The prevalence of overweight and obesity has increased worldwide in the past ∼50 yr, reaching pandemic levels. The Global Burden of Disease study reported that, between 1980 and 2013, the proportion of adults with overweight or obesity increased from 28.8 to 36.9% in men, and from 29.8 to 38.0% in women (Ng et al., 2014). A high BMI accounted for 4.0 million deaths globally, more than two-thirds of which were caused by cardiovascular disease (CVD) (Afshin et al., 2017), which is a frequent consequence of overweight or obesity in adults accompanied by dyslipidemia. According to the Report on the Status of Nutrition and Chronic Diseases of Chinese Residents (2015), the prevalence of overweight among Chinese adults was 30.1%, the prevalence of obesity was 11.9%, and the overall prevalence of dyslipidemia was 40.40% in 2012, which represents a substantial increase from the values recorded in 2002 (18.9, 2.9 and 18.6% respectively). The increase in serum cholesterol levels in the population will lead to an increase of approximately 9.2 million cardiovascular events in China from 2010 to 2030 (Moran et al., 2010). Therefore, effective control of overweight and obesity is of great significance for the prevention and control of CVD worldwide, and especially in China.

Gut microbiota have gained great attention for their complex interactions with many diseases of the host, including obesity. As early as 2005, the Gordon research team demonstrated that obesity could alter the gut microbiota ecology and that a significantly increased ratio of Firmicutes/Bacteroidetes was an obvious intestinal microbial characteristic of individuals with obesity (Ley et al., 2005; Turnbaugh et al., 2006). Thereafter, increasing evidence has demonstrated that interventions targeting the modulation of gut microbiota and its metabolites (such as probiotics and prebiotics, etc.) are effective and integrative for obesity and its comorbidities (Canfora et al., 2019; Green et al., 2020).

Xinjiang is a multiethnic region with a unique food culture characterized by a high proportion of mutton and low proportion of dietary fiber. Moreover, the genetic background of the Xinjiang population is complicated, with European and Asian characteristics (Feng et al., 2017; Yin et al., 2020). Studies have implied that dietary habits, ethnicity, and geography substantially impact the taxonomic composition of the microbiome (Deschasaux et al., 2018; He et al., 2018; Kolodziejczyk et al., 2019). In particular, the present knowledge on how the composition of the gut microbiome relates to patient health is principally based on investigations of European and North-American populations. Although this knowledge is useful for understanding the microbiota of individuals from such populations, it may not accurately extrapolate to individuals from distinct populations or living in distinct locations, especially in Xinjiang. Therefore, it is crucial to study the characteristics of gut microbiota and their metabolite profiles in the Xinjiang population, as they may help prevent and treat obesity more precisely using microbial targets.

Medications were also suggested to affect the gut microbiota, especially diabetes medications, such as acarbose and metformin (Gu et al., 2017; Wu et al., 2017). Orlistat is a common antiobesity drug that acts by inhibiting the absorption of triglycerides, whereas ezetimibe is a hypolipidemic drug that selectively inhibits the absorption of cholesterol by attaching to the brush border of the villi epithelium of the small intestine (Nakou et al., 2008). Therefore, the administration of either of the two drugs can result in the accumulation of triglycerides and cholesterol in the colon, respectively. Complex carbohydrates that have not been digested by upstream digestive tract enzymes are the main source of nutrients for gut microbes, and these carbohydrates can be fermented by gut microbiota to produce metabolites, such as SCFAs. Our previous study suggested that high-fat diets and high-cholesterol diets could differentially alter gut microbiota and their metabolites, including SCFAs and bile acids (Liang et al., 2021). However, whether these effects are caused by the high dietary lipid/cholesterol content or to the concomitant decrease of carbohydrate intake is unclear. The effects of orlistat and ezetimibe on gut microbiota are poorly understood, and the characteristics of the bacterial populations involved in lipid and cholesterol metabolism are unclear.

Therefore, our present study aimed to investigate the gut microbiota and SCFAs profiles of patients with overweight and obesity with dyslipidemia in Xinjiang and identify the core microbes that responded to orlistat and ezetimibe, thus suggesting the contributions of gut microbiota to host metabolism.

The use of germ-free animals and microbiota transplantation showed that the gut microbiota may play a causal role in the development of obesity and associated metabolic disorders. Differences in gut microbiota composition, functional genes, and metabolic activities between overweight/obese and normal-weight individuals were investigated in the Xinjiang region of China, and the microbial responses to antiobesity medications were examined in the present study.

Most studies in this field have found that the gut microbiota of obese people is characterized by an increase in the F/B ratio, the phylum Firmicutes, and a decrease in Bacteroidetes (Gomes et al., 2018). However, studies focusing on childhood obesity have reported controversial data about the F/B ratio and the proportion of Firmicutes and Bacteroidetes, which implied that F/B might be a rough tool to predict or examine the gut microbiota associated with obesity (Taylor, 2017). Furthermore, a cross-over study of 29 adults did not find differences in the proportion of Firmicutes and Bacteroidetes between obese and non-obese individuals using FISH (Duncan et al., 2008). No significant differences in Firmicutes to Bacteroides ratio was observed between healthy lean and overweight or obese whose dietary intakes and physical activity levels were comparable in a cohort study (Fernandes et al., 2014). Consistently, in the present study, no significant changes in the F/B ratio and the relative abundance of Firmicutes and Bacteroidetes were found between overweight/obese and normal-weight adults. Clostridium and Faecalibacterium prausnitzii were suggested as classical SCFAs producers and immunomodulators by many studies (Atarashi et al., 2013; Zhou et al., 2018). Previous works have demonstrated that Clostridium species are positively correlated with obesity, while the fecal Faecalibacterium prausnitzii amount was not significantly different between obese and lean subjects (Collado et al., 2008; Feng et al., 2014). Akkermansia muciniphila was a promising candidate among the various next-generation beneficial microbes, because it was inversely associated with obesity, metabolic disease, cardiovascular diseases, and even cancer (Cani and de Vos, 2017). A recent study found that Akkermansia muciniphila abundance was lower in cases with severe obesity, while other studies reported that not only overweight pregnant women, but also their infants, had a higher abundance of Akkermansia muciniphila than did normal-weight women and their infants (Collado et al., 2010; Dao et al., 2019). However, in the present study, no obvious changes in the abundance of Clostridium, Faecalibacterium prausnitzii, and Akkermansia muciniphila were found between normal-weight and overweight/obese adults, which might result from the regional and dietary differences of the Xinjiang population compared with those of other studies.

Conversely, in the present study, the microbiota composition of overweight and obese adults with dyslipidemia in Xinjiang was characterized by an overrepresentation of Lactobacillus and a decrease in the alpha diversity index and a reduction of Actinobacteria, Bacteroides, Bifidobacterium, and Bacteroides fragilis. Conversely, a significant increase in microbial diversity in overweight adults was found in a Japanese population using next-generation sequencing (Kasai et al., 2015). The microbial diversity was normally negatively related to the disease occurrence in adulthood, while the inverse relationship was detected in infancy and childhood because of the unstable state of gut microbiota during early life. Long et al. found that the phylum Actinobacteria, particularly the genera Mobiluncus, Corynebacterium and Bifidobacterium, was less abundant among non-diabetic obese subjects compared with normal-weight individuals (Long et al., 2017). Furthermore, research data suggest that a high concentration of Bifidobacterium during both infancy and adulthood protects against obesity, which was consistent with the results of our study (Kalliomäki et al., 2008; Collado et al., 2010). Bifidobacterium species may decrease fat absorption through the deconjugation of bile acids (Begley et al., 2006), which is possibly related to its reduction in obese subjects. Polysaccharide A of Bacteroides fragilis, which belongs to the genus Bacteroides, has been highlighted as directing microbiota and host interactions, such as anti-inflammation, and may enhance the efficacy of cancer therapies based on immune checkpoint inhibitors (Erturk-Hasdemir and Kasper, 2018). However, in contrast with the results of our study, Scheepers et al. demonstrated that Bacteroides fragilis was present at high concentrations in obese and overweight compared with lean infants or children and was positively correlated with the BMI Z-score (Scheepers et al., 2015). Furthermore, a recent clinical trial of overweight adults on a vegan diet suggested that a smaller reduction in Bacteroides fragilis was associated with a greater loss of body weight, fat mass, and visceral fat, and a greater increase in insulin sensitivity (Kahleova et al., 2020). Other studies demonstrated that Lactobacillus was positively correlated with weight gain, and a significantly higher Lactobacillus concentration was detected both in obese children and adults compared with lean control individuals (Armougom et al., 2009; Bervoets et al., 2013; Million et al., 2013), which was consistent with the results of the present study. These significantly altered microbes, all of which were significantly decreased in obese people, with the exception of Lactobacillus, were commonly considered to benefit the health status of the host. Therefore, it is suggested that the gut microbiota characteristics of overweight and obese people with dyslipidemia in Xinjiang were the obvious reduction of species diversity and of some potential beneficial microbes (such as Bifidobacterium) without the enrichment of opportunistic pathogens. These results also suggest that the gut microbiota composition of overweight and obese adults with dyslipidemia in Xinjiang, PR China, was mostly different from those reported previously and has its own characteristics. Despite this, the functional prediction of these distinctive microbiota still featured lipid metabolism, which was related to the metabolic phenotypes. This might be the consequence of the study subjects included in the present study, who were characterized by an abnormally elevated BMI and blood lipids vs. the subjects who had only a high BMI enrolled in other studies.

To explore further the gut microbial patterns and metabolic activities between obese and normal-weight individuals in Xinjiang, we also analyzed the enterotypes and SCFAs. Enterotypes were first proposed by Arumugam et al., in 2011 and were determined as clusters dominated by Prevotella (P-type), Bacteroides (B-type), and Ruminococcus (R-type) (Arumugam et al., 2011). The enterotypes were shown to be independent of age, gender, cultural background, and geographical location. Regardless of the geographical region, the P-type and B-type were more robust compared with the third type (i.e. Bifidobacterium, Faecalibacterium, and Ruminococcus) in the healthy human gut microbiome (Mobeen et al., 2018). Furthermore, these enterotypes have been found and involved in many diseases, especially in obesity (Christensen et al., 2018). For instance, the P-type, B-type, and R-type have been investigated in adults with severe obesity in France (Dao et al., 2019). However, a recent study performed in Mexico found that most normal-weight children had a gut bacterial community dominated by the R-type, while most obese children had a community dominated by the P-type (Maya-Lucas et al., 2019). In the present study, the enterotypes were P-type, G-type, and E-type, which were significantly different from the recent studies regardless of a lean or obese status.

Previous studies indicated that it is likely that differences in diets worldwide have created subgroups of enterotypes (Nishijima et al., 2016). A Bacteroides subtype with high abundance of Bifidobacterium is prevalent in Japan (Nakayama et al., 2015). The P-type was associated with a diet high in fiber and resistant starch, while the B-type was associated with a diet high in fats and low in fiber (Wu et al., 2011). A recent study reported that a high intake of carbohydrates and fat and a low intake of micronutrients, dietary fiber, and high-quality protein were the characteristics of the current diet in Xinjiang (Yin et al., 2020), which does not present the nutritional structure of a balanced diet. Therefore, we believe that the G-type and E-type found in Xinjiang might associated with a diet high in carbohydrates and fat and low in fiber and protein. Despite this, the existence and generalizability of these proposed enterotype subgroups warrant further investigation using a larger sample size, and more-refined enterotype predictors that model complex bacterial taxa distributions may be needed.

Many studies indicated that SCFAs levels are significantly different between overweight/obese and lean people (Canfora et al., 2015). However, a non-significant change in microbial metabolic activities characterized by a comparable SCFAs level was found in the present study. Bacterial species from the genera Bacteroides and Prevotella produce acetate and propionate, and those from Lactobacillus and Bifidobacterium genera mainly produce lactate and acetate (Koh et al., 2016). In the present study, although the abundance of these microbes was different between obese and lean individuals, the ability of these microbes to produce SCFAs was similar, which might be correlated with the relatively stable enterotypes found among these participants, especially those with the P-type.

Antiobesity medication may modulate the gut microbiota, thereby altering fecal SCFAs composition, which may improve metabolic health. Gu et al. demonstrated that acarbose increases the relative abundances of Lactobacillus and Bifidobacterium in the gut microbiota and depletes Bacteroides, which may beneficially affect metabolism in patients with T2D by changing the relative abundance of microbial genes involved in BA metabolism (Gu et al., 2017). Orlistat, as a classical antiobesity agent, was approved by the FDA in the US in 1999 based on its efficacy in the treatment of obesity (Heck et al., 2000). In the present study, the WC, TG, and FBG of patients with overweight/obesity with dyslipidemia decreased significantly after taking orlistat, which indicated that the drug was effective in improving metabolic health, as previously demonstrated, albeit without weight reduction.

However, the use of orlistat has been associated with several mild-to-moderate gastrointestinal adverse effects, such as oily stools, diarrhea, abdominal pain, and fecal spotting, which were related to the increased fat malabsorption (Filippatos et al., 2008). However, whether these adverse effects play roles in modulating the gut microbiota and its metabolites remains unclear. In the present study, the microbial diversity, predominant bacteria, enterotypes, and fecal SCFAs were not significantly altered by a 12-wk course of orlistat. These results were consistent with those of Morales et al., who performed an orlistat intervention study in 10 healthy subjects over 7 days (Morales et al., 2016). However, a recent study demonstrated that the microbiota of HFD-fed mice supplemented with orlistat were characterized by significantly increased Actinobacteria, Proteobacteria, Pseudomonas, Rhodococcus, Roseburia, and Acetivibrio, and significantly decreased microbial diversity, Firmicutes, and Deferribacteres (Ke et al., 2020). Nevertheless, in the present study, some bacterial populations with a low relative abundance were modified by orlistat, presenting as increased Proteobacteria and decreased Alloprevotella. Proteobacteria is a small-proportion phylum comprising many endotoxin-producing bacteria that might be pathogenic to the host, such as E.coli. A recent study found that a HFD diet promoted the flourishment of Proteobacteria, while Ganoderma lucidum mycelium significantly reversed this effect (Chang et al., 2015). Species of Alloprevotella, which belongs to Bacteroidetes, could produce moderate amounts of acetic acid and major amounts of succinic acid as end products of fermentation (Downes et al., 2013). We also found that the positively significant correlations between Alloprevotella and SCFAs became insignificant with the reduction of Alloprevotella caused by orlistat. Animal studies found that both the HFD and high-sucrose diet reduced the abundance of Alloprevotella, while the combination of the abrownin and the high-sugar diet promoted the reproduction of Alloprevotella (Kong et al., 2019; Yue et al., 2019). These results indicate that the colonic accumulation of dietary fat might result in the enrichment of the opportunistic pathogens with the depletion of potential beneficial microbes related to obesity. Furthermore, previous studies have found that the fecal bile acid level of obese subjects was slightly decreased or kept stable after the administration of orlistat, implying a lower potential for modulating microbiota composition (Trouillot et al., 2001; Ahnen et al., 2007). Here, the absence of significant changes in microbial metabolite levels after the orlistat intervention probably reflected the fact that only small changes were observed in the bacterial populations.

As mentioned above, in the present study, FBG, TG, and WC were significantly decreased by orlistat. Before the orlistat intervention, Coprococcus, Dialister, and Prevotella were negative modulators of FBG in overweight and obese patients with dyslipidemia; however, this correlation became insignificant with the slight increases in Coprococcus and Dialister, and the modest decreases of Prevotella that occurred after the intervention. In contrast, after the intervention, Actinomyces and Bacteroides were positively correlated with FBG, together with the slight decrease in two genera. A recent study reported that the genus Actinomyces showed a strong inverse association with the risk of diabetes (Long et al., 2017). In patients with obesity with metabolic disorders, Bacteroides has been found to be negatively correlated with BMI, WC, and LDL-C, whereas Coprococcus has been found to be negatively correlated with DBP, TG, and UA, and positively correlated with HDL-C and eGFR (Zeng et al., 2019). These results indicate that these microbes might be keystones that are associated with the hypoglycemic efficacy of orlistat, especially in relation to the abundance of Actinomyces and Bacteroides, which implies that the depletion of these genera might help orlistat to improve glucose metabolism. An animal study demonstrated that the enrichment of Pseudomonas, Rhodococcus, Roseburia, and Acetivibrio in HFD-fed mice supplemented with orlistat was positively correlated with metabolic pathways, including glucose, lipid metabolism, and inflammation pathways (Ke et al., 2020), suggesting that orlistat exerts beneficial effects on glycolipid metabolism by modifying gut microbiota, which may offer a novel mechanism of orlistat action. However, there was no significant correlation between TG, TC, and these core selected genera, regardless of the intervention. The significant correlations between HDL-C, LDL-C, and core microbes detected before the intervention became insignificant after the orlistat intervention. In addition, Clostridium XVIII was significantly negatively correlated with WC and SCFAs before the intervention; in contrast, this correlation was not statistically significant regarding the decrease in Clostridium XVIII afforded by orlistat. Clostridium XVIII is one of the SCFAs producers and was believed to induce Treg production, thereby participating in the anti-inflammation response of the host (Atarashi et al., 2013). These results indicate that the improvement in the lipid metabolism afforded by orlistat might be the consequence of the malabsorption of dietary fat, regardless of the modification of these selected microbes and SCFAs, although the network correlations between SCFAs and core taxa were dynamically altered by orlistat. Therefore, the benefits of glucose metabolism from orlistat were related to the reduction of microbes that responded to orlistat, especially Actinomyces and Bacteroides; however, a similar phenomenon was not found in lipid metabolism.

Hypolipidemic drugs are prescribed in most of cases for the treatment of overweight and obesity accompanied by dyslipidemia. Several studies have shown that the gut microbiota is able to regulate the host cholesterol metabolism (Wahlström et al., 2016; Kriaa et al., 2019; Le Roy et al., 2019). In the present study, not surprisingly, the serum TG, TC, and LDL-C were significantly decreased by ezetimibe, which is a widely approved hypolipidemic drug. Similar to orlistat, the alpha diversity, predominant bacteria, enterotypes, and SCFAs were not significantly altered by ezetimibe. Consistently, no change in total or major bacteria, despite a selective significant increase in Lactobacillus spp., was found in normocholesterolemic mice treated with ezetimibe (Catry et al., 2015).However, in the present study, bacteria with low abundance, presenting as Fusobacteria and Fusobacterium, rather than Lactobacillus, were significantly increased after ezetimibe therapy. Furthermore, the significant increase in the abundance of Fusobacterium afforded by ezetimibe was positively correlated with AST and negatively correlated with butyric acid. A previous study stated that Fusobacterium was responsible for bile acid desulfation (Robben et al., 1989), which further indicated the possible links between Fusobacterium and liver dysfunction. Previous evidence suggested a potential active role for Fusobacterium (Fusobacteria), specifically F. nucleatum, in colorectal cancer (Hussan et al., 2017). Our previous study found that Fusobacteria and Fusobacterium were decreased by the potential probiotic strain Bifidobacterium bifidum TMC3115 in middle-aged and elderly individuals, while the TC and LDL-C were also significantly reduced by TMC3115 (Wang et al., 2019). Therefore, these results indicate that the colonic accumulation of dietary cholesterol results in a slight flourishment of Fusobacterium, which has a low abundance and is potentially harmful to the host. Findings from an animal study suggested that the Lactobacillus spp., which are negatively correlated with the expression of genes related to cholesterol metabolism, were significantly increased by ezetimibe (Ke et al., 2020). Zhong et al. suggested that ezetimibe is more effective when microbiota is absent and speculated that this was because ezetimibe may be co-administered with antibiotics and/or probiotics clinically (Zhong et al., 2015). No studies have examined the effects of ezetimibe on gut microbiota and its metabolites in patients with metabolic disorders. To the best of our knowledge, the present study was the first analysis focusing on this topic, and detailed studies are required to address the gut microbiota alterations afforded by ezetimibe in the future.

Before the ezetimibe intervention, Dorea, Clostridium XVIII and Lachnospiracea incertae sedis were positively correlated with BW, BMI, HC, and WC, respectively. Conversely, Megasphaera was negatively correlated with BW and HC. In contrast, these correlations were not observed in patients enrolled in the orlistat intervention group, indicating that the microbial responses are dependent on the specificity of the medications. In patients with obesity with metabolic abnormalities, it was reported that Dorea was positively correlated with BMI, WC, TC, and LDL-C, while Blautia was positively correlated with WC, TC, and LDL-C (Zeng et al., 2019), which was partly consistent with the results of the present study. These genera were also predominant in the obesity group and were shared among patients with different metabolic disorders (Zeng et al., 2019). As an acetate producer, Blautia can drive insulin release and promote metabolic syndromes, such as hypertriglyceridemia, fatty liver, and insulin resistance (Zhang et al., 2015). Moreover, Dorea was negatively associated with insulin resistance (Brahe et al., 2015). However, a recent study reported that Megasphaera was associated with overweight and obesity (Kaplan et al., 2019). These results suggest that the relationship between gut microbiota and metabolic indicators in subjects who have overweight and obesity accompanied by dyslipidemia, such as those enrolled in the present study, was partly distinct from that detected in individuals who had a high BMI exclusively. After the intervention, although TG levels decreased significantly, no significant correlations with core taxa were found. In contrast, TC and LDL-C were negatively correlated with Clostridium XVIII and Lachnospiracea incertae sedis, while these two genera were slightly altered by ezetimibe. Recently, new bacterial phylotypes belonging to Lachnospiraceae have been associated with a high coprostanol level in healthy humans (Antharam et al., 2016). It is reported that coprostanol is one of the metabolites of cholesterol, and that the production of coprostanol can protect the gastrointestinal tract against the accumulation of cholesterol (Kriaa et al., 2019). Clostridium was involved in the various processes of BA metabolism, including deconjugation, oxidation, epimerization, 7-dehydroxylation, and desulfation (Kriaa et al., 2019). The absorption of dietary cholesterol accounts for approximately one-third of the pool of cholesterol in the body, which was inhibited by ezetimibe in this study, resulting in the reduction of serum lipid levels. Bile acids can shape the gut microbiota community by promoting the growth of bile-acid-metabolizing bacteria and inhibiting the growth of other bile-sensitive bacteria through its direct antimicrobial effects or its indirect effects via FXR and the immune system (Wahlström et al., 2016). Although the network correlations between SCFAs and these selected microbes were significantly modified by ezetimibe, the fecal SCFAs contents were not significantly altered, which was in accordance with the findings of the orlistat intervention. Therefore, we speculated that Clostridium XVIII and Lachnospiracea incertae sedis were ezetimibe-responsive microbes that reduced TC and LDL-C, which suggests the possibility of targeting selected gut microbiota to modulate cholesterol metabolism. However, the underlying mechanisms warrant further detailed study in the context of BA metabolism.

In conclusion, the gut microbial characteristics of patients with overweight and obesity with dyslipidemia in Xinjiang were partly distinct from those reported for subjects who had a high BMI exclusively. In addition, the particularity of the Xinjiang population indicated that the contribution of the gut microbiota to obesity should be analyzed by adapting to local conditions. The relatively stable gut microbiota and SCFAs recorded between before and after orlistat/ezetimibe intervention suggested the limited microbial modifications of anti-obesity medicines. The significant decreases of TG by orlistat and ezetimibe may not associate with the altered microbes but because of the medication induced malabsorption of lipid. Despite this, lowering effect of FBG was more sensitive in orlistat intervention, Actinomyces and Bacteroides might be the assistants in improving host glucose metabolism, while the reduction of TC and LDL-C were more effective in the ezetimibe intervention, Clostridium XVIII and Lachnospiracea incertae sedis might be targets for modulating the host cholesterol metabolism. These findings may provide scientific evidence of the precise prevention and control of metabolic diseases (obesity, etc.) in Xinjiang.

This study has four limitations: ① we did not recruit the age-matched healthy controls with the patients in the two drug intervention groups, which result in the sample size was relative small when compare the gut microbiota between control and patient group. ② we did not collect the dietary information which might influence the clinical efficacy of orlistat and ezetimibe through gut microbiota or conduct a randomized controlled trial to ensure the dietary intake in control. ③ we did not detect the bile acids (BAs), one of the mainly metabolites of gut microbiota, which was reported its association with metabolic disorders (including obesity). Since the patients had liver dysfunction and the enterohepatic circulation of BAs might be obstructed, something new might be found if we further detect the BAs. ④ the dose effects of the two drugs on gut microbiota and its metabolites were not considered because of the clinical trial, and animal study or in vitro fermentation should be considered.