CatherinaJTV - I'm assuming that you have read the paper in its entirety.
As is standard the authors discuss the limitations of the study. I don't think we can accuse them of presenting the paper as more important than it is when they comprehensively detail its limitations. Of course many people probably don't read as far as that but that is hardly the authors' fault.
One of the study's limitations was the relatively small and unequal sample numbers, especially in the IBD groups, although the pattern of aberrant gene expression was consistent with that described previously for these disease groups. The original goal of this preliminary study was to describe the molecular phenotype(s) for the ASDGI group ? the IBD cases were included for comparison. Despite the differences in group sizes, highly significant differences emerged, particularly between ASDGI cases and non-inflamed controls.
Additional factors known to influence human intestinal mucosal gene expression include, but are not limited to, age, gender, ethnicity, prescription medications, diet, and dietary supplements. Insofar as this is a retrospective study designed primarily to explore the relationship between the ASDGI phenotype and inflammatory bowel disease, these potential confounding factors could not be adequately controlled for. The variety of diets, medications, and nutritional supplements in the ASD-GI group is depicted in Table S1. For the most part ASD-GI children were on a diet that restricted ingestion of both gluten and casein, and in some instances also soy, whereas individuals in the control groups were not on restrictive diets. In addition, food auto-restriction, a common feature in autism, serves to further limit the variety of foods to which the bowel mucosa is exposed and could potentially impact mucosal gene expression. None of the ASDGI cases in this study were receiving medications known to impact inflammatory processes of the intestinal mucosa for at least four weeks prior to obtaining the biopsies. This includes NSAIDS, H2 blockers, proton pump inhibitors, corticosteroids, antibiotics, probiotics, and immune-suppressants. Well-designed prospective studies that would control for additional mitigating factors such as these are both important and necessary.
Because the cellular composition of biopsy material would be expected to influence gene expression, it is theoretically possible that greater levels of lymphoid tissue present in the younger ASDGI patients in the study group may have contributed to differences between them and the older patients in the other three groups. Although this may have been a factor in the terminal ileal gene expression (all of which demonstrated the histologic presence of LNH), this cannot be the case in the colonic analyses as only two of the colonic specimens contained histologic LNH. It therefore seems likely that LNH, by itself, does not contribute to DET variation in this study.
An important consideration when examining between-group differences in gene expression reported here is the disparity in mean age between the ASDGI group (5 years) and the non-ASD groups (12.5 years). Although this factor represents a limitation to this study, we are not aware of any gene expression studies that have reported a significant age-related pattern of expression in pediatric gastrointestinal tissue. In the absence of these data, we are making the assumption that, while there may be some differences in gastrointestinal tissue gene expression in children as they age from 5 to 13, these differences would not significantly skew the overall transcriptome profile. A second important limitation in the design of this study is the lack of gender-matched samples. Given the increased risk of ASD in boys, the ASD cases tend to be mostly male, whereas the non-ASD patients are more evenly distributed between male and female. Once again, although we are not aware of any report suggesting significant gender-related differences in gene expression in pediatric gastrointestinal tissue, we acknowledge that they may exist and if so, could confound the interpretation of these results. Importantly, the largest source of variation found via principal component analysis of gene expression in these 53 biopsy samples was disease state, i.e. the presence or absence of light microscopic histopathologic findings.