Indonesian Brown rice improves obesity parameters and dysbiosis of gut microbiota

Abstract

BACKGROUND:

Brown rice contains nutrients that significantly reduce the incidence of obesity. This study aimed to investigate the effect of brown rice as a functional food on the reduction of obesity incidence through the pathways of gut microbiota dysbiosis.

METHODS:

In this study, we used white rats (Rattus norvegicus albus), which were divided into five groups, i.e., Normal, High fructose feed diet (HFFD), HFFD + Brown rice (BR) I, HFFD + BR II, HFFD + BR III. The parameters were SCFA concentration, FFAR3 expression, and Firmicutes– Bacteroidetes ratio.

RESULTS:

The rats fed HFFD + BR III diet with a high intake of brown rice resulted in a greater reduction in abdominal circumference. The group of rats fed the HFFD had a higher BFI than the other rats. The brown rice intervention reduced the Lee index, a higher concentration of short-chain fatty acid (SCFA), and led to a higher reduction in Firmicutes– Bacteroidetes ratio. The brown rice intervention also increased the FFAR3 expression in the rat ileal L cells.

CONCLUSIONS:

Brown rice has significant benefits for reducing obesity, as evidenced by the improvement in the abdominal circumference, Lee index, and BFI through the improvement of intestinal dysbiosis and increase in SCFA concentration and FFAR3 expression.

Keywords

Brown rice FFAR3 gut microbiota obesity SCFA

1 Introduction

Obesity is a condition associated with excessive intake of energy and nutrients. It is an epidemic disease that is becoming a global health concern. Obesity is characterized by abnormal or excessive fat accumulation in the adipose tissue [1, 2]. Moreover, it harms the serum vitamin and mineral levels in the body [3, 4]. Obesity is a consequence of multifactorial etiologies that involve various factor of energy imbalance in te body.

According to the WHO database, the worldwide obesity rate (aged 18 years or older) in 2014 was double that in 1980. It is estimated that 2030, up to 57.8% of the world’s adult population will be overweight or obese [5, 6]. The prevalence of obesity respectively in the United States is very high, with adults and children accounting for 34.9% and 17% of the total population in 2012 [7]. It has been predicted that approximately 42% of the US population will be obese by 2030 [8]. The Indonesia Basic Health Research demonstrated that the prevalence of adult obesity increased from 14.8% (2013) to 21.8%, whereas central obesity increased from 18.8% in 2007 to 31% in 2018 [9].

New evidence suggests intestinal microbiota plays a potential role in the pathophysiology of obesity. The ratio of Firmicutes and Bacteroidetes, the two dominant gut phylum bacteria, increases in obesity [10–12]. The role of intestinal microbiota in the etiology of obesity is mediated through the mechanism of short-chain fatty acid (SCFA) production. An increasing number of Firmicutes leads to less energy expenditure. Fermentation of undigested carbohydrates and dietary fiber results in SCFA production, which can affect hormonal status through PYY and GLP-1 and G protein receptor 41/FFAR3, leading to alteration in satiety and food intake [13, 14].

One of the treatments for obesity is weight loss, in which the Firmicutes– Bacteroidetes ratio is decreased. Brown rice contains higher fiber and magnesium compare with white rice [15]. Aside from its low production cost, brown rice is easily available and accessible compared to red or black rice [16]. There are numerous studies on brown rice; however, how it affects obesity through the improvement of intestinal microbiota and dysbiosis has never been investigated. This study aimed to examine the effect of brown rice as a functional food on the reduction of obesity incidence through the pathway of intestinal microbiota dysbiosis.

2 Materials and methods

2.1 Study design

This study has a posttest-only control group design. The study protocol was approved by the research ethics committee of the Faculty of Medicine, Universitas Brawijaya Malang, Indonesia (Decree No. 210/EC/KEPK-S3/09/2018). Furthermore, the study was conducted from December 2018 to December 2019 at the Institute of Biosciences, Central Laboratory of Life Sciences, Mathematics, and Natural Sciences, Agricultural Product Technology, Anatomical Pathology, Biochemistry and Biomedics, Universitas Brawijaya Malang, and the Agricultural Technology Laboratory Gadjah Mada University, Yogyakarta.

2.2 Experimental animal

The experimental animals were 35 healthy male Sprague Dawley rats aged 2.5– 3 months old (70– 90 days postnatal) and weighed 150– 250 g. The number of samples were calculated according to Federer’s formula. The experimental test is (t-1)(n-1)≥15, where t is the number of intervention groups, and n is the number of repetitions or the number of samples in each group. The number of intervention groups in this study was 5 groups with the types of interventions given as follows:

Group	Intervention
I	Feed with normal diet (N)
II	High fat high fructose diet (HFFD)
III	High fat diet high fructose + brown rice I (HFFD + BR I)
IV	High fat high fructose diet + brown rice II (HFFD + BR II)
V	High fat high fructose diet + brown rice III (HFFD + BR III)

Note: (1) N group, rats given normal diet (AIN 93-Modification); (2) high-fat-and-fructose diet (HFFD) group, rats given high-fat diet (AIN 93-Modification) + 30% fructose drink; (3) HFFD + BR I group, rats given HFFD with dose 1 BR substitution (112.5-g/1000-g feed or 12.43% feed weight); (4) HFFD + BR II group, rats given HFFD with dose 2 brown rice substitution (225-g/1000-g feed or 24.86% feed weight); and (5) HFFD + BR III group, rats given HFFD with dose 3 brown rice substitution (337.5-g/1000-g feed or 37.29% feed weight).

Since the number of treatment groups (t) = 5, so the number of replications is at least 5 samples. It was added with a spare experimental unit in each group to anticipate unwanted possibilities such as death. The correction for the number of replications based on the Higgins formula is 1/(1-f), with an estimate of the experimental unit dropping out (f) of 25%. Thus, the number of rats needed was 35 rats.

The brown rice substitution was equivalent to the human meal portion of 125 g of rice at each meal. Dose 1 was equivalent to 1 portion of brown rice/day; dose 2 was 2 portions of brown rice/day; and dose 3 was 3 portions of brown rice/day). The energy and nutrient contents of the feed are presented in Table 1 [13, 14].

Table 1
Energy and nutrient contents of the 1000-g feed

Energy and nutrients Diet type

N HFFD HFFD + BR I HFFD + BR II HFFD + BR III

Carbohydrate (%) 53.40 27.34 27.34 27.34 27.34

Protein (%) 24.76 19.93 19.93 19.93 19.93

Fat (%) 21.84 52.73 52.73 52.73 52.73

Magnesium (ppm) 17.00 15.00 40.76 66.52 92.51

Fiber (g) 28.42 24.25 26.71 29.19 31.68

Energy (Cal) 4090.00 4820.00 4820.00 4820.00 4820.00

Energy density (Cal/g) 4.09 4.82 4.82 4.82 4.82

Energy and nutrients	Diet type
Carbohydrate (%)	53.40	27.34	27.34	27.34	27.34
Protein (%)	24.76	19.93	19.93	19.93	19.93
Fat (%)	21.84	52.73	52.73	52.73	52.73
Magnesium (ppm)	17.00	15.00	40.76	66.52	92.51
Fiber (g)	28.42	24.25	26.71	29.19	31.68
Energy (Cal)	4090.00	4820.00	4820.00	4820.00	4820.00
Energy density (Cal/g)	4.09	4.82	4.82	4.82	4.82

2.3 Administration of brown rice to obese rat

The study began with the induction of obesity in the experimental animals (Phase 1) by feeding them HFFD for 14 weeks in 4 groups, as defined above and it proceeded with the intervention research (Phase 2), which lasted 8 weeks. The brown rice intake was measured by weighing the amount of feed given and the remaining feed; body weight was measured by weighing once per week. Other anthropometric parameters, including abdominal circumference, Lee index, and body fat index (BFI), were compared before and after brown rice intervention. The Lee index was calculated using the following formula: [body weight (g)^1/3] / [body length (cm)×1,000]; the BFI was calculated based on the weight of white adipose tissue (WAT)/100-g body weight.

2.4 qPCR

After the sacrifice process, the Firmicutes– Bacteroidetes ratio was analyzed from the cecum via qPCR. DNA was previously isolated from the rats’ cecum digesta using the FastDNA Spin Kit for Soil Cat Number 116560000 from MP Bio with slight modification. qPCR with duplication was performed on the DNA isolation results in a 10-L volume using SsoFast EvaGreen Supermix (Bio-Rad Laboratories) on the Bio-Rad CFX 96 real-time PCR detection system. The system was run at a standard DNA concentration (50 ng/μL). The specific primers for 16 S rRNA from the bacterial taxa used are presented in Table 2.

Table 2
List of primers used in qPCR

Group Targeted Primer Sequence 5’ to 3’ Amplicon size

Bacteroidetes F CGATGGATAGGGGTTCTGAGAGGA 238

Bacteroidetes R GCTGGCACGGAGTTAGCCGA

Firmicutes F GCGTGAGTGAAGAAGT 126

Firmicutes R CTACGCTCCCTTTACAC

Universal Eubacteria F ACTCCTACGGGAGGCAGCAGT 174-199

Universal Eubacteria R GTATTACCGCGGCTGCTGGCAC

Group Targeted Primer	Sequence 5’ to 3’	Amplicon size
Bacteroidetes F	CGATGGATAGGGGTTCTGAGAGGA	238
Bacteroidetes R	GCTGGCACGGAGTTAGCCGA
Firmicutes F	GCGTGAGTGAAGAAGT	126
Firmicutes R	CTACGCTCCCTTTACAC
Universal Eubacteria F	ACTCCTACGGGAGGCAGCAGT	174-199
Universal Eubacteria R	GTATTACCGCGGCTGCTGGCAC

All primers presented in Table 2 were synthesized by Integrated DNA Technology, Canada, at an annealing temperature of 60°C. qPCR was performed under cycling conditions of denaturation at 98°C for 2 min. It was then followed by 39 cycles of denaturation at 98°C for 5 s and then an annealing temperature of 60°C for 31 s. The cover temperature used was 105°C. The result of the calculation of the Firmicutes– Bacteroidetes ratio was derived from the relative expression values for bacterial taxa normalized to the total bacteria present and amplified using universal Eubacteria primers.

2.5 SCFA concentration from the cecum was measured via gas-liquid chromatography

SCFA was analyzed using gas-liquid chromatography followed by Shimadzu GC 2010 Plus. The analysis was started by taking 4 mg of the sample centrifugation at 10,000 rpm for 15 min. The supernatant (2 mL) was added into a small 5-mL plastic tube; 30 mg of 5-sulfosalicylic acid was added also the solution was shaken, centrifuged at 3000 rpm for 10 min at 4°C, and filtered using a Millipore filter until a clear liquid was obtained. The supernatant (1μL) was injected into the GC apparatus using a microsyringe. After 9 min, the area of the compound was drawn on the recorder paper. Before the sample was injected, a mixture of acetate, propionate, and butyrate standard solutions with concentrations of 0.025%, 0.05%, and 0.3% were prepared. Next, the regression equation, which describes the association between the area of standard acetic, propionic, and butyric acids (Y) and the concentrations of standard acetic, propionic, and butyric acids (X), was calculated [20–23].

2.6 FFAR3 expression in the ileum measured via immunohistochemistry (IHC)

The embedding process was carried out on the ileal tissue sample after that tissue in the paraffin block was cut and placed on an object glass. Deparaffinization and rehydration were performed by dipping the organ correspondingly into xylol twice, graded alcohol (100%, 90%, 80%, 70%, and 30%) for once, and distilled water, followed by washing it with phosphate-buffered saline (PBS) (pH 7.4) for 3×5 min. The cells were also washed once with PBS. Subsequently, the cells were incubated at 3% H₂O₂ in PBS for 10 min at room temperature and then washed thrice with PBS. The cells were incubated in 1% bovine serum albumin (BSA) for 1 h at room temperature and then washed thrice with PBS. The primary antibody was diluted in goat serum/FBS/BSA at the desired concentrations and volumes (GPER41 Antibody (AA 320-360) ABIN685717 in goat serum or FBS from antibodies). The cells were incubated in primary antibodies at 4°C for 12 h or at room temperature for 2 h and then washed with PBS for 3×5 min. The biotin-labeled secondary antibody was diluted in PBS at the desired concentrations and volumes (biotin-labeled anti-rabbit IgG, SIGMA 1 : 500 in PBS), followed by incubation of the cells in secondary antibody for 1 h at room temperature. The slides were washed with PBS for 3×5 min and dripped with streptavidin horseradish peroxidase (1 : 500) in PBS for 40 min. After washing with PBS for 3×5 min, diaminobenzidine was dripped for 10 min. The slides were washed with distilled water for 3×5 min, followed by with Mayer’s hematoxylin for 10 min. Subsequently, the slides were washed with tap water, and then distilled water for 10 min, and left to stand at room temperature. Each slide was labeled, and the mounting medium (entellan) was dripped onto the slide. The cover glass was put on the preparation after being given a mounting medium.

After immunohistochemical staining, the preparations for each treatment and each replicate were photographed using Nikon E100, 400×photo magnification was performed using a Sony A7 camera, and calculation was carried out on 20° fields of view with 1000×and 400×photo magnifications. The results were then analyzed to determine the intensity of the color expression of GPR41/FFAR3 [17, 24, 25].

2.7 Statistical analysis

The data obtained were presented descriptively and analyzed statistically using paired t-test to compare the parameters pre-and post-design. The differences among the groups were analyzed via one-way analysis of variance, followed by a posthoc Tukey HSD test for parametric data and Mann– Whitney U test continued with Kruskal Wallis test for nonparametric data. The correlation between brown rice intake and anthropometric parameters was analyzed using Pearson’s correlation test. The results were considered significant if P < 0.05. Analysis was conducted using IBM SPSS Statistics 20.0 for Windows.

3 Results

3.1 Food intake

The food intake from this study that also has been previously reported [14] indicated that the average energy intake of the N group rats was the lowest compared with those in the other groups since they did not consume fructose solution (significant differences, P = 0.000). Among the rats’ given fructose solution drinks, the rats in the HFFD + BR III group had lower energy intake than those given HFFD and brown rice doses I and II; however, the differences were not significant (P > 0.05). The higher the dose of brown rice given, the lower the energy intake. The higher dose of brown rice contributed to the increase in fiber intake (Table 3).

Table 3
Food intake per 24 h during 8 weeks of intervention

Characteristic Group

N HFFD HFFD + BR I HFFD + BR II HFFD + BR III

Brown rice (% food weight) 0.00^a 0.00^a 0.145±0.02^b 0.23±0.03^c 0.33±0.03^d

Fiber (g) 3.86±0.79^c 2.75±0.28^a 2.86±0.28^b 2.89±0.32^c 3.17±0.28^c

Total Energy (Kcal) 55.62±11.38^a 80.47±7.75^b 86.82±7.01^b 78.07±5.25^b 76.21±5.03^b

Characteristic	Group
Brown rice (% food weight)	0.00^a	0.00^a	0.145±0.02^b	0.23±0.03^c	0.33±0.03^d
Fiber (g)	3.86±0.79^c	2.75±0.28^a	2.86±0.28^b	2.89±0.32^c	3.17±0.28^c
Total Energy (Kcal)	55.62±11.38^a	80.47±7.75^b	86.82±7.01^b	78.07±5.25^b	76.21±5.03^b

One-way ANOVA, significant differences marked by different letter notations (a, b, c, d).

3.2 Abdominal circumference

There were improvements in the obesity parameters, such as decreased abdominal circumference, Lee index, and BFI. Abdominal circumference is an anthropometric parameter that indicates the condition of visceral obesity. The rats from the N and HFFD groups experienced an increase in abdominal circumference, whereas those fed the brown rice diet experienced the opposite. The high intake of brown rice by the rats fed the HFFD + BR III diet resulted in a greater than reduction in abdominal circumference compare with that in the other two groups fed lower doses of brown rice; however, this difference was not significant. The changes in the abdominal circumference of the rats after brown rice intervention are presented in Fig. 1 by the group.

Fig. 1

Changes in the abdominal circumference of rats after brown rice intervention.

3.3 Lee index

Another parameter used to assess the success of obesity therapy in experimental animal models of obesity is the Lee index. The results indicated that the brown rice intervention led to a higher reduction in the Lee index compared with the Lee index of rats receiving only HFFD. Rats from the N and HFFD groups experienced an increase in the Lee index, whereas those fed HFFD + BR diet experienced the opposite. The higher the dose of brown rice, the lower the Lee index escalation. Compare with the initial Lee index before thhe brown rice intervension, the group of rats fed the brown rice diet had a lower Lee index. The changes in the Lee index are presented in Fig. 2 by the group.

Fig. 2

Changes in the Lee index of rats after brown rice intervention.

3.4 Body fat index

The same anthropometric parameter in humans is known as subcutaneous fat tissue, but in rats, it can be done by calculating BFI. BFI is calculated by summing up the total epididymal, perirenal, omental, and inguinal fat deposits, known as WAT/100 g of body weight. Fig. 3 demonstrates that the BFI of the rat group with the intervention doses of brown rice 1, 2, and 3 (HFFD + BR I, II, and III) was the same as that of the group of rats fed the N diet and significantly different from that of the group of rats fed the HFFD. The rats fed the HFFD had a higher BFI than the other rats.

Fig. 3

Body fat index of rats after brown rice intervention.

3.5 Intestinal microbiota dysbiosis

The results indicated that the group of rats fed HFFD had the highest number of Firmicutes, with an average of 3.97±0.89. Brown rice intervention was shown to reduce the number of Firmicutes; the higher the dose given, the lower the number of Firmicutes, which was observed in the HFFD + BR III diet group with an average of 0.46±0.19. In contrast to the number of Firmicutes, the number of Bacteroidetes in the HFFD group was lower than that in rats with brown rice intervention; the higher the dose of brown rice consumed by the rats, the higher the number of Bacteroidetes.

The calculation results of the Firmicutes– Bacteroidetes ratio indicated that the group of rats given HFFD had the highest ratio, whereas the addition of brown rice intervention to the HFFD could reduce this ratio. The higher the dose of brown rice given, the lower the Firmicutes– Bacteroidetes ratio. There were also improvements in intestinal microbiota dysbiosis, as indicated by a smaller Firmicutes– Bacteroidetes ratio (see Fig. 4).

Fig. 4

Firmicutes– Bacteroidetes ratio after brown rice intervention.

3.6 SCFA

SCFA analysis was conducted via gas-liquid chromatography. The most common SCFA are acetate, propionate, and butyrate. The results of the SCFA analysis on the cecum digesta of rats given the N diet, HFFD, and brown rice intervention doses 1, 2, and 3 are presented in Table 4.

Table 4
Concentration of acetate, propionate, butyrate, and total SCFA in the digesta cecum of rats by group

SCFA Group

N HFFD HFFD + BR I HFFD + BR II HFFD + BR III

Acetate (mMol/g) 28.9±10.32^b 14.55±4.05^a 20.67±4.04^ab 21.57±4.81^ab 28.66±7.22^b

Propionate (mMol/g) 6.76±2.93^a 4.94±1.52^a 5.44±1.34^a 6.05±2.52^a 820±1.78^a

Butyrate (mMol/g) 3.48±0.36^b 2.37±0.85^a 2.66±0.50^ab 8.20±0.50^ab 3.57±0.71^b

SCFA total (mMol/g) 39.02±12.87^b 21.85±4.40^a 28.77±5.36^ab 30.57±7.38^ab 40.42±7.25^b

SCFA	Group
Acetate (mMol/g)	28.9±10.32^b	14.55±4.05^a	20.67±4.04^ab	21.57±4.81^ab	28.66±7.22^b
Propionate (mMol/g)	6.76±2.93^a	4.94±1.52^a	5.44±1.34^a	6.05±2.52^a	820±1.78^a
Butyrate (mMol/g)	3.48±0.36^b	2.37±0.85^a	2.66±0.50^ab	8.20±0.50^ab	3.57±0.71^b
SCFA total (mMol/g)	39.02±12.87^b	21.85±4.40^a	28.77±5.36^ab	30.57±7.38^ab	40.42±7.25^b

Table 4 demonstrates that the group of rats given brown rice intervention had a significantly higher concentration of SCFA in the digesta cecum than those given HFFD.

3.7 FFAR3

SCFA is a ligand of the FFAR3 receptor. The FFAR3 expression was measured via IHC on the L cells of the distal ileal tissue using the GPER-1 antibody (AA 320-360) ABIN685717 from the antibodies. The highest levels of FFAR3 were observed in the distal small intestine (ileum). The immunohistochemical results were consulted by an anatomical pathologist. Furthermore, from these results, the number of cells was calculated or quantified using a Nikon E100 microscope, and 400× photo magnification was performed using a Sony-A7 camera. Calculations were performed at 1000× magnification with a 2° field of view. The immunohistochemical results of the FFAR3 expression are presented in Fig. 5. The calculation of the cells number of is presented in Table 5. In Fig. 5, blue indicates the result of FFAR3 staining, whereas brown indicates that FFAR3 is expressed due to SCFA. The arrows indicate FFAR3, which is secreted by SCFA.

Fig. 5

Image of the FFAR3 expression in the L cells of rat ileal tissue. Photos were taken using a Sony-A7 camera with 400×magnification. Arrows (brown) indicate the FFAR3 expression in rat villous ileal L cells. The immunohistochemical results indicated that brown rice intervention increased the FFAR3 expression in the rat ileal L cells. Note: (a) N group; (b) HFFD group; (c) HFFD + BR I group; (d) HFFD + BR II group; (e) HFFD + BR III group.

Table 5

The number of cells resulting from the FFAR expression

Number of cells per field of view	Group
	N	HFFD	HFFD + BR I	HFFD + BR II	HFFD + BR III
Average amount of cell	15.00±1.79^c	4.83±2.32^a	8.50±2.51^b	13.33±1.86^c	14.67±1.51^c

4 Discussion

HFFD is recommended in creating experimental animal models of obesity [23]. Based on research results, a high intake of foods and drinks containing high-fructose corn syrup (HFCS) increases the prevalence of central obesity. HFCS is widely used in a variety of food and beverage products, such as soft drinks, pastries, cookies, gums, jelly, and desserts, which are much enjoyed by the public. Many popular beverage brands use 55% – 65% of HFCS. Bodyweight gain due to HFCS (55% fructose) intake is accompanied by an increase in adipose tissue, especially in the visceral region [27, 28]. Another effect of fructose overconsumption is the expression of leptin resistance in the brain. Fructose decreases the expression of cholecystokinin and growth hormones in the ventromedial nucleus. Thus, long-term fructose consumption increases calorie intake due to the loss of satiety signals in the brain, which enventually leads to overweight ness and obesity [29]. Along with an increase in fructose consumption, an increase in fat consumption also occurs. An HFFD is likely to induce obesity, as demonstrated in vivo experimental studies, which is pathophysiological, like human diseases [33, 34]. The latest study showed that the intake of HFCS-containing drinking water in mice resulted in considerable changes in the colonic microbiota structure and body fat content [28]. Furthermore, HFCS can contribute to metabolic disorders and altered dopamine function independent of weight gain and high-fat diets [33].

There was a significant difference (P = 0,000) in the average energy intake per 24 h during the intervention between rats in the control group and the HFFD group (see Table 3). Despite the lower feed intake (grams), the rats in the HFFD group still had higher energy intake due to the consumption of fructose solution, which had an energy content of 1.125 calories/mL, resulting in a higher energy intake in the HFFD [34, 35]. Another research result proved that consuming sugary drinks is associated with excessive calorie intake [36]. The dietary fiber can increase energy expenditure but does not change energy intake in obese-induced rats. In the groups of rats given brown rice intervention, their food intake tended to be lower [37]. HFFD generally consists of ingredients with high carbohydrate content and low fiber [38]. Dietary fiber can induce gastric distension, decreased appetite, and increased satiety [39]. The consumption of dietary fiber reduces the secretion of leptin and gastric ghrelin but increases the concentration of GLP-1 in a high-fat diet [40].

We investigated the changes in abdominal circumference, Lee index, and BFI as the obesity parameters. We found that the abdominal circumference increased in the N and HFFD rat groups but decreased in rats given brown rice intervention (Fig. 1). This result is in line with the previous study that brown rice intervention significantly reduced body weight, BMI, body fat percentage, and abdominal circumference in type 2 diabetes mellitus patients [41]. Excessive calorie intake by rats fed HFFD due to fructose intake causes triglyceride accumulation in the liver and other adipose tissues as an energy reservoir, resulting in adipocyte hypertrophy [42]. Adipose tissue has an extra role as an endocrine organ involving glucocorticoid (steroid hormone) metabolism. Dysregulation of glucocorticoid metabolism promotes obesity, which commonly occurs in the visceral region, named central obesity [35, 43]. Several studies also demonstrated that HFFD induces central obesity rather than just weight gain [36, 44, 45]. The improvement in abdominal circumference is related to the dietary fiber in brown rice. A study demonstrated that a low-fiber diet significantly increased abdominal circumference [46]. Another study revealed that fiber intake exhibited a downward trend regarding abdominal circumference [47].

In addition, this study demonstrated improvement in the Lee index in rats given brown rice intervention. Rats fed with HFFD had a higher Lee index score than rats in the control group (Fig. 2). In this study, the mean body weight of the rats given HFFD tended to be higher than those given a normal diet. The body length of the rats given HFFD was significantly shorter than those given a normal diet. One of the causes of the shorter body length of rats given HFFD was the lower protein content of the feed (– 4.83%) compared with that of a normal diet. Protein has carbon, hydrogen, oxygen, and nitrogen elements, not fats and carbohydrates. Previous studies on experimental animals have confirmed the influence of nitrogen and essential amino acids, potassium, sodium, magnesium, zinc, phosphorus, and water on body length growth [48]. Another demonstrated that high protein intake, especially from animals, led to better height increase in 1– 9-year-old children [49]. This effect can be explained by the fact that brown rice, besides its high-fiber content, is also rich in manganese, magnesium, and potassium. There was an adverse relationship between the serum levels of manganese, magnesium, and potassium and the BFI measurements; the lower the serum levels of these minerals, the higher the BFI [50].

The BFI of rats in the HFFD group was significantly higher than in the N group. In contrast, the BFI of rats given brown rice intervention showed no significant improvement compared to the N group (Fig. 3). Likewise, the Shiitake mushroom administration with high fiber, soluble in water, and β-glucans could reduce WAT in rats given a high-fat diet [51]. This effect is related to brown rice’s fiber content, which can reduce lipogenesis in WAT. Administration of a high-fiber diet would reduce 30% of the lipogenesis in WAT and increase the PKA, PGC-1 α, and UCP1 protein expressions involved in the lipolysis process [52]. Dietary fiber also increases the protein expression of alpha and gamma receptors in PPAR, liver X receptor alpha, and ATP A1-binding transporter in the target tissue [53].

The number of Firmicutes obtained from the calculation of the log fold change shows that the rats in the HFFD group had the highest number of Firmicutes bacteria (Fig. 4). A high-fat diet creates an environment for Firmicutes bacteria where they can live comfortably in the intestine. Firmicutes utilize fat from foods to increase lipopolysaccharide in blood plasma, leading to obesity. In addition, Firmicutes induce lipoprotein lipase activity to increase fat absorption and storage in peripheral tissues. Likewise, high-fructose feeding will lead to the creation of an environment highly favored by Firmicutes. Brown rice intervention decreases the number of Firmicutes and increases the number of Bacteroidetes. The result of this study is following that restriction of fat intake increases the number of Bacteriodetes, which are initially 3% to 15%. Our study demonstrated that the Firmicutes– Bacteroidetes ratio in the group given the brown rice intervention decreased and had a lower ratio with the increase of brown rice doses given. The 16 S rRNA sequencing revealed that the diversity of intestinal microbiota was influenced by dietary fiber. Dietary fiber induces a decrease in the Firmicutes– Bacteroidetes ratio at the phylum level and an increase in the relative abundance of the Roseburia genera at the genus level. These findings suggest that dietary fiber increases energy homeostasis and prevents obesity by increasing intestinal microbiota diversity and beneficial bacterial colonization [37].

Diet and intestinal environment interact in complex ways with bacterial populations in the intestine. Our results indicated that rats in the HFFD + BR III group had higher SCFA concentrations (acetate, butyrate, and propionate) in the cecum digesta than those in the HFFD group. The SCFA level of the group given the highest dose of brown rice did not significantly differ from that of rats in the N group. Brown rice intervention increases SCFA production starting from the lowest dose. Previous studies showed that a high-fiber, low-fat diet leads to a higher amount of fecal SCFA than a low-fiber diet [54, 55]. Another study found that the fecal SCFA concentration of obese children was lower than those of children with normal weight (P = 0.04) [56]. The protective effect of SCFA on metabolic changes induced by a high-fat diet seems to depend on the regulation activated by PPARγ, thus promoting the change from lipid synthesis to lipid oxidation [57]. Another mechanism has been proposed to explain this effect: the activation of signaling pathways is mediated by protein kinases, such as AMP-activated protein kinases [57, 58]. MAPK-activated protein kinases [59]. Butyrate and propionate have been reported to induce intestinal hormone production, thereby reducing food intake [60]. Acetate has also been proven to reduce appetite, in this case, through interactions with the central nervous system [61].

Furthermore, SCFA affects the adipogenesis of adipose tissue through interactions with FFAR3 as a mediator of adipocyte development and differentiation, resulting in smaller adipocyte size and lower pro-inflammatory cytokine secretion. This observation may have important implications for obesity characterized by large adipocytes and systemic inflammation. SCFA is an endogenous ligand of G-protein-coupled receptor/GPR41 or free fatty acid receptor/FFAR3 [62]. The FFAR3 expression is related to the efficiency of calorie extraction from food, PYY secretion, and other anorexigenic hormones. Also, it regulates intestinal motility associated with inhibiting transit along the intestine. Our study’s results indicate that these receptors’ expression is higher in rats given brown rice intervention while significantly lower in rats fed HFFD. Provision of HFFD has been shown to reduce the L cell activity in the ileum, thus reducing the number of active FFAR3 receptors [63]. SCFAs are products from dietary fiber fermentation in the brown rice intervention diet food as a ligand from FFAR3 receptors on ileum L cells. The higher the dose of brown rice given, the higher the expression of FFAR3. SCFA also determines FFAR3 induction in PYY secretion, which can inhibit gastric emptying and cause longer intestinal transit time [64].

Brown rice was considered a functional food owing to its several advantages over white rice, such as its higher fiber content (soluble and insoluble fiber) and its mineral content (magnesium, potassium, and manganese) that, which is almost seven times higher [18]. Brown rice and black rice are less attractive to the public because they have a firm texture and an unpleasant taste [65]. This study followed the previous research to provide brown rice interventions for 8 weeks to adjust [66, 67].

5 Conclusions

This study demonstrated that the provision of brown rice intervention to obese rats induced by feeding HFFD for 14 weeks were given significant benefits for improving obesity. It was evidenced by improvement in the obesity parameters, such as abdominal circumference, Lee index, and BFI, through the improvement of intestinal dysbiosis and increase in SCFA concentration and FFAR3 expression. The dose of this study is representative of the amount of food intake of as much as 340-g brown rice/day for humans. The limitation of this study is the diversity of the gut microbiota was not performed in the obese rats model. Whereas, the diversity of microbiota is closely related to obesity and metabolic disorders caused by obesity. The pH of the cecum was also not measured, including the levels of PYY and GLP-1 as incretin hormones. The excretion is affected by FFAR3 expression, which were also not measured in this study. Furthermore, it is necessary to carry out further research and analysis on the relationship between intestinal microbiota dysbiosis and serum mineral levels associated with obesity. Hence, further research to prove the real benefits of brown rice is necessary.

Footnotes

Acknowledgments

The authors would like to thank the Ministry of Health of Indonesia for funding this study, as well as all parties who were directly or indirectly involved in this research.

Funding

This research was funded by the Ministry of Health of Indonesia (No.02.02/H.V/SK/169/2016).

Conflicts of interest

The authors have no conflicts of interest to report.

Authors’ contributions

ES and DH have given substantial contributions to the conception or the design of the manuscript. SS, XFH, and AR contributed to acquisition, analysis, and interpretation of the data. All authors have participated in drafting the manuscript, and author A revised it critically. All authors read and approved the final version of the manuscript. All authors contributed equally to the manuscript and read and approved the final version of the manuscript.

References

Apovian

. Obesity: definition, comorbidities, causes, and burden. Am J Manag Care. 2016;22:s176–185. PMID: 27356115.

Baqai

, Wilding

JPH

. Pathophysiology and aetiology of obesity. Medicine. 2015;43:73–76. doi: 10.1016/j.mpmed.2014.11.016.

Bird

, Ronnenberg

, Choi

S-W

, Du

, Mason

, Liu

. Obesity is associated with increased red blood cell folate despite lower dietary intakes and serum concentrations. J Nutr. 2015;145:79–86. doi: 10.3945/jn.114.199117.

Geiker

NRW

, Veller

, Kjoelbaek

, Jakobsen

, Ritz

, Raben

, et al. Effect of low energy diet for eight weeks to adults with overweight or obesity on folate, retinol, vitamin B12, D and E status and the degree of inflammation: a post hoc analysis of a randomized intervention trial. Nutr Metab. 2018;15:24. doi: 10.1186/s12986-018-0263-1.

Forse

, Krishnamurty

. Epidemiology and discrimination in obesity. In: Nguyen NT, Blackstone RP, Morton JM, Ponce J, Rosenthal RJ, editors. The ASMBS Textbook of Bariatric Surgery New York. NY: Springer New York; 2015. p. 3–12.

Singh

, Chang

H-W

, Yan

, Lee

, Ucmak

, Wong

, et

. Influence of diet on the gut microbiome and implications for human health. J Transl Med. 2017;15:73. doi: 10.1186/s12967-017-1175-y.

Ogden

, Carroll

, Kit

, Flegal

. Prevalence of obesity and trends in body mass index among US children and adolescents 1999–2010. JAMA. 2012;307:483–490. doi: 10.1001/jama.2012.40.

Finkelstein

, Khavjou

, Thompson

, Trogdon

, Pan

, Sherry

, et

. Obesity and severe obesity forecasts through Am J Prev Med. 2012;42:563–570. doi: 10.1016/j.amepre.2011.10.026.

Kemenkes-

. Hasil Utama Riskesdas [Main Results of Riskesdas [2018] [Main Results of Riskesdas 2018]. Jakarta: Balitbangkes Kemenkes RI; 2018.

10.

Bervoets

, Van Hoorenbeeck

, Kortleven

, van

Noten C

, Hens

, Vael

, et al. Differences in gut microbiota composition between obese and lean children: a cross-sectional study. Gut Pathogens. 2013;5:10. doi: 10.1186/1757-4749-5-10.

11.

Ismail

, Ragab

, Abd

ElBaky A

, Shoeib

ARS

, Alhosary

, Fekry

. Frequency of Firmicutes and Bacteroidetes in gut microbiota in obese and normal weight Egyptian children and adults. Arch Med Sci. 2011;7:501–507. doi: 10.5114/aoms.2011.23418.

12.

Jandhyala

, Talukdar

, Subramanyam

, Vuyyuru

, Sasikala

, Reddy

. Role of the normal gut microbiota. World J Gastroenterol. 2015;21:8787–8803. doi: 10.3748/wjg.v21.i29.8787.

13.

Kusumastuty

, Handayani

, Hanifa

, Lisan

, Sulistyowati

. The effect of brown rice on superoxide dismutase level and nonalcoholic fatty liver in an spraque-dawley rat model of high-fat-high-fructose diet-induced obesity. Open Access. Maced J Med Sci. 2020;8(A):873–880.

14.

Andarini S, Kiwari GL, Handayani D. Magnesium-Rich Indonesian Brown Rice “Sinta Nur” Improves Insulin Sensitivity in High Fat High Fructose Diet-Induced Obesity Spaque Dawley Rats. WSEAS Trans Syst. 2022;21:257-267.

15.

Handayani D, Ramadhan A, Anindyati RD, Inayah AM, Sulistyowati E, Kusumastuty I. Dietary Brown Rice on Carcass Composition and Nitric Oxide (NOx) Metabolite Levels in High-Fat High-Fructose Diet-Induced Spraque Dawley Rats as Obesity Model. J Phys Conf Ser [Internet]. 2020. Oct;1665:012007

16.

Delzenne

, Neyrinck

, Cani

. Gut microbiota and metabolic disorders: how prebiotic can work? Br J Nutr. 2013;109:S81–85. doi: 10.1017/S0007114512004047.

17.

Tolhurst

, Heffron

, Lam

, Parker

, Habib

, Diakogiannaki

, et al. Short-chain fatty acids stimulate glucagon-like peptide-1 secretion via the G-protein-coupled receptor FFAR2. Diabetes. 2012;61:364–371. doi: 10.2337/db11-1019.

18.

Sulistyowati

, Rudijanto

, Soeharto

, Handayani

. The identification of characteristic macro- and micronutrients and the bioactive components of indonesian local brown rice as a functional feed in obesity nutrition therapy. Curr Nutr Food Sci. 2020;16:494–500. doi: 10.2174/1573401315666190328223626.

19.

Lin

J-H

, Singh

, Chang

Y-T

, Chang

Y-H

. Factor analysis of the functional properties of rice flours from mutant genotypes. Food Chem. 2010;126:1108–1114. 10.1016/j.foodchem.2010.11.140

20.

Fernandes

, Su

, Rahat-Rozenbloom

, Wolever

TMS

, Comelli

, Adiposity gut microbiota and faecal short chain fatty acids are linked in adult humans. Nutr Diabetes. 2014;4:e121. doi: 10.1038/nutd.2014.23.

21.

Hald

, Schioldan

, Moore

, Dige

, Lærke

, Agnholt

, et al. Effects of arabinoxylan and resistant starch on intestinal microbiota and short-chain fatty acids in subjects with metabolic syndrome: a randomised crossover study. PloS One, 2016;11:e0159223.

22.

Heinritz

, Weiss

, Eklund

, Aumiller

, Heyer

CME

, Messner

et al. Impact of a high-fat or high-fiber diet on intestinal microbiota and metabolic markers in a pig model. Nutrients. 2016;8:E317.

23.

Salazar

, Dewulf

, Neyrinck

, Bindels

, Cani

, Mahillon

et al. Inulin-type fructans modulate intestinal Bifidobacterium species populations and decrease fecal short-chain fatty acids in obese women. Clin Nutr. 2015;34:501–507.

24.

Indra

. Kultur Adiposit dan Pemeriksaan Adipositokin 2015–Laboratorium Fisiologi, Fakultas Kedokteran, Universitas Brawijaya Malang.

25.

, Kokrashvili

, Mosinger

, Margolskee

. Gustducin couples fatty acid receptors to GLP-1 release in colon. Am. J. Physiol. Endocrinol. Metab. Am J Physiol-Endoc M. 2013;304:E651–660.

26.

Angelova

, Boyadzhiev

. A review on the models of obesity and metabolic syndrome in rats. Trakia J Sci. 2013;11:5–12.

27.

Bocarsly

, Powell

, Avena

, Hoebel

. High-fructose corn syrup causes characteristics of obesity in rats: increased body weight, body fat and triglyceride levels. Pharmacol Biochem Behav. 2010;97:101–106.

28.

Sulistyowati, Etik, Handayani, Dian, Soeharto, Setyowati, Rudijanto, Achmad. A high-fat and high-fructose diet lowers the caecal digesta’s weight and short-chain fatty acid level of a Spraque-Dawley rat model. Turkish J Med [Internet]. 2021;00(00):1-7. Available from: http://Journals.tubitac.gov.tr/medical/.

29.

Senaphan

, Kukongviriyapan

, Sangartit

, Pakdeechote

, Pannangpetch

, Prachaney

. et al. Ferulic acid alleviates changes in a rat model of metabolic syndrome induced by high-carbohydrate, high-fat diet. Nutrients. 2015;7:6446–6464.

30.

Axelsen

, Lademann

, Petersen

, Holstein-Rathlou

N-H

, Ploug

, Prats

et al. Cardiac and metabolic changes in long-term high fructose-fat fed rats with severe obesity and extensive intramyocardial lipid accumulation. Am J Physiol Regul. Integr Comp Physiol. 2010;298:R1560–1570.

31.

Lozano

, Vander Werf

, Bietiger

, Seyfritz

, Peronet

, Pinget

et al. High-fructose and high-fat diet-induced disorders in rats: impact on diabetes risk, hepatic and vascular complications. Nutr Metab. 2016;13:15.

32.

Wang

, Zhu

, Li

. et al. Effects of high fructose corn syrup on intestinal microbiota structure and obesity in mice. npj Sci Food 2022;6:17. https://doi.org/10.1038/s41538-022-00133-7.

33.

Meyers

, Mourra

, Beeler

. High fructose corn syrup induces metabolic dysregulation and altered dopamine signaling in the absence of obesity. PLoS ONE 2017;12(12):e0190206. https://doi.org/10.1371/journal.pone.0190206.

34.

Colley

, Castonguay

. (2015) Effects of sugar solutions on hypothalamic appetite regulation. Physiol Behav. 2015;139:202–209.

35.

Mamikutty

, Thent

, Sapri

, Sahruddin

, Mohd

Yusof MR

& Haji et al. The establishment of metabolic syndrome model by induction of fructose drinking water in male wistar rats. Biomed Res Int. 2014;2014:e263897.

36.

Tappy

, Lê

K-A

. Metabolic effects of fructose and the worldwide increase in obesity. Physiol Rev. 2010;90:23–46.

37.

Wang

, Hong

, Li

, Zang

, Wu

. Soluble dietary fiber improves energy homeostasis in obese mice by remodeling the gut microbiota. Biochem Biophys Res Commun. 2018;498:146–151.

38.

Albertson

, Reicks

, Joshi

, Gugger

. Whole grain consumption trends and associations with body weight measures in the United States: Results from the cross sectional National Health and Nutrition Examination Survey 2001–2012. Nutr J 2016;15:8.

39.

Requena

, Aguilar-González

, Barragán

LAP

, das

Graças Carneiro-da Cunha M

, Correia

, Esquivel

JCC

, et al. Dietary fiber: an ingredient against obesity. Emir JFood Agric. 2016;28:522–530.

40.

Wang

, Zuberi

, Zhang

, Macgowan

, Qin

, Ye

, et al. Effects of dietary fibers on weight gain, carbohydrate metabolism, and gastric ghrelin gene expression in mice fed a high-fat diet. Metab. Clin Ex. 2007;56:1635–1642.

41.

Handayani

, Kusumastuty

, Innayah

, Retnaningtyas

, Sulistyowati

, Sasiarini

, Rudijanto

. Substitution of local Indonesian varieties of brown rice on anthropometry and blood glucose level improvement in type 2 DM patients: a pilot project. J Public Health Res. 2021;11(1):2283.

42.

Zhou

, Qin

. Adipocyte dysfunction and hypertension. Am J Cardiovasc Dis. 2012;2:143–149. doi: 10.1371/journal.pone.0159223.

43.

Gerbaix

, Metz

, Ringot

, Courteix

. Visceral fat mass determination in rodent: Validation of dual-energy X-ray absorptiometry and anthropometric techniques in fat and lean rats. Lipids Health Dis. 2010;9:140. doi: 10.1186/1476-511X-9-140.

44.

Bray

. Soft drink consumption and obesity: it is all about fructose. Curr Opin Lipidol. 2010;21:51–57. doi: 10.1097/MOL.0b013e3283346ca2.

45.

Lee

, Jun

, Kim

, Jeon

, Nam

, Saeed

, et al. Histologic and metabolic derangement in high-fat, high-fructose, and combination diet animal models. Sci World J. 2015;2015;:e306326. doi: 10.1155/2015/306326.

46.

Romaguera

, Angquist

, Du

, Jakobsen

, Forouhi

, Halkjaer

, et al. Dietary determinants of changes in waist circumference adjusted for body mass index—A proxy measure of visceral adiposity. PloS One. 2015:5;e11588. doi: 10.1371/journal.pone.0011588.

47.

Al Hammadi

. Effects of dietary fiber intake on body weight and waist circumference. Arab J Nutr Exerc. 2017;1:77.

48.

Uauy

. Improving linear growth without excess body fat gain in women and children. Food Nutr Bull. 2013;34:257–260. doi: 10.1177/156482651303400222.

49.

Braun

, Erler

, Kiefte-de

Jong JC

, Jaddoe

, van

den Hooven EH

, Franco

, et al. Dietary intake of protein in early childhood is associated with growth trajectories between 1 and 9 years of age. J Nutr. 2016;146:2361–2367. doi: 10.3945/jn.116.237164.

50.

Sulistyowati

, Handayani

, Soeharto

, Rudijanto

. Serum mineral (Mg, Mn, and K) levels are associated with increasing the body mass index (BMI) and abdominal circumference. Obes Med. 2019;15:100107. doi: 10.1016/j.obmed.2019.100107.

51.

Handayani

, Chen

, Meyer

, Huang

. Dietary shiitake mushroom (Lentinus edodes) prevents fat deposition and lowers triglyceride in rats fed a high-fat diet. J Obes. 2011;2011:258051. doi: 10.1155/2011/258051.

52.

Han

S-F

, Jiao

, Zhang

, Xu

J-Y

, Zhang

, Fu

C-L

, et al. Lipolysis and thermogenesis in adipose tissues as new potential mechanisms for metabolic benefits of dietary fiber. Nutrition. 2017;33:118–124. doi: 10.1016/j.nut.2016.05.006.

53.

Han

, Jiao

, Zhang

, Xu

, Wan

, Zhang

, et al. Dietary fiber prevents obesity-related liver lipotoxicity by modulating sterol-regulatory element binding protein pathway in C57BL/6J mice fed a high-fat/cholesterol diet. Sci Rep. 2015;5:15256. doi: 10.1038/srep15256.

54.

Cuervo

, Salazar

, Ruas-Madiedo

, Gueimonde

, González

. Fiber from a regular diet is directly associated with fecal short-chain fatty acid concentrations in the elderly. Nut Res. 2013;33:811–816. doi: 10.1016/j.nutres.2013.05.016.

55.

, Carbonero

, Zoetendal

, DeLany

, Wang

, Newton

, et al. Diet, microbiota, and microbial metabolites in colon cancer risk in rural Africans and African Americans. Am J Clin Nutr. 2013;98:111–120. doi: 10.3945/ajcn.112.056689.

56.

Barczyńska

, Litwin

, Sliżewska

, Szalecki

, Berdowska

, Bandurska

, et al. Bacterial microbiota and fatty acids in thefaeces of overweight and obese children. Pol J Microbiol. 2018;67:339–345. doi: 10.21307/pjm-2018-041.

57.

den Besten

, Bleeker

, Gerding

, van

Eunen K

, Havinga

, van Dijk

, et al. Short-chain fatty acids protect against high-fat diet-induced obesity via a PPARγ-dependent switch from lipogenesis to fat oxidation. Diabetes. 2015;64:2398–2408. doi: 10.2337/db14-1213.

58.

Gao

, Yin

, Zhang

, Ward

, Martin

, Lefevre

, et al. Butyrate improves insulin sensitivity and increases energy expenditure in mice. Diabetes. 2009;58:1509–1517. doi: 10.2337/db08-1637.

59.

Jung

T-H

, Park

, Jeon

W-M

, Han

K-S

. Butyrate modulates bacterial adherence on LS174T human colorectal cells by stimulating mucin secretion and MAPK signaling pathway. Nutr Res Pract. 2015;9:343–349. doi: 10.4162/nr2015.9.4.343.

60.

Lin

, Frassetto

, Kowalik

, Nawrocki

, Lu

, Kosinski

, et al. Butyrate and propionate protect against diet-induced obesity and regulate gut hormones via free fatty acid receptor 3-independent mechanisms. PloS One. 2012;7:e35240. doi: 10.1371/journal.pone.0035240.

61.

Frost

, Cai

, Raven

, Otway

, Mushtaq

, Johnston

. Effect of short chain fatty acids on the expression of free fatty acid receptor 2 (Ffar2), Ffar3 and early-stage adipogenesis. Nutr Diabetes. 2014;4:e128. doi: 10.1038/nutd.2014.25.

62.

Kasubuchi

, Hasegawa

, Hiramatsu

, Ichimura

, Kimura

. Dietary gut microbial metabolites, short-chain fatty acids, and host metabolic regulation. Nutrients. 2015;7:2839–2849. doi: 10.3390/nu7042839.

63.

Lee

E-Y

, Zhang

, Miyamoto

, Kimura

, Taknaka

, Furusawa

, et al. Gut carbohydrate inhibits GIP secretion via a microbiota/SCFA/FFAR3 pathway. J Endocrinol. 2018;239:267–276. doi: 10.1530/JOE-18-0241.

64.

Greiner

, Bäckhed

. Microbial regulation of GLP-1 and L-cell biology. Mol Metab. 2016;5:753–758. doi: 10.1016/j.molmet.2016.05.012.

65.

Kristamtini

. Potensi Pengembangan Beras Merah sebagai Plasma Nutfah [Potential Development of Brown Rice as Germplasm] Yogyakarta. Bogor: Departemen Pertanian; 2019.

66.

Dong

, Xiang

, Shen

, Liu

, Zhu

, Ma

. Oat products modulate the gut microbiota and produce anti-obesity effects in obese rats. J Funct Foods. 2016;25:408–420. doi: 10.1016/j.jff.2016.06.025.

67.

Moran-Ramos

, He

, Chin

, Tovar

, Torres

, Slupsky

, et al. Nopal feeding reduces adiposity, intestinal inflammation and shifts the cecal microbiota and metabolism in high-fat fed rats. PloS One. 2017;12:e0171672. doi: 10.1371/journal.pone.0171672.