Article Text

Gut epithelium: food processor for the mucosal immune system?
  1. P W BLAND
  1. Division of Molecular and Cellular Biology,
  2. Department of Clinical Veterinary Science,
  3. University of Bristol, Langford House,
  4. Langford, Bristol BS40 5DU, UK

Statistics from

See article on page 538

To those outside the field of mucosal immunology it must seem incredible that we do not yet understand something so basic as the interaction of protein antigens with the intestinal epithelium. Although we now know in some detail the mechanisms by which the mucosal immune system is regulated and we can construct elegant adjuvant–peptide fusion proteins with systems capable of precise delivery in conjunction with what we think are appropriate cytokines, we know very little of what happens to even the most simple protein antigens during transport. Conversely, the emergence of several good rodent models of intestinal inflammation has had a major impact on our understanding of inflammatory bowel disease pathogenesis, but without further information on T cell–epithelial cell–antigen interactions, the vital mechanisms of the initial “barrier” defect will continue to elude us. Are antigens selectively degraded during transport across the epithelium? If so, is this essential to the maintenance of immune homeostasis? How is the relative transport of intact protein, peptides and amino acids controlled and is the maintenance of a “normal” flora important in this control? All these are vital questions to which we have few sensible responses.

Macromolecular transport by the adult gut epithelium was established in the 1970s, in particular by the extensive work from Allan Walker’s laboratory.1 Some answers (but even more questions!) have been provided in the past decade by workers analysing antigen processing and presentation by isolated gut epithelial cells.2 These studies, which were stimulated by the discovery of MHC antigen expression by absorptive enterocytes, give some indication that antigens are modified by enterocytes, but prolonged analysis of isolated enterocytes in primary culture is impossible. Studies in vivo, with induction of mucosal tolerance as a readout, have shown that tolerance after feeding is dependent on gut processing of fed antigen at some level and indeed that gut processed antigen which has been absorbed into the serum has distinct characteristics and can transfer tolerance,3 4 but the role of the epithelium in this processing has not been investigated further. The reasons are complex, but largely boil down to a lack of appropriate in vitro systems for analysing such a complex tissue. The development of polarised intestinal epithelial cell lines has supplied potential models, but the questions asked must be appropriate to the model and the predictions made must be cautious.

In this issue (see page 538), Terpend et al have used HT-29 epithelial cell monolayers in Ussing chambers to determine the transport of horseradish peroxidase (HRP) and its products across a model intestinal epithelium. They show that under control conditions, HRP is transported from apical to basolateral surfaces as intact molecule, peptides and amino acids. The peptide products have been characterised by molecular weight using steric exclusion HPLC as a peak containing a heterogenous group of degradation products and a separate peak comprising peptides of around eight amino acids. This latter group represents peptides capable of binding to MHC molecules for presentation to T cells.

Based on these results, we can speculate that if HT-29 in vitro mimics the transport characteristics of small bowel epithelium in vivo, then the epithelium is capable of taking an active part in processing dietary proteins to immunologically relevant peptides. Other work from this same group suggests that the balance which the epithelium maintains between absorbed native protein and absorbed degradation products may be important in preventing food allergy.5However, the “if” is considerable in a tissue in which the maintenance of cellular differentiation, polarity, microvillus development, ion channel integrity, tight junction complexity, basement membrane interaction, and cytokine profile (to name but a few known parameters) are essential to stable function. The authors have gone to considerable lengths to control for the limitations of the system and to analyse peptides, monolayer integrity and proteolytic enzymes using appropriate methodology— unfortunately, not universal practice among the increasing numbers of workers using epithelial cell lines, often inappropriately, for analysis of gut epithelial function. We can only make the leap from such relatively simple systems to confident statements on the enormously complex in vivo situation if such appropriate methods and assumptions are used.

The results also show, not surprisingly, that basolateral interferon γ (IFN-γ) causes the monolayer to become leaky and, interestingly, that this induced leakiness does not affect the relative proportions of transported intact HRP (presumed paracellularly) and degraded HRP (presumed transcellularly). Recent work by other groups has shown that transcellular and paracellular HRP transport and the relation of native to degraded antigen are differentially modulated in rat intestine after mast cell degranulation,6 or in biopsy tissue from individuals with atopic eczema.7 Clearly, the effects of different types of local immune activation on the degradation of antigen by enterocytes and on tight junction effectiveness are complex.

The increased transport of HRP with IFN-γ treatment in Terpendet al’s study was not reflected by changes in the intracellular proteases analysed. More detailed analysis of these enzymes and changes in their expression due to proinflammatory cytokines is both suggested by these results and essential to understand the epithelial degradative capacity. The lack of effect of IFN-γ on cathepsin expression in HT-29 confirms the recent report of Hershberg et al,8 but the Hershberg group showed different effects with different cell lines—cathepsin B was increased in T84 cells after IFN-γ treatment, but not cathepsins L or H or aspartic proteases. Again, this suggests that we must use caution in the wider interpretation of studies using these model systems.

HLA-DR was only present after IFN-γ induction in the HT-29 cells. This might indicate that as the proportions of native to degraded protein are similar in control and IFN-γ treated cells, MHC class II proteins are not important in providing a protected transport route for certain peptides.9 However, the role of MHC class II in enterocytes— it is expressed constitutively in normal human and rodent intestinal epithelium—is a complex issue and will only be resolved by analysis of dietary peptides bound to epithelial MHC class II proteins, as the authors suggest.

This work adds significantly to our understanding of the transport and processing capacities of the intestinal epithelium. Together with previous work with ex vivo epithelial cells and in vivo analyses, it allows us to predict that the gut epithelium is an active participant in the generation and regulation of immune responses to enteric antigens and that these responses can, in turn, feed back and modulate epithelial cell function.

See article on page 538


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles