Children with autism are commonly affected by gastrointestinal problems such as abdominal pain, constipation and diarrhea. In recent years, there has been a growing interest in the use of probiotics in this population, as it hypothetically may help to improve bowel habits and the behavioral and social functioning of these individuals. The gut microbiome plays an important role in the pathophysiology of organic as well as functional gastrointestinal disorders. Microbial modification with the use of antibiotics, probiotics, and fecal transplantation have been effective in the treatment of conditions such as recurrent Clostridium difficile infection, pouchitis, and irritable bowel syndrome. The present review presents a number of reported clinical, immunological and microbiome-related changes seen in children with autism compared to normally developed children. It also discusses gut inflammation, permeability concerns, and absorption abnormalities that may contribute to these problems. Most importantly, it discusses evidence, from human and animal studies, of a potential role of probiotics in the treatment of gastrointestinal symptoms in children with autism.

The influence of the enteric microbiota on the human body has only started to be unveiled. Its impact is wide, as it has been shown to affect a number of processes including the immune response, metabolism, and neurologic function[1-3]. The disruption of the normal commensal microbial community in humans, also called “dysbiosis”, is associated with an increasing number of disorders such as inflammatory bowel disease, irritable bowel syndrome, obesity, hypertension, diabetes, and autism[4-8]. The aim of the present review is to synthetize current data on the association between microbiota dysbiosis and autism, and to assess if its modification could have a beneficial effect in children with autism.

Autism is a neurodevelopmental disorder which affects social interaction, verbal and non-verbal communication, and behavior. A recent report from the Centers for Disease Control and Prevention indicates a rise in the prevalence of autism in children to one in 68 children in the United States (78% increase since 2007)[9].

Children with autism spectrum disorders (ASD) are among the populations that are most often referred to the Pediatric Gastroenterology clinic. During a two-year period, 3% (121/4013) of children seen by 4 pediatric gastroenterologists for various abdominal complaints in our clinic had an underlying ASD (C. Bearden, U.T. Bioinformatics, personal communication 9-24-2016). The true prevalence of gastrointestinal symptoms (GIS) in ASD is not known, but available data suggest a figure approximately 40%[10]. Wang et al[11] reported data obtained from families with children with ASD registered in the Autism Genetic Resource Exchange (AGRE). In their study of 589 affected children, 42% had GIS. Increased autism symptom severity was associated with higher odds of having GIS[11]. Abdominal pain, constipation, diarrhea, nausea, and bloating were the most common symptoms. In the largest study, Mazurek et al[12] reported that of 2973 children in an ASD network, 42% reported GIS lasting > 3 mo. A wide range of gastrointestinal (GI) problems have been reported, including feeding abnormalities, gastroesophageal reflux, abdominal pain, diarrhea, fecal incontinence, constipation, and alternating diarrhea and constipation have been reported in one out of three children in the autism spectrum[13,14]. More recently, based on a large epidemiological study, eosinophilic esophagitis in children with ASD and dysphagia has been added to the list of disorders with increased risk in this population, compared to the general population[15]. This group of children with autism reportedly also has severe anxiety, irritability and social withdrawal symptoms, which may overshadow their GI complaints[16].

Some researchers such as Pusponegoro et al[17] have reported no differences between children with autism and controls with regard to gastrointestinal symptoms, intestinal inflammation (based on fecal calprotectin), microbiota (based on urinary D-lactate) or intestinal permeability (based on urinary lactulose/mannitol ratio). However, this group reported an increased urinary I-FABP (marker of enterocyte damage) in children with autism who had severe behavioral abnormalities, compared with autistic children with mild maladaptive behavior and compared with normal children[17].

A number of recent studies have suggested that the GIS in ASD may be a manifestation of an underlying inflammatory process. Systemic inflammation has been suggested by an excessive accumulation of receptors for advanced glycation end products (RAGE) in blood and their proinflammatory ligand S100A9 in the plasma of individuals with ASD[18]. The level of S100A9 in plasma correlated with the autism severity score. Another study hypothesized that the inflammation may be pathophysiologically related to an abnormal microbiota. They compared the metagenomic profile of ileal and colonic biopsies in children with ASD, ulcerative colitis (UC), and Crohn’s disease (CD). These investigators found that the transcriptome profiles of these tissues of children with ASD segregated apart from normal controls and alongside those with CD and UC when they used principal components analysis, as would be seen with an inflamed colon[19]. However, the authors did not identify why these tissues of ASD children had different transcriptional profiles; for example, they did not look for evidence of inflammation by assessing serum cytokines or fecal inflammatory markers such as calprotectin or interleukin-8. Other groups studying ASD have failed to show changes in gut biopsy cytokine levels[20] or changes in fecal calprotectin[21]. One must keep in mind that these studies were small, and measurable abnormalities were observed in a significant subset of with ASD (approximately 25% of those studied).

Enhanced T cell activation, heightened immunoglobulin and cytokine profiles, as well as histologic changes assessed in intestinal biopsies such as infiltration of lymphocytes, monocytes, natural killer cells and eosinophils have been described in children with autism[22-26]. These findings can be present in other gastrointestinal conditions such as food allergies and immunodeficiency[27]. In contrast, other laboratory measures of intestinal health, such as fecal levels of calprotectin, lactoferrin, secretory IgA, and elastase have found to be normal in children with autism[21,28]. In addition, reports of intestinal permeability (IP) in children with autism have been conflicting. Studies have reported abnormal IP in these children compared to controls[29,30]. Some have also reported increased IP to occur in first degree relatives of patients with autism et al[31]. In contrast, our group as well as others (mostly in small series) have found that the intestinal permeability of children with autism was not different from normal controls[17,32-34].

A recent report indicated that children with autism also have an abnormal carbohydrate digestion based on significant decrease in the expression on their intestinal biopsies of disaccharidases (sucrose-isomaltase, maltase-glucoamylase, and lactase), as well as the hexose transporters (SGLT1 and GLUT-2)[35], a finding which agreed with a previous uncontrolled study[36]. This finding was not supported by extensive observations of Kushak et al[37] from a center that performs many intestinal biopsies. These investigators had originally found that more than half of a group of children with autism had low levels of the enzyme lactase in duodenal biopsies[38]. However, in a follow-up study which included neurotypical controls, mucosal disaccharidase activity was not different comparing autistic and nonautistic individuals. Interestingly, even though the disaccharidases were within the normal range, the investigators found that children with ASD had evidence of mucosal inflammation on intestinal biopsy. Standard fecal indicators of gut inflammation, fecal calprotectin and lactoferrin were similar in both groups. A measure of gut permeability, lactulose/rhamnose ratio in urine after oral administration, was also not statistically different in patients with and without autism. Larger controlled studies are required to determine if the gastrointestinal symptoms in children with autism are in fact related to reproducible, “organic” findings, such as intestinal inflammation, to differences in nutrient digestion, or to an abnormal intestinal permeability[27].

Gastrointestinal symptoms in ASD may be simply a reflection of sensory over-responsivity to abdominal signals. However, in the authors’ opinion, the most common gastrointestinal complaints in children with ASD resemble those of adults and teens with functional bowel diseases such as irritable bowel syndrome (IBS). Irritable bowel syndrome is characterized by symptoms of diarrhea and/or constipation, typically with the relief of pain accompanying the passage of a stool, symptoms which fulfill the Rome III criteria[39]. Many children with ASD have diffuse abdominal pain and an irregular stool pattern with either diarrhea or constipation, or alternating diarrhea and constipation. We have postulated that a significant proportion of children with ASD and chronic GIS, have a form of IBS. However, the Rome III criteria are validated in adults with normal IQ but are somewhat difficult to apply to normal children, and even more so in those with ASD. When compared to GI symptom scores in ASD, which have been useful but are not validated, there is much broader experience in quantifying autistic behavior changes, such as irritability as measured by the Aberrant Behavior Checklist[40]. As mentioned, studies have shown that the presence and severity of GI symptoms correlate with the severity of underlying autism[11,28,41].

Trillions of microbes and 500-1000 species of microorganisms are natural inhabitants of our gastrointestinal tract, wherein the phyla Firmicutes, Bacteriodetes, and Actinobacteria are the most common. Anaerobic bacteria, yeasts, viruses, and bacteriophages (viruses which reside and proliferate within bacteria) also influence the gut microbial diversity[42,43]. The gut microbiome has a symbiotic interaction with the various organ systems of our body, and it is known to contribute to many GI functions, such as maintaining the integrity of the epithelial barrier, stimulating immune interactions, participating in gastrointestinal motility, and regulating drug and nutrient metabolism[44]. This normal interaction can be disturbed by a number of events, such as infections, gastrointestinal diseases, dietary changes, and neurologic disorders. Drugs such as acid suppressants, antibiotics, and corticosteroids have also been reported to perturb this homeostatic equilibrium. This dysbiosis contributes to the pathophysiology of many gastrointestinal conditions such as inflammatory bowel disease, functional gastrointestinal disease, food allergy, obesity, and liver disease[45].

The enteric microbiome of children with ASD is different from that of typically developed children. Abnormal colonization could be related to diverse factors, including a more restricted diet and exposure to more antibiotic early in life. For example, two studies found that children with ASD were more likely to be treated with antibiotics for otitis media[46,47]. Finegold et al[48] reported different levels of bacterial phyla in children with ASD by pyrosequencing. When comparing autistic children with controls there were changes in phyla Firmicutes (63% vs 39%, respectively), Bacteriodetes (30% vs 51%), Actinobacteria (0.7% vs 1.8%), and Proteobacteria (0.5% vs 3.1%)[48]. In a different study, this same group also reported the presence of non-spore-forming anaerobes and microaerophilic bacteria in gastric and duodenal aspirates from children with autism, organisms which were not present in control children[48].

As mentioned, a less diverse microbial community in gut of children with autism with lower levels of some genera (Prevotella, Coprococcus and Veillonellaceae) has been reported. Interestingly, these particular species are known to be versatile carbohydrate metabolizers; and in a controlled trial, reduced colonization correlated with autistic symptoms but not with diet pattern[49]. Other differences in individuals with ASD include the overgrowth of Clostridium species, including Clostridium histolyticum (linked to the presence of GI symptoms in one study), and low levels of Bifidobacteria, a species known to have anti-inflammatory effects[48,50,51].

Overgrowth of other bacteria such as Desulfovibrio species has also been found in children with autism and their relatives, compared to controls[52]. Additionally, higher levels of Caloramator, Sarcina, Alistipes, Akkermansia, Sutterellaceae and Enterobacteriaceae were found in children with autism compared with typically developed children[53,54]. Kang et al[49] reported a less diverse fecal microbiome by pyrosequencing of 16S rDNA in children with autism. Despite these studies, it should be noted that when bacteria tag-encoded pyrosequencing was used, Gondalia et al[55] did not find differences in the gut microbiome, comparing children with autism with their siblings.

Much work needs to be done in determining the metabolic consequences of an abnormal microbiota in ASD. Bacterial by-products are the likely mediators of systemic effects that could lead to alterations in the children’s behavior. Some investigators have hypothesized that the abnormal microbiota in children with ASD produces changes in behavior via a mechanism involving excessive production of short chain fatty acids (SCFA), such as propionate and butyrate, which represent the major anions of human feces. These SCFA can produce behavioral changes in rodents when injected into the brain ventricles or systemically via intermediates such as p-cresol that alter dopamine metabolism[56]. Ongoing investigations have begun to highlight the importance of SCFA in ASD[57,58].

Gastrointestinal symptoms in children with autism are common and are often linked to the children’s abnormal behavior and social interactions. Probiotics are hypothesized to positively impact gut microbial communities and alter the levels of specific potentially harmful metabolites in children with ASD. Whether probiotics improve behavior and these markers has yet to be determined. Although the evidence presented in this review does not confirm benefit of probiotics in this population, it provides a solid rationale for the design of larger prospective trials.