Th1 and Th2 mediate acute graft-versus-host disease, each with distinct end-organ targets

STAT4 and STAT6 are transcription factors that play crucial roles in responding to IL-12 and IL-4, respectively. STAT4 gene knockout (STAT4^–/–) mice have markedly reduced Th1 responses and enhanced Th2 responses. STAT6^–/– mice show the inverse phenotype. We compared the ability of bone marrow transplantation (BMT) with the inclusion of spleen cells from STAT6^–/–, STAT4^–/–, and wild-type (WT) mice to produce graft-versus-host disease (GVHD) in lethally irradiated MHC-mismatched recipients. Acute GVHD mortality was more rapid when induced by cells from STAT6^–/– mice than when induced by STAT4^–/– cells. However, cells from STAT4^–/– and STAT6^–/– donors both induced delayed GVHD mortality compared with WT controls, or compared with combined STAT4^–/– and STAT6^–/– cells, indicating a contribution of both Th1 cells and Th2 cells to acute GVHD. Recipients of STAT6^–/– BMT showed evidence of acute GVHD with severe diarrhea and marked weight loss. Recipients of STAT4^–/– BMT showed signs of GVHD with only initial transient weight loss and later development of severe skin GVHD. Histopathology showed that Th2 responses were required for the induction of both hepatic and severe skin GVHD. In contrast, both Th1 cells and Th2 cells were capable of causing intestinal pathology of GVHD. Our studies demonstrate an additive role for Th1 and Th2 cells in producing acute GVHD, and suggest a cytokine-directed approach to treating end-organ manifestations of GVHD.

Despite major advances, the full therapeutic potential of bone marrow transplantation (BMT) has not yet been achieved. Graft-versus-host disease (GVHD), arising from an attack of donor T cells against alloantigens of the recipient, is a major complication of allogeneic BMT, and prohibits the successful application of BMT across extensive human leukocyte antigen barriers. Despite the morbidity and death it causes, GVHD has been associated with 2 clinical benefits. Donor T cells exert graft-versus-leukemia effects by destroying residual leukemic cells, and also enhance the engraftment process, probably by destroying residual host T cells that survive intensive conditioning regimens (1). Therefore, separation of the beneficial engraftment-promoting and anti-leukemic effects of donor T cells from their ability to induce GVHD is a major goal of research in BMT.

Two distinct functional cytokine secretion patterns, termed Th1 and Th2, have been defined for murine as well as human T helper cells. The Th1 response is characterized by the production of IFN-γ, IL-2, and TNF-β, by the stimulation of delayed-type hypersensitivity reactions, and by the production of high-affinity Fcγ receptor–binding Abs (IgG_2a in the mouse) (2). Characteristic of the Th2 response in the mouse is the production of IL-4 and IL-5, and the stimulation of the production of IgE and IgG₁ Abs (2). These cell types are cross-regulatory in vitro, and the balance of these cells in vivo determines the character of cell-mediated immune and inflammatory responses (2). Recently, it has been suggested that the Th1/Th2 polarization of T helper cell subsets plays an important role in the development of GVHD. In both mice and humans, a “cytokine storm” related to the Th1 phenotype has been found to correlate with the development of GVHD (3, 4). Studies have shown that in vivo injection of IL-2 plus IL-4 into donor mice resulted in the generation of CD4⁺ IL-4– and IL-10–producing Th2 cells that, upon transfer into nonirradiated F₁ or sublethally irradiated mice, inhibited GVHD lethality (5, 6). Furthermore, production of Th2 cytokines induced by G-CSF administration has been associated with inhibitory effects on the development of GVHD (7). In another study, transplantation of in vitro–polarized murine Th2 cells resulted in a significant increase of survival after BMT across minor histocompatibility antigenic barriers (8).

However, some other studies of the therapeutic approach of using a Th1 to Th2 switch for the prevention or treatment of GVHD have suggested that the roles of Th1 and Th2 cells in GVHD may be far more complicated than was previously believed. Altering the development of GVHD or treating ongoing GVHD by direct administration of Th2 cytokines in vivo has not been successful. In 1 study, recipients of IL-4 showed increased mortality and equivalent, histopathological changes in the skin and liver compared with mice receiving no cytokine. It has also been shown that the exogenous administration of IL-10 was not toxic per se, but did accelerate GVHD lethality (9, 10). Furthermore, recent studies using IFN-γ or IL-4 gene knockout mice as donors showed that the absence of IFN-γ is still associated with severe acute GVHD (11, 12), whereas the inability of donors to produce IL-4 can be associated with less severe GVHD (12).

To elucidate the contribution of T helper cell subsets, we initiated a study using gene knockout mice that were capable of mounting only Th1 or Th2 responses as donors. Interactions between cytokines and their receptors leads to the activation of multiple signaling molecules, including the family of signal transducer and activator of transcription (STAT) proteins. STAT4 is a member of this family that is essential for responsiveness to IL-12. Mice with a disrupted STAT4 gene (STAT4^–/– mice) have impaired IL-12 responsiveness of NK and T cells, lack Th1 responsiveness, and have enhanced Th2 cell function (13, 14). STAT6 is activated in response to IL-4 and IL-13. Therefore, T lymphocytes from STAT6^–/– mice have impaired IL-4 and IL-13 responsiveness, and almost completely lack Th2 responses, but show enhanced Th1 responses (15–17). We assessed the role of Th1 cells and Th2 cells in GVHD by transferring bone marrow cells (BMC) and spleen cells from STAT4^–/– and STAT6^–/– mice to lethally irradiated, fully MHC-mismatched BMT recipients. Our results demonstrated that both Th1 cells and Th2 cells contribute to the development and severity of acute GVHD.

Both Th1 cells and Th2 cells contribute to GVHD. To evaluate the separate contributions of Th1 and Th2 cells to the induction of GVHD in a fully MHC-mismatched and multiple minor histocompatibility antigen–mismatched strain combination, we compared the ability of BMC plus spleen cells from BALB/c–backcrossed H-2^dSTAT4^–/– mice, STAT6^–/– mice, or WT control BALB/c mice to induce GVHD in lethally irradiated C57BL/6 recipients. The number of donor CD4⁺ and CD8⁺ T cells administered in WT, STAT4^–/–, and STAT6^–/– inocula was similar in all experiments, as measured by flow cytometry analysis of each inoculum. As shown in Figure 1, the most severe mortality occurred in the group receiving WT splenocytes, which are capable of producing both Th1 and Th2 responses (median survival time [MST] = 22 days). STAT6^–/– BMT, which is capable of producing high Th1 responses and no Th2 responses, induced mortality (MST = 31.5 days) that was less rapid in onset than that induced by WT BMT (P < 0.01). Mortality caused by the injection of STAT4^–/– cells, which lack Th1 and produce enhanced Th2 responses, was consistently less rapid (MST = 79 days) than that caused by both WT (P < 0.0001) and STAT6^–/– cells (P < 0.001). Figure 1 shows the mortality of 117 animals from 6 separate experiments. We observed that the rapidity of onset of GVHD varied significantly from experiment to experiment, but in each experiment the GVHD mortality was most rapid when induced by WT cells, followed by STAT6^–/– cells and finally STAT4^–/– cells. Furthermore, when GVHD was induced in the reverse strain combination, using C57BL/6–backcrossed H-2^bSTAT4^–/–, STAT6^–/–, or WT control C57BL/6 mice to induce GVHD in lethally irradiated BALB/c recipients, the same hierarchy of rapidity of GVHD mortality was observed: mortality induced by WT cells was the most rapid, followed by STAT6^–/– cells and then STAT4^–/– cells, the slowest (data not shown).

Figure 1

Both Th1 cells and Th2 cells contribute to GVHD mortality. Lethally irradiated B6 mice received 5 × 10⁶ T-cell–depleted (TCD) B6 BMC and either syngeneic spleen cells (SYN) (n = 18), BALB/c WT BMC and spleen cells (WT) (n = 38), STAT4^–/– (Th2) BMC and spleen cells (n = 31), or STAT6^–/– (Th1) BMC and spleen cells (n = 30). Each BALB/c (WT, STAT4^–/–, or STAT6^–/–) inoculum contained 13 × 10⁶ spleen cells plus 10 × 10⁶ BMC. Pooled data from 6 experiments, involving a total of 117 animals, are shown.

To further demonstrate that both Th1 cells and Th2 cells contribute to acute GVHD, we compared the severity and mortality of GVHD induced by STAT4^–/– cells, STAT6^–/– cells, and combined STAT4^–/– plus STAT6^–/– cells. Mice received either a full dose (13 × 10⁶) or a half dose (6.5 × 10⁶) of both STAT4^–/– and STAT6^–/– spleen cells plus BMC, or 13 × 10⁶ WT, STAT4^–/–, or STAT6^–/– cells alone. Whereas recipients of the full dose of combined STAT4^–/– and STAT6^–/– spleen cells developed accelerated GVHD (MST = 6 days) compared with recipients of WT inocula (MST = 16 days), there was no difference in GVHD induced by 13 × 10⁶ WT spleen cells compared with the combination of 6.5 × 10⁶STAT4^–/– cells and 6.5 × 10⁶STAT6^–/– cells (MST = 15.5 days). Recipients of 13 × 10⁶STAT6^–/– or STAT4^–/– spleen cells alone showed delayed GVHD compared with recipients of either 13 × 10⁶ WT cells or 6.5 × 10⁶STAT4^–/– cells and 6.5 × 10⁶STAT6^–/– cells, confirming that both Th1 cells and Th2 cells contribute to acute GVHD.

The addition of STAT4^–/– cells to WT cells does not protect against GVHD. It has previously been shown that transplantation of polarized Th2 cells, after 7 days in culture with IL-2 and IL-4, reduced the GVHD mortality caused by nonpolarized cell populations (27). We therefore attempted to determine whether naive Th2 cells could ameliorate GVHD induced by Th1 cells by adding STAT4^–/– BMC and spleen cells to WT cells. The addition of STAT4^–/– (Th2) cells to WT cells not only failed to prevent or reduce the severity of acute GVHD (MST = 18 days), but it accelerated GVHD mortality compared with that induced by WT cells alone (MST = 31 days) (Figure 2 shows one of two similar experiments). This result further emphasizes that Th2 cells play a pathogenic role in GVHD, and shows that they do not effectively inhibit GVHD mediated by naive cells that can differentiate into Th1 cells.

Figure 2

The addition of STAT4^–/– cells to WT BMC plus spleen cell inocula does not protect from GVHD. Lethally irradiated B6 mice received TCD B6 BMC and either syngeneic spleen cells (n = 3), BALB/c WT BMC and spleen cells (WT) (n = 6), STAT4^–/– (Th2) BMC and spleen cells (n = 8), or WT BMC and spleen cells plus STAT4^–/– BMC and spleen cells (STAT4^–/– + WT) (n = 3). Each inoculum contained 13 × 10⁶ BALB/c spleen cells plus 10 × 10⁶ BALB/c BMC (WT-induced control GVHD); 13 × 10⁶ BALB/c STAT4^–/– spleen cells plus 10 × 10⁶ BALB/c STAT4^–/– BMC (STAT4^–/–-induced GVHD); or 6.5 × 10⁶ BALB/c WT spleen cells plus 6.5 × 10⁶ BALB/c STAT4^–/– spleen cells plus 5 × 10⁶ BALB/c WT BMC plus 5 × 10⁶STAT4^–/– BMC (STAT4^–/– + WT cells). Results of one of two similar experiments are shown.

Unique clinical and pathological syndromes associated with Th2 and Th1 GVHD. We monitored transplanted animals for clinical signs of GVHD. Figure 3 shows the mortality curve and weight curve from a representative experiment. GVHD scores for generalized signs (hunching, ruffled fur, and periorbital edema) of WT recipients were the most severe in intensity and were the earliest in onset: on day 8, mean GVHD score was 9.8 ± 0.6 out of a possible maximum score of 10. The recipients of STAT6^–/– BMT and STAT4^–/– BMT had GVHD scores that were lower at this time point (on day 8, mean score for recipients of STAT4^–/– BMT was 2.3 ± 1.2, and for recipients of STAT6^–/– BMT it was 5.4 ± 0.5) and were similar to each other. However, the time required to achieve the maximal score for generalized signs of GVHD was earlier in recipients of STAT6^–/– BMT (on day 12, mean score was 6.7 ± 0.6) than in STAT4^–/– BMT recipients (on day 16, mean score was 7 ± 1). A weight curve, which provides a sensitive indicator of GVHD, is shown in the lower panel of Figure 3. Animals receiving WT cells rapidly lost weight before their deaths by day 13. Animals receiving Th1 (STAT6^–/–) and Th2 (STAT4^–/–) BMT showed initial weight loss during the first 10 days after BMT. However, the clinical signs subsequently diverged markedly in the two groups. Whereas recipients of STAT6^–/– BMT developed severe diarrhea and began to show mortality around day 25, recipients of STAT4^–/– BMT did not have diarrhea, and had only mild weight loss. As shown in the upper panel of Figure 3, recipients of STAT4^–/– BMT began to show mortality around day 30. These mice developed extremely severe skin GVHD, which began around the eyes, mouth, and ears and progressed to involve the whole body, resulting in hair loss and profound scaling. This striking clinical picture was observed in all six experiments, with time of onset ranging from 15 to 50 days after BMT. Whereas these clinical changes were not observed in any experiments in the animals that received BMT from WT or STAT6^–/– mice and survived until day 100, they were present in the majority (11 of 14) of STAT4^–/– BMT recipients that survived to 10 weeks after BMT.

Figure 3

Weight loss and mortality for WT-, Th1-, and Th2-mediated GVHD in a single representative experiment. GVHD mortality is shown in the upper panel, and GVHD-induced weight loss is shown in the lower panel. Lethally irradiated B6 mice received 5 × 10⁶ TCD B6 BMC and either syngeneic spleen cells (n = 3), BALB/c WT BMC and spleen cells (WT) (n = 6), STAT4^–/– (Th2) BMC and spleen cells (n = 8), or STAT6^–/– (Th1) BMC and spleen cells (n = 8). Each BALB/c (WT, STAT4^–/–, or STAT6^–/–) inoculum contained 13 × 10⁶ spleen cells plus 10 × 10⁶ BMC.

Different localization of pathological GVHD-associated changes in recipients of WT, STAT4^–/–, and STAT6^–/– BMT. After death, carcasses were fixed in formaldehyde and the tissues were processed and scored in a blinded fashion for lesions associated with GVHD. All analyzed organs were harvested from animals that died at similar times after transplantation from two experiments where the severity of GVHD was similar (MSTs in days for experiment 1 and 2: WT = 12 and 19, STAT6^–/– = 35 and 29, STAT4^–/– = 79 and 155). The liver, spleen, lungs, kidney, and large intestine were examined. There were no pathological changes in the appearance of liver (Figure 4a), skin (Figure 4e), or large intestine (Figure 4i) in recipients of syngeneic BMT. Moderate to severe GVHD-induced liver changes were observed in 5 of 5 recipients of WT BMT, which died on days 19, 24, 27, and 54 after BMT. Figure 4b shows typical liver lesions found in these mice; portal areas were diffusely infiltrated by a mixed population of leukocytes including neutrophils, lymphocytes, macrophages, and a few plasma cells. There was associated multifocal, single-cell hepatocellular necrosis, consistent with severe hepatic GVHD. Similar changes were observed in 4 of 4 recipients of STAT4^–/– BMT, which died on days 20, 39, 47, and 106 after BMT (Figure 4d). In striking contrast, liver sections of STAT6^–/– BMT recipients that died at similar times (22, 23, 25, 26, 27, 29, and 38 days after BMT) appeared normal, without any infiltrate or evidence of cholestasis (Figure 4c).

Figure 4

Histological evidence of GVHD in the liver, skin, and large intestine. Liver, skin, and large intestine sections were obtained from lethally irradiated B6 recipients reconstituted with syngeneic BMT (a, e, i), WT BMT (b, f, j), STAT6^–/– BMT (c, g, k), or STAT4^–/– BMT (d, h, l). Lethally irradiated B6 mice received 5 × 10⁶ TCD B6 BMC and either syngeneic spleen cells, BALB/c WT BMC and spleen cells (WT), STAT4^–/– (Th2) BMC and spleen cells, or STAT6^–/– (Th1) BMC and spleen cells. Each BALB/c (WT, STAT4^–/–, or STAT6^–/–) inoculum contained 13 × 10⁶ spleen cells plus 10 × 10⁶ BMC. These samples, with the exception of those from recipients of syngeneic BMT, were harvested from animals that died during the third week after BMT from two experiments in which the severity of GVHD was similar. Animals that died at later time points showed similar pathological changes. Syngeneic BMT recipients were sacrificed 200 days after BMT. Photographs were taken through a ×50 objective lens.

All recipients of WT BMT showed signs of slight skin hyperkeratosis. In some mice (3 of 7) that died on days 19 and 27 after BMT, there were very mild focal mononuclear inflammatory cell infiltrates located mostly in the dermis (Figure 4f). In accordance with the clinical picture, recipients of STAT4^–/– BMT (3 of 4) that died on days 20, 39, and 106 after BMT showed marked skin abnormalities. In some animals, pustules and crusting with neutrophil infiltration was observed on the skin’s surface (Figure 4h). In all animals, the dermis was infiltrated by a mixed population of leukocytes, including lymphocytes, macrophages, and neutrophils (Figure 4h). In several places the cells formed granulomas. There was also a suggestion of lysis along the epidermal-dermal junction. Consistent with the clinical picture, inflammatory changes were absent in all 12 recipients of STAT6^–/– BMT that died on days 5, 22, 23, 25, 26, 27, 29, 38, and 44 after BMT (Figure 4g). Furthermore, these severe skin changes were not observed in mice with GVHD induced by STAT6^–/– that had long-term survival (4 animals that died at 74, 122, 149, and 150 days after BMT). Their skin showed signs of mild hyperkeratosis, but without inflammatory changes.

Recipients of WT and STAT6^–/– BMT showed clinical signs of severe colitis manifested by profuse diarrhea and weight loss. The pathological examination revealed that in 7 of 7 recipients of WT inocula, which died on days 19, 24, 27, 39, and 54 after BMT, there was a thickening of colonic mucosa with mild to severe infiltration of mononuclear inflammatory cells. The epithelium was hyperplastic with elongated, tortuous, and branched crypts (Figure 4j). The colon sections of recipients of STAT6^–/– BMT showed similar changes in 9 of 9 examined animals, which died on days 5, 22, 23, 25, 26, 27, 29, 38, and 44 after BMT. Three of 9 animals had extremely severe chronic colitis, with lumens that were filled with exudate and severely ulcerated mucosa (Figure 4k). The changes observed in recipients of STAT6^–/– BMT were more severe than changes observed in recipients of WT BMT.

Although the recipients of STAT4^–/– BMT lacked clinical signs of colitis or weight loss, in 3 of 4 animals that died on days 39, 47, and 106 after BMT, there were histological signs of colitis, with a thickening of the mucosa and diffuse cellular infiltrate (Figure 4l). However, these changes were less severe than those observed in recipients of WT or STAT6^–/– BMT.

There were no pathological changes in the appearance of lungs or kidneys in any of the 3 groups of recipients (not shown).

STAT6^–/– cells and STAT4^–/– cells exhibit stable Th1 and Th2 phenotypes, respectively, in vivo. Cytokines modulate B-cell function by differentially stimulating specific Ig isotype production. Th1 responses are characterized by an increase in the level of IgG_2a production (2). Similarly, immune responses characterized by high levels of IL-4 lead to elevated serum IgE and IgG₁, and suppress the production of IgG_2a (2). Increases in serum IgE have been reported in GVHD in man and mouse (28, 29). We followed serum IgE levels in the GVHD mice over time. Normal nontransplanted C57BL/6 and BALB/c age- and sex-matched mice contained an average of 300 ng/mL IgE in their sera at the age of 20 weeks. IgE levels in sera of the WT and syngeneic BMT recipients did not differ significantly from those of controls (Figure 5a). In contrast, the recipients of STAT4^–/– BMT–induced GVHD showed significantly increased serum IgE levels as early as 2 weeks after BMT (Figure 5a). These levels increased over time, becoming greater than 4,000 ng/mL in GVHD mice sacrificed 5 months after BMT. The recipients of STAT6^–/– BMT showed almost no detectable IgE in the serum at the time of sacrifice 5 months after BMT (Figure 5a). This profile of high levels of IgE in STAT4^–/– BMT–induced GVHD, and IgE deficiency in the recipients of STAT6^–/– BMT, indicates that donor-derived T cells retain their Th1 or Th2 phenotype in GVHD mice.

Figure 5

Serum IgE (a) and IgG₁ (b) levels in GVHD induced by WT, STAT6^–/–, and STAT4^–/– cells. C57BL/6 animals were transplanted with syngeneic, WT BALB/c, STAT4^–/–, or STAT6^–/– cells. Lethally irradiated B6 mice received 5 × 10⁶ TCD B6 BMC and either syngeneic spleen cells, BALB/c WT BMC and spleen cells (WT), STAT4^–/– (Th2) BMC and spleen cells, or STAT6^–/– (Th1) BMC and spleen cells. Each BALB/c (WT, STAT4^–/–, or STAT6^–/–) inoculum contained 13 × 10⁶ spleen cells plus 10 × 10⁶ BMC. For syngeneic BMT and for WT- and STAT6^–/–-induced GVHD, animals were sacrificed at 5 months after BMT. Recipients of STAT4^–/– cells were bled and sacrificed at 2 weeks, at 3 months, and at 5 months after BMT. Sera were collected and Ig levels were measured by ELISA. For each time point for each group, 2–4 animals were analyzed. MSTs in days for this experiment were 26 for recipients of WT cells, 54 for recipients of STAT4^–/– cells, and 35 for recipients of STAT6^–/– cells. Concentrations of IgE and IgG₁ in sera are expressed as mean ± SEM.

We also followed the IgG₁ levels in the recipients of STAT4^–/–, STAT6^–/–, WT, and syngeneic BMT. The levels of IgG₁ in sera of these animals followed the IgE pattern. The recipients of STAT4^–/– GVHD inocula had significantly higher levels of IgG₁ in their sera than did the recipients of WT, STAT6^–/–, or syngeneic inocula, and this level increased over time (Figure 5b). In contrast, the recipients of STAT6^–/– GVHD inocula had almost undetectable levels of serum IgG₁ (Figure 5b). Total IgG levels were similar in the recipients of STAT6^–/– BMT (312.9 ± 152.0 μg/mL) sacrificed at 5 months after BMT and in recipients of STAT4^–/– cells (470.3 ± 38.0 μg/mL) sacrificed at 2 weeks after BMT, despite the difference in IgG₁ levels cited above.

Donor engraftment and T helper phenotype in recipients of WT, STAT4^–/–, and STAT6^–/– BMT. Recipients of STAT4^–/– cells were sacrificed at 2 weeks, 3 months (87 days), and 5 months (136 days) after BMT. Recipients of syngeneic and WT cells were sacrificed at 2 weeks and at 3 months (87 days) after BMT. For STAT6^–/–-induced GVHD, animals were sacrificed only at 5 months (136 days) after BMT. MLCs were prepared with host-type (B6) irradiated stimulator cells. FACS^® analysis of peripheral blood lymphocytes and bone marrow showed that these animals were full donor-type hematopoietic chimeras at all 3 time points (data not shown). The splenocytes from these animals showed mixed lymphocyte response reactivity both toward host C57BL/6 cells and third-party strain A/J, while being nonreactive toward donor antigens (data not shown). Thus, there was constant, ongoing activation and proliferation of donor T cells toward host antigens. By quantifying the levels of IFN-γ, IL-2, IL-4, and IL-10 in MLC supernatants from WT and STAT4^–/– BMT recipients, distinct patterns emerged. In supernatants obtained from spleen cells of mice receiving STAT4-deficient cells, the amounts of IL-4 and IL-10 (i.e., the Th2-specific cytokines) were increased compared with those in supernatants of spleen cells from WT recipients (see Figure 6). Similar results were obtained at all 3 time points after BMT. IFN-γ and IL-2, which are both induced during the Th1 response, were detected at higher levels in supernatants of spleen cells from recipients of WT BMT than in recipients of STAT4^–/– BMT (see Figure 6). Although produced at lower levels than in the recipients of WT BMT, there was measurable production of IFN-γ in the recipients of STAT4^–/– BMT (see Figure 6). Due to profound atrophy of the spleen at 5 months after BMT, only 1 recipient of STAT6^–/– BMT was available for analysis. This mouse showed the expected Th1 phenotype, with large amounts of IFN-γ (781 pg/mL) and IL-2 (591 pg/mL), low amounts of IL-10 (228 pg/mL), and undetectable IL-4 in the collected supernatants.

Figure 6

Cytokine profile of splenocytes from mice with WT- or STAT4^–/–-induced GVHD. Lethally irradiated C57BL/6 host mice transplanted with 5 × 10⁶ TCD B6 BMC and either 13 × 10⁶ WT spleen cells plus 10 × 10⁶ WT BMC (n = 2) or 13 × 10⁶STAT4^–/– spleen cells plus 10 × 10⁶STAT4^–/– BMC (n = 4) were sacrificed at 2 weeks after BMT, and splenocyte suspensions were prepared. Their donor-derived phenotype was demonstrated by FACS^®. The cells were cultured with irradiated splenocytes from C57BL/6 mice. Supernatants were harvested 72 hours after initiation of MLC, and cytokine levels were determined by ELISA. MSTs for this experiment were 26 days for recipients of WT cells and 54 days for recipients of STAT4^–/– cells. Concentrations of IFN-γ, IL-2, IL-4, and IL-10 in cell culture supernatants are expressed as mean ± SEM (pg/mL). ^AP < 0.01 vs. WT group.

Our study indicates that the role of Th1 cells and Th2 cells in GVHD is more complicated than previously thought. We have shown, contrary to other studies (6, 8, 27, 30), that both Th1 cells and Th2 cells contribute to acute GVHD. However, each subset causes injury of specific tissues, resulting in a unique clinical picture when Th1 or Th2 responses separately induce GVHD. Most strikingly, Th2 cells appear to be absolutely required for the induction of hepatic GVHD and severe skin GVHD. If confirmed in humans, these observations could have important implications for the therapy of GVHD, because they suggest that Th2-targeted approaches might be used to most effectively and specifically target liver and skin manifestations of acute GVHD, which are frequently seen without intestinal GVHD in clinical BMT.

We consistently observed that WT cells induced the most severe GVHD when compared with that induced by STAT6^–/– (Th1) or STAT4^–/– (Th2) cells. In other studies, we have observed that GVHD induced by IL-4^–/– cells (wherein the disease results mostly from Th1 cells, but with some Th2 component due to residual signaling through IL-13) was of a severity intermediate between that induced by WT cells and STAT6^–/– cells (B. Nikolic and M. Sykes, unpublished data). In further support of the hypothesis that both Th1 cells and Th2 cells contribute to GVHD, our group and others have shown that acute GVHD can occur in the absence of IFN-γ (11, 12), and that GVHD induced by donors that cannot produce IL-4 is milder than that induced by WT cells (12). Furthermore, the administration of anti–IL-4 mAb can ameliorate acute GVHD (31). Previous studies have shown that systemic administration of the Th1-inducing or Th1-associated cytokines IL-2 (18), IL-12 (32), or IFN-γ (33) can inhibit GVHD. Paradoxically, in other models, the administration of anti–IL-2 or anti–IL-12 mAbs (31, 34) has also been shown to prevent acute GVHD. Apparently, cytokines play complex roles in the induction and inhibition of GVHD, and the timing, the route of administration, and the strain combination may be critical in determining their effect (11, 35–38).

Although the experiments described herein were performed under similar conditions, with careful matching of the ages of donors and recipients and of the number of injected cells and dose of irradiation, we have always observed wide variability in the severity of GVHD from experiment to experiment, particularly over long time periods. There seem to be periods of months or years in which a given T-cell dose produces GVHD of a certain tempo in a particular model, and then this tempo later changes for reasons that cannot be identified. To minimize the effect of this variability and enable interpretation of data, we included every control group in every experiment. GVHD induced by WT cells serves as a standard for the severity of GVHD, and syngeneic BMT acts as a negative control. In this study, we consistently observed that GVHD mortality was more rapid when induced by cells from STAT6^–/– mice than that induced by STAT4^–/– cells. Nevertheless, the observed variability in the severity of GVHD and differences between human and murine GVHD mandate additional studies before direct translation of these results to human transplantation should be considered.

Although our results are at variance with previous studies suggesting that Th2 cells do not cause, and even prevent acute GVHD, these studies used spontaneous or LPS-induced mortality as an endpoint, and did not show an absence of GVHD pathology in recipients of Th2 cells (8, 27). Another possible reason for the inconsistency of our results with others that implicated Th1 cells but not Th2 cells in GVHD is that some of the previous studies relied on polarization of T helper cells by incubation with cytokines for several days in vitro (5, 8, 27). Polarized T cells were then transferred to irradiated recipients. This procedure might result in incomplete polarization, partial anergy, or a failure of subsequent proliferation, and could allow secondary conversion of T cells to a mixed phenotype upon transfer to BMT recipients. Indeed, examination of T-cell subsets in these BMT models revealed that T-cell populations were not polarized into “classical” Th1 or Th2 cytokine phenotypes 2 weeks after BMT (8), and that intermediate cytokine profiles of both CD8⁺ and CD4⁺ T-cell subsets developed, as has been described in other systems (30). STAT4- and STAT6-deficient mice have the advantage over these experimental models of allowing more pure Th1 or Th2 responses to be examined, and of not requiring a priming step before T cells are transferred for the induction of GVHD (15, 17, 39, 40). Although STAT6^–/– lymphocytes are completely impaired in their ability to generate IL-4–secreting cells, we were able to detect some IFN-γ in supernatants of MLC from STAT4^–/–-induced GVHD (Figure 6). A recent study showed that under conditions that promote Th1 cell differentiation, it is possible to detect low levels of production of IFN-γ via a STAT4-independent pathway (41). However, whereas mice that are both STAT4^–/– and STAT6^–/– are able to produce significant amounts of IFN-γ and mount Th1 cell–mediated delayed-type hypersensitivity, STAT4^–/– mice retain their predominantly Th2 phenotype, presumably through dominant IL-4 activation of STAT6 or Th2-mediated inhibition of Th1 cytokine production, or both (41).

Perhaps in keeping with our data showing that Th1 cells are associated with severe intestinal GVHD, a requirement for Th1 responses in inducing pathology in 2 different models of experimental colitis was recently demonstrated (42). Treatment with anti–IL-12 Abs or bone marrow reconstitution of T cell– and NK cell–deficient Tgε26 mice from STAT4-deficient mice resulted in modulation of the severity of disease (42). Thus, the reduced intestinal pathology in Th2 recipients may have been an important determinant in the observed prolonged survival of recipients of STAT4^–/– BMT. This may also explain data from other experiments in which mortality of recipients of in vitro–polarized Th2 cells was diminished or even absent (5, 7, 8, 27). In some of these studies, models were used that tend to induce milder GVHD, perhaps accounting for the failure to observe the contribution of Th2 cells to acute GVHD. Consistent with the failure of Th2 cells to induce acute GVHD across minor histocompatibility barriers without MHC differences (8), a recent study demonstrated that Th2 cells were unable to reject MHC-matched, minor histocompatibility antigen–mismatched cells, but were able to rapidly reject fully MHC-mismatched grafts (43). Thus, the presence of full MHC disparities in our model may partially explain the apparent discrepancy with the results of Krenger et al. (8, 27).

Consistent with our data showing a requirement for Th2 cells in hepatic and skin GVHD, a recent study has shown that although mortality from GVHD is substantially diminished after infusion with donor in vitro–polarized Th2 cells, clinical signs of GVHD are still present in surviving mice, and target organ pathology persists. Significant cutaneous and hepatic histopathological damage was evident 7 weeks after BMT in mice receiving polarized donor Th2 cells, and was equivalent to that in animals receiving naive donor T cells (8). Furthermore, after induction of GVHD with IFN-γ^−/− cells, extensive mononuclear cell infiltrates were seen in livers (12); treatment with anti–IL-4 mAb markedly reduced IgE and IgG₁ levels, and liver disease was ameliorated (31). Increased levels of IgE correlated strongly with the histological grades and stages of liver disease in parental mice receiving F₁ spleen cell transfer, and treatment with anti–IL-4 mAb prevented both phenomena (44). In humans, predominant activation of Th2 cells, indirectly measured by elevated serum IgE levels, was associated with liver disease in GVHD (28).

T1 (Tc1) and T2 (Tc2) subsets of CD8⁺ T cells secrete polarized patterns of cytokines similar to the Th1 and Th2 CD4 subsets. Although both T1 and T2 CD8⁺ T cells exhibit comparable levels of perforin-mediated cytolysis (45), T1 CD8⁺ T cells exhibit significantly higher Fas-mediated cytolysis (45). Fas-mediated anti-recipient cytotoxicity has been implicated in the development of hepatic and cutaneous GVHD against minor histocompatibility antigens (46), but this was not reproduced in a full haplotype-mismatched strain combination (47). In another study, administration of an anti–Fas ligand inhibitory mAb resulted in a reduction of the weight loss and mortality associated with GVHD directed against a full haplotype mismatch, but without improvement in skin or liver lesions (48). Therefore, Tc1-mediated cytolysis may not play a major role in hepatic and skin GVHD caused by a full MHC disparity. In accordance with these data, we have observed that only Th2 and Tc2 cells induced liver and skin GVHD in a fully MHC-mismatched strain combination.

The preferential injury of gut-associated tissue in Th1-induced GVHD, and of skin and liver in GVHD caused by Th2 cells, may be the result of different homing properties of the different subsets. Indeed, there is growing evidence that in addition to effector function, activated Th1 and Th2 lymphocytes acquire different migratory capacities (49–52). Chemokines and chemokine receptors are probably important for the selective migration of T-cell subsets (50). Distinct profiles of chemokine receptors are acquired by T cells after polarization; these are modulated by cytokines (50–52). Indeed, in vivo studies have confirmed this concept. Mouse Th1 cells express the functional ligand for P-selectin and E-selectin, PSGL-1, which promotes their migration into inflamed tissue (53). It was recently shown that Th1 cells could efficiently enter inflamed sites, such as sensitized skin or arthritic joints (53). In our model, preferential migration of Th2 cells, but not Th1 cells, into the liver and skin was observed, suggesting that other receptors or chemokines may play a decisive role in this migratory process.

There are potential dangers associated with therapies based on immune deviation from Th1 responses to Th2 responses. Th2 responses are responsible for the pathology of allergic diseases and at least 1 autoimmune condition (54–57). Our data indicate that manipulation of the Th1/Th2 balance as an approach to the prevention of GVHD may not be ideal. Possible severe side effects of such treatment should be considered before proceeding with human trials. However, our results suggest that the potential to treat hepatic and skin GVHD with anti–Th2 cytokine therapy deserves exploration.

This work was supported by National Institutes of Health (NIH) Grant HL-49915. B. Nikolic was supported in part by The Daland Fellowship for Research in Clinical Medicine (American Philosophical Society) and by an American Society of Transplant Physician–Sandoz Fellowship in Transplantation Award. M.J. Grusby was supported by NIH Grant AI-40171 and a gift from The Mathers Foundation. M.J. Grusby is a scholar of The Leukemia Society of America. S. Lee is currently studying at the Division of Clinical Immunology, Huddinge University Hospital, Karolinska Institute, Sweden. We thank Diane Plemenos for assistance in preparing the manuscript. We also thank David Anderson, Henry Winn, and Hugh Auchincloss for helpful review of the manuscript. We thank David H. Sachs for helpful advice.

1. Butturini, A, Gale, RP. T cell depletion in bone marrow transplantation for leukemia: current results and future directions. Bone Marrow Transplant 1988. 3:185-192.
   View this article via: PubMed Google Scholar
2. Abbas, AK, Murphy, KM, Sher, A. Functional diversity of helper T lymphocytes. Nature 1996. 383:787-793.
   View this article via: PubMed CrossRef Google Scholar
3. Antin, JH, Ferrara, JLM. Cytokine dysregulation in acute graft-versus-host disease. Blood 1992. 80:2964-2968.
   View this article via: PubMed Google Scholar
4. Blazar, BR, Korngold, R, Vallera, DA. Recent advances in graft-versus-host disease (GVHD) prevention. Immunol Rev 1997. 157:79-109.
   View this article via: PubMed CrossRef Google Scholar
5. Fowler, DH, Kurasawa, K, Smith, R, Eckhaus, MA, Gress, RE. Donor CD4-enriched cells of Th2 cytokine phenotype regulate graft-versus-host disease without impairing allogeneic engraftment in sublethally irradiated mice. Blood 1994. 84:3540-3549.
   View this article via: PubMed Google Scholar
6. Fowler, DH, Kurasawa, K, Husebekk, A, Cohen, PA, Gress, RE. Cells of the Th2 cytokine phenotype prevent LPS-induced lethality during murine graft-versus-host reaction. Regulation of cytokines and CD8+ lymphoid engraftment. J Immunol 1994. 152:1004-1013.
   View this article via: PubMed Google Scholar
7. Pan, L, Delmonte, J, Jalonen, CK, Ferrara, JLM. Pretreatment of donor mice with granulocyte colony-stimulating factor polarizes donor T lymphocytes toward type-2 cytokine production and reduces severity of experimental graft-versus-host disease. Blood 1997. 86:4422-4429.
   View this article via: PubMed Google Scholar
8. Krenger, W, et al. Transplantation of polarized type 2 donor T cells reduces mortality caused by experimental graft-versus-host disease. Transplantation 1996. 62:1278-1285.
   View this article via: PubMed CrossRef Google Scholar
9. Blazar, BR, Taylor, PA, Smith, S, Vallera, DA. Interleukin-10 administration decreases survival in murine recipients of major histocompatibility complex disparate donor bone marrow cells. Blood 1995. 85:842-851.
   View this article via: PubMed Google Scholar
10. Krenger, W, Snyder, K, Smith, S, Ferrara, JLM. Effects of exogenous interleukin-10 in a murine model of graft-versus-host disease to minor histocompatibility antigens. Transplantation 1994. 58:1251-1257.
    View this article via: PubMed Google Scholar
11. Yang, Y-G, Dey, B, Sergio, JJ, Pearson, DA, Sykes, M. Donor-derived interferon γ is required for inhibition of acute graft-versus-host disease by interleukin 12. J Clin Invest 1998. 102:2126-2135.
    View this article via: JCI PubMed CrossRef Google Scholar
12. Murphy, WJ, et al. Differential effects of the absence of interferon-γ and IL-4 in acute graft-versus-host disease after allogeneic bone marrow transplantation in mice. J Clin Invest 1998. 102:1742-1748.
    View this article via: JCI PubMed CrossRef Google Scholar
13. Kaplan, MH, Sun, Y-L, Hoey, T, Grusby, MJ. Impaired IL-12 responses and enhanced development of Th2 cells in STAT4-deficient mice. Nature 1997. 382:174-177.
    View this article via: PubMed CrossRef Google Scholar
14. Thierfelder, WE, et al. Requirement for STAT4 in interleukin-12-mediated responses of natural killer and T cells. Nature 1996. 382:171-174.
    View this article via: PubMed CrossRef Google Scholar
15. Takeda, K, et al. Essential role of STAT6 in IL-4 signalling. Nature 1996. 380:627-630.
    View this article via: PubMed CrossRef Google Scholar
16. Kaplan, MH, Schindler, U, Smiley, ST, Grusby, MJ. STAT6 is required for mediating responses to IL-4 and for the development of Th2 cells. Immunity 1996. 4:313-319.
    View this article via: PubMed CrossRef Google Scholar
17. Shimoda, K, et al. Lack of IL-4-induced Th2 response and IgE class switching in mice with disrupted STAT6 gene. Nature 1996. 380:630-633.
    View this article via: PubMed CrossRef Google Scholar
18. Sykes, M, Romick, ML, Sachs, DH. Interleukin 2 prevents graft-versus-host disease while preserving the graft-versus-leukemia effect of allogeneic T cells. Proc Natl Acad Sci USA 1990. 87:5633-5637.
    View this article via: PubMed CrossRef Google Scholar
19. Dialynas, DP, et al. Characterization of murine T cell surface molecule, designated L3T4, identified by monoclonal antibody GK1.5: similarity of L3T4 to human Leu3/T4 molecule. J Immunol 1983. 131:2445-2451.
    View this article via: PubMed Google Scholar
20. Sarmiento, M, Glasebrook, AL, Fitch, FW. IgG or IgM monoclonal antibodies reactive with different determinants on the molecular complex bearing Lyt2 antigen block T cell-mediated cytolysis in the absence of complement. J Immunol 1980. 125:2665-2672.
    View this article via: PubMed Google Scholar
21. Sykes, M, Abraham, VS, Harty, MW, Pearson, DA. IL-2 reduces graft-versus-host disease and preserves a graft-versus-leukemia effect by selectively inhibiting CD4⁺ T cell activity. J Immunol 1993. 150:197-205.
    View this article via: PubMed Google Scholar
22. Sykes, M, Romick, ML, Hoyles, KA, Sachs, DH. In vivo administration of interleukin 2 plus T cell-depleted syngeneic marrow prevents graft-versus-host disease mortality and permits alloengraftment. J Exp Med 1990. 171:645-658.
    View this article via: PubMed CrossRef Google Scholar
23. Unkeless, JC. Characterization of a monoclonal antibody directed against mouse macrophage and lymphocyte Fc receptors. J Exp Med 1979. 150:580-596.
    View this article via: PubMed CrossRef Google Scholar
24. Ozato, K, Mayer, NM, Sachs, DH. Monoclonal antibodies to mouse major histocompatibility complex antigens. IV. A series of hybridoma clones producing anti-H-2d antibodies and an examination of expression of H-2d antigens on the surface of these cells. Transplantation 1982. 34:113-120.
    View this article via: PubMed Google Scholar
25. Wang, M-G, Szebeni, J, Pearson, DA, Szot, GL, Sykes, M. Inhibition of graft-versus-host disease (GVHD) by IL-2 treatment is associated with altered cytokine production by expanded GVH-reactive CD4+ helper cells. Transplantation 1995. 60:481-490.
    View this article via: PubMed Google Scholar
26. Ohdan, H, Yang, Y-G, Shimizu, A, Swenson, KG, Sykes, M. Mixed chimerism induced without lethal conditioning prevents T cell- and anti-Galα1,3Gal-mediated graft rejection. J Clin Invest 1999. 104:281-290.
    View this article via: JCI PubMed CrossRef Google Scholar
27. Krenger, W, Snyder, KM, Byon, JCH, Falzarano, G, Ferrara, JLM. Polarized type 2 alloreactive CD4+ and CD8+ donor T cells fail to induce experimental acute graft-versus-host disease. J Immunol 1995. 155:585-593.
    View this article via: PubMed Google Scholar
28. Heyd, J, Doonenberg, AD, Burns, WH, Saral, R, Santos, GW. Immunoglobulin E levels following allogeneic, autologous and syngeneic bone marrow transplantation: an indirect association between hyperproduction and acute GVHD in allogeneic bone marrow transplantation. Blood 1988. 72:442-446.
    View this article via: PubMed Google Scholar
29. Claman, HN, Spiegelberg, H. Immunoglobulin dysregulation in murine GVHD: a hyper IgE syndrome. Clin Immunol Immunopathol 1990. 56:46-53.
    View this article via: PubMed CrossRef Google Scholar
30. Kelso, A. Th1 and Th2 subsets: paradigms lost? Immunol Today 1995. 16:374-379.
    View this article via: PubMed CrossRef Google Scholar
31. Ushiyama, C, et al. Anti-IL-4 antibody prevents graft-versus-host disease in mice after allogeneic bone marrow transplantation. The IgE allotype is an important marker of graft-versus-host disease. J Immunol 1995. 154:2687-2696.
    View this article via: PubMed Google Scholar
32. Sykes, M, Szot, GL, Nguyen, PL, Pearson, DA. Interleukin-12 inhibits murine graft-versus-host disease. Blood 1995. 86:2429-2438.
    View this article via: PubMed Google Scholar
33. Brok, HPM, Heidt, PJ, Van der Meide, PH, Zurcher, C, Vossen, JM. Interferon-gamma prevents graft-versus-host disease after allogeneic bone marrow transplantation in mice. J Immunol 1993. 151:6451-6459.
    View this article via: PubMed Google Scholar
34. Williamson, E, Garside, P, Bradley, JA, More, IAR, Mowat, AM. Neutralizing IL-12 during induction of murine acute graft-versus-host disease polarizes the cytokine profile toward a Th2-type alloimmune response and confers long term protection from disease. J Immunol 1997. 159:1208-1215.
    View this article via: PubMed Google Scholar
35. Dey, B, Yang, Y-G, Szot, GL, Pearson, DA, Sykes, M. IL-12 inhibits GVHD through a Fas-mediated mechanism associated with alterations in donor T cell activation and expansion. Blood 1998. 91:3315-3322.
    View this article via: PubMed Google Scholar
36. Sykes, M, Abraham, VS, Harty, MW, Pearson, DA. Selective inhibition of CD4 graft-versus-host activity in IL-2-treated mice. Transplant Proc 1992. 25:1225-1226.
    View this article via: PubMed Google Scholar
37. Sykes, M, Abraham, VS. Mechanism of IL-2-mediated protection against GVHD in mice. II. Protection occurs independently of NK/LAK cells. Transplantation 1992. 53:1063-1070.
    View this article via: PubMed Google Scholar
38. Sykes, M, Harty, MW, Szot, GL, Pearson, DA. Interleukin-2 inhibits graft-versus-host disease-promoting activity of CD4+ cells while preserving CD4- and CD8-mediated graft-versus-leukemia effects. Blood 1994. 83:2560-2569.
    View this article via: PubMed Google Scholar
39. Huang, H, Paul, WE. Impaired interleukin 4 signaling in T helper type 1 cells. J Exp Med 1998. 187:1305-1313.
    View this article via: PubMed CrossRef Google Scholar
40. Akimoto, T, et al. Abrogation of bronchial eosinophilic inflammation and airway hyperreactivity in signal transducers and activators of transcription (STAT)6-deficient mice. J Exp Med 1998. 187:1537-1542.
    View this article via: PubMed CrossRef Google Scholar
41. Kaplan, MH, Wurster, AL, Grusby, MJ. A signal transducer and activator of transcription (STAT)4-independent pathway for the development of T helper type 1 cells. J Exp Med 1998. 6:1191-1196.
42. Simpson, JS, et al. T cell-mediated pathology in two models of experimental colitis depends on the interleukin 12/signal transducer and activator of transcription (STAT)-4 pathway, but is not conditional on interferon γ expression by T cells. J Exp Med 1998. 187:1225-1234.
    View this article via: PubMed CrossRef Google Scholar
43. Xian, LC, Zand, MS, Li, Y, Zheng, XX, Strom, TB. On histocompatibility barriers, Th1 to Th2 immune deviation, and the nature of the allograft responses. J Immunol 1998. 161:2241-2247.
    View this article via: PubMed Google Scholar
44. Takai, S, et al. Increased IgE level as a marker of host-versus-graft disease: inhibition of this HVGD with a monoclonal antibody to IL-4. Cell Immunol 1993. 149:1-10.
    View this article via: PubMed CrossRef Google Scholar
45. Carter, LL, Dutton, RW. Relative perforin- and Fas-mediated lysis in T1 and T2 CD8 effector populations. J Immunol 1995. 155:1028-1031.
    View this article via: PubMed Google Scholar
46. Baker, MB, Altman, NH, Podack, ER, Levy, RB. The role of cell-mediated cytotoxicity in acute GVHD after MHC-matched allogeneic bone marrow transplantation. J Exp Med 1996. 183:2645-2654.
    View this article via: PubMed CrossRef Google Scholar
47. Tsukada, N, Kobata, T, Aizawa, Y, Yagita, H, Okumura, K. Graft-versus-leukemia effect and graft-versus-host disease can be differentiated by cytotoxic mechanisms in a murine model of allogeneic bone marrow transplantation. Blood 1999. 93:2738-2747.
    View this article via: PubMed Google Scholar
48. Miwa, K, et al. Therapeutic effect of an anti-Fas ligand mAb on lethal graft-versus-host disease. Int Immunol 1999. 11:925-931.
    View this article via: PubMed CrossRef Google Scholar
49. Sallusto, F, Mackey, CR, Lanzavecchia, A. Selective expression of the eotaxin receptor CCR3 by human T helper 2 cells. Nature 1997. 227:2005-2007.
50. Ward, SG, Bacon, KB, Westwick, J. Chemokines and T lymphocytes: more than an attraction. Immunity 1998. 9:1-11.
    View this article via: PubMed CrossRef Google Scholar
51. Xu, D, et al. Selective expression of a stable cell surface molecule on type 2 but not type 1 helper T cells. J Exp Med 1998. 187:787-794.
    View this article via: PubMed CrossRef Google Scholar
52. Bonecchi, R, et al. Differential expression of chemokine receptors and chemotactic responsiveness of type 1 T helper cells (Th1s) and Th2s. J Exp Med 1998. 187:129-134.
    View this article via: PubMed CrossRef Google Scholar
53. Austrup, F, et al. P- and E-selectin mediate recruitment of T-helper-1 but not T-helper-2 cells into inflamed tissues. Nature 1997. 385:81-83.
    View this article via: PubMed CrossRef Google Scholar
54. Lafaille, JJ, et al. Myelin basic protein-specific T helper 2 (Th2) cells cause experimental autoimmune encephalomyelitis in immunodeficient hosts rather than protect them from the disease. J Exp Med 1997. 186:307-312.
    View this article via: PubMed CrossRef Google Scholar
55. Pakala, SV, Kurrer, MO, Katz, JD. T helper 2 (Th2) T cells induce acute pancreatitis and diabetes in immune-compromised nonobese diabetic (NOD) mice. J Exp Med 1997. 186:299-306.
    View this article via: PubMed CrossRef Google Scholar
56. O’Garra, A, Steinman, L, Gijbels, K. CD4+ T-cell subsets in autoimmunity. Curr Opin Immunol 1998. 9:872-883.
    View this article via: PubMed CrossRef Google Scholar
57. Chilosi, M, et al. CD30 cell expression and abnormal soluble CD30 serum accumulation in Omenn’s syndrome: evidence for a T helper 2-mediated condition. Eur J Immunol 1996. 26:329-334.
    View this article via: PubMed CrossRef Google Scholar