The present study shows that children with untreated CD have lower expression of genes involved in IGN in duodenal biopsies compared with children with a normal intestinal mucosa. Decreased expression is correlated to a higher Marsh score and, to a lesser degree, tTG antibody levels. These results suggest that CD patients could have an impaired function of IGN, either as a consequence of chronic intestinal inflammation in untreated disease or due to an impaired metabolic pathway shared with other chronic metabolic diseases suggested by the large number of traits shown to be associated with the GLS/STAT1/STAT4 gene region. In the long-term perspective, a lower expression of these genes, which are central for IGN and thus important for metabolic homeostasis, may lead to other chronic diseases like NAFLD and diabetes if not recognized.

This study provides the first evidence of the downregulation of intestinal expression of G6PC, GPT1, SLC6A19, and PPARGC1A in CD. Glucose-6-phosphatase, the protein product of G6PC, has an essential role in gluconeogenesis by converting glucose-6-phosphate, which cannot be transported out of the cell, into glucose, which can then be released into the bloodstream, making its downregulation an especially important finding. Decreased expression of PCK1, FBP1, and solute carriers has previously been shown in adult CD patients but until now has not been put in the context of IGN [19,20,21]. Decreased expression of GLS is consistent with our previous finding of its downregulation in the GENEX material [1]. G6PC3 and GOT1 showed no significant changes in the expression; however, G6PC and GPT1, which catalyze the same reactions in IGN, are more critical for these respective functions [22, 23].

These data thus imply that the ability of the small intestine to perform gluconeogenesis and release glucose from the intestinal enterocyte might be severely decreased in patients with untreated CD. Decreased expression of GLS, GPT1, PCK1, FBP1, and G6PC suggest an impairment of the IGN pathway from the start of using glutamine in gluconeogenesis to releasing glucose into the blood. If only GLS, GPT1, and PCK1 were downregulated, using glycerol, the second most important substrate of IGN [5] could still be possible since it enters gluconeogenesis in later steps, but since FBP1 and G6PC, the protein products of which catalyze crucial final steps in gluconeogenesis, are also downregulated, it stands to reason that the whole pathway of IGN is impaired. We propose that the decreased expression of PPARGC1A might provide an explanation for this at a regulatory level since its protein has a key role in regulating hepatic gluconeogenesis. Even though it is unknown if it has a similar role in the intestine, it does not seem entirely implausible. Decreased expression of SLC6A19 suggests the ability to absorb glutamine, as well as other neutral amino acids transported by the SLC6A19 protein (also known as B0AT1), from the intestinal lumen is impaired. Lower expression of SLC5A1 (SGLT1) and SLC2A2 (GLUT2) indicate decreased capacity for glucose transport.

The metabolic effects of impaired IGN in humans are not entirely clear. Studies on IGN in animal models show that an increase in glucose levels in the portal vein provides signals that increase satiety and improve energy homeostasis. Induced high levels of IGN appear to offer protection against metabolic disease, while impairment leads to signs of dysregulated glucose control and hepatic steatosis [7, 10].

CD patients are at increased risk of NAFLD, with the highest risk seen during the first years after diagnosis and the largest relative risk increase seen in patients with a normal BMI [24, 25]. We speculate that impaired IGN could provide an explanation for the increased risk of NAFLD in CD patients. Our study does not examine whether expression of IGN-related genes return to normal in CD patients treated with a gluten-free diet, but the correlation with Marsh scores suggests a lower degree of inflammation might improve IGN. If IGN is normalized when CD is treated, perhaps this could be part of the explanation for why the risk of NAFLD is at its highest in the first year after CD diagnosis, when the gluconeogenetic capability of the intestine perhaps has not fully recovered. Such a recovery might also be suggested by a 1968 study of glutaminase enzymatic activity, which found lower levels in untreated CD patients that seemed to recover in patients on treatment with a gluten-free diet [26]. Studying the expression of IGN-related genes in patients before and after treatment would be an important next step.

This study has several limitations. The expression of the selected genes in CD cases were compared with disease controls, i.e., these were children referred for an upper endoscopy investigated for other intestinal diseases affecting the gut. It cannot be excluded that the disease controls may have had conditions that can affect the expression of the selected genes. However, all disease controls had normal mucosal findings, and children with inflammatory bowel disease and Helicobacter pylori infections were excluded prior to analysis. Another limitation was that cases and controls were not age- and sex-matched. Still, when adjusting for age and sex, the results remained significant. The strength of the study is that children were enrolled from four sites by pediatric gastroenterologists with long clinical experience in diagnosing and treating children with CD. Enrollment of study participants occurred in 2012 or earlier, e.g., when the intestinal biopsy was the golden standard for diagnosis of CD, meaning that also children with very high levels of IgA-tTG were included in the cohort. Moreover, we previously had all intestinal biopsies reviewed and scored histologically by a single pathologist blinded to the clinical and serological data before the analysis to reduce observation bias and potential risk of discordant classification of cases and controls between the sites.

The results from the present study raise several questions. It is not clear whether the downregulation of the target genes is specific for CD or related to intestinal inflammation in general. The association between the gene region containing GLS and other autoimmune traits, many of which show an increased risk of metabolic disease, could suggest that the downregulation of IGN could also be present in other inflammatory diseases. Thus, further studies of IGN in other diseases are warranted. Also, the study does not answer if impaired IGN is involved in the risk of developing the disease or if it is a response to other disease-initiating mechanisms in CD. In addition, we have not explored if changes in the microbiota could possibly have an effect on the gene expression or if the IGN expression is affected by body mass index and glucose levels. Moreover, while we see a significant correlation between decreased IGN gene expression and the degree of damage in the mucosa, the study does not answer if treatment with a gluten-free diet leading to the healing of the intestinal mucosa restores the expression of genes involved in IGN. Furthermore, it could be of interest to explore whether these genes are downregulated in other intestinal mucosal diseases such as immune deficiency disorders or autoimmune enteropathy.