Is the Th1/Th2 paradigm of immune regulation applicable to IBD?

Some of the most important initial advances in our understanding of inflammatory bowel disease (IBD) (Crohn's disease and ulcerative colitis, CD and UC, respectively) have come from the study of the many murine models of mucosal inflammation resembling these diseases.1 Perhaps the key finding to emerge from these earlier studies is that in most, if not all, cases the mucosal inflammation can be shown to be due to either an excessive Th1 T cell response or an excessive Th2 T cell response, the former being characterized by increased IL-12, IFN-γ, and TNF-α production and the latter by increased IL-4 and/or IL-13 production. This was demonstrated in the parallel murine models of colitis induced by the administration of haptenating agents such as trinitrobenzene sulphonic acid (TNBS) or oxazolone.2,3 Thus, in TNBS-colitis one sees a dense transmural inflammation as seen in CD associated with a Th1 response dominated by IL-12 and other Th1 cytokine production (IFN-γ and TNF-α). In addition, prevention of inflammation or a reversal of established inflammation can be achieved by administration of anti-IL-12 (p40 chain) antibody or other Th1-inhibiting factors. Conversely, in oxazolone colitis a relatively superficial inflammation is observed marked by epithelial cell disruption associated with Th2 response, in this case initially dominated by IL-4 production and subsequently superceded by an IL-13 cytokine response.4 The histopathological features of TNBS colitis and oxazolone colitis bear a resemblance to CD and UC, as do other models of Th1 and Th2 immune responses, respectively. It is thus of great interest to determine if the 2 forms of human IBD also conform to the Th1/Th2 paradigm. This question is not a trivial one, since many students of IBD are still uncertain whether CD is a Th1 T cell-mediated inflammation or UC is a Th2 T cell-mediated inflammation. Thus, it is well worth analyzing the available evidence to demonstrate whether human IBD can be organized immunologically in the same way as in the murine models. Over the past 10–15 years a great deal of evidence has accumulated that supports the view that the inflammation in CD is associated with and, indeed, is due to a Th1 response. This evidence begins with observations in the early 1990s showing that the T cells isolated from lesional tissues were producing high amounts of IFN-γ, and below-normal amounts of IL-4, the canonical Th1 and Th2 cytokines, respectively.5 This initial evidence of a Th1 response was then augmented by studies showing that antigen-presenting cells (APCs) isolated from CD lamina propria produce greatly increased amounts of IL-12,6 a cytokine that can direct the differentiation of Th1-producing effector T cells (IFN-γ and TNF-α). The above evidence of increased IL-12 production in CD is reaffirmed by reports showing that CD is associated with increased IL-12 signaling. In particular, it has been demonstrated that nuclear extracts of T cells from CD lesions contain Th1-related transcription factors that are detected by electrophoretic mobility shift assays (EMSAs). T cells from CD patients, but not UC patients, contain activated (intranuclear) STAT4 and T-bet, the former a factor that is activated directly by IL-12 and the latter a factor activated indirectly by IL-12 via the production of IFN-γ and stimulation of cells via STAT1.7 A final piece of evidence that IL-12 signaling is a feature of CD but not UC arises from the fact that cells isolated from CD tissues bear the IL-12Rβ2 chain, the signaling chain of the IL-12 receptor that is upregulated during exposure to IL-12.8 The above data supporting the presence of a Th1 response in CD (but not in UC) provide evidence that this form of IBD is associated with a Th1 response; nevertheless, the data fall short of proving that CD is caused by this dysregulated Th1 response. So far, in this discussion of the immunological basis of CD we have been focusing on IL-12, a cytokine long thought to be the only cytokine capable of initiating a fully fledged Th1 response. In recent years, however, it has become evident that a second T cell-inducing cytokine, termed IL-23, may also be an important player in the Th1 response. IL-23 is similar to IL-12 in that the two share an identical chain, the p40 chain. However, whereas in IL-12 p40 is linked to a second chain, known as p35, in IL-23 p40 is linked to another second chain, known as p19. Evidence for a role of IL-23 in the pathogenesis of autoimmune-mediated animal models is shown in the publications demonstrating that in Th1 animal models of inflammation, experimental autoimmune encephalomyelitis and the SCID T-cell transfer colitis model wherein blockade of IL-23 response prevents development of disease.9 Recent studies have shown that these cytokines IL-12 and IL-23 although both produced by APCs and/or dendritic cells formulate the differentiation of rather distinct T cell subsets (reviewed in Ref.10). Thus, IL-12 induces the aforementioned classical IFN-γ producing Th1 T cells, whereas IL-23 is involved in the possible maintenance and/or expansion of another polarized T-cell population, namely Th17. This latter T-cell population is characterized by the secretion of IL-17, IL-6, and TNF-α. These cytokines are intimately involved in innate host defense but also can play a primary role in the occurrence of tissue inflammation. Thus, IL-17 may act on cell populations to induce secretion of inflammatory chemokines that have a prompt role in rapid neutrophil recruitment. TNF-α can cause direct tissue injury, while IL-6 can enhance resistance to T cell-activated cell death (apoptosis) and survival of such inflammatory effector cells. If we turn our attention back to human studies, an interesting and important feature of the human anti-IL-12 p40 trial of CD patients (as discussed earlier) was not only that it had a therapeutic effect but that the effect of therapy on patient immunologic responses was due to downmodulation of both IL-12 and IL-23. In these studies the main findings were that successful treatment correlated with a dramatic decrease in both IL-12 and IL-23 production by macrophages as well as in a decrease in IFN-γ, IL-17, IL-6, and TNF-α production by T cells.11,12 Overall, these data established that the clinical response of patients with CD to anti-IL-12 p40 therapy is grounded in the ability of this agent to subdue the immunologic mechanisms in play causing disease potentiation, namely, the IL-12 and IL-23-driven T-cell response. Thus, it is fair to say that the results of this trial serves as an indicator for the concept that CD may not be a classical Th1 T-cell response but both a Th1-driven inflammation and a Th17-driven inflammation and that effective therapies may need to address both inflammatory arms for long-term resolution of disease. Further human studies to ascertain which may be the overriding force driving in the inflammatory process are forthcoming. To a greater extent than CD, the immunopathogenesis of UC has been a mystery owing to the fact that it did not fit neatly into a niche of the Th1/Th2 paradigm as either a Th1 or Th2 disease. In particular, neither IFN-γ, a major Th1 cytokine, nor IL-4, the major Th2 cytokines was found in excess.5 In fact, IL-4 production was found to be decreased in cells extracted from UC tissue and only the fact that an additional Th2 cytokine IL-5 secretion by these cells was somewhat increased hinted that the disease may have a Th2 character. A step forward in our understanding of the UC paradigm came from the aforementioned hapten-induced colitis, oxazolone-colitis.3,4 Intrarectal administration of this hapten led to a very unique colitis marked by an edematous bowel wall associated with an exudative inflammatory response and epithelial ulceration. Thus, although the disease was acute and short-lived, it could be said to resemble some characteristics of UC rather than CD. This cytokine response was also unique: initially it was dominated by an IL-4 response and indeed anti-IL-4 antibody administration prevented disease. However, if the disease was allowed to progress so that the inflammation lasted longer, the IL-4 response was observed to decline and to be replaced by another Th2 cytokine, IL-13. The latter was also shown to be important in disease pathogenesis, by administration of anti-cytokine antibody, in this case, anti-IL-13. The next question to be addressed in any disease process was the cellular origin of the cytokine response, particularly IL-13 production and targets. Interestingly, this turned out to be a natural killer T (NKT) cell, i.e., a CD4+ T cell bearing NK markers that take part in innate immune responses through their capacity to recognize foreign lipid antigens presented in the context of an atypical presenting molecule (class I MHC molecular known as CD; in humans, CD1d). A characteristic of NKT cells that is relevant here is that these cells can function as cytotoxic cells and can produce IFN-γ or IL-4/IL-13 under various circumstances. Taken together, these various studies of the murine oxazolone colitis model strongly suggested that the pathogenesis of the colitis was due to the generation of NKT cells that secrete IL-13. From these data it was inferred that oxazolone-colitis could be due to either a direct effect of IL-13 on epithelial cells or a (cytolytic) effect of NKT cells on epithelial cells. The above studies of oxazolone-colitis led one to ask the question whether the same immune parameters pertained to its presumptive counterpart in human disease. Here it was found that lamina propria cells isolated from inflamed UC tissue stimulated in vitro produced increased amounts of IL-13 as compared to cells from both control individuals and patients with CD.13 Moreover, depletion of cells bearing a marker found on NKT cells from cell population extracted from UC tissue, followed by stimulation and culture in vitro, led to greatly decreased IL-13 production. Further studies to marshal the link between NKT cells and IL-13 production in UC came from flow cytometric analyses that revealed that UC mononuclear cells contained increased numbers of cells bearing a marker for NKT cells were producing IL-13. These similarities between oxazolone colitis and UC suggest that NKT cells associated with an increased IL-13 secretion may be one of the key pathogenic mechanisms of inflammation in these disorders. One last, but important, finding in the study of UC patients was that lamina propria cells enriched for NKT cells from the patients could be shown to be cytotoxic for epithelial cells and such cytotoxicity was enhanced by IL-13. On this basis, the possible immunopathologic mechanism operating in UC is that antigen in the mucosal microflora activate NKT cells that, in turn, cause cytolysis of epithelial cells and the characteristic ulcerations associated with the disease. Furthermore, it has been recently shown that IL-13 could have direct cytotoxic effects on epithelial cells that also compromise epithelial function and integrity.14 Thus, these findings buttress the results of the oxazolone-colitis model and point in the direction that UC is indeed mediated by a Th2-type response. The message that emerges from the data provided by both murine and human IBD is that mucosal immune function can resolve itself into dominant activity of specific T cell effector pathways. This conclusion has important theoretical implications for the understanding of the mechanism of IBD, as well as practical implications for its therapy. The fact is that, if the pathogenesis of IBD is ultimately “channeled” into these effector T-cell pathways, we can focus on treatment of these diseases on the cytokines controlling these mechanisms. Although currently not thought of as a pure “Th1” classical response, CD may ultimately depend on the elaboration of the IL-12/IL-23 selective pathways. Conversely, since Th2 responses ultimately depend on IL-4 and IL-13 production, as noted, then these molecules or a molecule immediately downstream or upstream of IL-13 becomes a primary therapeutic target in therapy of UC. We await further human clinical studies to validate these basic principles of the Th1/Th2 paradigm as they pertain to human disease.