Using a novel mass spectrometry method, we report the CS/DS-GAG composition, or sulfation code, of the human intestine. To our knowledge, this analysis has not previously been reported, nor has our finding that relative to healthy controls, the intestinal CS/DS-GAG composition differs significantly in pediatric and young adult patients with active colonic IBD. Specifically, we report that in active IBD, pro-inflammatory isomers are increased and stabilizing matrix isomers are decreased, a pattern that is largely reversed upon achieving CR with IBD therapy. We also show that in patients with active IBD, the magnitude of these changes in intestinal CS/DS-GAG composition varies in proportion to disease severity, based on clinical as well as histochemical criteria. Although causality cannot be inferred from these relationships, our findings are consistent with a model whereby IBD pathogenesis involves pathological changes in ECM composition, and future studies are warranted to test this hypothesis.
The pattern created by repeating CS/DS isomers bound to core proteins such as aggrecan in the ECM creates a sulfation code that influences many cellular processes and can potentially contribute the pathogenesis of specific diseases.45–47 A key goal of the current work was to report the sulfation code of the intestines for the first time. We found that in young individuals with endoscopically and histologically normal intestinal tissue, CS-A is by far the most abundant isomer, and this predominance increases as one moves from the small intestine to in the colon. Conversely, the abundance of CS-C, the second most abundant isomer, is higher in proximal small bowel than terminal ileum or colon. Stated differently, the ratio of CS-C to CS-A increases as one moves distally from the proximal small bowel to the colon. Although the functional significance of this pattern requires additional study, it may reflect a permissive role for CS-C to facilitate epithelial cell turnover, a process that is more rapid in small intestine than colon.48,49 Furthermore, the flatter mucosal surface of the colonic epithelium, where CS-A content is highest, may benefit from a less plastic and more stable network of CS/DS isomers compared to the villous structure of the small intestine. Additional studies are warranted to investigate these possibilities, as well as the potential impact on CS/DS isomer profiles of gut microbiota, including bacteria that express chondroitinase enzymes that digest CS/DS-GAGs,50 which are likely more abundant in colon than the small intestine.
Next, we sought to determine how the intestinal CS/DS-GAG profile is impacted in patients with clinical IBD symptoms and active IBD mucosal inflammation. This work was motivated in part by evidence of a role for aberrant CS/DS-GAG composition in the pathogenesis and progression of IBD. Specifically, the overall abundance of CS/DS-GAGs has been shown to be reduced in intestinal samples from patients with UC and CD compared to controls, and the areas of bowel with the highest GAG disruption have been shown to correlate with the highest concentration of macrophages that are positive for tumor necrosis factor alpha (TNFɑ), a key mediator of the abnormal immune response implicated in the pathogenesis of IBD.51 Pathological changes in the composition of specific CS/DS isomers is also reported in other disease states, including Alzheimer’s dementia, spinal cord injury, and cardiac reperfusion injury.28,52–54 We therefore sought to determine whether the CS/DS-GAG sulfation signature is altered in association with mucosal IBD lesions. Relative to healthy control tissue, we found marked differences in the CS/DS-GAG profile of mucosal biopsies taken from individuals with both active IBD symptoms at diagnosis and evidence of endoscopic and histologic inflammation. Specifically, the abundance of pro-inflammatory isomers CS-C and CS-E were increased in IBD mucosal lesions, while that of pro-stabilizing isomers CS-A and DS was decreased.
One potential confounding factor in comparing the CS/DS isomer profile in control patients to individuals with IBD is that our control population tended to be somewhat older. To address this concern, we investigated whether CS/DS isomer composition varied by age in normal subjects. We report that in healthy patients ranging in age from 12 to 67, intestinal CS/DS isomer composition did not vary significantly with age. It is therefore unlikely that the differences in CS/DS isomer profiles observed between our control patients and patients with IBD were confounded by an effect of age. This is distinct from our current knowledge about CS/DS-GAG sulfation patterns in the brain, which are shown to change throughout the lifespan as the brain matrix “stiffens” with age and becomes less plastic,32 implying that the intestinal matrix maturation process is complete by early adolescence. Further studies are needed to investigate the intestinal CS/DS-GAG profiles in younger individuals to determine the “critical period” for ECM maturation in the intestines.
We next sought to determine whether these CS/DS-GAG sulfation code changes represent an underlying predisposition to IBD or instead are features of active disease. To this end, we compared the sulfation signature among individuals with IBD in CR at restaging endoscopic evaluation while on effective IBD therapy to that of patients with active IBD at diagnosis. This analysis revealed that in patients in CR, the ratio of pro- and anti-inflammatory isomers was restored to a value similar to that of healthy controls. Therefore, the observed changes in the isomer profiles appear to develop only in the setting of active IBD pathology and do not represent an underlying predisposition to developing disease. Furthermore, compared to patients in CR, patients in true ER who achieve deep mucosal healing show no significant CS/DS isomer profile changes, which suggests that the CS/DS isomer composition is not merely a byproduct of active inflammation but is related to IBD symptomatology and is altered by effective IBD treatments.
Notably, although the CS/DS isomer composition in individuals with IBD in CR at follow up is similar to that of controls, there is persistent hypersulfation present even when CR is achieved, mostly driven by a less dramatic reduction in the pro-inflammatory disulfated CS-E isomer (4S6S). The degree of sulfation therefore may serve as a disease biomarker. Further studies are warranted to follow individuals longitudinally and determine whether the degree of sulfation or the abundance of CS-E is associated with a higher likelihood of disease relapse. Our results also point to a threshold abundance of CS-A above 77% as predictive of CR. CS-A abundance, therefore, may serve as a biomarker for response to therapy and sustained CR, although additional studies are indicated to probe this association further. There is clear potential for the CS/DS-GAG sulfation signature to serve as a disease biomarker in IBD, which can guide treatment decisions and timing of repeat endoscopic evaluation.
A key question raised by these findings is whether observed changes in CS/DS isomer composition contribute to IBD pathology and are therefore also a potential therapeutic target, or instead are only a byproduct of active IBD inflammation. While additional work is needed to establish causality, it is noteworthy that the abundance of pro-stabilizing CS/DS isomers (CS-A and DS) correlated inversely with both pathologic severity and clinical IBD scores, while the abundance of pro-inflammatory isomers (CS-C and CS-E) correlated positively with these markers of IBD activity. These observations are suggestive of an intricate link between altered intestinal ECM structure and function (and CS/DS isomers, in particular) and the underlying disease process. Preclinical data from rodent studies provides further support that CS as part of the ECM plays an active role in IBD. Specifically, in rats, oral CS administration (a combination of CS isomers, predominantly CS-A), improves clinical and histologic outcomes in a dextran sodium sulfate colitis (DSS) model of IBD.55 Notably, this effect was more robust and effective compared to animals receiving 5-aminosalysilic acid administration, a current first line therapy for mild to moderate UC.55 Additionally, DSS colitis outcomes in mice were found to be improved in response to an siRNA-based strategy that altered sulfotransferase enzymatic activity, thereby reducing the abundance of pro-inflammatory CS-E in the colon.36 Similar outcomes are reported in cardiac tissue, where targeted administration of the stable DS isomer following a cardiac ischemic event was effective in preventing reperfusion inflammation injury.54
In summary, our mass spectrometry analysis decodes the CS/DS-GAG sulfation signature of the human intestine for the first time and reveals pathogenic changes of CS/DS-GAG composition in areas of inflamed bowel in IBD, with pro-inflammatory isomers displacing pro-stabilizing isomers. As these ECM changes are both strongly associated with disease activity and reversed by effective IBD treatment, a priority for future studies is to determine 1) the extent to which reversal of pathogenic intestinal ECM compositional changes can improve IBD outcomes without immunosuppression, and 2) whether the intestinal CS/DS-GAG sulfation signature can be monitored as a biomarker of IBD activity.