The Association of Gut Microbiota with Resting Metabolic Rate in Overweight/Obese Women: A Case-Control Study

Resting compare the Firmicutes/Bacteroidetes ratio and the relative abundance of, Prevotellaceae, Faecalibacterium prausnitzii, bidobactrium spp, lactobacillus spp, Akkermansia muciniphila, Bacteroides fragilis, and Escherichia coli in two groups of people with normal and low RMR in overweigh/obese women in

role of gut microbiota as a crucial factor affecting energy metabolism [32,33], whilst some others have assessed the gut microbiota composition of obese and normal-weight subjects [34][35][36][37]. Regarding energy expenditure, in 2013, P. KOCEŁAK et al. [38] evaluated the association between gut microbiota and REE, where higher bacterial load was found in obese compared to normal weight individuals, but no independent association was seen with REE. In contrast, several animal studies have found that some bacteria, such as Akkermansia muciniphila [39], Lactobacillus reuteri [40] and Lactobacillus gasseri [41] could increase energy expenditure. However, the role of gut microbiota in the context of energy homeostasis, especially RMR, has not been clearly demonstrated in humans so far. Therefore, we sought to compare the abundance of some bacteria in overweight/obese women with normal and low RMR. To the best of our knowledge, this study would be the rst in this context. Moreover, we also aimed to determine the association between the abundance of such bacteria and RMR among our studied population.

Study population
In the current matched case-control study, we enrolled 36 healthy women (18 cases and 18 controls), who had been referred to a nutritional laboratory according a public invitation in Tehran. We used a Telegram bot and a total of 1300 participants were signed up among which, 122 individuals were eligible samples. Finally, after assessing their medical history, 85 volunteers were quali ed for the study (Figure-1). After a follow-up period of 2-3 months, to ensure no medication and supplementation, as well as no uctuations in weight and other inclusion criteria were maintained, individuals were referred to the Nutrition Laboratory of the Faculty of Nutrition and Dietetics located at the School of Health of Tehran University of Medical Sciences for further evaluation. With the purpose of random distribution, the participants in the case and control groups were divided into paired blocks in terms of age, BMI, and RMR level. Each sample in the case group (low RMR) was randomly matched with a sample in the control group (normal RMR) with a maximum difference in age ± 2 years and BMI ± 2 units. Volunteers were randomly selected for the study based on the following inclusion criteria (for both cases and controls): 25≤ BMI <40 (obesity and overweight), aged 18-50 years old. To ensure comparable data, we included the following exclusion criteria: use of antibiotics (within the previous 3 months) [79], use of alcohol, smoking, signi cant infection, history of diabetes, coronary, thyroid diseases or other hormonal disease and cancer. Use of medications or treatments effective on their RMR, use of supplementary vitamins and minerals, being pregnant, lactating or menopausal, daily or irregular intake of probiotics within the previous 2 months, history of digestive diseases, such as in ammatory bowel disease, irritable and constipation. Also those with gastrointestinal surgery, use of dietary supplements for weight loss during the past 6 months [38] were excluded. The inclusion and exclusion criteria of the samples were based on the individual's medical history or their own statements. The study protocol was approved by the Ethics Commission of Tehran University of Medical Sciences (IR.TUMS.VCR.REC.1398.562) and prior to the study, all subjects signed a written informed consent.

Demographic questions
Demographic questions were used to collect data about characteristics such as age, smoking status, education level, lifestyle, marital status, menopause, medical history, taking medication, supplement use etc.

Anthropometric assessment
For each participant, body composition, including weight, BMI, body fat mass (BFM), fat free mass (FFM), body fat percentage (%), waist to hip ratio and waist circumference were measured. All measurements performed by using a multi-frequency bioelectrical impedance analyzer, InBody 770 scanner (Inbody Co., Seoul, Korea). For all participants, a very low, safe electrical signal was sent from four metal electrodes through both hands and feet.
The electrical signal passes quickly through water that is present in hydrated muscle tissue but meets resistance when it hits fat tissue. This resistance, known as impedance, can be measured to infer the proportion of fat free mass and fat mass. Measurement was conducted according to the manufacturer's guidance.
Resting Metabolic Rate measurements RMR was measured by indirect calorimetry (MetaLyzer®3B, made in Germany). The RMR was assessed in the morning after a requested overnight fast (10-12 hour). It was measured under "resting conditions," which included no prior severe exercise, and abstinence from alcohol and caffeine. The indirect calorimetry device was calibrated before each assessment. To measure RMR (m-RMR), after 20 minutes of rest, the patients assumed the supine position without movement for 30 min, and the middle 20 min was used for calculation (the rst and the last 5 min were ignored). We strati ed participants based on measured RMR (m-RMR) and predicted RMR values (p-RMR) [80]. Patients were de ned as "hypometabolic" when their measured RMR was less than 85 % of the predicted RMR, based on the Harris and Benedict equation [81], or "normometabolic" when it was within ± 15% of the predicted RMR. During the luteal phase of the menstrual cycle, RMR typically rises and is lower during menstruation. For this purpose, indirect calorimetry performed for all premenopausal women during the follicle process [82-84].

Dietary assessments
Dietary intake was assessed by a semi-quantitative 147-item food frequency questionnaire (FFQ) that has been validated in previous work [85]. Questionnaires were completed in the presence of a trained nutritionist, and participants reported the intake frequency of each food item over the past year. Household measurements and servings were then converted into weight (grams per day). Dietary intakes were analyzed using NUTRITIONIST 4 (First Data Bank, San Bruno, CA) software for estimating energy and nutrient intake.

Physical activity assessment
For evaluating physical activity in the form of metabolic equivalent hours per week (MET-h/wk), the short form of the International Physical Activity Questionnaire (IPAQ) was used, whose Persian language version has been validated by Moghaddam et al. [86]. Scores were calculated according to the frequency and time spent on light, moderate, high, and very high-intensity activities, based on a list of common daily activities.

Blood sampling
Following a 12-hour overnight fasting, 10 cc of venous blood sample was drawn between 7 and 10:30 a.m., and immediately divided. Half of each sample was kept for 30 minutes at room temperature until clotting, then, blood samples were centrifuged at 3000 g for 20 minutes, and decanted into several separate clean micro-tubes and stored in a freezer at −80°C for further analysis.
Biochemical and hormonal assessments All hormones were determined by using enzyme-linked immunosorbent assay (ELISA) method. Leptin, ghrelin and insulin levels were assessed using a LDN kit (Nordhorn, Germany) with a sensitivity of 0.50 ng/ml, a Crystal Day Christian Day kit with a sensitivity of 0.01 ng/ml, and an IBT kit (in nitum biotech, IBT; Netherland) with a sensitivity of 0.11 μU / ml. Intra-and inter-assay coe cients of variation (CV) reported by the manufacturer for leptin, ghrelin, and insulin were 3.7-5% and 5.9-5.8%, CV<8% and CV<10%, and 3.7-4.2%, and 3.7-4.2%, respectively.

Fecal sampling and DNA extraction
Participants were asked to collect their stool samples in a conventional laboratory plastic container dedicated to fecal sampling. The samples were moved immediately to the laboratory in ice packs and stored at -80 °C ( ash frozen) upon arrival (within 2 h.), before further processing. Extraction of total bacterial DNA from 200 mg of each stool sample has been done using QIAamp The Fast DNA Stool Mini Kit (51604) (Qiagen, Hilden, Germany) was used according to manufacturer's instructions. The purity and concentration of the extracted DNA were determined by Nanodrop spectrophotometer (Thermo Scienti c NanoDrop, USA) [87]. The extracted DNAs were stored at −20 °C until further analysis. By using the nucleotide BLAST in NCBI, the speci city of the primers was evaluated. The speci c sequences of primers are shown in Table 1.

q-PCR analyses
The abundance of bacteria was analyzed using Quantitative real-time PCR based on SYBER green method (LightCycler® 96 SW 1.1; Roche, Germany) [35,87,88]. Each 10 μl of q-PCR reaction was composed of SYBR Premix Ex Taq II (RR820L; Takara, Japan), 0.5 μl of each of the speci c 16s rRNA primers [35,[87][88][89][90][91][92][93][94] (Table 1), and 1 μl of the DNA template. The q-PCR reactions were carried out in duplicate using LightCycler® 8-Tube Strips (clear; Roche). An appropriate annealing temperature was used for designing the ampli cation program: 1 cycle of 95 °C for 1 min, 40 cycles of denaturation at 95 °C for 5 s, then annealing at 55 °C for 30 s, and extension at 72°C for 30 s. Finally, melting curve analysis was performed after ampli cation to con rm the speci city of PCR reactions, followed by 1 cycle at 95 °C for 5 s, 60 °C for1 min, and 95 °C for 1 s. Moreover, there was no difference between energy intake and dietary intake of micro/macro nutrients and food groups among cases and controls (presented in Table-3) (P> 0.05).    (Table-5). Further, after adjustment for body fat percent as a confounder (because it was signi cantly different between groups), the above-mentioned associations remained statistically signi cant (P<0.05).

Discussion
The nding of the current study shows that the abundance of B. fragilis, F. prausnitzii, and Firmicutes were higher in those with normal RMR compared to those with low RMR. Moreover we found that with an increasing abundance of B. fragilis, F. prausnitzii, lactobacillus, and Firmicutes, RMR concurrently increases. In order to discern the relationship between gut microbiota an obesity, numerous animal and human-based research [30,35,[42][43][44] has been conducted [45]. Accordingly, studies have shown that energy homeostasis imbalances, lowgrade in ammation, and insulin resistance result from dysbiosis, which can contribute to negative host metabolism regulation [46]. Moreover, since the rst paper on the role of intestinal microbiota in host metabolism, including body weight regulation [31], the gut microbiome has been shown to have important functions, such as impacting on dietary energy harvesting and the control of anti-in ammatory and metabolism; indeed, any alteration in the composition of the gut microbiota may trigger and develop obesity, or vice versa [46,47]. The rise in Frimicutes and decrease in Bacteroidetes concentrations (increase in F/B ratio) in obese vs. normal subjects has been shown in various animal and human studies [30,[48][49][50]. In the current study, there was no association between F/B ratio and RMR, which is concordant with the results of the previous study conducted on both normal and obese individuals [38]. However, we observed higher abundance of Firmicutes phylum in control groups (with normal RMR) and a positive association between this phylum and RMR. Dominancy of Firmicutes in people with normal RMR could be explained by some genus belonging to this phylum, such as lactobacillus spp. and Faecalibacterium prausnitzii. In fact, in this study we observed signi cant positive correlation between lactobacillus spp. and Faecalibacterium prausnitzii with RMR. Moreover, Bacteroides fragilis showed a positive correlation with RMR as well. There is a dearth of studies that have assessed the correlation between gut microbiota and RMR. Nevertheless, F. prausnitzii has been shown to be a butyrate producer that exhibits antiin ammatory and protective effects against obesity [51]. Previous studies have demonstrated that consumption of fermentable carbohydrates could in uence production of butyrate [52,53]. Indeed, diets high in non-digestible carbohydrates can promote the abundance of some butyrate-producing bacteria, and thus contribute to higher levels of butyrate in plasma. Moreover, in animal studies, it has been shown that high circulation of butyrate in plasma may improve insulin sensitivity and increase energy expenditure [54,55]. In addition, it has been proved that obese participants have reduced levels of plasma butyrate [56]. Intestinal gluconeogenesis, via the intestinebrain circuit, is activated by butyrate and propionate, thus promoting glucose regulation and metabolic bene ts expressed by body weight [57]. The other mechanism by which butyrate can increase energy expenditure is via its capacity to promote thermogenesis and fatty acid oxidation by activation of 5'-AMP-activated protein kinase (AMPK), phosphorylation of peroxisome proliferator-activated receptor-gamma coactivator-1α (PGC1α) in muscle and liver, and up regulation of PGC-1α and mitochondrial uncoupling protein-1 (UCP-1) in brown adipose tissues (BAT) [58,59]. AMPK can inhibit acetyl CoA carboxylase via phosphorylation, and consequently reduces malonyl CoA synthesis [59]. frequency and RMR. In fact, our study is concordant with some animal studies [40] showing that some bacteria belonging to this family can promote energy expenditure. This study veri ed that L. retueri 263 may have an impact on increasing energy expenditure by the up-regulation of Uncoupling protein-1 (Ucp1), presented in white adipose tissue (WAT), by a process called browning. The browning remodels the energy metabolism of WAT and converts it to a beige adipose tissue with higher energy consumption and greater thermogenesis and respiratory rate [64][65][66]. In other words, browning of WAT is a proposed mechanism to increase energy expenditure and decrease obesity. Moreover, Bungo Shirouchi et al. [41] asserted that Lactobacillus gasseri SBT2055 (LG2055) can elicit anti-obesity effects via increases in carbohydrate oxidation and contribute to the increase of energy expenditure. Indeed, the increase in butyrate production by LG2055 can activate G protein-coupled receptor41 (GPR41) and increase energy expenditure following enhancement of glucose tolerance. Indeed, research suggests that in the host, GPRs serve as receptors for free fatty acids (FFAs), among which, free fatty acid receptor 3 (FFA3/GPR41) can be activated via SCFAs, such as propionate, butyrate, and acetate [67][68][69][70]. Interestingly, animal studies have illustrated that GPR41 knockout mice have decreased energy expenditure and less glucose tolerance [71]; suggesting that short chain fatty acids (SCFAs) are able to increase the sympathetic nervous activity by GPR41, and consequently increase energy expenditure in animal models [72]. As we observed signi cant difference between Bacterides fragelis abundance in both cases and controls with higher RMR, it seems that, in contrast to previous studies proposing obesogenic role of B. fragilis [73,74], it may impose protective effects against obesity by improving RMR. As we discussed earlier, butyrate can increase energy expenditure via aforementioned mechanisms, and B. fragilis, that promotes mucin production, is a propionate producer, which is a precursor of butyrate [75].

Conclusions
In conclusion, the results of this study showed, for the rst time, that despite the absence of signi cant differences in abundance of, A. muciniphila, Bi dobactrium, Prevotella, and E. coli between subjects with normal RMR and low RMR, the abundance of F. prausnitzii, B. fragilis and Firmicutes phylum were signi cantly higher in fecal samples of individuals with normal RMR. In addition, based on the correlation of studied bacteria and RMR, our ndings showed a positive association between F. prausnitzii, B. fragilis and Firmicutes phylum, as well as lactobacillus, with RMR. Finally, by determining the gut microbiota effectiveness on host energy expenditure, proper strategies can be designed to manage obesity based on each targeted population. The main strength of the current study is that, for the rst time, the relationship between gut microbiota and resting metabolic rate in Iranian overweight /obese women by matching two important confounders, age, and BMI, has been investigated. Moreover, we assessed dietary intake, body composition, leptin, ghrelin, and insulin levels, and physical activity of all participants, allowing a comprehensive pro le of all participants. Nevertheless, despite the strengths noted above, there are some limitations of the present study. Regarding the case-control design of this study, we are not able to discern any cause and effect association between gut microbiota and energy expenditure, and further prospective longitudinal studies and clinical trials are needed. Although we matched cases and controls by age and BMI, there are likely some other confounders which might affect our ndings.

Declarations
Ethics approval and consent to participate The study protocol was approved by the Ethics Commission of Tehran University of Medical Sciences (IR.TUMS.VCR.REC.1398.562). Then, written informed consent was obtained from all patients. All methods were performed in accordance with the declaration of Helsinki.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Authors' contributions
KHM and SDS supervised the study; SM and SH collected the data; SM, SAB and AM did experimental analysis; SM and MSY analyzed the data; SM wrote the rst draft with contributions from the other authors and CCTC, revised the manuscript. All authors reviewed and commented on subsequent drafts of the manuscript. See image above for gure legend