Increased abundance of Sutterella spp. and Ruminococcus torques in feces of children with autism spectrum disorder

1Autism spectrum disorder (ASD) is a neurodevelopmental disorder where there is evidence of gastrointestinal (GI) disturbance in many affected individuals. Several studies have demonstrated an altered GI microbiota in children with ASD compared with controls. Recently, Williams et al. reported a significantly higher prevalence of Sutterella spp. in biopsies taken from the GI tract of ASD children with GI disturbance compared to controls with GI disturbance. Bacteria of the genus Sutterella have been identified in canine and human feces. Sutterella wadsworthensis is noteworthy as it has been associated with some GI infections in humans. In one study using culture-based techniques, less than 1% of individuals had detectable fecal levels of S. wadsworthensis, even among individuals with bowel disorders. However, in another study, where it was shown that the abundance of S. wadsworthensis in colonic biopsies of adults with inflammatory bowel disease (IBD) was not different from controls, Sutterella appeared to be more prevalent in the human gut than previously reported.

2A previous study by our group used quantitative real-time PCR (QPCR) to compare the abundance of a range of bacteria in feces of children with ASD, their siblings and community controls, and demonstrated a low relative abundance of Bifidobacterium spp. and the mucolytic bacterium Akkermansia muciniphila in children with ASD. We have carried out further QPCR analyses of samples from that study to examine whether the abundance of Sutterella spp. is also altered in the feces of children with ASD relative to controls. We have also examined whether fecal numbers of some other mucosa-associated bacteria, namely Ruminococcus torques and Ruminococcus gnavus, are altered in individuals with ASD. As with A. muciniphila, mucosal numbers of Ruminococcus species are altered in individuals with IBD.

3Fecal samples were collected from 23 children with ASD (3 without siblings), 22 typically developing siblings (SIB) and 9 community controls (CON) without a family history of ASD. The study complied with National Health and Medical Research Council (Australia) guidelines relating to ethical conduct in human research. Participants’ caregivers completed a functional GI disorder (FGID) questionnaire. Procedures for fecal sample collection and processing, as well as microbial DNA extraction, have been described previously. QPCR was used to estimate the abundance of Sutterella spp., R. torques and R. gnavus in feces using previously described primers. Methods for QPCR, data analysis and primers for total bacteria were also as reported previously. For Sutterella, SsoFast™ Probes Supermix was used instead of EvaGreen® Supermix and an updated version of Bio-Rad CFX Manager software (Version 2.0) was used. Statistical analyses were carried out using SPSS for Windows™ (version 17, SPSS Inc., Chicago, IL, USA). Analyses were carried out using analysis of variance (ANOVA) followed by Dunnett’s test or by Student’s t-test. A P value of less than 0.05 was considered statistically significant.

4The absolute numbers (per gram of feces) and relative numbers of bacteria in feces of individuals from the ASD, SIB and CON groups are presented in Table 1. Sutterella was detected in the majority of samples analyzed (ASD 21/23; SIB 19/22; CON 9/9). ANOVA revealed significantly different absolute and relative numbers of Sutterella in ASD children compared with sibling and community controls (P = 0.044 and P = 0.05, respectively). A post hoc analysis showed significantly increased absolute and relative numbers of Sutterella in individuals with ASD compared with community controls (P = 0.028 and P = 0.03, respectively). Moreover, absolute fecal numbers of Sutterella tended to be higher in siblings than community controls (P = 0.047). R. torques and R. gnavus were detected in all fecal samples. Absolute numbers of R. torques were not significantly higher in ASD children than in other groups (ANOVA, P = 0.08) but were higher in siblings relative to community controls (P = 0.046). Absolute and relative numbers of R. gnavus did not differ significantly between groups. Significantly elevated absolute numbers of R. torques were evident in children with ASD whose caregivers reported them having a FGID (nine individuals) compared to those without reported FGID (14 individuals) (Student’s t-test, P = 0.008). There was no significant difference in fecal numbers of Sutterella or R. gnavus between ASD children with and without reported FGIDs (data not shown).

5Our results demonstrate that numbers of Sutterella are elevated in the feces of children with ASD relative to community controls. This confirms and builds on the findings of Williams et al., who showed high detection rates of these bacteria in GI biopsies taken from ASD children with a GI disturbance. In our study there was no evidence of a difference in numbers of Sutterella between ASD children with or without caregiver-reported FGID. It is not yet evident what the consequences of an increase in fecal Sutterella populations means but it is possible that under specific conditions these bacteria could cause infection. Our findings add to the growing list of bacterial groups whose abundance appears to be altered in the GI tract of children with ASD. Our results also indicate that fecal samples may suffice for the detection and quantification of Sutterella in the human gut, including children with ASD. This contrasts with a previous estimation of detectability and abundance in human fecal samples by Engberg et al.. Previously it has been difficult to culture and isolate anaerobic and microaerophilic bacteria, such as Sutterella, using traditional microbiological techniques and it is likely the QPCR method we employed was more sensitive than the culture method used by the previous study. However, our detection rates of Sutterella in humans are comparable with detection rates in colonic biopsies from human adults.

6In this study we also extended our analysis of mucus-degrading bacteria in feces of ASD children, having previously shown a low abundance of the mucolytic bacterium A. muciniphila. We measured the abundances of R. torques and R. gnavus, both of which can degrade mucus and have been associated with GI disturbance. We have shown there is a trend of increased fecal numbers of R. torques in ASD children. This pattern of decreased numbers of A. muciniphila but increased numbers of R. torques in feces of ASD children is congruent with the pattern observed in the mucosa of adults with IBD. In the latter study it was also found that growth of A. muciniphila was inhibited by co-culture with MUC2, the predominant form of mucin in the large bowel, whereas this mucin augmented the growth of other bacteria such as R. torques. Although it is likely that the populations and activities of microbes that adhere to the mucus lining the large bowel may differ substantially from those in the feces, it is also likely that some changes in gut mucosal microbial populations are reflected in, and hence detectable in, feces. Changes in the amount of mucus produced may drive changes in mucus-degrading microbes in both the mucosa and in feces, as mucus can be incorporated into the fecal stream. Individuals with IBD or ASD may have changes in large bowel mucus production that could impact (or are a result of changes in) the mucosal barrier of the gut. Indeed, increased gut permeability has been reported in a subgroup of children with ASD. Further studies that look at gut mucus production and turnover in children with ASD are warranted.

7In summary, we show further evidence of changes in the gut microbiota of children with ASD, now demonstrating that fecal abundances of Sutterella and R. torques are altered (the latter in those reported to have FGID). The role that these particular bacteria play in ASD is not yet understood but they may simply reflect larger scale shifts in the microbial populations as a result of the condition.

8ANOVA: Analysis of variance; ASD: Autism spectrum disorder; CON: Control group; FGID: Functional gastrointestinal disorder; GI: Gastrointestinal; IBD: Inflammatory bowel disease; QPCR: Quantitative real-time polymerase chain reaction; SIB: Sibling group.

9The authors declare that they have no competing interests.

10CTC, JPG, LW, MAC, MJS and MTA contributed to the design and planning of the study. LW conducted the experiment and analytical methods. CTC, LW and MAC contributed to the analysis of the results, LW, MAC and MTA prepared the manuscript, and all authors read and approved the manuscript.