Intestinal epithelial barrier function and Crohn’s disease are intimately related. An impaired barrier in association with active inflammation leads to increased exposure of the mucosal innate and acquired immune system to proinflammatory molecules. This has been implicated as a major driving force for mucosal inflammation. In active disease, macromolecules can permeate the barrier at an increased rate via, for example, breaks in the integrity of the epithelium (ulceration, erosions, or nests of apoptotic cells)¹ via increased pinocytotic uptake of luminal proteins,^2,3 and via increased proportion of M cells in the terminal ileum.⁴ Impaired barrier function may precede the clinical development of the disease and might represent a marker of increased susceptibility to Crohn’s disease. This notion derives from observations made in several centres using different techniques that paracellular permeability is abnormal in 10–20% of first degree relatives of patients with Crohn’s disease.^8,9,11 That this reflects exposure to an environmental factor is supported by the observations from three studies that increased intestinal permeability is also found in 13–36% of spouses of patients with Crohn’s disease.^8,9,11

Tumour necrosis factor α (TNF-α) and Crohn’s disease are intimately related. Mucosal levels and inflammatory cell production of TNF-α are elevated. Therapy with monoclonal antibodies to TNF-α, such as infliximab, leads to a rapid reduction in inflammation and healing in many patients, and remission can be maintained with ongoing therapy.^12–14 The spectacular effect of infliximab has been attributed not only to its ability to mop up TNF-α but also to induction of apoptosis in activated T cells.¹⁵

TNF-α and intestinal epithelial barrier function are also intimately related. TNF-α increases paracellular permeability via an effect on the tight junctions.¹⁶ While it is theoretically possible that TNF-α may also increase paracellular permeability by induction of epithelial apoptosis, evidence is scant. TNF-α suppressed the expression and activity of the intestinal p-glycoprotein MDR-1 in an intestinal epithelial cell line,¹⁷ an effect that would impair the efflux of xenobiotics taken up by epithelial cells.¹⁸ It also may play a role in facilitating bacterial translocation across the epithelium, as recently demonstrated in glutamine starved Caco-2 cells.¹⁹ In this issue of Gut, Söderholm and colleagues²⁰ extend the role of TNF-α by demonstration that expression of mRNA for TNF-α in mucosa correlates with endosomal uptake of horseradish peroxidase (HRP) in resected ileal mucosa mounted in Ussing chambers (see page 1813). A causal relationship was suggested by experiments with T84 cell monolayers where relatively low concentrations of TNF-α increased uptake of HRP. In other words, evidence now implicates TNF-α in global impairment of intestinal barrier function, including leakier tight junctions, increased uptake of proteins from the lumen, and less efficient efflux of foreign substances from the cells, all favouring increased permeation of luminal macromolecules to the lamina propria.

All of these associations raise the key issue of whether TNF-α is the link between the leaky gut and Crohn’s disease. Intestinal permeability in patients with active Crohn’s disease normalises following successful anti-TNF (infliximab) therapy.²¹ However, this observation may simply reflect healing of the injured mucosa independently of the mechanism by which it was achieved, and does not resolve this issue. The real answers may be found in understanding the mechanisms responsible for the reduced barrier that occurs across the intact epithelium, such as that observed in first degree relatives and spouses of patients with Crohn’s disease.

It is first necessary to consider the more general issue of what aspects of barrier function clinical measures of intestinal permeability are actually assessing. Permeability probes used all measure predominantly the efficiency of the paracellular route.²² There are no readily applicable clinical measures of transcellular permeability. Yet the luminal proinflammatory molecules and bacteria, which induce mucosal inflammation, are likely to traverse the physically intact epithelium via the transcellular rather than paracellular route. Despite the likelihood that these routes of uptake are independently controlled, relatives of patients with Crohn’s disease who have elevated paracellular permeability exhibit evidence of excessive antigen exposure, suggesting that more than just paracellular pathways are leaky.²³ In other words, it appears that measures of paracellular permeability in this setting are reflecting general barrier dysfunction not just that between cells, and that a common mechanism needs to be invoked. The findings of Söderholm et al that histologically normal mucosa, albeit in patients with severe Crohn’s lesions nearby, exhibited elevated endosomal uptake of the marker protein the marker prtoeinin association with an increase in TNF-α mRNA and presumably epithelial exposure to TNF-α,²⁰ might suggest TNF-α as a candidate. A similar situation of increased intestinal epithelial exposure to TNF-α might also occur in first degree relatives with elevated paracellular permeability as evidence for subclinical inflammation of the bowel in this population has been reported.²⁴ It is feasible that interactions between the lumen and epithelium in association with, for example, microbial dysbiosis or changes in the soluble component of luminal contents, might lead to induction of TNF-α production and secretion by subepithelial lamina propria cells, with a subsequent diffuse increase in exposure of TNF-α to intestinal epithelium, but without clear histological abnormalities.

Another issue that arises is how the epithelial effects of TNF-α fit into the mode of action of infliximab in inducing remission, in healing the mucosal lesions, and in maintaining remission. That TNF-α is a cytokine of major pathogenic significance in Crohn’s disease is incontrovertible, as shown by the powerful and rapid healing response to infliximab and maintenance of remission in many patients with moderately severe disease. It has been generally believed that the remission inducing effect of infliximab is the result of taming activated T cells, possibly via induction of their destruction by apoptosis.¹⁵ In active Crohn’s disease, it is likely that epithelium distant from the major inflammatory lesion is being exposed to increased amounts of TNF-α and that therapy with anti-TNF-α regimens will minimise that exposure and correct potential abnormalities of permeability. Reducing a widespread leaky gut might have a greater effect on well being than we currently appreciate. The mechanisms by which anti-TNF-α regimens maintain remission may also involve epithelial effects rather than or in addition to direct effects on immune and inflammatory mechanisms.

Thus recent studies indicate a potentially more expansive role for TNF-α in intestinal homeostasis, barrier function, and pathogenesis of intestinal inflammation in Crohn’s disease. Caution however must be exercised in the interpretation of many of the recent observations, particularly as they are very dependent on results from experimental studies in cell line models of epithelium or from observations in tissue from patients with advanced disease. It will be of key importance to assign an in vivo relevance to many of these findings before speculation can be converted to real understanding. Nevertheless, it is an appealing unifying concept that TNF-α is a major mechanism underlying global barrier dysfunction in Crohn’s disease, both early and late in its pathogenesis.