Variable responses of human and non-human primate gut microbiomes to a Western diet
As hypothesized, our data indicate that the composition of the human and vervet gut microbiota is distinct on a Western diet. However, we did not observe differences in the relative abundance of fewer microbial taxa and genes in the vervets compared to the humans. Instead, the effect of a Western diet on the vervet gut microbiota was driven by distinct patterns in microbial taxa and genes compared to humans.
Many of the patterns we detected in both vervets and humans have been reported in other studies. For example, we observed elevated microbial richness and higher relative abundances of Prevotella in non-Western humans and elevated relative abundances of Bacteroides in Western humans. These results are concordant with a study of the human gut microbiota that associates diets high in protein and animal fat with high levels of Bacteroides and diets high in plant carbohydrates with high levels of Prevotella [23, 33]. Similarly, all published studies of Western and non-Western humans to-date report higher microbial richness and higher relative abundances of Prevotella in non-Western populations [2–6, 34]. In contrast, vervets on a Western diet showed similar microbial richness and higher relative abundances of Prevotella, a change that is mirrored in the macaque model [12].
In addition, we detected an elevated relative abundance of Collinsella in humans on a Western diet, and Collinsella has been associated with obesity in other studies of humans [35]. Similarly, while decreased relative abundances of Bacteroidetes coupled with increased relative abundances of Firmicutes, like those we observed in Western humans from the USA, are not reported in many studies comparing Western and non-Western humans [4, 5, 34]; they have been associated with obesity in human and mouse studies [36]. In our TWD-fed vervets, we measured the opposite patterns, and these findings are analogous to previous work contrasting a high-fat diet to a captive control (chow) diet in a macaque model (Macaca fuscata) where a shift to a high-fat diet also resulted in significant increases in Bacteroidetes, especially Prevotella and no significant variations in Firmicutes [12].
These commonalities suggest that our results are not unique to the human and vervet populations that we examined in this study. The human gut microbiota and its response to an easy-to-digest Western diet that is low in fiber differ fundamentally from the non-human primate gut microbiota. These findings provide the first evidence implying a specialization of the human gut microbiota. Therefore, while non-human primates may serve as comparative biomedical models for other aspects of human physiology [28], we suggest that the non-human primate gut microbiota may not provide an ideal direct model for understanding the effect of the human gut microbiota on host metabolism and nutrition in the context of a Western diet.
It is important to note that both the human and vervet data used in this study compare distinct populations, and therefore it is impossible to control for potential non-diet influences on the gut microbiota such as host genetics, exposure to local microbial pools, and antibiotic use. Likewise, future studies must control for the potential effects of non-human primate captivity such as host social contact networks and early life influences on the gut microbiota. However, it is unlikely that the results we present here are primarily driven by these potential confounds. Patterns in confounding factors such as antibiotic use across populations of the same host are likely to be similar for both humans and vervets, with Western humans and captive vervets generally having more exposure to antibiotics than non-Western humans and wild vervets. Therefore, while diet may not be the only factor causing differences between populations of the same host species, the comparisons of patterns between the two host species remains valid. Additionally, the captive population of vervets was taken from St. Kitts between 1975 and 1980. As a result, the two populations are only separated by 3–8 generations, reducing the impact of host genetics as a potential confound for the vervets.
Despite the comparison of multiple datasets in this analysis, studies generally indicate that host diet has the strongest effect on the gut microbiota compared to other factors [37], reducing the potential for confounding factors associated with different host populations to drive the observed patterns. For example, when human studies broadly control for host ethnic backgrounds, they continue to illustrate strong impacts of Western and non-Western diets, suggesting that differences in gut microbiota composition between populations of the same host species with distinct diets are unlikely to be genetically-driven [6]. Furthermore, although the vervet Western diet is distinct from a human Western diet, it is important to note that Western and non-Western human diets vary markedly across populations (e.g., USA vs. Italy and Malawi vs. Venezuela vs. Tanzania; [2, 3]). Despite differences within these diet categories, Western and non-Western human populations still cluster together in terms of gut microbiota composition, and other studies comparing Western and non-Western humans consistently report similar patterns [2–6, 34], suggesting that differences in the human and vervet gut microbiota are not driven by subtle variation in the composition of either a Western or non-Western diet or by idiosyncrasies in the selected datasets. Likewise, while the data we present were not all generated using the same methodology, the potential effects of distinct DNA extraction protocols, PCR primers, or sequencing platforms on gut microbiota data have been shown to be small [38, 39] compared to the effects of diet. Additionally, all DNA extractions utilized bead-beating step, which reduces extraction bias [40], and our use of the same closed-reference OTU picking pipeline (see methods) on rarefied data also reduces the potential effect of sequencing error and read depth on the results [41]. Therefore, while we cannot completely eliminate the biases of study-specific methodology on this dataset, we are confident that the patterns presented are biologically-driven.
Although data for other non-human primate taxa must be collected, we propose two explanations for the distinct responses of the human and non-human primate gut microbiota to diet. First, it is possible that the unique properties observed in the human gut microbiota are simply the result of unique human diet and physiology (namely an easy-to-digest, meat-heavy diet and increased brain size, reduced gut size, increased fat deposition, and decreased muscle mass [17–19]). Since both host diet and physiology drive gut microbiota composition [42], evolutionary changes in human diet and physiology could have easily led to a distinct gut microbiota. Another possible explanation is that, in addition to other factors such as diet, unique properties of the human gut microbiota contributed to the evolution of human physiology. For example, if over time the human gut microbiota shifted in a way that confers an increased capacity for energy production and storage, it could have promoted increased brain size during human evolution. Similarly, because the human gut microbiota plays a role in regulating host energy intake and fat production [43], and an increased capacity to store energy as fat has been hypothesized to have enabled humans to develop larger brains [18, 19], changes in the gut microbiota that affected host metabolic pathways could have contributed to the evolution of the human brain as well. To distinguish between these two alternatives, further studies are necessary that measure the metabolic potential of the human and non-human gut microbiota and more directly compare the physiological consequences of consuming a Western diet. However, a recent study suggests that large-brained primates endure seasonal periods of food limitation more successfully than small-brained primates [44]. Together with evidence that gut microbes compensate for periodically reduced energy intake in some wild primates [45], these data could indicate a role for the gut microbiota in buffering hosts nutritionally against energetically expensive physiological adaptations.
Finally, our data indicated that, in addition to the gut microbiota, the physiological responses of humans and non-human primates to a Western diet may be distinct. Although vervets have an adverse metabolic reaction to a Western diet and can become obese in captivity [46, 47], most of the TWD-fed vervets did not gain weight during the study (Additional file 1: Table S8). It is possible that two factors restricted vervet weight gain: (1) finite food availability and (2) study duration (6Â months). We regard these factors as unlikely to be limiting because the vervets were provided with enough food to result in a 10Â % daily surplus, and no obvious reduction in food intake was observed [47]. Moreover, a sufficiently high-protein, high-fat diet can impact the gut microbiota almost immediately in humans, and changes are similar to those observed across populations with distinct diets [24], suggesting that a 6-month interval was sufficient. Although the captive vervets weighed more than the wild vervets both before and after the diet challenge, indicating an effect of captivity on body weight, this effect is likely a result of increased food availability and decreased activity, both of which would be expected to exacerbate weight gain and negative health outcomes on a Western diet, not mitigate them. While additional research is necessary to confirm and examine this pattern in more detail, we suggest that a side effect of the proposed unique human gut microbiota may be an increased susceptibility to obesity and metabolic disorders, particularly when hosts are consuming a high-protein, high-fat diet. If the non-human primate gut microbiota does, in fact, possess properties that make it resistant to obesity when subjected to a Western diet, it may open new avenues of exploration for translational metabolic therapies. What might confer this resistance and whether it can be maintained over generations of Western diet consumption remains to be investigated, but understanding these factors could help develop treatments to improve human resistance to obesity via the gut microbiota.