Could selected gut microorganisms be diagnostic biomarkers for autism spectrum disorders? Study based on a commercial microbiota test

The early-life modications of intestinal microbiota may impact children's subsequent emotional and cognitive development. Stool samples of three groups of children (including probiotics and non-probiotics) were analyzed. In cases where probiotics were not used, Bidobacterium spp. levels differed between the ASD and ALG. In healthy non-probiotics use children, strong positive correlations were observed in E. coli and Enterococcus spp. and Bacterioides and Klebsiella spp. In the non-probiotic use ASD group, a strongly negative correlation was observed in Lactobacillus spp., and Faecalibacterium prausnitzii and Akkermansia muciniphila levels. The presented results did not show conclusive evidence that changes in the level of selected intestinal microorganisms are symptomatic for children with ASD. results using descriptive correlation for factor analysis of component loadings for multi-presence, multiple-bacteria to determine homogenous subgroups. 12 strains factor wise


Introduction
The human gut microbiota is a complex, non-homogenous ecosystem represented by 10 13 -10 14 microbes, with over a thousand different species, which possess a hundred fold more genes than found in the human genome. Strict anaerobic bacteria are the primary microcrobes found in the gut, but protozoa, fungi, archaea, and viruses are also detectable. Microbiological colonization of the intestines begins during childbirth. The type of delivery (vaginal or Caesarean section) and gestational age of birth (pre-term or full-term) may play a signi cant role in post-natal development, as well as in the maturation of endocrine, immune, and nervous systems [2][3][4] . According to recent studies, early-life modi cations of intestinal microbiota may affect subsequent emotional and cognitive development. The diversity of gut microbiota may be crucial for the successful implementation of behavioral skills and proper brain development [5][6][7] .
In recent years, an increasing number of studies reported that gut microbiota might participate in the process of maintaining human homeostasis through the regulation of mood and well-being, involvement in enteric and central nervous system development, and controlling appetite and metabolism. A bidirectional communication pathway exists between intestinal microorganisms and the brain, known as the gut-brain axis, enabling gut microbes to communicate with the brain, while also acting in an inverse neurological changes such as autism, schizophrenia or Parkinson's disease 6,16 and can impact the severity of psychiatric disorders including depression, stress or anxiety 8, 17,18 .
Autism spectrum disorders (ASD), referenced by neurological and developmental dysfunction, are manifested by de ciencies in social communication skills, lack of reciprocal social interactions, and unusual repetitive behaviors. Generally, ASD includes different developmental disorders: the classic form of autistic disorder, Asperger's Syndrome, and Pervasive Developmental Disorder -Not Otherwise Speci ed 19 . Various studies have shown that genetic implications and environmental factors (chemicals, drugs, diet, prenatal viral infections) can be associated with ASD etiopathogenesis [20][21][22] . Moreover, according to recent ndings, the abundance of some bacterial species, as well as intestinal microbiota composition, may play a crucial role in ASD development and gastrointestinal (GI) problems, which is characteristic for individuals with autism and can be due to gut dysbiosis. A link between alterations in gut microbiota composition and ASD is not well established [23][24][25][26][27][28] .
Several studies have reported children with ASD present more frequently with gastrointestinal problems such as abdominal pain, constipation or diarrhea, bloating, and/or gastroesophageal re ux than in healthy individuals. In turn, chronic Gl disturbances may also aggravate behavioral problems, such as frustration and aggression, and are speculated to correlate with the severity of autism [29][30][31][32][33] . The cause of these intestinal disorders is unknown. However, they appear to be associated with a disarrangement of gut microbiota, particularly, in the excessive growth of pathogenic bacteria (e.g. Clostridium spp.) and the decrease of bene cial microorganisms (such as Lactobacillus and Bi dobacterium) 28, 34 The purpose of this study was to examine selected gut microorganisms, both bene cial and pathogenic, in the feces of three groups of children: healthy, with allergies (ALG), and with ASD. Children with allergies were treated as the positive control group. Multiple studies have reported that gut microbiota of allergic patients shows, similar to ASD children, a signi cant abnormalities in the composition of gut microorganism [40][41][42] . In our study, we evaluated 19 gut microorganisms by comparing their composition in ASD subjects to neurotypical children. We also considered differences and similarities between groups, trying to determine if microbiota imbalances could be the basis for manifestation of, or a marker for, ASD.
Our analysis based on the diagnostic intestinal microbiota test enables the detection and identi cation of foundation and keystone bacterial species in the intestinal ecosystem. Nowadays, diagnostic tests for gut microorganisms are becoming easily available and increasingly used in commercial diagnostics. In this way, we tried to emphasize the practical application of scienti c research and link it with the diagnostic process.

Results
Taking into consideration the Klebsiella spp. level between ASD, ALG, and healthy group participants, we observed signi cant differences (p = 0.0055) regardless of probiotic usage. The Dunn post-hoc test showed a signi cantly higher median level of Klebsiella spp. in the healthy group compared to ALG children (p = 0.0199), with a wider range in children with allergies. No other differences were found regardless of the bacteria species. Taking into consideration probiotics usage, a signi cant difference in the Klebsiella spp. level (p = 0.004) was observed. The Dunn post-hoc test showed that the levels differed between the healthy and ALG groups (p = 0.028). In children who used probiotics, the difference was not signi cant (p = 0.593, Fig. 1). The range was highest in the ALG group, but the median value was slightly higher in the healthy control.
When probiotics were not administered, there was a signi cant difference in Bi dobacterium spp. (p = 0.029) in the ASD and ALG group (p = 0.036) as determined by the Dunn post-hoc test. When probiotics were used, the difference was not signi cant (p = 0.278, Fig. 2). The range and median were highest in the healthy controls.
Agglomeration analysis using Euclidian distances (single linkage analysis) Ward's method, was computed for bacteria in the stool. Additionally, the correlation matrix for factor analysis of principal component loadings for multi-presence, multiple-bacteria species with unrotated factor rotation was computed to obtain homogenous subgroups. As presented in Fig. 3, differences in bacteria species presence, distribution, and coexistence in ASD, ALG and healthy group is shown. Although the joining tree shows similar coexistence of different species of bacteria, however, the two way-clustering differs between the health status in some cases. It is especially evident in the variation level of bacteria marked with red boxes and could point dysbiosis. Bi dobacteria and Bacterioides vary the most, but as shown in Fig. 3, in the ASD and the ALG children's stool, the Faecalibacterium prausnitzii level was highly variable. Factor analysis of principal component loadings for multi-presence, multiple-bacteria species highlights the changes in bacterial coexistence. Some species, due to lack of variation in the subgroups, were excluded from the analysis. The adjacent species and the Voronoi tessellation lines show the different presence of the bacteria in the stool and de ne adjacent polygons, including all peaks in the output plot.
As expected, Lactobacillus spp. and H 2 O 2 Lactobacillus were in a similar distance, always in the same quadrant and near Bacterioides. Surprisingly, E. coli and C. albicans were near in the ALG and ASD group but not in healthy children. Akkermansia muciniphila was distant from Bacterioides in all analyzed groups. It was not possible to perform the analysis based on probiotic usage due to an insu cient number of cases in those divisions.
Descriptive statistics as a summary of various microorganism species' concentration investigated in the stool of children were shown in Supplementary materials (Table S1., Table S2., and Table S3.). The Supplementary data show values in the whole group, health status divided, and probiotic usage categorized. The Clostridium di cile and molds, due to a lack of conclusive data for all cases, were excluded from statistical analysis.
In the non-probiotic, healthy control group, the Spearman's rank correlation showed a signi cant, positive, and strong correlation between E. coli and Enterococcus spp. and Bacterioides and Klebsiella spp.

Discussion
Increasing evidence suggests the balance and diversity within the bacterial population are essential in maintaining proper function of the gastrointestinal tract and immune system as well as human homeostasis. On the other hand, there are a wide range of indicators that propose an imbalance of the gut microbial ecosystem may lead to in ammation and immune activation in several disorders such as gastrointestinal diseases, cardiovascular disease, metabolic and psychiatric disorders, allergy, or asthma 6,12,15,16,41 .
The pathogenesis of ASD is complex, and apart from genetic factors, environmental, factors such as the intestinal community, may play a key role in the symptomology of ASD. The composition of gut microorganisms that increase susceptibility to autism development, as well as evidence linking autism symptoms and intestinal dysbiosis, have yet to be fully explained [19][20][21]25,43 . However, frequent occurrence of GI symptoms in ASD children suggest the involvement of the gut microbiota in gastrointestinal pathophysiology which then constitute potential diagnostic and therapeutic targets. It was suggested that dietary intervention (gluten-, casein-, and soy-free diet), probiotic/prebiotic treatment, microbiota transfer therapy, or targeted antibiotic therapy could be a new strategy for treatment. It could help children with chronic gastrointestinal disorders and may reduce ASD symptoms by improving language, cognitive skills, and behavioral de cits (Doenyas 2018 Moreover, the anaerobic bacteria Clostridium and Bacteroides are sources of short-chain fatty acids (SCFA), such as propionic, acetic, butyric, and valeric acid, usually produced during ber fermentation. These metabolites are believed to be involved in gut immune system function, modulation of the nervous system through the gut-brain axis, and host cell gene expression 26,34,51 . SCFAs can induce widespread effects on the human organism, but an imbalance in their levels may change intestinal homeostasis and cause peripheral in ammation. SCFAs reach the brain through blood circulation and affect its development by modulating production of serotonin and dopamine. High concentrations of propionic acid, a signi cant neurotoxic metabolite, may disrupt brain function, resulting in developmental delay or regression 26,52−54 .
However, the presence of speci c Clostridium species, clusters, and their content in the intestinal microbiome of ASD children is still under discussion. Moreover, current results are often inconclusive, and the contribution of selected species in ASD etiology have yet to be fully explained. Our analysis showed no signi cant differences in the levels of Clostridium spp. within the groups. The results, therefore, are similar to those of Wang and Iovene 38,55 but are not consistent with other reports that found increased Clostridium in the stool of ASD children 34,50,56,57 .
Bacterioides spp. and Clostridum spp. are de ned as bacteria associated with ber fermentation and SCFA/propionic acid production. It has been suggested that neurodevelopmental disorders in ASD patients correlate with impaired propionic acid metabolism and changes in propionate producing bacteria 54,58 . In our analysis, the level of Bacteroides is unchanged in all analyzed groups. These ndings are comparable with those of Parracho et al. and Ma et al. 56,59 but contrasts other studies where increased Bacteroides in ASD patients has been reported 34,51 . Moreover, in the microbiome of the ASD group, a signi cant increase of Proteobacteria phylum, particularly species belonging to Enterobacteriaceae, was observed 34,51,60 . However, our analysis showed no signi cant changes in the abundance of this family in stool samples of any studied groups, except a higher level of Klebsiella species in the healthy group. This result is compatible with Adams' observation 29 .
Additionally, in healthy controls, a signi cantly positive and strong correlation of Escherichia coli and Enterococcus spp. was noted. It is well-known that certain strains of E. coli and Enterococcus have probiotic properties and can activate the gut mucosal immune system by increasing antibody quantities

Some studies have also indicated varying levels of probiotic bacteria such as Lactobacillus and
Bi dobacterium in the intestines of ASD and neurotypical subjects 27,29,30,34,38,51 . Our results showed lower levels of Bi dobacterium in ASD group, which is compatible with several other studies 29,34,38,51 . We speculate this may be due to a derangement of the probiotic bacteria population in the intestines of autistic children. On the other hand, similar numbers of Bi dobacterium in ALG and healthy groups may be a compensatory mechanism. Allergies are a chronic in ammatory diseases, and this group of bacteria shows strong anti-in ammatory properties. Recently published studies have reported that Bi dobacterium strains may inhibit the in ammatory response and exert an immunomodulatory effect by stimulating IL-10 or IL-12 synthesis by dendritic cells 66,67 . Furthermore, the presence of both probiotic bacteria in the intestines contributes to maintenance of the epithelial barrier integrity and protects against an overgrowth of pathogens 55,68 . Additionally, they can impact the metabolism of toxins, drugs, and some dietary compounds as well as gut epithelial cell proliferation 68-70 . Interestingly, both genera may produce γ-aminobutyric acid (GABA), the primary inhibitory neurotransmitter in the brain 26,69 . According to some studies, lower GABA levels are correlated with anxiety and social disorders in ASD individuals 71,72 .
Moreover, some Lactobacillus and Bi dobacterium strains are the main components of probiotic supplements. Growing clinical evidence suggests the consumption of oral probiotics reduce GI discomfort, modulate the stress response, and improve mood and anxiety symptoms in patients with ASD 29,30,73,74 . However, in our analysis, we observed a strong negative correlation between probiotic bacteria and Akkermansia muciniphila and Faecalibacterium prausnitzii levels in ALG and ASD groups using probiotics. We assume this may be due to the dominant role of some probiotic strains or as a result of nutrient competition. Both A. muciniphila and F. prausnitzii are considered biomarkers of healthy intestinal ora and modulators of immune system 75,76 . Additionally, Faecalibacterium may regulate the expression of interferon-gamma (IFNγ), which plays an indirect role in neuroplasticity and synapse formation 26,77 . Based on these factors, it can be assumed in children with these associated diseases, selection of appropriate probiotic strains is important, and probiotic therapy should be performed on the basis of previous microbiota analysis.
In our studies, we estimated the content of fecal fungi, especially Candida genus, in ASD children. The healthy gut is colonized by yeast and good bacteria living in balance with one other. Most Candida species are harmless commensals, but when intestinal homeostasis is disturbed, they can cause infections called candidiasis. Yeast infections have been rarely investigated in autistic individuals. Our studies have shown no signi cant differences between groups. However, some investigators report substantial growth of Candida, particularly Candida albicans, in ASD patients 27,55,78,79 , Contrary to these results, Adams et al. did not con rm these ndings 29 . The potential role of the Candida species in ASD etiology is unclear, and further studies are needed. It is believed that an overgrowth of Candida spp. may induce autistic behavior through excessive production of ammonia which then is converted to betaalanine, a non-essential amino acid structurally similar to the inhibitory neurotransmitter GABA 79,80 . Additionally, a high abundance of yeast may impair the absorption of both carbohydrates and mineral elements, as well as affect the release and accumulation of toxins 78,80 . Moreover, enormous growth of Candida in the gut of autistic individuals may aggravate GI abnormalities by dysregulation of cytokine release 27 .

Conclusions
The results of our study do not fully support the hypothesis that the composition of the gastrointestinal microbiota, the presence of certain species, or signi cantly altered ratios of these microbes change susceptibility to ASD development of children. The formation of intestinal microorganisms is in uenced by many factors such as the type of delivery or feeding, child's diet later in life, and even geographical location. Typical research methods are often heterogeneous and do not include this additional information. However, it cannot be excluded that ASD etiopathogenesis is likely multifactorial and involves multiple etiopathogenic mechanisms. Despite the complexity of this issue, it can be assumed that the increased abundance of certain harmful bacterial species, as well as reduction of bene cial ones, in autistic individuals may result in intensi ed gastrointestinal problems. For these reasons, an analysis of intestinal microbiota along with an exclusion diet enriched with probiotic/prebiotic supplementation could help alleviate GI symptoms and improve the quality of life of ASD children.

Participants
This study aimed at comparing 89 stool specimens from children who were enrolled to gut microbiota test at the Institute of Microecology (Poznan). The children's parents completed a self-reporting questionnaire consisting of a set of questions regarding their child's: age, sex, body mass, height, health status, probiotic or/and antibiotic supplementation, and radiotherapy/chemotherapy treatment. Based on this information, three children's groups were distinguished: without existing illnesses, with allergies, and with ASD. The individuals with autoimmune diseases, Lyme disease, diabetes, cancer, and children who had undergone antibiotic therapy within three months, radiotherapy, and chemotherapy were excluded from further analysis. Neurotypical children and allergy sufferers were not related to any degree with ASD children. The children were enrolled regardless of any gastrointestinal disturbances. According to the exclusion criteria, 73 children were included in further statistical analysis: 16 healthy children, 24 with allergies, and 33 with ASD. In the ASD group, 9 out of 33 children (27%) and 7 out of 27 (29%) of ALG group reported gastrointestinal symptoms, while none of the children in the healthy group. The descriptive statistics for the study group was shown in Table 1. The healthy children and those with allergic diseases were treated as negative and positive controls, respectively. The data were analyzed between the illness status groups as well as probiotic usage. Informed consent was obtained from all parents or legally authorized representatives, and identifying information was removed from each sample. The study protocol was in accordance with the Declaration of Helsinki and was approved by the Ethical Committee of Poznan University of Medical Sciences.

Materials and Procedures
Collection and preparation of stool samples Stool samples were collected in sterile stool tubes at the participants' homes and delivered to the Institute of Microecology (Poznan), where they were analyzed for selected intestinal microorganisms. The analysis was carried out following the KyberKompakt Pro protocol and included both microbiological cultures and quantitative polymerase chain reactions (qPCR). All counts were expressed as the numbers of log 10 CFU (colony-forming unit) per 1 g of sample.
Microbiological identi cation of selected microorganisms Before microbial culture, 0.25 g of each sample was diluted ten times in 0.85% sterile NaCl solution, suspended by vortexing, and subsequently plated on selective and differential agar medium plates. Both cultures and microscopic observations determined the presence of fecal fungi. Samples were suspended in 0.85% sterile NaCl solution containing trypsin and antibiotics (penicillin-streptomycin). Afterward, samples were inoculated into two Sabouraud agar plates with antibiotics (gentamicin and chloramphenicol). After 2-5 days, yeast colonies were transferred to differential plates were assigned to the taxonomy species group. Molds were examined by morphological observation after 5-7 days of incubation.
To analyze anaerobic, unculturable bacteria such as Akkermansia muciniphila and Faecalibacterium prausnitzii and determine Clostridium di cile numbers, quantitative polymerase chain reaction was used.

DNA isolation and quantitative PCR analysis
Bacterial DNA from stool samples was extracted using RIDA® Xtract kit in accordance with the manufacturer's instructions. To estimate bacterial quantity, qPCR was performed with the use of RIDA®GENE Akkermansia muciniphila, RIDA®GENE Faecalibacterium prausnitzii, and RIDA®GENE Clostridium di cile kits in accordance with the manufacturer's instructions (R-Biopharm AG, Darmstadt, Germany). The thermal pro le was as follows: initial denaturation (1 min, 95°C), then 45 cycles of denaturation (15 sec, 95°C) and annealing/extension (30 sec, 60°C). The total reaction mixture volume was 25 µL containing 19.9 µL reaction mix, 0.1 µL Taq Polymerase, and 5 µL DNA-extract. The standard curve was generated with standard DNA A: 5 × 10 2 copies/reaction, standard DNA B: 5 × 10 4 copies/reaction, and standard DNA C: 5 × 10 6 copies/reaction. The reaction was performed in the RotorGene thermal cycler (QIAGEN, Manheim, Germany). The nal number of bacteria/gram of stool was obtained by multiplying by 200 due to the dilution factor of the stool sample during extraction. Statistical analyses Several statistical analyses were performed using Statistica ver. 13 software (TIBCO Software, Tulsa, USA). The distributions of the continuous variables were assessed with the Shapiro-Wilk test. As the data was not normally distributed, a nonparametric, 2-sided Kruskal-Wallis test with Dunn's post-hoc test for multiple comparisons was used. Nonparametric Spearman's rank correlation test was applied to determine the strength of a link between microbe species. For individual comparisons, a p-value of < 0.05 was considered signi cant. The results were described using nonparametric descriptive statistics. The correlation matrix for factor analysis of principal component loadings for multi-presence, multiplebacteria strains was computed to determine homogenous subgroups. It was calculated for 12 strains with unrotated factor rotation. Missing data were case wise deleted, 80 cases were processed, and 73 valid cases were accepted. Based on the multi-presence of microorganisms, using Ward's method cluster joining and Euclidean distances, the bacteria genera were clustered.

Institutional Review Board Statement
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Poznan University of Medical Sciences.

Informed Consent Statement
Informed consent was obtained from all parents or legally authorized representatives, and identifying information was removed from each sample. Bi dobacterium spp. level in the stool of children with ASD, allergies, and in the healthy group divided by probiotic usage