The human gut bacteria Christensenellaceae are widespread, heritable, and associated with health

1The composition of the human gut microbiome is now well established as a factor important to human health conditions, including metabolic, pathogen, and immune-related diseases. Its composition varies substantially between individuals and populations due to local, personal, and stochastic factors. The high inter-individual variability of the gut microbiome has challenged efforts to define what constitutes a healthy versus an unhealthy microbiome. Indeed, community composition alone is generally not a good predictor of disease state. The contribution of specific taxa, their metabolic pathways, and their interactions to human health is a new priority for microbiome research, and this deeper understanding of the microbiome will be necessary for the development of evidence-based microbial therapeutics. Given that thousands of microbial species and strains live in the gut, one challenge is to identify targets for further investigation and development.

2Here, we focus on the family Christensenellaceae, within the Firmicutes phylum of Bacteria, due to its emergence as a health-related group. First encountered from 16S rRNA gene sequences alone, the family was named in 2012 after an isolate named Christensenella minuta (pictured in Fig. 1), cultivated from the feces of a healthy Japanese male. Members of this family of Firmicutes are, with a few exceptions, increasingly revealing themselves as associated with a healthy phenotype in humans. Because of the relatively recent naming and phylogenetic placement of the Christensenellaceae family (Box 1), it was not discussed in the literature prior to a few years ago. And since representatives of this family were only recently isolated (Box 2), little is known about its ecology outside of what can be inferred from its associations with host factors and other microbiota (Box 3). Here, we review the literature to date, focusing on consistent trends that associate Christensenellaceae with parameters of human health. Taken together, these various observations strongly argue for further investigation into the Christensenellaceae.

3Most of what is known about the family Christensenellaceae comes from 16S rRNA gene surveys of the microbiome obtained from feces of humans and other animals. Given that Christensenellaceae 16S rRNA gene sequences were relatively recently included in reference databases, only microbiome studies published since 2013 report this taxon. Two cultured isolates, Christensenella minuta and Catabacter hongkongensis, have published genomes, and genomes constructed during metagenomic assemblies are increasingly available. At the time of writing this review, there are 11 Christensenellaceae genomes in the Genome Taxonomy Database and 89 genomes for the order Christensenellales (Box 1). A phylogeny of 9 members of the Christensenellaceae, based on full length 16S rRNA gene sequences available in NCBI, is shown in Fig. 2. Surveying the post-2013 literature, it is evident that members of the Christensenellaceae are cosmopolitan inhabitants of the animal gut (Table 1), with a likely preference for the distal colon, which is consistent with its fermentative activities (detailed in Box 3).

4In humans, the family comprises on average 0.01% of the fecal microbiota. Its fine-scale distribution along the human gastrointestinal tract remains to be clarified; but in addition to feces, Christensenellaceae has been detected in human colonic mucosa, ileum, and appendix, and there is also suggestive evidence of airway colonization. The family Christensenellaceae is widespread across human populations, and is reported from subjects inhabiting North America, South America, Europe, Asia, Africa, and Australia.

5Within human populations, traits associated with different relative abundances of Christensenellaceae include ethnicity and sex. For instance, a recent study of > 2000 individuals with various ethnicities residing in Amsterdam, Deschasaux et al. reported that Dutch subjects harbored the greatest relative abundances of Christensenellaceae. Similarly, Brooks et al. compared microbiome variation between ethnicities in 1673 people residing in the USA and reported that Christensenellaceae was overall less represented in fecal samples of Asian-Pacific Islanders relative to other ethnicities. A greater relative abundance of Christensenellaceae in women compared to men was also observed, and similar observations have been reported in animals. The underlying causes of these ethnic and sex differences are unclear.

6Christensenellaceae has been associated with human longevity, based on the observation that the relative abundance of Christensenellaceae is greater in centenarians and supercentenarians in comparison to younger individuals in populations in China, Italy, and Korea. Positive associations of Christensenellaceae with age have also emerged from studies with relatively young individuals across multiple geographic locations (Table 2). Given that none of these studies followed the same individuals over time, the association with age could reflect a cohort effect rather than an age effect. For example, dietary patterns that vary by age may influence this association (see below), or individuals born earlier may have always harbored greater levels of Christensenellaceae compared to those born later.

7Host genotype is estimated to influence 30–60% of the variation in the relative abundance of Christensenellaceae across individuals. Of the hundreds of taxa in the gut, the family Christensenellaceae is consistently identified as among the most highly heritable. This means that a significant proportion of the variance in the relative abundance of the family across a population can be attributed to genetic factors. Heritability refers to the genetic predisposition of a quantitative trait: for example, height is heritable, because this trait is largely genetically determined. Heritability calculations take into account quantitative measures of the trait (such as relative abundance) and should not be confused with whether the Christensenellaceae are inherited (i.e., vertically transmitted) from family members, which is not known.

8Goodrich et al. first identified the Christensenellaceae as heritable in a well-powered (n = 977) study of monozygotic and dizygotic twins from the UK. A remarkable 40% of the variation between individuals in the relative abundance of the family Christensenellaceae could be attributed to host genetic factors. A more fine-grained analysis of species-level operational taxonomic units (OTUs) showed that just a few Christensenellaceae OTUs were driving the heritability of the family. Other studies of heritability employing the same population have observed similar results, whether the analysis was specific to species-level OTUs or when analyzing modules of co-occurring microbes. So far, the Christensenellaceae have not been included in the analysis of heritability based on shotgun metagenome data, due to the absence of genomes for this family in the reference databases used.

9The high heritability of the Christensenellaceae has been corroborated in other human populations. Goodrich et al. had confirmed its heritability in two previous studies involving twin pairs from the USA. Additionally, Lim et al. evaluated microbiome heritability in a Korean cohort of 655 individuals and identified Christensenellaceae as heritable. In a Canadian cohort (n = 270), it was again identified as among the most highly heritable taxa. Together, these observations across multiple populations indicate that the heritability of the Christensenellaceae is a widely shared trait. That individuals are genetically predisposed to harbor a high or low relative abundance of these bacteria may be a generalizable human trait.

10So far, attempts to identify the genetic factors that account for the high heritability of Christensenellaceae by genome-wide association (GWA) have not succeeded. These studies are generally underpowered, given the millions of tests conducted simultaneously (i.e., testing all genetic variants against all microbiome traits), and the necessity to correct for false positives. An alternative to GWA is to take a candidate gene approach, restricting the analysis to genes with interesting functions. For instance, Zakrzewski et al. examined the relationship between a SNP in the interleukin 23 receptor (ILR23) gene and the microbiome of mucosal biopsies from the ileum and rectum. The A allele of this variant has been associated with a reduced risk of ileal Crohn’s disease (CD). Within a population of individuals with no signs of CD or other gastrointestinal disorders, a significantly greater relative abundance of Christensenellaceae was detected in the feces of individuals harboring the protective allele (AG genotype) compared to the population with the GG genotype. How the IL23R genotype may affect members of the gut microbiota remains to be clarified.

11Christensenellaceae has also been associated with the fucosyltransferase 2 (FUT2) gene, which encodes an enzyme responsible for ABO blood group antigens that are expressed on the intestinal surface as well as secreted. Non-secretors (AA genotype) have an elevated risk for CD, while secretors (AG or GG) are less likely to develop CD. A re-analysis of healthy individuals studied in showed that secretors harbored relatively more of this family compared to non-secretors (n = 24). It is important to note that in this case a targeted approach was used, and subsequent studies associating the microbiome with FUT2 do not reach this same conclusion. When Davenport et al. also did this analysis in UK twins (n = 1503), where heritability of Christensenellaceae was first reported, no link between Christensenellaceae and secretor status was found, which is consistent with the results of Turpin et al. in a cohort of 1190 healthy individuals.

12The Christensenellaceae may interact with host genetic status to affect risk of colorectal cancer (CRC). Le Gall et al. reported elevated Christensenellaceae in healthy controls relative to individuals with CRC (n = 50 age- and sex-matched individuals per group), yet Yazici et al. observed that the relative abundance of Christensenellaceae in stool was higher on average in African-American CRC patients compared to controls. Furthermore, using tumor and healthy mucosal tissue biopsies from 44 patients with five different loss-of-function mutations in CRC, Burns et al. observed that the association of Christensenellaceae with CRC was dependent on the type of mutation present. These findings may offer an explanation for the inconsistent patterns of Christensenellaceae abundance with respect to CRC status. However, whether the Christensenellaceae participate in CRC pathology remains to be ascertained. While associations between Christensenellaceae and host genotypes remain to be reproduced, they suggest that health/disease promotion by these genotypes may be mediated in part through promotion of the Christensenellaceae.

13Body mass index (BMI) was the first host phenotype associated with the relative abundance of Christensenellaceae in the gut. Goodrich et al. observed that Christensenellaceae was significantly enriched in individuals with a normal BMI (18.5–24.9) compared to obese individuals (BMI ≥ 30). Since this initial observation, the association of Christensenellaceae with a normal BMI has been corroborated repeatedly in populations from a number of countries that included adult men and women of various ages (Table 3). Consistent with its association with leanness, Christensenellaceae have been shown to increase after diet-induced weight loss. Although obese and lean subjects can often be differentiated using aspects of microbial ecology of the gut, these aspects (e.g., alpha-diversity, or abundances of phyla) have differed between studies: the link between Christensenellaceae and BMI therefore stands as the strongest corroborated association between the gut microbiome and BMI.

14BMI is a proxy for adiposity, and consistent with reports linking levels of Christensenellaceae with BMI, studies in which adiposity is more directly measured have also noted strong associations with the abundance of Christensenellaceae in the gut. For instance, Beaumont et al. correlated adiposity measures, determined using dual x-ray absorptiometry (DEXA), with the microbiome in a study of 1313 UK twins. At the family level, the most significant association was with Christensenellaceae, which negatively correlated with visceral fat mass, a type of fat that is considered a cardiometabolic risk factor. A similar observation was made by Hibberd et al., who reported significant negative correlations of Christensenellaceae with trunk fat and android fat. Additionally, Christensenellaceae has been negatively correlated with waist circumference and waist to hip ratio, which are direct markers of central adiposity.

15In addition to its association with body fat measures, Christensenellaceae is negatively correlated with serum lipids in several studies. In the Dutch LifeLines DEEP cohort (n = 893), Fu et al. reported a negative correlation of Christensenellaceae with BMI, together with a strong association with low triglyceride levels and elevated levels of high density lipoprotein (HDL, or “good cholesterol”). Other groups have also reported that Christensenellaceae is associated with reduced serum triglycerides. Similarly, this family is also negatively associated with total cholesterol, low density lipoprotein (LDL; or “bad cholesterol”), and apolipoprotein B, a component of LDL particles.

16Christensenellaceae is reported as depleted in individuals with metabolic syndrome (MetS) compared to healthy controls. In addition to excess visceral fat, MetS includes other risk factors such as dyslipidemia and impaired glucose metabolism, and is a risk factor for type 2 diabetes and cardiovascular disease. Christensenellaceae was identified in a cohort of 441 Colombians as positively associated with a lower cardiometabolic risk score, and others report it is negatively correlated with blood pressure, which is often elevated in MetS. Christensenellaceae has also been associated with healthy glucose metabolism and Christensenellaceae OTUs are reduced in individuals with pre-type 2 diabetes. Given that a high BMI, impaired glucose metabolism, dyslipidemia, and other aspects of MetS are comorbidities, it is not surprising that Christensenellaceae inversely tracks with many of these conditions. The mechanism underlying its negative association with MetS remains to be elucidated.

17Metabolic disorders are often linked to dietary patterns. The Christensenellaceae appear to be responsive to diet, and evidence points to a role in protein and fiber fermentation. On a coarse level, large-scale diet studies have associated Christensenellaceae with healthy dietary habits low in refined sugar and high in consumption of fruit and vegetables. Christensenellaceae is reported higher in relative abundance in humans with an omnivorous diet, relative to vegetarians, and has also been associated with dairy consumption. In a more direct link, Christensenellaceae has been shown to respond rapidly to an increase in animal products in the diet. Furthermore, Christensenellaceae has been positively associated with gut metabolites typical of protein catabolism and dietary animal protein. Christensenellaceae has also been reported to increase in human dietary interventions involving prebiotic fibers such as resistant starch 4, galacto-oligosaccharide, and polydextrose. Similar observations have also been made in rodent models. Taken together, these studies indicate that the association of Christensenellaceae with health parameters may in part be due to its association with a diet high in protein and fiber.

18To test for a causal role for Christensenellaceae in metabolic disease while controlling for diet, Goodrich et al. selected an obese human donor based on almost undetectable levels of Christensenellaceae in the microbiome, and performed fecal transfers to germfree mice that were fed the same fiber-rich chow, but otherwise only differed by whether or not the obese human microbiome inoculum was amended with C. minuta. These experiments showed that amendment with C. minuta reduced the adiposity gains of mice compared to those that received unamended stool (or stool amended with heat-killed C. minuta). The mechanism underlying the protective effect of C. minuta against excess adiposity gain remains to be elucidated, but may involve re-modeling the microbial community, as a shift in diversity was observed when C. minuta was added. These experiments demonstrated that the activity of C. minuta in the gut microbiome can affect host body composition even when diet is controlled for, possibly via interactions with other members of the microbiota. Indeed, the ecological role of members of the Christensenellaceae and their function in the gut in general remains to be better understood (Box 3).

19In a meta-analysis of inflammatory bowel disease (IBD) that included over 3000 individuals, Mancabelli et al. reported Christensenellaceae as one of five taxa considered a signature of a healthy gut. Indeed, Christensenellaceae were consistently depleted in individuals with Crohn’s disease and ulcerative colitis, the two major sub-types of IBD. In irritable bowel syndrome (IBS), a gastrointestinal disorder characterized by abdominal pain and abnormal bowel movements, a higher relative abundance of Christensenellaceae in healthy controls relative to individuals with IBS has been reported in several studies. Several studies have also noted a positive correlation of Christensenellaceae and longer transit time or even constipation. Thus, the Christensenellaceae appear to be depleted in conditions associated with inflammation and fast transit time.

20Given Christensenellaceae’s link with transit time, it is perhaps not surprising that the family has been linked to affective disorders that impact gut motility. For instance, gastric dysfunction, particularly constipation, affects approximately two-thirds of patients with Parkinson’s disease (PD) and multiple sclerosis (MS). Studies have noted a greater relative abundance of Christensenellaceae in PD and MS patients relative to healthy controls. Since diet is also related to gut transit time, the effects of diet, host status, and host genetics remain to be carefully disentangled to better understand how levels of the Christensenellaceae are controlled.

21The family Christensenellaceae is a relatively recently described bacterial family that is highly heritable and shows compelling associations with host health. Its strong ties to host health have warranted the suggestion that cultured representatives of the Christensenellaceae, such as C. minuta, should be considered for use as a therapeutic probiotic for the improvement of human health. However, the functional role of Christensenellaceae in the gut remains to be understood. While the collection of associations between Christensenellaceae and host health parameters continues to grow, allowing inferences about the role of these bacteria, they remain to be studied experimentally. Genomes offer a powerful platform for generating hypotheses regarding the metabolic capacity of the Christensenellaceae, but further functional characterization in vitro and in vivo will be necessary to fully characterize the role of Christensenellaceae in the gut. The ecological role of members of the Christensenellaceae, their interactions with other members of the microbiome and with the host and host diet, all remain to be better understood if these intriguing microbes are to be harnessed fully to improve human health.