Polyphenols as Prebiotics in the Management of High-Fat Diet-Induced Obesity: A Systematic Review of Animal Studies

Obesity is a disease growing at an alarming rate and numerous preclinical studies have proven the role of polyphenols in managing this disease. This systematic review explores the prebiotic effect of polyphenols in the management of obesity among animals fed on a high-fat diet. A literature search was carried out in PubMed, Scopus, CINAHL, Web of Science, and Embase databases following the PRISMA guidelines. Forty-four studies reported a significant reduction in obesity-related parameters. Most notably, 83% of the studies showed a decrease in either body weight/visceral adiposity/plasma triacylglyceride. Furthermore, 42 studies reported a significant improvement in gut microbiota (GM), significantly affecting the genera Akkermansia, Bacteroides, Blautia, Roseburia, Bifidobacteria, Lactobacillus, Alistipes, and Desulfovibrio. Polyphenols’ anti-obesity, anti-hyperglycaemic, and anti-inflammatory properties were associated with their ability to modulate GM. This review supports the notion of polyphenols as effective prebiotics in ameliorating HFD-induced metabolic derangements in animal models.


Introduction
Obesity is a public health concern as it is a risk factor for life-threatening diseases such as type II diabetes [1], cardiovascular diseases [2], certain cancers [1], kidney diseases [3], and neurodegenerative diseases [4]. In the year 2017, overweight and obesity caused over four million deaths, with more than 140 million 'global deaths and disability-adjusted life years' [5]. The interplay between genetic-epigenetic-environmental factors is often linked to obesity [6,7], and of late, the role of gut microbes (GM) in the development of obesity has been extensively explored [8][9][10].
The GM-obesity relationship was first highlighted by Bäckhed et al. [11] using germfree mice, whereby conventionalisation of such mice with a normal microbiota increased body fat by 60% and elicited insulin resistance (IR) within 14 days. This finding was supported by Turnbaugh et al. [12], where the administration of microbes of obese or lean mice into germ-free mice, saw higher adiposity in the former. GM's role was further reinforced when germ-free mice were protected against a high-fat diet (HFD)-induced obesity [13]. Collectively these studies proved the role of GM in energy harvest and obesity. Consequently, subsequent studies proposed dietary intervention for GM's modulation, and substantial work involving probiotics, prebiotics or synbiotics in the management of obesity had emerged.
Polyphenols (PP) are secondary metabolites found abundantly in plants, fruits and nuts [14]. A growing body of research indicates that PP possess various biological activities

Study Characteristics
Twenty-two studies used pure single phenolic compounds, while two studies used a combination of phenolic compounds. The remaining 22 studies investigated phenolic-rich extracts. Twenty-six studies assessed the effectiveness of flavonoids, and the remaining studies investigated stilbenes or phenolic acid or a combination of flavonoids and other polyphenol classes. Most studies (n = 38) used mice, while only seven used rats. All the studies included in this review used male animals. The study duration ranged from 2 to 24 weeks. A majority of the studies used either 60% HFD or 45% HFD. Thirty-three studies were based on a preventive model, i.e., the intervention and HFD intake were initiated simultaneously. Ten studies used the treatment model approach; i.e., treatment began a few weeks after HFD intake. This information was not clearly stated in two studies. Figure S2. Sequence generation, allocation of concealment, random outcome assessment and blinding (detection) were unclear in all the studies. Baseline characteristics were reported in 98% of the studies, and incomplete outcome data was low in 96% of the studies. The animals were randomly housed in 73% of the studies while this was not carried out in 27% of the studies. The study quality was also assessed using the GSP checklist, which provided a quantitative assessment of each study. As depicted in Figure S3, five studies scored above 13 with the highest score being 14, while 40 studies scored between 7-12 with the lowest score being 9. Most studies did not address sample-size calculation, concealment of allocation, the animal's weight, time of intervention, method and time of sampling, number, and reason of excluded animals. The detailed GSPC scores are presented in Table S1.

Effect of Polyphenols on Food/Energy Intake, Obesity-Related Parameters, and Markers-Associated with Inflammation
The results are presented as significantly high (SH), significantly low (SL), or as not significant (NS) which refers to either a parameter being significantly increased, decreased or not-significant compared to HFD group. Studies that reported more than a compound or tested various dosages are presented as separate studies.

Effect of Polyphenols on Gut Microbiota
We report the changes in GM as alpha diversity, beta-diversity, Firmicutes: Bacteroidetes ratio (F:B ratio), overall changes in phyla, family/genus and frequently modulated genera.

Effect of Polyphenols on Gut Microbiota
We report the changes in GM as alpha diversity, beta-diversity, Firmicutes: Bacteroidetes ratio (F:B ratio), overall changes in phyla, family/genus and frequently modulated genera.

Discussion
Polyphenols are emerging as prebiotics due to their ability to impro dysbiosis and health parameters [29,30]. Numerous preclinical studies have the role of polyphenols in improving various health conditions in animal comprehensive review was carried to obtain a complete and exhaustive su preclinical studies investigating the effect polyphenols in ameliorating HFD sity.
In general, 98% of the studies reviewed indicated significant improv metabolic derangements inflicted by HFD intake after consuming pure ph pounds or its extracts. These changes were related to the modulation of GM

Discussion
Polyphenols are emerging as prebiotics due to their ability to improve intestinal dysbiosis and health parameters [29,30]. Numerous preclinical studies have investigated the role of polyphenols in improving various health conditions in animal models. This comprehensive review was carried to obtain a complete and exhaustive summary of the preclinical studies investigating the effect polyphenols in ameliorating HFD-induced obesity.
In general, 98% of the studies reviewed indicated significant improvement in the metabolic derangements inflicted by HFD intake after consuming pure phenolics compounds or its extracts. These changes were related to the modulation of GM at the genus level. Our analysis also indicates that all 23 studies that tested pure PP and 21 out of 22 studies used extracts observed improvement in metabolic parameters and gut microbiota profiles. Thus, we can conclude that both pure PP and extracts were equally effective as prebiotics. However, this review's findings will need to be interpreted with some caution as the quality score for the majority of studies was below 12. This was because a few essential criteria such as sequence generation, allocation of concealment, random outcome assessment, blinding (detection), and sample size calculation were not addressed in most studies.

Effect of Polyphenols on Obesity-Associated Parameters, Adipocytokines, and LPS/LBP
It is evident that phenolic intervention significantly reduced one or more metabolic parameters such as weight gain, VAT, FBG, FI, GTT, HOMA-IR, TC, TAG, LDL, LPS, TNFα and IL-6 in 98% of the studies included in this review.
The most prominent observation in this review was the reduction in body weight/VAT/ TAG reported in 83% of the studies. Among these, 40% of studies showed a decrease in all three parameters, i.e., body weight, VAT and TAG. Therefore, it can be assumed that the reduction in body weight is probably related to reduced VAT and TAG. The mechanisms highlighted for TAG improvement in some of these studies include downregulation of Ag-pat2 [60], reduction in taurine-conjugated bile acids causing activation of nuclear farnesoid X receptor (FXR) in the liver, consequently, suppressing lipogenic pathway in the liver [61], inhibition of hepatic adipogenic genes such as Fas, PPARγ, and SREBP-1c [71,72], and upregulation of fasting-induced adipose factor (FIAF) expression in the intestine, hence, suppressing TAG deposition [81]. Our observation indicates that the mode of action varies between different PP.

Effect of Polyphenols on Glucose Homeostasis
We also observed that studies reporting a reduction in weight, VAT, and or TAG revealed a significant improvement in glucose homeostasis (either FBG, GTT, FI or HOMA-IR), and its inflammatory state. However, whether the improvement in glucose homeostasis is a cause or effect of a decreased inflammatory state is unclear. Although many studies have hypothesised that low-grade systemic inflammation in white adipose tissue disrupts glucose homeostasis, a clear mechanism of HFD-induced hyperglycaemia or IR remains elusive [84][85][86]. Few other studies mention that IR precedes WAT inflammation. Shimobayashi et al. [87] reported that the accumulation of macrophages in WAT preceded IR in HFD and mTORC2-knockout animal models. They further stated that IR in WAT induced MCP-1 production responsible for the recruitment of monocytes and activation of macrophages in WAT. Lee et al. [88], on the other hand, observed IR and glucose intolerance as early as day 3 following HFD intake, and this was associated with lipid overload in the adipose tissue. Chronic inflammation was identified as a critical mechanism for IR once obesity is established. Apart from this, ellagic acid and its metabolites were demonstrated to mitigate IRs by improving mitochondrial function by activating 5 -AMP-activated protein kinase [89][90][91]. Some PPs such as geraniin, corilagin, and ellagic acid are also shown to exhibit insulin-like activity, causing 2-NBDG uptake in the absence of insulin by preadipocytes [92].
Few PPs only improved glucose intolerance/IR without affecting weight/VAT/TAG. The exact mechanisms for this were not well elaborated in the papers included in this review. Nevertheless, the mechanism of actions was discussed extensively in a few reviews [14,[93][94][95]. Some of which include; inhibition of salivary and pancreatic α-amylase and α-glucosidase activities. These enzymes are responsible for metabolising carbohydrate and its inhibition of reduced carbohydrate metabolism and therefore, glycaemic spike [14,94]. This effect has been reported with punicalagin [96], ferulic acid, and isofer-ulic acid (α-glucosidase inhibitors) [97]. Certain polyphenols were also shown to inhibit glucose transporters in the gut, therefore reducing glucose uptake. For instance, quercetin, chamomile and green tea were shown to inhibit glucose uptake in the intestine by inhibiting GLUT2, and GLUT5 [98,99]. Some polyphenols such as isovanillic acid 3-O-sulfate (metabolites of cyanidin 3-O-glucoside) [100] and epicatechin [101] have been shown to increase glucose intake by upregulating GLUT4 expression in tissues.

Effect of Polyphenols on Gut Microbiota
Approximately 93% of the studies included in this review reported significant changes in one or more parameters of GM modulation. It is unclear whether the remaining 7% of the studies (three studies) performed any statistical analysis to prove GM modulation and only reported their findings in relative abundance changes. Nevertheless, these studies did report positive changes with phenolic interventions.
The alpha-diversity was reported in 29 studies (64%); among these nine studies reported a significant increase in this diversity. Beta-diversity was presented as a PCoA plot in most studies. Among the 35 studies that reported this diversity, 26 plots showed distinct cluster/clustering closer to normal diet groups. The phenolic intervention did not improve alpha-diversity in most studies, which could be due to the shorter study duration [102]. Furthermore, diet-diversity was previously observed to affect beta and not alpha-diversity [103].
F:B ratio was reported in 28 studies, with 17 studies indicating a significant reduction of the ratio. This raises the question of phyla level relevancy often reported in obesity studies. Higher firmicutes and lower bacteroidetes (high F:B ratio) are often associated with obesity [104][105][106]. On the contrary, few studies showed a low F:B ratio [107][108][109]. A few meta-analyses have also shown a lack of differences in F:B ratio between obese and lean individuals [110][111][112]. It was suggested that the F:B ratio may not be a robust indicator of dysbiosis in obesity [113]. It was further highlighted that the disparities that exist between studies could be related to methodological differences such as sample collection, DNA extraction, hypervariable regions of interest, sequencing methods and bioinformatics data analysis software (QIIME, Mothur) [113].
Nevertheless, polyphenols significantly improved relative abundance of GM at the family/genus level in 42 studies. A few genera were consistently observed to be modulated irrespective of the type of intervention. The commonly improved genera include Akkermansia, Bacteroides, Blautia, and Roseburia, and to a lesser extent, Bifidobacteria, Lactobacillus, and Alistipes. Desulfovibrio, on the other hand, was reduced in three studies. The genus Akkermansia was significantly increased with the intake of vanillin, anthocyanin, kudingcha dicaffeoylquinic acids, EGCG, quercetin-resveratrol, total flavonoid, blueberry extract, tea polyphenols, cranberry extracts, cloudberry extract, alpine bearberry extract, green tea extract, and olive leaf extract. Akkermansia muciniphila (A. muciniphila), is a gramnegative bacterium belonging to the Verrucomicrobia phylum. This bacterium produces mucin degrading enzymes causing fermentation of mucin to acetic, propionic acid, and sulphate [114]. A preclinical study reported that A. muciniphila reduced plasma TAG, improved fatty liver, and gut dysbiosis [115]. In a randomised human trial, administration of A. muciniphila for 30 days improved plasma insulin, HOMA-IR, Dipeptidyl-peptidase 4 (DPP4) (inflammation) and LPS concentration in overweight/obese individuals [116]. Administration of procyanidins, resveratrol, Rubus occidentalis, tea polyphenols, kudingcha tea, fuzhuan brick tea, green tea polyphenol, and black tea polyphenol improved Bacteroides. Some Bacteroides strains were categorized as next-generation probiotics [117], including B. uniformis, B. acidifaciens, and B. dorei. These strains were shown to ameliorate HFD-induced metabolic derangements [117].
In contrast, some strains of B. fragilis were identified as a risk factor for the development of colon cancer [118]. It can therefore be concluded that the health benefit of the genus is strain specific. Blautia was improved with grape seed proanthocyanidin extract, resveratrol, Rutgers scarlet lettuce, green tea, oolong tea and black tea, rapeseed oil with sinapine, tea polyphenols. This genus is a butyric producer, reported to ameliorate intestinal inflammation, IR [119], and exhibited an inverse association with visceral adiposity [120]. Therefore, it may play a crucial role in the management of obesity and obesity-related diseases. Roseburia was improved with grape seed proanthocyanidin extract, a combination of resveratrol and sinapic acid, Rutgers scarlet lettuce, grape seed extract, tea polyphenols, grape pomace extract. Roseburia also a butyrate producer and this genus has been implicated in the control of intestinal inflammation and was shown to be reduced in patients with type II diabetes [121]. The conventional probiotics such as Bifidobacterium and Lactobacillus were improved in fewer studies. Administration of kudingcha dicaffeoylquinic acids, blueberry extract, and tea polyphenols improved Bifidobacterium, while Lactobacillus was improved with the intake of hydroxytyrosol, pandanus tectorius fruit extract, pomegranate polyphenols. Some of the health benefits of Bifidobacterium include improvement in celiac disease with the administration of B. infantis strain NLS [122], irritable bowel syndrome with B. lactis DN-173010 [123]. Conversely, Lactobacillus was significantly improved in three studies, while six studies showed a significant reduction.Since only the changes up to genus level were reported in all the studies, it is uncertain the type of species/strains affected following phenolic interventions. Literature suggests that the health benefits of Lactobacillus are species/strain-dependent. For instance, weight loss was reported with L. gasseri [124], and L. reuteri JBD301 [125]. Whilst, weight gain was reported with L. acidophilus [126]. The abundance of Alistipes improved with the administration of total flavonoids of quzhou, tea polyphenols, kudingcha and fuzhuan brick tea. Similar to Lactobacillus, the health attributes of this genus are species/strain specific. Among patients with liver cirrhosis, lower A. shahii and A. putredinis levels were observed compared to healthy controls [127], while A. finegoldii was shown to promote colorectal cancer [128]. Desulfovibrio is known to produce a cytotoxic compound; hydrogen sulfide. This bacterium was observed to increase in patients with ulcerative colitis [129].

Is LPS the Only Linking Factor between Metabolic Derangements and Gut Dysbiosis?
Consumption of HFD was proven to increase fat deposition, elicit IR, and inflammation in adipose tissue, liver and skeletal muscle [130,131]. LPS was discovered as one of the linking factors between HFD intake and metabolic derangements. The administration of LPS into lean mice on normal-diet caused similar changes as the HFD-fed mice [130]. Therefore, we explored the relationship between GM, LPS, and metabolic parameters. Among the 23 studies that reported a significant reduction in LPS/LBP, 20 stated significant modulation in GM and metabolic derangements following PP administration. Two studies [35,74] reported significant improvement in metabolic parameters and LPS concentration; however, they did not indicate GM's statistical significance. Thus, the relationship between GM modulation and LPS is unclear for these two studies. Interestingly, Cheng et al., 2016 [62], reported non-significance for LPS even though GM modulation and significant improvement in OGTT were reported with Rutgers scarlet lettuce. This suggests that other factors may also play a role in the amelioration of metabolic derangements such as bile acids, and short-chain fatty acids (SCFA). Gut microbiota composition has been shown to alter the type of secondary bile acids, and consequently influence glucose homeostasis [132]. It has also been shown to improve SCFA production; for instance, the administration of butyrate was shown to improve rodent's insulin sensitivity [133]. Considering the studies included in this review, we conclude that PP's administration improved GM composition and consequently, the metabolic parameters. This was achieved partly by reducing LPS concentration; however, the involvement of other mechanisms cannot be discounted.

Limitations
In general, PP interventions significantly improved HFD-induced metabolic derangements in 98% of the animal studies and these changes were shown to be related to the GM modulation. Nevertheless, there are some limitations to this review, which will be discussed here.

Experimental Animals and Their Environment
A variety of experimental animal species (C57BL/6J, C57BL/6N, C57BL/6 wild type Sprague Dawley rats, Wistar rats) of varying ages were used. These experimental animals were exposed to different types of a high-fat diet, water source, bedding materials, and were either housed alone or in groups.

Types of Intervention
Different types of phenolic compounds were tested, either in its pure form or as an extract at varying doses. These compounds were also administered via various modes (gavage/within diet or drinking water).

Methodology to Assess GM
GM modulation may have been affected due to variations in the methodology employed between studies, including the use of faecal samples or intestinal contents, faecal sampling time, and the difference in the hypervariable regions, sequencing methods, and bioinformatics tools utilised.

Outcomes Reported
The parameters reported for metabolic derangements and GM modulation differed, i.e., not all studies reported the same parameters. These variations greatly influenced the reported outcomes, thus posing a challenge in summarising polyphenols' prebiotic effect in preclinical studies.

Conclusions
Polyphenols appear to be a promising phytochemical in the management of dietinduced obesity. This review, havedemonstrated that the administration of polyphenols to obese animals improved metabolic derangements, particularly, weight gain, visceral adiposity, plasma TAG, and glucose homeostasis inflicted by HFD intake. These improvements were established to be the result of improved gut dysbiosis as presented in all the studies. The findings can form the basis for the development of randomised clinical trials required to conclusively place polyphenols as prebiotics. Moreover, Nonetheless, discrepancies in study designs, metabolic markers measured, and the methodology of gut microbiota analysis in preclinical studies need to be streamlined to make a more robust conclusion.
Supplementary Materials: The following are available online at https://www.mdpi.com/2304-8 158/10/2/299/s1, Figure S1. Search strategy (PubMed); Figure S2. Risk of bias using SYRCLE's Risk of Bias tool; Figure S3. Study quality using Gold Standard Publication Checklist, Figure S4. Effect of polyphenols on energy/food intake, and body weight; Figure S5. Effect of polyphenols on adiposity; Figure S6. Effect of polyphenols on lipid profile; Figure S7. Effect of polyphenols of glucose homeostasis; Figure S8. Adipocytokines, C-Reactive Protein (CRP), and LPS/LBP; Figure S9. Beta-diversity of (a) pure phenolic compounds and (b) phenolic extracts; Table S1. Gold Standard Publication Checklist of study quality; Table S2. Alpha diversity of pure phenolic compounds and extracts.