or 5, the middle, and at week 10, the end of the study period. Dietary data were analyzed through NUT 4 software to assess the fixed calorie consumption during the study period.

Participants were asked to avoid doing physical activities 48 h before the blood collection day. After overnight fasting, 12-mL venous blood samples were taken from 7:30 to 9:00 a.m. to prevent circadian variation after a 5-min rest. Leukocytes were isolated from 2-mL heparinized whole blood samples in aliquots using a red blood cell lysis buffer (Biolegend, CA, USA), according to the manufacturer’s instructions. The remaining 10 mL were left to clot to be centrifuged for serum separation. The isolated leukocytes and serum tubes were stored at –80°C until the final analysis.

Transcript expression was assessed by quantitative real-time polymerase chain reaction (qPCR) following total RNA extraction from leukocytes using a High Pure RNA Isolation Kit (Roche-life Science, Mannheim, Germany) containing the DNase digestion step to remove the potential genomic DNA residual trace. Following RNA sample quality assessment by spectrophotometry (NanoDrop^TM 2000; Thermo Fisher Scientific, USA), total RNA samples were transcribed in reverse using a BioFact kit (BioFactTM 2X RT Series Pre-Mix; Daejeon, South Korea). The synthesized complementary DNA was used to amplify the gene expression by qPCR (StepOnePlus^TM Real-Time PCR System; Thermo Fisher Scientific, USA). In this step, SYBER Green BioFACT^TM 2X Real-Time PCR Master Mix (Daejeon, South Korea) was consumed. The gene-specific sequences for amplification included:

TLR4 forward: 5′- AAGCCGAAAGGTGATTGTTG -3′, reverse: 5′- CTGAGCAGGGTCTT-CTCCAC-3′; MyD88 forward: 5′- CTCTCTCCAGGTGCCCATCA -3′, reverse: 5′- AGGCGAGTCCAGAACCAAGA -3′; TRIF forward: 5′- GTCCAGGTGTTGGCTCTGTT -3′, reverse: 5′- CTGGACGAACACTCCCAGAT -3′; TRAF6 forward: 5′- CCTTTGGCAAATGTCATCTGTG -3′, reverse: 5′- CTCTGCATCTTTTCATGGCAAC -3′; NF-κB forward: 5′- GCACGACAACATCTCATTGG -3′, reverse:5′- TCTGCTCCTGCTGCTTTG -3′; IRF3 forward: 5′- TCGTGATGGTCAAGGTTGT -3′, reverse: 5′-TACTGGTCGGAGGTGAGG-3′; A20 forward: 5′- ACAGAAGAGCAACTGAGATC -3′, reverse: 5′- GTTGGGATGCTGACACTC -3′; and GAPDH as housekeeping gene forward: 5′- GTGAAGGTCGGAGTCAACGG -3′, reverse: 5′- CCTGGAAGATGGTGATGGGAT -3′.

All of the experiments were amplified in triplicate, and every run included a nontemplate reaction. The transcript expression values of the target genes were determined in relative units based on the threshold cycle (Ct) of each gene normalized according to GAPDH expression and quantified based on the 2^–ΔΔCt method.

The serum cytokine assay was established as a valid approach to assess the final product quantification of the biological pathway. Quantikine ELISA kits were applied to measure the serum levels of TNFα (BioLegend Inc., USA) and IFNγ (PeproTech Inc., USA).

The data distribution was determined by Q-Q plots and the Shapiro-Wilks test; nonnormal data were transformed into a natural logarithm (log10). The Levene’s test analyzed the homogeneity of variances. Between-group differences at baseline were assessed using an independent t test for normally distributed data or the Mann-Whitney U test for nonnormal distributed data. The dependent variables’ responses to the intervention were evaluated by ANCOVA analysis, while the pretesting values were input as the covariate factor. The data were analyzed using a mean differences t test or the Mann-Whitney U test, where the ANCOVA assumptions were violated. The paired sample t test was used to verify within-group changes. The effect size was interpreted via ∆% calculation to explore the magnitude of the intervention effect. The Spearman correlation coefficient test was used to assess the correlation of TLR4 and its related adaptors genes with the anthropometrical and cardiorespiratory fitness variables. p < 0.05 was considered statistically significant. All data are reported means ± SD and were analyzed using IBM SPSS Statistics 22 and GraphPad Prism 8.0.1.

The data were analyzed per protocol. Four participants (2 from each group) did not complete the training intervention due to family, health, and personal problems, leaving 13 participants per group (Fig. 1). The baseline comparisons of both groups are shown in Table 1. Although there were no statistically significant differences in preintervention anthropometrical variables (p > 0.05), the values of systolic blood pressure and VO2_max revealed significant differences (p = 0.02 and p = 0.01, respectively). At baseline, the upper and lower levels of mRNA expression of all of the participants were related to TRIF and IRF3, respectively (Table 3).

There was no statistically significant difference (p > 0.05) in between- or within-group comparisons in terms of the daily calorie intake (Fig. 3a) and physical activities (Fig. 3b) during the study period.

Participants in the EG exhibited a statistical significant decrease in weight (p = 0.006, Δ% = –2.58), BMI (p = 0. 01, Δ% = –2.51), BF% (p = 0.001, Δ% = –5.25), WC (p = 0.001, Δ% = –3.49), and WHR (p = 0.001, Δ% = –2.56) either during pre- to posttraining or in comparison to CG, where the HC changes were not statistically significant compared to the preintervention (p = 0.05) and CG (p = 0.12; Table 2).

The initial value of VO_2max was recorded at preintervention and postintervention; they were 30.73 ± 5.29 and 38.19 ± 4.76 mL/kg/min, respectively. Thus, a statistically significant increase of 24.48% (p < 0.01) was found in EG, while no statistically significant improvement was observed in CG. When VO_2max was compared in both groups, a significant statistical increase was revealed (p = 0.01). No statistically significant differences were observed in SPB and DBP between the CG (p = 0.59 and p = 0.6, respectively) and pretesting (p > 0.05). Moreover, the resting heart rate improved by 12.48%, which was statistically significant in both within- (p < 0.01) and between-group (p = 0.001) comparisons (Table 2).

TNFα levels decreased significantly in the EG from preintervention to posttesting (p < 0.01, Δ% = –72.04) and compared to the CG (p = 0.001). Although the IFNγ concentration decreased posttesting compared to preintervention (Δ% = –13.86), this was not statistically significant (p > 0.05) even when compared to CG (p = 0.59) values (Table 2).

According to Figure 4, the CHIIT intervention induced a lower expression of TLR4 in the EG compared to both preintervention (p < 0.05) and the CG (p = 0.001). Moreover, a significant statistical decrease was observed at the MyD88 level in the EG compared to the CG (p = 0.04). However, no statistically significant changes in TRIF or TRAF6 mRNA expression were found in any of the comparisons. NF-κB and IRF3 gene expression significantly decreased after intervention in the EG compared to preintervention (p < 0.01 and p < 0.05, respectively) and the CG (p = 0.001). Also, the statistically significant reduction in the A20 mRNA level was found following the intervention in comparison with either the CG (p < 0.01) or pretesting (p = 0.002; Table 3).

According to Table 4, bivariate comparative analyses demonstrated statistically significant positive correlations between TLR4 gene expression changes and BF% (r = 0.5, p < 0.01), WC (r = 0.43, p < 0.05), and WHR (r = 0.45, p < 0.01). There was a statistical significant positive correlation between IRF3 gene expression changes and weight (r = 0.4, p < 0.05), BMI (r = 0.39, p < 0.05), BF% (r = 0.61, p < 0.01), and WHR (r = 0.55, p < 0.01). Moreover, changes in NF-κB gene expression levels revealed a statistically significant positive correlation with BF% (r = 0.64, p < 0.01) and a negative correlation with VO_2max (r = –0.45, p < 0.05). A statistically positive association was found between the A20 gene level and BF% (r = 0.62, p < 0.01) and WHR (r = 0.42, p < 0.05). Also, there was a statistically significant positive correlation between the TNFα serum concentration and BF% (r = 0.68, p < 0.01) and WHR (r = 0.49, p < 0.05) values. A statistically significant correlation was found between TRIF, MyD88, TRAF6 genes expression, and all of the variables, as well as the serum concentration of IFNγ.

To the best of our knowledge, this is the first study to assess the impact of a non-weight-bearing combined all-extremity HIIT protocol on the TLR4 pathway (MyD88- and TRIF-dependent downstream cascades) and its identical negative regulator. In the present study, we showed that CHIIT downregulated TLR4, MyD88, NF-κB, IRF3, and A20 gene expression, while there were no statistically significant differences in TRIF and TRAF6 mRNA expression following 10 weeks of exercise. Serum levels of TNFα and IFNγ decreased after the proposed exercise training protocol. Additionally, we found that VO_2max was improved, and weight, BMI, BF%, WC, and WHR were reduced, after CHIIT.

The TLR4 pathway-related adaptors and negative regulator’ responses to our CHIIT are described in their sequential sense in the following paragraphs.

Animal studies have revealed that high-fat diets increase TLR4 gene levels compared to normal chow [20]. Although the acute protocols have not significantly changed TLR4 mRNA levels, the increase in the duration of training protocols appears to shift these results toward a reduction of TLR4 levels [33]. According to human studies, both acute, though not all, types of exercise (HIIT) [23] and long-term interventions downregulate TLR4 in the obese, healthy, and type 2 diabetes subjects [24, 34, 35]; this corresponds with the findings here. On the contrary, TLR4 levels have remained unchanged or elevated after RT and aerobic protocols [36, 37]. TLR4 expression decreased in macrophages pretreated with irisin, regarded as a potent anti-inflammatory myokine [38], suggesting the exercise-induced anti-inflammatory myokine as one of the possible mechanisms suppressing TLR4 expression in this study. Accordingly, more experiments are required to resolve these existing contradictions in TLR4 responses to exercise.

As TLR4 recruits both MyD88 and TRIF molecules and consequently activates 2 independent cascades, it is recognized as a prerequisite for obese AT pathology [2, 11]. Some RT interventions have reported a reduction in mRNA and protein levels of TRIF [24, 39, 40], while no statistically significant changes were observed in the current study. On the contrary, the MyD88 gene expression level was downregulated in our study, which is supported by previous findings [24, 39, 40], but not all [36]. It has been reported that the suppression ability of irisin on the TLR4 pathway is associated with inhibition of MyD88 activities [38], suggesting our CHIIT-induced immunomodulatory myokine.

Recruitment of TRAF6 in the activation of 2 TLR4 downstream cascades infers its effects on TLR4 inflammatory power [13]. Consistent with some available studies [39], our protocol did not affect TRAF6 expression, suggesting that this CHIIT might not be strong enough to mediate this adaptor mRNA expression. Upregulation of TRAF6 following exercise was reported by Lira et al. [41] in lipopolysaccharide-treated adipocytes. Accordingly, these conflicts over TRAF6 responses could reveal a regulatory pathway over TRAF6 actions and/or interactions with other signaling molecules. In this sense, A20 is a negative intracellular regulator of the TLR4 pathway, which induces inhibitory effects on both TRIF- and MyD88-dependent cascades by suppressing TRAF6 recruitment [14]. Unlike TRAF6, A20 levels were reduced following our intervention. People with nonalcoholic fatty liver disease and obesity have shown a higher level of A20 expression in the liver in comparison with healthy control individuals [42, 43]. In contrast to our findings, an acute strenuous exercise intervention has increased the A20 expression levels [44]. To the best of our knowledge, this is the first study to assess A20 responses to a long-term exercise intervention. As adiponectin is one of the proposed proteins inducing beneficial metabolic impacts through A20 [45], it is assumed that some adipo/myokines interfere with A20 expression. More studies are required to explain the exact responses and responsibilities of A20. Therefore, A20 is a possible adaptor in regulating TLR4-induced inflammation in circulating leukocytes during obesity, while its actions and/or interactions in the regulation of TLR4 signaling have remained obscure in response to exercise.

Our intervention reduced the mRNA levels of NF-κB and IRF3, confirming the identical effects of CHIIT on both MyD88- and TRIF-dependent cascades. In the TLR-4 activation process, the expression of different signaling cascade molecules eventually reaches the NF-κB transcription factor which induces production of proinflammatory cytokines such as IL-1β, IL-6, and TNFα [3]. So, the reduction of NF-κB and IRF3 gene expression in our study may be due to the decreased TLR-4 expression. Studies have indicated that the gene expression levels of IRF3, NF-κB, and its inhibitors are downregulated in response to both RT and combined interventions [24, 39, 40], even though continuous aerobic exercise and interval training do not change NF-κB levels [36]. Consistent with our findings the reduction of NF-κ B has been demonstrated in macrophages pretreated with irisin, suggesting a mechanism for exercise in reducing TLR4/NF-kB pathway activity [38]. Moreover, increased secretion of catecholamines, cortisol, and IL-6 by contracting muscles downregulates the NF-κB pathway [19].

In the present study we found that TNFα levels were decreased after the proposed exercise training protocol. TNFα has a central role in metabolic dysfunction, and anti-TNF approaches apply as therapy in chronic inflammatory diseases [1]. As we mentioned before, one of the final products of the TLR-4 signaling pathway is TNFα, which acts a proinflammatory cytokine. So, it makes sense that by reducing the gene expression of the TLR-4 and then reducing the expression of its downstream molecules in the signaling pathway, such as IRF3 and NF-κB, the expression of TNFα as the end product of this pathway will also decrease. A reduction in TNFα serum levels has been shown following long-term protocols [39, 46-48], which is in line with our study, while others have reported increased or unchanged levels [22, 34, 49]. Studies have shown that in the presence of IFNγ macrophages shift their phenotype toward M1 and produce proinflammatory cytokines, including IL-12, TNF-α, and IL-1β [50, 51]. Also, this cytokine is known as the progenitor of local AT inflammation. Increased IFNγ levels during obesity have been associated with the development of M1 polarization and impairment of AT homeostasis during high-fat-diet-induced immunometabolic dysfunction in the early stages [5]. Consistent with previous findings, the IFNγ level was decreased following our exercise training protocol [44, 52]. Although this decrease was not statistically significant, it seems that IFNγ reduction effectively decreases the production of TNF-α as an inflammatory cytokine. The other possible mechanism for the decrease in TNFα and IFNγ is that contractile muscle-increased IL-6 develops the production of IL-10. This cytokine as an anti-inflammatory cytokine causes a decrease in TNFα and IFNγ levels, suggesting a reduction of M1 polarization in AT, simultaneously [19].

The significant associations between BF% and the inflammatory variables, including TNFα serum level and TLR4, NF-κB, and IRF3 gene expression, were revealed in this study. Considering that all-cause mortality rates increase by 30% per 5-kg/m² increase in BMI [3], the reduction in body composition values in our study revealed the potential benefits of this all-extremity CHIIT for the mortality rate – something that is not supported by an aerobic all-extremity protocol [27]. As high-intensity training has been reported to be essential to increase fat oxidation capacity [26], this discrepancy may be due to the differences in the subjects’ features and/or the resistance exercises in our protocol.

Regarding an inverse relation between aerobic fitness and inflammation [25], we found a moderate negative correlation between VO_2max and NF-κB levels. As the improvement of aerobic fitness per 3.5 mL/kg/min reduces the CVD mortality risk to 19% [53], the increased level of VO_2max in this study (i.e., 24%, 5.25 mL/kg/min) indicates a potential power of our protocol to reduce the CVD mortality risk. These findings are consistent with other studies [22, 47]. Therefore, prescribing short-duration exercise training can promote health-related physiological aspects and increases the possibility of individuals’ participation in exercise training programs.

• Because the main focus of this study was on designing a new, potent, time-efficient protocol, the authors did not compare this protocol with traditional exercise training methods.

• Concerning the importance of the same phases of the menstrual cycle for pre- and posttesting results, some participants were in the follicular phase and some were in the luteal phase, and the remaining were in other phases. In 10 weeks, these populations confronted abnormalities in their menstrual cycle, preventing control of the same menstrual phase in the testing days.

• Provision of an identical meal for pre- and postintervention blood testing would be more reliable than the subjects’ self-reports.

Overall, our findings point to the immunomodulatory and anti-inflammatory impacts of our all-extremity combined HIIT protocol on metainflammation. These effects were associated with the TLR4 signaling pathway, including downregulation of TLR4, its protein adaptors, transcriptional factors, and negative regulator gene expression or reduction of TNFα serum levels. The identical effects of this protocol on both NF-κB and IRF3 were in accordance with A20 downregulation, suggesting that the negative regulators in the TLR4 pathway are the possible mechanisms of exercise for improvement of the inflammatory state in obesity (Fig. 5).

The authors, express their appreciation to Dr. Hossein Khanahmad for his valuable assistance and advice. The support of Mrs. Mohadeseh Toghyani and Mrs. Fahimeh Hosseininasab, from the Immunology Department of the Isfahan University of Medical Science, is appreciated as well.

This study was approved by the Research Affairs Office, Research Committee of the University of Isfahan, Iran, according to the policy of the Ethics Committee (No. IR.UI.REC.1396.039). All of the subjects underwent a verbal and written check-up by a reliable physician. They were fully informed about the experiment and possible risks before signing a written consent form.

There are no competing conflicts of interests to declare.

This study was funded by the University of Isfahan, Isfahan, Iran. Some kits for laboratory experiments were purchased with this financial support.

All authors contributed to this study conception. N.S., S.M.M., and N.E. designed this study and performed the statistical analysis. N.S., N.E., and M.K. conducted the laboratory experiments and analyses. N.S. drafted the initial version of this paper, and N.E. edited the final version. All of the authors approved this work.