Persistence of pathogenic CD4+ Th1-like cells in vivo in the absence of IL-12 but in the presence of autoantigen.

Despite recent successful treatment of murine autoimmune disease with anti-IL-12 mAb, it has not yet been addressed whether anti-IL-12 mAb can also be effective in late stages of disease and whether it can provide lasting protection against recurrence, especially during continued presence of autoantigen. We used a newly developed psoriasis model in scid/scid mice, which allows easy tracking of pathogenic T cells, to show that when anti-IL-12 mAb is given for 2 wk (1 mg/wk) in the late stage of severe disease, inflammation is greatly reduced, as measured by ear thickness and histology (scores, 1.1 +/- 0.1 vs 2.0 +/- 0.4). Moreover, prolonged treatment (4 wk) of chronic psoriatic mice with high doses of mAb (1 mg/wk; prolonged active anti-inflammatory treatment (PAAIT)) results in the almost complete resolution of lesions (scores, 0.3 +/- 0.1 vs 2.7 +/- 0.2). Surprisingly, however, despite these significant treatment results, the psoriasis-like lesions return soon after the anti-IL-12 mAb treatment is discontinued. This rapid relapse of disease may be attributed to large populations of activated CD4(+) T cells present in the lymph nodes of PAAIT animals still expressing an effector/memory phenotype (CD45RB(low), L-selectin(low)). Upon stimulation in vitro such PAAIT lymph node cells secrete high amounts of IFN-gamma (129 ng/ml); when transferred into naive scid/scid animals they are able to rapidly induce disease without costimulation. Our data indicates an alternative IL-12-independent pathway for pathogenic Th-1-like cells in vivo during the chronic phase of disease that allows these cells to persist and maintain their pathogenicity in the draining lymph tissue of the autoimmune site.

T helper type 2 cell differentiation occurs in the presence of interleukin 12 receptor beta2 chain expression and signaling.

The differentiation of CD4(+) T cells into T helper type 1 (Th1) cells is driven by interleukin (IL)-12 through the IL-12 receptor beta2 (IL-12Rbeta2) chain, whereas differentiation into Th2 cells is driven by IL-4, which downregulates IL-12Rbeta2 chain. We reexamined such differentiation using IL-12Rbeta2 chain transgenic mice. We found that CD4(+) T cells from such mice were able to differentiate into Th2 cells when primed with IL-4 or IL-4 plus IL-12. In the latter case, the presence of IL-4 suppressed interferon (IFN)-gamma production 10-100-fold compared with cells cultured in IL-12 alone. Finally, in studies of the ability of IL-12 to convert Th2 cells bearing a competent IL-12R to the Th1 cells, we showed that: (a) T cells bearing the IL-12Rbeta2 chain transgene and primed under Th2 conditions could not be converted to Th1 cells by repeated restimulation under Th1 conditions; and (b) established Th2 clones transfected with the IL-12Rbeta2 chain construct continued to produce IL-4 when cultured with IL-12. These studies show that IL-4-driven Th2 differentiation can occur in the presence of persistent IL-12 signaling and that IL-4 inhibits IFN-gamma production under these circumstances. They also show that established Th2 cells cannot be converted to Th1 cells via IL-12 signaling.

Tissue specificity of E- and P-selectin ligands in Th1-mediated chronic inflammation.

The demonstrated role of E- and P-selectin ligands in the recruitment of Th1 cells raises the question of tissue specificity determination by pathogenic T cells. We took advantage of the fact that chronic Th1-mediated inflammation in the scid/scid CD4+CD45RBhigh T cell transfer model can occur at multiple tissue sites, resembling inflammatory bowel disease in the colon and psoriasis in the skin. We show that the majority of infiltrating effector T cells from psoriatic skin expresses high levels of functional P-selectin ligand (87 +/- 3%), detected by P-selectin-Ig (PIg), while a significantly smaller subset of T cells from colitic lesions expresses this ligand (24 +/- 2%). Similarly, E-selectin ligand is preferentially expressed on CD4+ T cells infiltrating the skin (24 +/- 2%), but only on very few CD4+ T cells infiltrating the colon (CIT; 1.3 +/- 0.8%). In contrast, CD4+ T cells infiltrating the skin express alpha4beta7 at a significantly lower level than CIT (mean fluorescence intensity, 28 vs 61, respectively), although, interestingly, alphaEbeta7 was expressed at high levels on both populations. Analysis of the disease-inducing potential of PIg+ and PIg- CD4+ CIT cells revealed that both populations not only express similar levels of the gut-homing molecule alpha4beta7 (mean fluorescence intensity, 50 vs 56, respectively), but do not differ in their capacity to express IFN-gamma. Furthermore, CIT depleted of cells expressing functional P-selectin ligand were able to induce colitis upon transfer, suggesting that induction of colitis in this model may be independent of E- and P-selectin. These results indicate that adhesion molecule expression and the homing pattern of inflammatory T cells are regulated by the local environment independently of their inflammatory capacity.

Role of IL-12 in intrathymic negative selection.

Cytokines are central regulatory elements in peripheral lymphocyte differentiation, but their role in T cell ontogeny is poorly defined. In the present study, we evaluated the role of IL-12 in thymocyte selection more directly by determining its role in two models of in vivo negative selection. In initial studies we demonstrated that abundant intrathymic IL-12 synthesis occurs during OVA peptide-induced negative selection of thymocytes in neonatal OVA-TCR transgenic mice, and such synthesis is associated with increased IL-12R beta2-chain expression as well as STAT4 intracellular signaling. In further studies, we showed that this form of negative selection was occurring at the alphabetaTCRlowCD4lowCD8low stage and was prevented by the coadministration of anti-IL-12. In addition, the IL-12-dependent thymocyte depletion was occurring through an intrathymic apoptosis mechanism, also prevented by administration of anti-IL-12. Finally, we showed that IL-12 p40-/- mice displayed aberrant negative selection of double positive CD4+CD8+ thymocytes when injected with anti-CD3 mAb. These studies suggest that intact intrathymic IL-12 production is necessary for the negative selection of thymocytes occurring in relation to a high "self" Ag load, possible through its ability to induce the thymocyte maturation and cytokine production necessary for such selection.

IL-12, independently of IFN-gamma, plays a crucial role in the pathogenesis of a murine psoriasis-like skin disorder.

The onset of acute psoriasis and the exacerbation of chronic psoriasis are often associated with a history of bacterial infection. We demonstrate that while only few scid/scid mice develop disease when CD4+CD45Rbhigh T cells are transferred alone, coadministration of LPS plus IL-12 or staphylococcal enterotoxin B into scid/scid mice 1 day after CD4+CD45Rbhigh T cell transfer greatly enhances disease penetrance and severity. Most importantly, the skin lesions induced by this method exhibit many of the histologic hallmarks observed in human psoriasis. Skin infiltrating CD4+ T cells were predominantly memory/effector cells (CD45Rblow) and exhibited a highly polarized Th1 phenotype. To test whether the development of pathogenic T cells was dependent on their production of IFN-gamma, we transferred IFN-gamma-/- CD4+CD45Rbhigh T cells into scid/scid or into T, B and NK cell-deficient scid/beige mice. Surprisingly, the incidence of psoriasis was similar to scid/scid animals that received IFN-gamma+/+ T cells, although acanthosis of the skin was attenuated. In contrast, the development of psoriasis was abolished if anti-IL-12 mAb was administered on day 7 and 35 after T cell transfer. Skin-derived IFN-gamma-/- inflammatory cells, but not cells from anti-IL-12-treated animals, secreted substantial amounts of TNF-alpha, suggesting that the inflammatory effect of IFN-gamma-/- T cells may be partly exerted by TNF-alpha and that the therapeutic effect of anti-IL-12 may depend on its ability to down-regulate both TNF-alpha and IFN-gamma. Overall, these results suggest that IL-12, independently of IFN-gamma, is able to induce pathogenic, inflammatory T cells that are able to induce psoriasiform lesions in mice.

Administration of mAb against alpha E beta 7 prevents and ameliorates immunization-induced colitis in IL-2-/- mice.

We previously demonstrated that 2,4,6-trinitrophenol (TNP)-OVA immunization leads to a transmural colitis in the IL-2-/- mouse that is caused by IL-12-driven CD4+ Th1 T cells and resembles human Crohn's disease. The integrin alpha E beta 7 is highly expressed on colonic intraepithelial lymphocytes and has been suggested to function as a homing or retention molecule for intraepithelial lymphocytes. To evaluate the role of alpha E beta 7 in colitis, we administered a mAb against alpha E beta 7 to IL-2-/- mice that were immunized at the same time with TNP-OVA in CFA. To our surprise, this treatment resulted in a significantly reduced colitis severity score, 0-2 vs 3-4, that was associated with a significant reduction in CD4+ lamina propria lymphocyte subpopulation (p < 0.01). In contrast, the total number of splenic CD4+ T cells of treated animals was significantly elevated compared with that of untreated animals (3.2 +/- 0.6 x 107 vs 1.2 +/- 0.2 x 107; p < 0.05). Similarly, functional studies revealed that IFN-gamma production by lamina propria lymphocytes isolated from IL-2-/- TNP-OVA-immunized mice treated with anti-alpha E beta 7 was significantly lower than in untreated IL-2-/- TNP-OVA-immunized mice. In contrast, IFN-gamma production by splenic cells isolated from treated IL-2-/- TNP-OVA-immunized mice was significantly higher than in untreated mice. Finally, TNP-OVA-immunized IL-2-/- mice that were treated after the colitis had been established also showed a significant decrease in mucosal inflammation after alpha E beta 7 mAb administration. Thus, the above findings demonstrate that the onset and maintenance of inflammatory bowel disease depends on the colonic localization of lamina propria CD4+ lymphocytes expressing alpha E beta 7.

Active Wegener's granulomatosis is associated with HLA-DR+ CD4+ T cells exhibiting an unbalanced Th1-type T cell cytokine pattern: reversal with IL-10.

Wegener's granulomatosis (WG) is a granulomatous vasculitis that affects the upper respiratory tract, lung, and kidney. Since T cells make up a significant proportion of cells infiltrating granulomatous lesions in WG, we investigated the proliferative response and cytokine profile of T cells from these patients. PBMCs were isolated from 12 patients with active WG, 7 patients with inactive disease, and 12 healthy normal donors. PBMCs from clinically active WG patients exhibited increased proliferation following stimulation with either PMA/ionomycin or anti-CD2 and anti-CD28, when compared with normal donors. In addition, these PBMCs exhibited increased secretion of IFN-gamma, but not of IL-4, IL-5, or IL-10. Furthermore, TNF-alpha production from PBMCs and CD4+ T cells isolated from patients with WG was elevated, when compared with healthy donors. In further studies, we investigated the ability of WG patients' monocytes to produce IL-12 and showed that both inactive and active patients produced increased amounts of IL-12. Finally, the in vitro IFN-gamma production by WG PBMC is inhibited in a dose-dependent manner by exogenous IL-10. These data suggest that T cells from WG patients overproduce IFN-gamma and TNF-alpha, probably due to dysregulated IL-12 secretion, and that IL-10 may therefore have therapeutic implications for this disease.

TGF-beta production regulates the development of the 2,4,6-trinitrophenol-conjugated keyhole limpet hemocyanin-induced colonic inflammation in IL-2-deficient mice.

A severe, Th1-mediated experimental colitis with similarities to inflammatory bowel disease in humans can be induced by a single injection of 2,4,6-trinitrophenol (TNP)-substituted protein plus adjuvant in IL-2-/- mice. To determine the early events involved in the pathogenesis of IL-2-/-colitis, we compared the function of lamina propria (LP) T cells from IL-2-/- and IL-2+/+ mice subjected to disease-inducing (TNP-conjugated keyhole limpet hemocyanin [TNP-KLH]) and disease-inhibiting (anti-CD3) immunization protocols. We show that LP T cells in TNP-KLH-immunized IL-2-/- mice fail to produce TGF-beta early (day 2), whereas LP T cells in TNP-KLH-immunized IL-2+/+ mice exhibit an approximately eightfold rise in TGF-beta secretion. The critical importance of local TGF-beta production was further substantiated by the following findings. 1) LP T cells from TNP-KLH-immunized IL-2-/- mice administered anti-CD3 (i.p.) exhibit a significant rise in TGF-beta, production but fail to produce IFN-gamma, and such mice do not develop colitis. 2) TNP-KLH-immunized IL-2-/- mice administered anti-CD3 and coadministered anti-TGF-beta mAb again give rise to IFN-gamma-producing LP cells, and such mice develop colitis. 3) TNP-KLH-immunized IL-2+/+ mice administered anti-TGF-beta mAb exhibit pockets of mononuclear cell infiltrates in the LP. These results indicate that the disposition of IL-2-/- mice to develop chronic colonic inflammation is due to a Th1 cell response in the LP that is not appropriately counter-regulated by the production of the suppressor cytokine, TGF-beta.

Dysregulated intrathymic development in the IL-2-deficient mouse leads to colitis-inducing thymocytes.

Gene-targeted mice lacking the IL-2 gene (IL-2 -/- mice) develop various forms of autoimmunity as well as severe colitis, either spontaneously in a conventional environment or after immunization with 2,4,6-trinitrophenol (TNP)-conjugated keyhole limpet hemocyanin (KLH) in a specific pathogen-free environment. We show here that the induction of colitis with TNP-KLH induces a change in the thymocyte population characterized by decreased numbers of double positive (DP; CD4+CD8+) thymocytes (IL-2 +/+, 45.2 x 10(6) vs IL-2 -/-, 23.6 x 10(6)) and increased numbers of single positive (SP; CD4+CD8- or CD4-CD8+) thymocytes (IL-2 +/+, 5.3 x 10(6) vs IL-2 -/-, 20.9 x 10(6)). The latter also bear activation markers. In addition, thymocytes from TNP-KLH-immunized IL-2 -/- mice produce more IFN-gamma and less IL-4 than similarly immunized IL-2 +/+ mice. These defects in thymocyte maturation and lymphokine production are IL-12 driven, since they are prevented when immunized IL-2 -/- mice are coadministered with anti-lL-12. Furthermore, we demonstrate that IL-2 -/- mice exhibit decreased cortical apoptosis as determined by thymocyte numbers and detection of apoptotic cells in situ. Finally, we show that colitis-inducing thymocytes are generated in the immunized IL-2 -/- thymus, since IL-2 +/+ mice develop colitis following injection of small numbers of single positive thymocytes from immunized IL-2 -/- mice but not from IL-2 +/+ mice. Taken together, these data indicate that, in the absence of IL-2, thymocyte maturation is abnormally directed by IL-12 toward the generation of mature, activated Th1-type thymocytes that are capable of mediating colitis.

Induction and prevention of colonic inflammation in IL-2-deficient mice.

Gene-targeted mice deficient for IL-2 (IL-2 -/- mice) are free of apparent disease when maintained under germfree conditions but develop colitis and autoimmunity in a conventional environment. Here we show that colitis can be reproducibly induced in IL-2 -/- mice, but not in IL-2 +/+ mice, by i.p. immunization with Ag in CFA; thus enabling the systematic study of the immunopathogenesis of the colitis. We found that TNP-KLH or TNP-OVA had the most significant effect in inducing colitis, and while TNP-KLH immunization leads to the early appearance of activated T cells in the colons of both IL-2 -/- and IL-2 +/+ mice, only lamina propria cells of IL-2 -/- mice produced high amounts of INF-gamma. Moreover, both infiltrating colon CD4+ (69%) and CD8+ (6%) T cells secrete large amounts of IFN-gamma; however, only the depletion of CD4+ T cells leads to abrogation of the inflammation. In further analysis, we showed that the high IFN-gamma production is IL-12 driven, since colonic tissues of IL-2 -/- mice but not IL-2 +/+ mice show the presence of heterodimeric IL-12 and co-administration of anti-IL-12 with TNP-KLH completely prevented colitis and significantly reduced IFN-gamma production. Finally, we demonstrate that IL-2 -/- mice are deficient in their ability to induce Th2 responses after TNP-KLH challenge and that such immunization also leads to autoimmune-like phenomena in other organs of IL-2 -/- mice. These findings suggest that in the absence of IL-2 systemic administration of Ag induces primarily Th1 cells driven by overexpression of heterodimeric IL-12.

Differential activation requirements of isotype-switched B cells.

In the present studies, we compared the activation requirements of sIgM+/sIgD+ B cells with those of isotype-switched sIgM-/sIgA+ B cells. We found that whereas sIgM+ B cells respond to T cell-independent (TI) and T cell-dependent (TD) Ag with no significant bias toward one stimulus, sIgA+ B cells were deficient in their ability to respond to antigen receptor cross-linking but responded remarkably well to TD stimuli. Thus, dextran-conjugated anti-IgA antibody (anti-IgA-dextran), anti-kappa-dextran, or various immobilized anti-IgA antibodies (Ab) induced only low-level IgA B cell proliferation and no IgA secretion in the presence of various lymphokines; in marked contrast, sIgA+ B cells responded to cognate and noncognate T cell stimulation as well as to stimulation by CD40 ligand-bearing fibroblasts by secreting large amounts of IgA (up to 240 000 ng/ml per 10(5) cells). This pattern of sIgA+ B cell responsiveness was noted with both germinal center peanut agglutininhi (PNAhi) and non-germinal center PNAlo B cells. In confirmation of these results, whole Peyer's patch or lamina propria cell populations containing less than 15% sIgA+ B cells stimulated with a noncognate T cell stimulus or T cell membranes secreted mainly IgA (68%-94% of the total Ig secreted) and relatively little IgM. The strict T cell dependence of IgA B cell activation and differentiation provides important insights into immune responses of mucosal tissues and must be considered in the development of vaccines, particularly those designed to stimulate mucosal tissues containing large numbers of isotype-switched B cells.

Reciprocal regulation of mucosal surface IgA+ B cells by Ig receptor cross-linking and CD40 ligand.

In the present study, we analyze the role of Ig receptor cross-linking in T cell-dependent stimulation of both preswitch (surface IgM+ (sIgM+/sIgD+) B cells and postswitch (sIgA+) B cells. We demonstrate that purified sIgA+ B cells pretreated with anti-IgA-dextran at low concentrations (10 and 100 ng/ml) exhibited an increased response to activated T cells, whereas pretreatment with higher doses (1 and 10 micrograms/ml) led to a profound suppression of IgA secretion (> or = 90%). The suppressive effect of anti-IgA-dextran was accentuated in the presence of IL-2 and attenuated in the presence of IL-4. Anti-IgA-dextran pretreatment had no effect on sIgA+ B cell survival. sIgM+/sIgD+ B cells pretreated with anti-IgD-dextran or anti-IgM-dextran did not show significant inhibition. The increased susceptibility of sIgA+ B cells, but not of sIgM+/sIgD+ B cells, to Ig cross-linking-mediated suppression was confirmed in cross-linking studies with the same Ab (anti-kappa-dextran). Importantly, anti-IgA-dextran-mediated suppression could be reversed by stimulation of sIgA+ B cells with fibroblasts expressing CD40L; such a reversal required persistent exposure to cells expressing high levels of CD40L. These studies imply that Ig receptor cross-linking renders postswitch sIgA+ B cells unresponsive to subsequent stimulation via activated T cells, but this unresponsiveness is overcome by a persistent high level CD40L signal.

New insights into the immunopathology of chronic inflammatory bowel disease.

Inflammatory bowel diseases (IBDs) are chronic inflammatory disorders of the intestine encompassing Crohn's disease (CD) and ulcerative colitis (UC). The inflammation in CD can extend anywhere along the alimentary canal, whereas in UC the inflammation is limited to the large bowel. In CD the inflammation is typically patchy, transmural, and contains noncaseating granulomata, whereas in UC the inflammation is relatively superficial with ulcer formations. Nonetheless, 10% of the cases are clinically and histologically identical. Both disorders are idiopathic, because no etiological agent can be found, and both show autoimmune characteristics. However, the underlying immunological mechanism is still very poorly understood and controversial. Therefore the treatment of IBD, has, thus far, been unsatisfying, because it involves the administration of rather unspecific immunosuppressive agents. To find more specific therapies, it is necessary to study the underlying immunopathological mechanism of the disease.

Effect of transforming growth factor (TGF)-beta 1 on IgA isotype expression. TGF-beta 1 induces a small increase in sIgA+ B cells regardless of the method of B cell activation.

Transforming growth factor-beta (TGF-beta) has been reported to play an important role in IgA isotype expression when B cells are stimulated with LPS. The goal of the present study was to determine whether TGF-beta has similar effects on IgA isotype expression under more physiologic conditions utilizing a variety of B cell activation systems. As previously reported, in the LPS system TGF-beta caused a small, but significant, absolute increase in surface IgA (sIgA) expression and a very definite increase in IgA secretion; these effects were enhanced by IL-2 and IL-5. In the case of B cell stimulation with another B cell stimulant, the thymus-independent type II mitogen, anti-IgD-dextran, TGF-beta led to a similar small increase in sIgA expression, but caused suppression of IgA secretion. Using the Th2 cell clones CDC35 and D10 to stimulate resting B cells in a cognate and non-cognate T cell-dependent fashion, respectively, TGF-beta again increased sIgA expression to a similar small extent. TGF-beta at low doses (0.1 ng/ml) did not increase IgA secretion significantly and, at higher doses (1.0 ng/ml) caused significant suppression of IgA secretion. Addition of various cytokines (IL-2, -4, -5, D10sup) other than TGF-beta to stimulated B cells did not increase sIgA expression, but did give rise to increased amounts of IgA secretion, especially when activated D10 T cells were used as the B cell stimulant. Finally, the addition of an antibody against TGF-beta to cultures containing TGF-beta on day 2 led to partial or complete reversal of the inhibitory effects of TGF-beta on IgA secretion. In conclusion, TGF-beta causes a consistent, but small increase in sIgA+ B cells, in cultures of B cells stimulated by a variety of T cell-dependent or independent stimuli. In contrast, TGF-beta either promotes or inhibits B cell survival and IgA secretion, depending on the method of B cell activation. These results are most consistent with the view that TGF-beta provides only a partial or incomplete IgA switch signal but that additional factors are involved in IgA isotype switching and differentiation.