Interferon-α drives T cell-mediated immunopathology in the intestine

Authors

  • Giovanni Monteleone,

    1. Division of Infection, Inflammation and Repair, University of Southampton, School of Medicine, Southampton General Hospital, Southampton, GB
    Search for more papers by this author
  • Sylvia L. F. Pender,

    1. Division of Infection, Inflammation and Repair, University of Southampton, School of Medicine, Southampton General Hospital, Southampton, GB
    Search for more papers by this author
  • Neville C. Wathen,

    1. Joint Academic Department of Obstetrics and Gynecology, St. Bartholomew's and The Royal London School of Medicine and Dentistry, London, GB
    Search for more papers by this author
  • Thomas T. MacDonald

    Corresponding author
    1. Division of Infection, Inflammation and Repair, University of Southampton, School of Medicine, Southampton General Hospital, Southampton, GB
    • Division of Infection, Inflammation and Repair, School of Medicine, Southampton General Hospital, MailPoint 813, Level E, South Academic Block, Tremona Road, Southampton SO16 6YD, GB Fax: +44-2380-796-603
    Search for more papers by this author

Abstract

The ability of interferon (IFN)-α to induce autoimmunity and exacerbate Th1 diseases is well known. We have recently described enhanced expression of IFN-α in the mucosa of patients with celiac disease (CD), a gluten-sensitive Th1-mediated enteropathy, characterized by villous atrophy and crypt cell hyperplasia. Previous studies from this laboratory have shown that T cell activation in explant cultures of human fetal gut can also result in villous atrophy and crypt cell hyperplasia. We have, therefore, examined changes that take place in explant cultures of human fetal gut after activation of T cells with anti-CD3 and/or IFN-α. We show that activation of T cells with anti-CD3 alone elicits a small IFN-γ and TNF-α response with no tissue injury. Similarly, no changes are seen in explants cultured with IFN-α alone. However, addition of IFN-α with anti-CD3 results in enhanced Th1 response and crypt cell hyperplasia. This is associated with enhanced phosphorylation of STAT1, STAT3, and Fyn, a Src homology tyrosine kinase, which interacts with both TCR and IFN-α signal components. Together these data indicate that IFN-α can facilitate activation of Th1-reactive cells in the gut and drive immunopathology.

Abbreviations:
CD:

Celiac disease

STAT:

Signal transducers and activators of transcription

MMP:

Matrix metalloproteinase

KGF:

Keratinocyte growth factor

RT-PCR:

Reverse transcriptase

PCR TIMP:

Tissue inhibitor of MMP

p-STAT1:

Phosphorylated STAT1

1 Introduction

The interferons (IFN) are a family of cytokines, which include two major subgroups based on their properties and cellular responses, termed type I (IFN-α and IFN-β) and type II (IFN-γ), respectively 1. IFN-α, mostly synthesized by antigen-presenting cells (APC), plays an important role in modulating immune responses to foreign and self antigens 1, 2. In this context, IFN-α has been shown to facilitate differentiation of naive T cells along the Th1 pathway, to enhance antibody production and to support the functional activity and survival of memory T cells 14. Similar biological responses have also been documented in patients with malignant and inflammatory diseases receiving IFN-α therapy. Indeed, after systemic administration, IFN-α transduces a systemic and persistent Th1 immune response 1, 2, 5. Consistent with this, evidence has emerged in recent years to show that IFN-α can induce autoimmunity and exacerbate Th1-mediated diseases [1, 2, 6–8].

Celiac disease (CD), a gluten-sensitive enteropathy, is characterized by villous atrophy and crypt cell hyperplasia 9. In CD, the intestinal lesion is associated with exaggerated Th1 response, which occurs in the absence of interleukin (IL)-12, the major Th1-inducing cytokine 10, 11. We have recently demonstrated that IFN-α is expressed in the mucosa of patients with untreated CD, but not in controls 12. In addition, we and others have documented the development of clinical CD in patients receiving IFN-α, suggesting a possible role for this cytokine in promoting the local immune response in CD 12, 13.

Previous studies from this laboratory have clearly demonstrated that activation of T cells in the lamina propria of fetal human gut can result in tissue injury. Activation of lamina propria T cells with pokeweed mitogen or anti-CD3 and IL-12 results in enhanced Th1 response, production of matrix metalloproteinases (MMP) and degradation of lamina propria extracellular matrix 1417. If lamina propria T cells are activated with the superantigen staphylococcal enterotoxin B (SEB), there is less cytokine production, less MMP production, and the tissue response mainly involves villous atrophy and an increased rate of epithelial cell proliferation. This is caused by the cytokine-induced production of the epithelial mitogen keratinocyte growth factor (KGF) by stromal cells 18, 19. This ex vivo model thus represents an ideal system to determine if an excess of IFN-α drives Th1 responses and results in immunopathology.

In this study we hypothesized that, in the gut mucosa, abnormal expression of IFN-α can enhance activation of Th1 cells with the downstream effect of causing mucosal damage.

2 Results

2.1 Cytokine profiles in IFN-α and anti-CD3-stimulated fetal intestinal explants

To determine if IFN-α had any effect on mucosal T cell responses, explants of human fetal gut were cultured with anti-CD3 and/or graded doses of IFN-α and the amount of cytokine transcripts measured by quantitative reverse transcriptase (RT)-PCR. On day 4 of culture, the levels of transcripts for IFN-γ and TNF-α were low or absent in unstimulated fetal gut explants as well as in those stimulated with IFN-α alone (Fig. 1). Anti-CD3 alone induced a few IFN-γ or TNF-α transcripts (Fig. 1). However, addition of IFN-α with anti-CD3 elicited a dose-dependent increase in IFN-γ transcripts (Fig. 1). Stimulation with IFN-α and anti-CD3 also produced a small but significant dose-dependent increase in the amount of transcripts for TNF-α compared to that obtained in explants stimulated with anti-CD3 alone (p<0.03) (Fig. 1). Similar results were observed in RNA samples extracted from explants cultured for 24 h (data not shown).

IL-4 and IL-5 transcripts were very low in both unstimulated and stimulated explants (data not shown).

Figure 1.

 Transcripts for IFN-γ and TNF-α in fetal gut explants stimulated with anti-CD3 alone, IFN-α alone or anti-CD3 and graded doses of IFN-α. Results are mean + SEM of four separate experiments.

2.2 Stimulation of human fetal gut explants with anti-CD3 and IFN-α results in crypt cell hyperplasia

As previously reported 14, 15, stimulation of explants with mitogenic anti-CD3 antibody caused a mild crypt cell hyperplasia. This was documented by an increase in the number of Ki67+ proliferating cells in the intestinal crypts (Fig. 2). No morphological change was observed in explants cultured in the presence of IFN-α alone, indicating that IFN-α by itself has no direct effect on the epithelial cells (Fig. 2). Addition of graded doses of IFN-α and anti-CD3 resulted in a more pronounced crypt cell hyperplasia (p<0.04) (Fig. 2). Although as little as 10 IU/ ml IFN-α enhanced crypt cell proliferation, maximal effects were obtained using 100 IU/ml IFN-α. Therefore, 100 IU/ml IFN-α was used for all subsequent experiments. The addition of a neutralizing IFN-α antibody to the organ cultures co-stimulated with 100 IU/ml IFN-α and anti-CD3 dramatically inhibited the crypt cell hyperplasia, whereas no inhibitory effect was seen with a control antibody (not shown). To prove that the morphological changes seen in explants stimulated with anti-CD3 and IFN-α were secondary to T cell activation, in two separate experiments we added 10 μg/ml cyclosporin A at the onset of culture with anti-CD3 and IFN-α. The addition of cyclosporin A completely prevented the crypt cell hyperplasia (p=0.003) (Fig. 2).

Figure 2.

Percentage of Ki67+ cells in the crypts of Lieberkuhn in fetal gut explants. Explants from two different fetuses were cultured with medium alone or soluble anti-CD3 in the presence or absence of IFN-α at the indicated concentrations for 4 days. The percentage of Ki67+ cells per crypt was evaluated in at least six crypts for each of ten explants per sample. Data are expressed as mean + SEM.

We have previously shown that, in the fetal gut model, KGF is an important mediator of the crypt cell hyperplasia induced by SEB-stimulated T cell activation 18. Therefore, we examined whether the increased rate of epithelial cell proliferation induced by anti-CD3 and IFN-α was associated with enhanced KGF. As shown in Fig. 3, KGF RNA transcripts were expressed at low level in fetal gut explants cultured with medium alone for 4 days. Addition of anti-CD3, but not IFN-α, to the explant cultures increased the expression of KGF (Fig. 3). Moreover, the stimulation of explants with anti-CD3 and IFN-α resulted in a small increase in KGF RNA in comparison to that seen in explants stimulated with anti-CD3 alone (Fig. 3). Similar results were observed in fetal gut explants cultured for 24 h (not shown).

Figure 3.

(A) Southern blot analysis of RNA transcripts for KGF (upper blot) and β-actin (lower blot) in fetal gut explants cultured with medium alone (unst), IFN-α (100 IU/ml), anti-CD3 (aCD3) or anti-CD3+IFN-α for 4 days. The experiment is representative of four separate experiments. (B) Ratio between KGF and β-actin RNA transcripts in fetal gut explants cultured with medium alone (unst), IFN-α (100 IU/ml), anti-CD3 (aCD3) or anti-CD3+IFN-α for 4 days. The level of RNA transcripts was measured by laser densitometry and the relative values expressed in arbitrary units. The numbers indicate the mean ± SD of four separate experiments.

Since, in the fetal gut system, Th1 cell activation has been associated with the synthesis of extracellular matrix-degrading MMP 1618, we then determined whether activation of T cells in the lamina propria of fetal gut with anti-CD3 and IFN-α resulted in the induction of specific MMP. On day 4 of culture, unstimulated fetal gut explants contained transcripts for interstitial collagenase (MMP-1), gelatinase A (MMP-2) and tissue inhibitor of MMP (TIMP)-1 and -2, whereas the amount of RNA transcripts for stromelysin-1 (MMP-3) and gelatinase B (MMP-9) was low (Table 1). Addition of anti-CD3, but not IFN-α (100 IU/ml) alone, elicited a significant increase in the amount of transcripts for MMP-1, MMP-2, MMP-3, and TIMP-1 (p<0.03) (Table 1). Addition of IFN-α to the fetal gut explant cultures stimulated with anti-CD3 did not, however, result in any significant increase in the amount of transcripts for MMP and TIMP compared with that measured in explants stimulated with anti-CD3 alone (Table 1).

Table 1. MMP and TIMP RNA transcripts in fetal gut explants stimulated with anti-CD3 and/or IFN-α for 4 days. Controls were cultured with medium alonea)
StimulusMMP-1MMP-2MMP-3MMP-9TIMP-1TIMP-2
  1. a) Numbers are transcripts/μg total RNA×103 and indicate mean  ±  SD of three separate experiments.

Control  18 ± 2    16 ± 2.4 5.3 ± 1.2  1 ± 0.4   40 ± 8.4 1,050 ± 40  
IFN (100 IU/ml)  19.6 ± 1.6 18.6 ± 1.6 6.9 ± 2.40.78 ± 0.6 32.1 ± 9.1 950.3 ± 120 
Anti-CD3 95.9 ± 10  29.6 ± 3.147.6 ± 4.4 3.5 ± 0.2145.5 ± 9.2 1,284 ± 90  
Anti-CD3+IFN (10 IU/ml)  105 ± 9.2 28.1 ± 1.740.9 ± 6.1 3.8 ± 0,2167.4 ± 6.8 1,182 ± 98.6
Anti-CD3+IFN (50 IU/ml)102.3 ± 7   31.1 ± 4.146.6 ± 5.4 3.7 ± 0.3133.7 ± 9.3 1,010 ± 76.3
Anti-CD3+IFN (100 IU/ml)  100 ± 9.8 30.9 ± 6.550.9 ± 4.8 3.4 ± 0.2160.9 ± 8.7 905.6 ± 87.6

2.3 IFN-α enhances phosphorylation of STAT1 and STAT3 in fetal gut after T cell activation with anti-CD3

The co-ordinated action of TCR-stimulated and cytokine signaling pathways plays a central role in the control of lymphocyte proliferation, survival and functional activity. These signals are in part integrated through activation of signal transducers and activators of transcription (STAT) 20, 21. Among these, STAT1, which is essential for IFN-α-induced functions in vivo, has recently shown to be involved also in TCR-mediated gene regulation 22, 23. Therefore we examined whether the synergistic effect of anti-CD3 and IFN-α in the fetal gut explants was associated with activation of STAT1. Total proteins were prepared from fetal gut explants after 2, 4 and 6 h of stimulation with anti-CD3 and/or IFN-α, and separated by SDS-PAGE. Phosphorylation of STAT1 (p-STAT1) was then monitored by Western blotting analysis using antibodies that specifically recognize STAT1 phosphorylation on tyrosine 701 (p-Tyr-STAT1) or serine 727 (p-Ser-STAT1), respectively. As shown in Fig. 4 (upper panel), incubation of fetal gut explants with anti-CD3 or IFN-α alone failed to induce tyrosine phosphorylation of STAT1 over a 6-h time period. However, anti-CD3 stimulation in the presence of IFN-α resulted in the appearance of a band corresponding to p-Tyr-STAT1α as early as 4 h of culture (Fig. 4, upper panel). In unstimulated fetal gut explants, STAT1 was weakly phosphorylated on the serine 727 group (Fig. 4, middle panel), consistently with previous studies showing that quiescent T cells have low basal STAT1 serine 727 phosphorylation 24. Addition of anti-CD3 and/or IFN-α to the explant cultures enhanced STAT1 serine phosphorylation, and the intensity of the band seen in explants stimulated with anti-CD3 and IFN-α in combination was greater than that induced by anti-CD3 or IFN-α alone (Fig. 4, middle panel). Reprobing the blots with a pan-STAT1 antibody demonstrated that equal amounts of STAT1α and STAT1β were present in all the samples (Fig. 4, lower panel).

Figure 4.

Activation of T cells in fetal gut explants with anti-CD3 and IFN-α results in enhanced STAT1 phosphorylation. Total proteins, extracted from fetal gut explants cultured for 2, 4 and 6 h, were run on SDS-PAGE, and subsequently analyzed for p-Tyr-STAT1 (upper blot), p-Ser-STAT1 (middle blot) and total STAT1 (lower blot). The example is representative of four separate experiments.

Figure 5.

Active (tyrosine phosphorylated) and total STAT3 (A), STAT4 (B), and STAT5 (C) in fetal gut explants cultured in the presence or absence of anti-CD3 antibody and/or IFN-α for 6 h. The example is representative of three separate experiments.

To examine if activation of T cells with anti-CD3 and IFN-α associates with activation of other STAT molecules in the fetal gut, we first analyzed the tyrosine phosphorylation of STAT3, a molecule which has reported to directly interact with IFN-α receptor and be activated by TCR ligation 25, 26. As expected, anti-CD3 or IFN-α alone enhanced the phosphotyrosine STAT3 signal in comparison to unstimulated explants (Fig. 5A). The addition of IFN-α in the presence of anti-CD3 resulted in an increase in the level of p-Tyr-STAT3 in comparison to the explants stimulated with anti-CD3 or IFN-α alone (Fig. 5A). This was evident at 2, 4 and 6 h of culture (not shown). Since there is also evidence that IFN-α enhances tyrosine phosphorylation of STAT4 and STAT5 in anti-CD3-activated T cells 27, we set out to explore the possibility that stimulation of fetal gut explants with anti-CD3 and IFN-α may also result in the activation of STAT4 and STAT5. Immunoreactivity for phosphorylated STAT4 was seen in unstimulated fetal gut explants (Fig. 5B). IFN-α, but not anti-CD3, increased the STAT4 phosphorylation, and no further increase was seen when anti-CD3 and IFN-α were used in combination (Fig. 5B). In contrast, signaling via CD3, but not IFN-α enhanced STAT5 phosphorylation (Fig. 5C). Similar results were seen at 2, 4 and 6 h of culture (not shown).

2.4 PP2, a Src homology protein inhibitor, prevents STAT1 and STAT3 phosphorylation, and crypt cell hyperplasia in fetal gut explants stimulated with anti-CD3 and IFN-α 

To investigate if the phosphorylation of STAT1 and STAT3 by anti-CD3 and IFN-α was dependent on the activation of up-stream kinases, we first analyzed the expression and phosphorylation of Fyn, a Src-related kinase, which associates with both TCR and Tyk-2, a protein involved in the IFN-α signaling 28, 29. Total proteins extracted from fetal gut explants were immunoprecipitated with anti-Fyn and subsequently incubated with a phosphotyrosine antibody. No band corresponding to the phosphorylated form of Fyn (p-Fyn) was seen in unstimulated fetal gut explants (Fig. 6). Addition of IFN-α to the explant cultures of fetal gut induced p-Fyn only after 6 hours of culture (Fig. 6). In contrast, anti-CD3-stimulated fetal gut explants exhibited a phosphorylation of Fyn as early as 4 h and continued to be evident after 6 h of culture (Fig. 6). Similarly, p-Fyn was evident in explants stimulated with anti-CD3 and IFN-α. The intensity of this band was however greater than that in explants stimulated with anti-CD3 or IFN-α alone (Fig. 6). Equal expression of Fyn and uniform loading of the gel was assured by stripping the membrane and re-probing it with an antibody recognizing both the active and inactive forms of Fyn as demonstrated in Fig. 6.

Figure 6.

Active (phosphorylated, p-Fyn) and total Fyn in fetal gut explants cultured for 2, 4 and 6 h in the presence or absence of anti-CD3 antibody and/or IFN-α. Both phosphorylated(p) (upper blot) and total Fyn (lower blot) were analyzed as indicated in Sect. 4. The example is representative of three separate experiments.

Figure 7.

Lack of tyrosine phosphorylated-Fyn (p-FYN) (A), tyrosine and serine phosphorylated-STAT-1 (p-Tyr-STAT1 and p-Ser-STAT1) (B), and tyrosine phosphorylated-STAT3 (p-Tyr-STAT3) (C) signal after treatment with the Src homology protein inhibitor, PP2. Fetal gut explants were cultured with anti-CD3 and IFN-α (100 IU/ml) in the presence or absence of PP2 (10 μM) for 6 h. Two of three representative experiments are shown.

Figure 8.

PP2, a Src homology tyrosine kinase inhibitor, prevents crypt cell hyperplasia induced in fetal gut explants by anti-CD3 and IFN-α. Explants from two different fetuses were cultured with medium alone or soluble anti-CD3 and IFN-α (100 IU/ml) in the presence or absence of PP2 (10 μM) for 4 days. Data are expressed as mean + SEM.

To show that activation of Src-related kinases are involved in the phosphorylation of STAT1 and STAT3 induced by IFN-α plus anti-CD3, we cultured fetal gut explants with anti-CD3 and IFN-α in the presence or absence of PP2, a Src homology protein inhibitor. As expected, PP2 completely abrogated the tyrosine phosphorylation of Fyn (Fig. 7A). In the same explants, PP2 also inhibited both tyrosine and serine phosphorylation of STAT1 (Fig. 7B), as well as the tyrosine phosphorylation of STAT3 (Fig. 7C). In addition, when PP2 was added at the onset of culture with anti-CD3 and IFN-α, it inhibited the crypt cell hyperplasia induced by anti-CD3 and IFN-α (Fig. 8)

3 Discussion

In this study we show that addition of IFN-α and soluble CD3 antibody to explant cultures of fetal gut results in activation of mucosal T cells and in a rapid and profound increase in the rate of crypt epithelial division. The effect of IFN-α was dose dependent and could be neutralized by an IFN-α antibody. No morphological change was observed in fetal gut explants stimulated with IFN-α alone. In addition, the crypt cell hyperplasia induced by anti-CD3 and IFN-α was prevented by cyclosporin A, clearly indicating that T cell activation plays a major role in the enhanced epithelial cell proliferation.

Originally considered as simple antiviral molecule, IFN-α has recently been reconsidered as important immunomodulatory cytokine that is capable of controlling the differentiation of helperT cell toward specific subsets 1, 2. In human but not mouse cells, IFN-α induces IFN-γ production and promotes Th1 cell development 3, 30. Consistent with these reports, we observed that, in the fetal gut, activation of T cells with anti-CD3 and IFN-α elicited IFN-γ expression. This response was dependent on the amount of IFN-α added and was associated with low expression of IL-4 and IL-5 RNA. Taken together, these results confirm and expand on previous data from our own laboratory showing that aconsequence of Th1 cell activation in the fetal gut mucosa is crypt hyperplasia and villous atrophy 1418, 31.

The mechanism by which IFN-α and anti-CD3 cause crypt cell hyperplasia in the fetal gut remains unknown. We previously showed that the enhanced rate of epithelial cell proliferation seen in explant cultures of fetal gut stimulated with SEB was due to the cytokine-stimulated fibroblasts synthesis of KGF 18. Consistent with this, we here report that activation of T cells in the lamina propria of fetal gut with anti-CD3 and IFN-α enhances the expression of KGF RNA. However, we would like to emphasize that the final rate of epithelial cell renewal can be modulated through activation of several pathways 3234.

We also provide evidence that addition of IFN-α to the fetal gut explant cultures stimulated with anti-CD3 does not significantly increase the expression of the extracellular matrix degrading MMP. In this context, the down-stream events associated with the Th1 cell activation induced by anti-CD3 and IFN-α differ from those observed after anti-CD3 and IL-12 17. Indeed, stimulation of fetal gut explants with anti-CD3 and IL-12 elicited a strong production of stromelysin-1, and the addition of a stromelysin-1 inhibitor at the onset of the cultures with anti-CD3 and IL-12 prevented the mucosal degradation 17. In the IL-12-mediated mucosal damage, TNF-α has proven to play a major role, given that TNF-α production correlated with the injury, and a TNFR-IgG fusion protein prevented the synthesis of stromelysin-1 and limited the mucosal damage induced by anti-CD3 and IL-12 17. Since the amount of TNF-αelicited by anti-CD3 and IFN-α is approximately ten times less than that in explants stimulated with anti-CD3 and IL-12, it is tempting to speculate that the documented differences in the induction of MMP rely on the different ability of IL-12 and IFN-α to promote TNF-α synthesis.

Characterization of the signaling pathways employed by IFN-α for the induction of gene expression led to the discovery of a family of transcription factors termed STAT proteins 3537. Upon stimulation, these proteins become activated by tyrosine and serine phosphorylation, multimerize, and translocate to the nucleus, where they bind DNA sequences found in promoters of STAT target genes 20. In fetal gut culture explants, engagement of both the TCR and IFN-α receptor resulted in a strong phosphorylation of both STAT1 and STAT3. This is consistent with the recent demonstration that STAT1 and STAT3 are components in TCR signal transduction, and that TCR signals can synergize with IFN-α to increase STAT-dependent transcriptional response and modulate one or several of the biological outcomes in T cells 2227. Data in our work also indicate that, in fetal gut stimulated withanti-CD3 and IFN-α, Src kinases are required for STAT1 and STAT3 phosphorylation. Indeed, we show that stimulation of T cells with anti-CD3 and IFN-α resulted in enhanced phosphorylation of the Src homology protein, Fyn, and that the Src kinase inhibitor, PP2, inhibited STAT1 and STAT3 phosphorylation, and prevented crypt cell hyperplasia. As PP2 also inhibits activation of Lck, another Src homology protein, we cannot exclude the possibility that Lck is also involved in the TCR/IFN-α signaling pathway. Our data are, however, consistent with recent studies showing that, in vascular smooth muscle cells, phosphorylation of STAT1 by angiotensin II is dependent on Fyn activation 38, and that serine phosphorylation of STAT1 and tyrosine phosphorylation of STAT3 induced in T cells via TCR is inhibited by Src homology protein inhibitors 23, 26. How Src homology proteins work in this pathway is not fully known. Fyn can, however, interact with both TCR and IFN-α-induced signal components. It is thus likely that Fyn serves as a docking site for the latent STAT, which brings TCR and IFN-α signal components in close physical proximity with STAT1 and 3, facilitating their phosphorylation 38.

In the gut there is a balance between the need to recognize pathogenic antigens with the need to prevent abnormal immune responses to dietary and normal flora antigens. The site at which this occurs is in the organized lymphoid tissue of the Peyer's patches 39. We have recently demonstrated that T cells in human Peyer's patches are sensitized to dietary antigens, and that the response is predominantly Th1 in type due to the local production of IL-12 40. Following sensitization in the Peyer's patches, Th1 cells leave via the lymphatics, enter the blood and migrate back to the lamina propria, where they rapidly undergo apoptosis 41. Intestinal homeostasis is also maintained through production of immune-suppressive molecules (e.g. transforming growth factor-β, IL-10, PGE2) and low expression of costimulatory molecules on APC 39. However, some individuals become hyper-reactive to dietary proteins and develop an exaggerated mucosal Th1 response that eventually leads to the intestinal lesion. This occurs in CD, in which the lesion is precipitated by ingestion of dietary gluten 9. The reason why this happens in some but not in other individuals remains unclear.

We have recently demonstrated an overexpression of IFN-α in the mucosa of untreated CD patients 12. In addition, CD has been described in patients receiving IFN-α therapy 12, 13, suggesting a role for this cytokine in driving the immune response in CD. IFN-α has the ability to rescue activated T cells from apoptosis, to maintain sensitized T cells after resolution of the immune challenge and to increase costimulatory molecules on APC, all mechanisms which may eventually lead to chronic inflammation 4, 4244. Consistently, it has been shown that overexpression of IFN-α in specific tissue microenvironments or IFN-α administration can break immune tolerance against self antigens or trigger tissue-damaging Th1 reactions 1, 2, 68, 45. It is thus conceivable that, in the intestinal mucosa, elevate concentrations of IFN-α may in genetically susceptible individuals facilitate the persistence of gluten-reactive Th1 cells and cause disease.

4 Materials and methods

4.1 Organ culture of human fetal gut

Human fetal intestinal tissue was obtained within 2 h of surgical termination from the Homerton Hospital, London, GB. All the specimens were aged between 15 and 16 weeks gestation. Fetal gut explants were cultured in the presence of recombinant human IFN-α2b (final concentration range 10–500 IU/ml) (National Institute for Biological Standards and Control, Potters Bar, GB) and/or soluble CD3 antibody (OKT3 culture supernatant, 5% final dilution) for 4 days as previously described 14, 15. In blocking experiments, neutralizing rabbit anti-human IFN-α polyclonal antibody (1 μg/ml, Immunokontact, Frankfurt, Germany) or control rabbit IgG (1 μg/ml) (Dako, High Wycombe, GB) or PP-2 (10 μM, Calbiochem-Novabiochem, Nottingham, GB) or cyclosporin A (Sandoz) (10 μg/ml final concentration) were added at the onset of culture with IFN-α (final concentration 100 IU/ml) and anti-CD3. At the beginning of culture and at various time points thereafter, the explants were snap frozen in liquid nitrogen, and stored at –80°C. The study received ethical approval from the Hackney and District Health Authority (London, GB).

4.2 Morphological assessment of the fetal gut and immunohistochemistry

Frozen sections (6 μm) were stained with anti-cytokeratin antibody to highlight the epithelium and with Ki67, which recognizes a nuclear antigen in all dividing human cells (both used at 1:50, Dako) by the indirect peroxidase method 14, 15. The proliferation of epithelial cells in the crypts of Lieberkuhn was evaluated as previously reported 14, 15, 19.

4.3 RT-PCR

RNA was extracted from fetal gut explants cultured for 1 and 4 days and used for preparing complementary DNA (cDNA) as previously indicated 17. IFN-γ, TNF-α, IL-4, IL-5, MMP, and TIMP RNA transcripts were determined by quantitative RT-PCR as previously reported 17.

KGF RNA was assessed semiquantitatively by Southern blotting as previously reported 46. In preliminary experiments we established the optimal number of cycles to obtain a PCR product within the linear phase of the amplification. cDNA (1 μl), prepared from 0.5 μg total RNA, was amplified using specific primers for β-actin (for 18, 20, 22 and 25 cycles) or for KGF (for 23, 24, 26 and 28 cycles). The sequence of both β-actin and KGF primers has been published previously 18, 46. For Southern blot experiments, cDNA samples were amplified with β-actin primers for 20 cycles and with KGF primers for 24 cycles.

4.4 Western blot analysis

Total proteins were extracted from fetal gut explants cultured for 2, 4 and 6 h as previously reported 46. Total proteins (300 μg) were separated by SDS/PAGE on an 8% gel and analyzed for p-STAT1 using a mouse anti-human antibody that specifically recognizes STAT1 phosphorylation on tyrosine 701 (1:1,000 final dilution, Santa Cruz Biotechnology, Santa Cruz, CA) or a rabbit anti-human that specifically recognizes STAT1 on serine 727 (1:5,000 final dilution, Upstate Biotechnology, Lake Placid, NY). Rabbit anti-mouse or goat anti-rabbit antibody conjugated to horseradish peroxidase (1:2,500 dilution, Dako) was used as secondary antibody, and the reaction was developed with ECL-plus kit (Amersham Pharmaceuticals, Amersham, GB). After detection of p-STAT1, blots were stripped by incubation for 30 min at 50°C in stripping medium (2% SDS, 0.05 M Tris, pH 6.8, 0.1 mM 2-mercaptoethanol) and subsequently incubated with a rabbit anti-human STAT1 polyclonal antibody (Santa Cruz Biotechnology) followed by a goat anti-rabbit antibody conjugated to horseradish peroxidase (1:2,500 dilution, Dako). For the detection of phosphorylated STAT3 (p-Tyr-STAT3), 250 μg of total proteins were separated by SDS/PAGE on an 8% gel. The membrane was incubated with a mouse anti-human antibody that specifically recognizes STAT3 phosphorylation on tyrosine 705 (1:1,000 final dilution, Santa Cruz Biotechnology) followed by a rabbit anti-mouse antibody conjugated to horseradish peroxidase (1:2,500 dilution, Dako). After detection of p-Tyr-STAT3, blots were stripped andincubated with a mouse anti-human STAT3 antibody (1:5,000 final dilution, Transduction Laboratories, Lexington, GB) followed by a rabbit anti-mouse antibody conjugated to horseradish peroxidase (1:2,500 dilution, Dako). To examine STAT4 and STAT5 phosphorylation, total proteins (500 μg/sample) were incubated with rabbit anti-human STAT4 or mouse anti-human STAT5 (2 μg/ml, Santa Cruz Biotechnology) at 4°C for 2 h. Immune complexes were collected by incubation with protein A/G agarose (Santa Cruz Biotechnology). Immunoprecipitates from extracts containing the same amount of protein were analyzed by Western blotting. An antibody against phosphotyrosine (p-Tyr; 1:1,000 final dilution; Santa Cruz Biotechnology) followed by a horseradish peroxidase-conjugated rabbit anti-mouse IgG mAb (1:10,000 dilution, Dako) was used to detect phospho-STAT4 and phospho-STAT5. After the analysis of phosphorylated STAT4 and STAT5, blots were stripped and incubated with an antibody againstSTAT4 or STAT5 (Santa Cruz Biotechnology) followed by a horseradish peroxidase-conjugated goat anti-rabbit or rabbit anti-mouse IgG antibody (1:2,500 dilution, Dako).

To examine Fyn expression, total protein (500 μg/sample) was incubated with an anti-Fyn antibody (2 μg/ml, Santa Cruz Biotechnology) at 4°C for 2 h and immune complexes analyzed by Western blotting. An Ab against phosphotyrosine (1:1,000 final dilution, Santa Cruz Biotechnology) followed by a horseradish peroxidase-conjugated goat anti-mouse monoclonal antibody (1:10,000 dilution, Dako) were used. After the analysis of phosphorylated Fyn, blots were stripped and incubated with a mouse anti-human Fyn antibody (1:300 dilution, Santa Cruz Biotechnology) followed by a rabbit anti-mouse antibody conjugated to horseradish peroxidase (1:2,500 dilution, Dako).

4.5 Statistical analysis

Differences between groups were compared using either the Mann-Whitney U test, if the data were not normally distributed, or the Student's t-test, if the observations were consistent with a sample from a normally distributed population.

Acknowledgements

This work was supported by the European Union grant (ERBFMRXCT9).

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

  7. 7

    WILEY-VCH

  8. 8

    WILEY-VCH

Ancillary