Frontline: Interferon regulatory factor-1 as a protective gene in intestinal inflammation: role of TCR γ δ T cells and interleukin-18-binding protein

Abstract

The transcription factor IFN regulatory factor-1 (IRF-1) regulates production and activity of many inflammatory mediators and cells. Here, we investigated the role of IRF-1 in intestinal inflammation using clinical and histologic scores; inflammatory mediators were also measured in colonic tissue. Dextran sulfate sodium (DSS) or trinitrobenzene sulfonic acid (TNBS) was administered to wild-type (WT) or IRF-1 knockout (KO) mice. DSS or TNBS led to a dramatic increase in lethality and colitis severity in IRF-1 KO compared with WT mice. Reduced levels of IFN-γ and IL-18-binding protein (IL-18BP) were observed in the colon of IRF-1 KO mice, whereas levels of inducible nitric oxide synthase, cyclooxygenase-2, phosphorylated STAT-3, chemokines, TNF-α, IL-1β, IL-15, and IL-18 were not significantly changed. Intestinal inflammation was not altered in IFN-γ KO mice or in WT mice given neutralizing anti-IFN-γ antibodies, but was increased in mice lacking TCR γ δ lymphocytes, a population significantly decreased in the intestine of IRF-1-deficient mice. Administration of IL-18BP reversed the increased susceptibility of IRF-1 KO mice to DSS. These results suggest a protective role for IRF-1 in intestinal inflammation, with a possible anti-inflammatory and/or restorative role. IL-18BP and TCR γ δ cells appear to be critical factors inthe anti-inflammatory effects of IRF-1.

See accompanying article http://dx.doi.org/10.1002/eji.200425351

Abbreviations:

COX-2:

Cyclooxygenase-2

DSS:

Dextran sulfate sodium

IL-18BP:

IL-18-binding protein

iNOS:

Inducible nitric oxide synthase

IRF-1:

IFN regulatory factor-1

KO:

Knockout

TNBS:

Trinitrobenzene sulfonic acid

WT:

Wild-type

1 Introduction

Interferon regulatory factor-1 (IRF-1) is a transcription factor originally identified as a regulator of IFN-β gene expression 1. It has since been shown that IRF-1 gene expression is induced by IFN-α/β and IFN-γ and by other cytokines in a positive feedback loop that amplifies IFN effects 2–4.

A Th1-polarizing property of IRF-1 is suggested by the prominent role of this transcription factor in autoimmune diseases 5. Consistently, mice lacking IRF-1 show a reduced incidence and severity of type II collagen-induced arthritis and experimental allergic encephalomyelitis 6. The Th1-inducing ability of IRF-1 is also of particular interest in the context of intestinal inflammation, particularly Crohn's disease, which is considered a Th1-polarized condition 7, 8. Furthermore, altered expression of IRF-1 has been reported in the nasal and intestinal epithelia of mice with cystic fibrosis, suggesting that a dysregulation of this transcription factor contributes to the chronic inflammation typical of this condition 9.

Beside its regulatory function on Th1 responses, under certain conditions IRF-1 induces the transcription of several gene products known to contribute to experimental colitis, such as caspase-1, inducible nitric oxide synthase (iNOS), and cyclooxygenase-2 (COX-2). IRF-1 is also required for the normal development of cell populations such as TCR γ δ, NK, and CD8⁺ cells 10–12. Moreover, IRF-1 directly or indirectly regulates production of a variety of cytokines, including IFN-γ, IL-1β, IL-12 p40, IL-15, IL-18 and the natural IL-18 inhibitor, IL-18-binding protein (IL-18BP) 5, 13–16. Despite these extensive actions of IRF-1 on diverse aspects of the inflammatory response, the involvement of this transcription factor in intestinal inflammation has not been studied to date.

Intestinal inflammation induced by oral administration of dextran sulfate sodium (DSS) to mice is an established model of acute or chronic colitis in which an initial epithelial cell damage is followed by development of colitis and, eventually, by a relatively slow mucosal repair process 17. Hence, this model can reveal the role played by different factors during colonic healing, a process that may be altered in patients with inflammatory bowel disease. Rectal administration of trinitrobenzene sulfonic acid (TNBS) in ethanol is another model of experimental colitis, in which initial disruption of the epithelial barrier leads to activation of intestinal immune cells 18. On the basis of the established functions of IRF-1 stated above, we expected to observe a reduced severity and a decreased repair of colitis in mice deficient in IRF-1. To our surprise, we observed a dramatically increased susceptibility of IRF-1 knockout (KO) mice to DSS or TNBS. The mechanisms underlying this increased susceptibility were investigated by the assessment of different mediators in the colon of wild-type (WT) and IRF-1 KO mice and by evaluating the role played by IFN-γ, IL-18BP, and TCRγ δ cells.

2 Results

2.1 Induction of IRF-1 in the colon of DSS-exposed mice

To assess whether acute DSS-induced colitis is associated with an induction of IRF-1, the colon of control and DSS-treated WT mice was examined by Western blot analysis. IRF-1 was significantly up-regulated in DSS-exposed WT compared with control mice at day 5, whereas no IRF-1 was detected in DSS-fed IRF-1 KO mice (Fig. 1).

Figure 1. Induction of IRF-1 after exposure to DSS. WT and IRF-1 KO mice were exposed to 3.5% DSS or regular drinking water for 5 days and then killed. IRF-1 induction was evaluated in colon homogenates by Western blot analysis; each lane represents an individual mouse.

2.2 Increased susceptibility of IRF-1 KO mice to DSS

There are two models of DSS-induced colitis — acute and chronic. To evaluate the response of IRF-1 KO mice to DSS, chronic colitis was induced by administering 2.5% DSS for 5 days followed by 5 days of regular drinking water and repeating this cycle three times, resulting in a 30-day experimental period. This same protocol has previously been used to effectively induce chronic colitis in C57BL6 mice, the background strain of the IRF-1 KO mice used in this study 19. In contrast to WT mice, IRF-1 KO mice failed to recover after the first DSS cycle and their condition progressively deteriorated. Consequently, the second DSS cycle was not initiated. Instead, the mice were observed for a total of 20 days (5 days of DSS followed by 15 days of regular drinking water). Whereas all WT mice survived this modified treatment, an unexpected 80% lethality rate was observed in IRF-1 KO mice (Fig. 2).

Given the increased susceptibility of IRF-1 KO mice to DSS, the model of acute, rather than chronic, colitis — as described in Section 4.2 — was employed for further evaluation. Starting from day 5 of DSS administration, IRF-1 KO mice exhibited a significantly higher disease score compared with WT mice (Fig. 3A). In IRF-1 KO mice, lethality was observed in three out of five separate experiments of acute DSS performed (a total of 27 mice were examined), yielding an average lethality of 32%. In sharp contrast, no lethality was observed in WT mice. As described previously, colon length correlates with DSS-induced colitis severity 17. Consistent with the clinical data described above, a 35.6% reduction in colon length was observed in DSS-exposed IRF-1 KO mice, compared with a 19.0% reduction in WT mice (Fig. 3B). These effects became even more apparent when the histologic score of colon sections was determined. Although DSS-fed WT mice had a score of 3.4±0.5, DSS-exposed IRF-1 KO mice presented with a score of 5.9±0.1 (Fig. 3C); in this model, 6 is the maximum score. Representative staining of colon sections from DSS-treated WT and IRF-1 KO mice is shown in Fig. 3D.

Figure 2. Lethality in IRF-1 KO mice receiving DSS. WT and IRF-1 KO mice were exposed to 2.5% DSS for 5 days followed by 15 days of regular drinking water as part of the protocol for inducing chronic colitis. The survival rate in both groups is shown. There were n=10 mice per group.

Figure 3. Increased susceptibility to DSS in IRF-1 KO mice. WT and IRF-1 KO mice were exposed to 3.5% DSS for a total of 5 days followed by 5 days of regular drinking water. (A) Disease score. (B) Colon length. At day 10, the colon length of DSS-exposed WT and IRF-1 KO mice and controls was measured. (C) Histologic score. At day 10, segments of the transverse colon were stained with hematoxylin and eosin, and the histologic score was determined. (D) Representative histology. In this figure, data are mean±SEM, n =10 mice per group for panel A; 5 WT and 5 IRF-1 KO for panels B and C. ***p<0.001 vs. DSS WT.

2.3 Increased susceptibility of IRF-1 KO mice to TNBS

To verify whether increased susceptibility to colitis in IRF-1 KO mice was also observed in other experimental models, the response of WT and IRF-1 KO mice to TNBS was evaluated. As shown in Fig. 4A, similar to the results obtained using DSS, a significantly higher lethality was observed in IRF-1 KO compared with WT mice receiving TNBS. The increased susceptibility of IRF-1 KO mice to TNBS was confirmed by histologic analysis of colonic tissue obtained from surviving mice (Fig. 4B).

Figure 4. Increased susceptibility of IRF-1 KO mice to TNBS-induced colitis. WT and IRF-1 KO mice were treated with TNBS or vehicle. (A) Lethality; (B) histologic score. Segments of the ascendant colon were stained with hematoxylin and eosin, and the histologic score was determined. Data are mean±SEM, n=10 mice per group. **p<0.01 vs. TNBS WT.

2.4 IFN-γ is not responsible for the increased susceptibility of IRF-1 KO mice to DSS

As expected, colon culture supernatants of IRF-1 KO mice receiving DSS contained significantly reduced levels of IFN-γ compared with WT colons (54.3±16.6 vs. 188.2±37.2 pg/mg protein in IRF-1 KO vs. WT, respectively; p<0.01). IRF-1 deficiency also leads to reduced cellular responses to IFN-γ itself 20. Therefore, the role of IFN-γ in DSS-induced colitis was investigated in IFN-γ KO mice. In agreement with previous reports, IFN-γ deficiency did not significantly alter the response to DSS 21, 22, whereas IRF-1 KO mice developed significantly more severe disease compared with either WT or IFN-γ KO mice (Fig. 5). Similar results were obtained in WT mice treated with a neutralizing anti-IFN-γ antibody (data not shown).

Figure 5. Lack of a role for IFN-γ in DSS-induced colitis. WT, IRF-1 KO and IFN-γ KO mice were exposed to 3.5% DSS for a total of 5 days followed by 5 days of regular drinking water. Data are mean±SEM, n=10 mice per group. ***p<0.01 vs. WT or IFN-γ KO

2.5 Role of TCRγ δ T cells in DSS-induced colitis

We confirmed the observation of Ohteki et al. 10 by demonstrating that the percentage of TCRγ δ⁺ intraepithelial lymphocytes was reduced in the intestine of IRF-1 KO compared with WT mice (14.0% vs. 34.7% in IRF-1 KO vs. WT mice, respectively). A direct comparison of the susceptibility to DSS of IRF-1 KO vs. γ δ KO mice was performed. As shown in Fig. 6 and in agreement with previous reports 23, 24, γ δ KO mice developed significantly more disease compared with WT mice when treated with DSS. Again, DSS-induced colitis was significantly more severe in IRF-1 KO than in γ δ KO mice.

Figure 6. Role of TCR γ δ cells in DSS-induced colitis. WT, IRF-1 KO and γ δ KO mice were exposed to 3.5% DSS for a total of 5 days followed by 5 days of regular drinking water. Data are mean±SEM, n=10 mice per group. ap<0.05; aap<0.001 vs. WT mice; bp<0.001 vs. γ δ KO mice.

2.6 Reduced IL-18BP in IRF-1 KO and the increased susceptibility to DSS

The IL-18BP promoter contains an IRF-1 site 14; IRF-1 KO mice have undetectable serum IL-18BP levels and strongly reduced IL-18BP mRNA in the liver and spleen 13, 14. Using RT-PCR, we confirmed reduced expression of IL-18BP in the colon of IRF-1 KO compared with WT mice (Fig. 7).

To study whether lack of IL-18BP could be responsible for the increased susceptibility of IRF-1 KO mice to DSS, WT and IRF-1 KO mice received DSS with or without concomitant administration of recombinant IL-18BP (0.25 mg/kg, twice daily). As shown in Fig. 8 and in agreement with previous data 25, 26, neutralization of IL-18 in WT mice significantly reduced disease severity throughout the treatment course. Furthermore, in IRF-1 KO mice, administration of IL-18BP was effective at late time points, when it restored susceptibility to a level comparable to that observed in WT mice. Disease score data were confirmed by the observed shortening of the colon, which was significantly less severe in IL-18BP-treated compared with vehicle-treated IRF-1 KO mice (7.17±0.11 vs. 6.14±0.21 cm in IL-18BP-treated vs. vehicle-treated IRF-1 KO mice, respectively; p<0.001).

Figure 7. Expression of IL-18BP and IL-15 in the colon of IRF-1 KO mice. Total RNA was extracted from the colon of control and DSS-treated (day 10) WT and IRF-1 KO mice, and RT-PCR was performed; each lane represents an individual mouse.

Figure 8. Effect of IL-18BP administration on DSS-induced colitis in WT and IRF-1 KO mice. The WT and IRF-1 KO mice were exposed to 3.5% DSS for a total of 5 days followed by 7 days of regular drinking water. Mice received i.p. injections of either human recombinant IL-18BP (0.25 mg/kg) or vehicle twice daily starting on day 1. Data are mean±SEM, n=7 mice per group. ap<0.01 IRF-1 KO vs. WT; bp<0.001 IRF-1 KO vs. IRF-1 KO + IL-18BP; cp<0.001 WT vs. WT + IL-18BP.

2.7 Expression of other inflammatory mediators in the colon of IRF-1 KO mice

Expression of IL-15 in bone marrow-derived cells is transcriptionally regulated by IRF-1 16, and IL-15 KO mice show increased susceptibility to DSS-induced intestinal inflammation 27. Therefore, we compared expression of IL-15 in the colon of IRF-1 and WT mice. As shown in Fig. 7, basal and DSS-induced steady-state mRNA levels of IL-15 did not differ between WT and IRF-1 KO mice. These data were confirmed by immunohistochemistry of colonic sections, which indicated comparable expression of IL-15 in epithelial cells of WT and IRF-1 KO mice (data not shown).

There were no significant differences between IRF-1 KO and WT mice in the production of IL-6, MIP-1α, MIP-2, IL-10, IL-1β, and IL-18 from colon cultures (data not shown). Levels of TNF-α and IL-4 were below the detection limit in cultures of both WT and IRF-1 KO mice. Expression of COX-2 and iNOS — two IRF-1-regulated genes — as well as STAT-3 phosphorylation were also evaluated. However, both COX-2 and iNOS were comparably up-regulated by DSS in both WT and IRF-1 KO mice. Furthermore, STAT-3 was phosphorylated to a similar extent in the colon of DSS-treated WT and IRF-1 KO mice (Fig. 9).

Figure 9. COX-2, iNOS and STAT-3 in WT and IRF-1 KO mice receiving DSS. (A) COX-2 and iNOS expression; (B) Total and phosphorylated STAT-3. Colon homogenates were tested from DSS-exposed WT and IRF-1 KO as well as control mice; each lane represents an individual mouse.

3 Discussion

In the present study, we identify IRF-1 as a critically protective gene in DSS- or TNBS-induced colitis. Because IRF-1 KO mice have a reduced incidence and severity of antigen-induced autoimmune diseases and are resistant in models of acute inflammation 6, 28, our initial hypothesis was that a lack of IRF-1 should also be associated with protection from DSS- or TNBS-induced colitis. However, lack of IRF-1 markedly worsened disease severity on both colitides.

A number of IRF-1-regulated genes are involved in the inflammatory process of colitis. In the DSS model, iNOS KO mice and mice treated with an NO inhibitor show a delayed onset and reduced severity of colitis 29. However, in the present study we found that colonic expression of iNOS as well as colonic levels of NO (data not shown) were equally up-regulated by DSS in WT and IRF-1 KO mice, suggesting that NO per se does not account for the differences in colitis severity observed in these mice. COX-2 plays a pathological role in protecting the intestinal mucosa, since COX-2 KO mice as well as mice treated with a selective COX-2 inhibitor are highly susceptible to DSS 30. COX-2 can be induced in an IRF-1-dependent manner 31. However, we found COX-2 to be similarly induced by DSS in the colon of WT and IRF-1 KO mice, suggesting that COX-2 can be induced independently from IRF-1 and that a lack of COX-2 induction is not responsible for the increased susceptibility of IRF-1 KO mice to DSS. A phenotype similar to the one observed in IRF-1 KO mice was reported in mutant mice with gp130 defective STAT-3 signaling 21. However, similar levels of phosphorylated STAT-3 were present in the colon of DSS-treated WT and IRF-1 KO mice, thus likely excluding a role for STAT-3 in the observed effects.

Bone marrow cell-derived IL-15 production is IRF-1-dependent 16 and IL-15 KO mice show increased susceptibility to DSS 27. However, we demonstrated comparable IL-15 gene expression in the colon of WT and IRF-1 KO mice; colonic IL-15 was expressed by epithelial cells. This is in agreement with reports indicating that non-immune-derived IL-15 is IRF-1-independent 32 and that intestinal epithelial cells constitutively express IL-15 33. Our data suggest that lack of IL-15 is not responsible for the increased susceptibility of IRF-1 KO mice to intestinal inflammation, although the possibility remains that immune cell-derived IL-15 vs. epithelial cell-derived IL-15 might play differential roles in the regulation of inflammation. Therefore, we cannot exclude that the lack of IL-15 expressionin bone marrow-derived cells of IRF-1 KO mice could contribute — either directly or indirectly — to their increased susceptibility to DSS and TNBS.

A possible role for IFN-γ as a protective factor in DSS-induced colitis was investigated in the present study. WT mice showed a significant increase in IFN-γ production after DSS exposure whereas — as expected — IRF-1 KO mice produce almost no IFN-γ in colon cultures. Furthermore, IRF-1 deficiency renders cells unresponsive to IFN-γ itself 20. The role of IFN-γ in experimental models of colitis is quite controversial. Although inhibition of IFN-γ is protective in models of T cell transfer into T cell-deficient mice 34, blockade of IFN-γ is not protective in the models of TNBS-induced colitis as well as in TCRα KO or IL-10 KO mice 35–37. In the present study, we confirm that IFN-γ KO mice and mice treated with neutralizing anti-IFN-γ antibodies are equally susceptible to DSS, thus excluding decreased IFN-γ as accounting for the increased susceptibility ofIRF-1 KO mice to intestinal inflammation.

TCRγ δ T cells, a major subpopulation among intraepithelial lymphocytes, are markedly reduced in IRF-1 KO mice 10. Several studies indicate a protective function of TCRγ δ T cells in the models of TNBS- or DSS-induced colitis 23, 24, 38 We confirmed these observation by evaluating the response to DSS of γ δ KO mice. However, when studied in parallel, IRF-1 KO mice developed significantly more severe disease compared with γ δ KO mice. We conclude that TCRγ δ cells could contribute to the increased susceptibility of IRF-1 KO mice to DSS, but are unlikely to be the sole mechanism.

Both IL-18 and its natural inhibitor, IL-18BP, are expressed in colonic tissue in health and are up-regulated in the mucosa of Crohn's disease patients 39, 40. Neutralization of IL-18 is protective in various models of colitis 25, 26, 41. The promoter of IL-18BP contains an IRF-1 site 14, and IRF-1 KO mice produce extremely low levels of IL-18BP 13, 14. Since colonic tissues from DSS-treated IRF-1 KO mice produce the expected levels ofIL-18, a lack of IL-18BP would allow IL-18 to exert its pro-inflammatory activities without opposition by its natural inhibitor. We observed that IL-18BP expression is reduced in the colon of IRF-1KO compared with WT mice. Importantly, IRF-1 KO mice treated with IL-18BP during DSS-induced colitis exhibited a level of disease severity comparable to that in WT mice. Thus, administration of IL-18BP reverses the increase in disease severity of IRF-1 KO mice. Moreover, IL-18BP reduced the disease in WT mice with DSS-induced colitis, an expected response similar to other strategies to neutralize IL-18.

In conclusion, these results indicate that IRF-1 plays a protective role in models of intestinal inflammation. Reduced TCRγ δ T cells and IL-18BP appear to be two of the important conditions that account for the increase in inflammatory effects of systemic IRF-1 deficiency. It remains to be evaluated whether tissue-specific, direct effects of IRF-1 could also contribute to the modulation of intestinal inflammation.

4 Materials and methods

4.1 Mice

Animal protocols were approved by the animal studies committee of the University of Colorado Health Sciences Center. Male or female IRF-1 KO mice (6–8 weeks old) were backcrossed for seven generations to C57BL/6 mice 42. C57BL/6 (WT), C57BL/6-Ifnγ^tm1Ts (IFN-γ KO), and C57BL/6-Tcrd^tm1Mom (TCRδ KO) mice were obtained from The Jackson Laboratories (Bar Harbor, ME, USA). The animals were housed at a controlled temperature, with light-dark cycles, were fed standard mice chow pellets, had access to water, and were acclimatized before being entered into the experiments.

4.2 Induction of acute DSS-induced colitis, and treatment with anti-IFN-γ or with IL-18BP

Mice were fed 3.5% DSS (molecular weight 40 kDa; ICN, Aurora, OH, USA) dissolved in sterile, distilled water ad libitum from day 1 to day 5, followed by 5 days of regular drinking water: this regimen results in a 10-day experimental period 19. For neutralization of IFN-γ activity, mice received 200 μg of a neutralizing rat monoclonal anti-mouse-IFN-γ antibody (clone RMMG-1, from PBL, New Brunswick, NJ, USA) on days 1, 3, 5 and 7. Human recombinant IL-18BP was a kind gift of Serono International S.A. (Geneva, Switzerland) and was administered i.p. at 0.25 mg/kg twice a day, beginning on the first day of DSS administration and continuing throughout the whole length of the experiment. This dose of IL-18BP was chosen based on its ability to significantly inhibit LPS-induced IFN-γ production in WT mice.

4.3 Induction of TNBS-induced colitis

Mice received two intrarectal administrations of TNBS (2 mg/mouse) in 50% ethanol on days 1 and 7 as previously described 19. Control mice received an equal amount of vehicle.

4.4 Clinical and histological assessment of colitis

Mice were weighed daily and monitored for the appearance of diarrhea and blood in the stools. A disease scoring system was used. Weights were ranked by assigning points as follows: 0 = 0–5% weight loss; 1 = 6–10% weight loss; 2 = 11–15% weight loss; 3 = 16–20% weight loss; 4 = >20% weight loss. Stool quality was scored as follows: 0 = well-formed pellets; 2 = pasty and semiformed stools that did not adhere to the anus; 4 = liquid stools that did adhere to the anus. Appearance of blood in the stools was scored as follows: 0 = no blood using hemoccult (Beckman Coulter, Palo Alto, CA, USA); 2 = positive hemoccult; 4 = gross bleeding. The total scores given for wasting, stool quality and stool blood were added and divided by 3, for a maximal disease score of 4. Post mortem the entire colon was excised and a 1-cm segment of the transverse colon was fixed in 10% buffered formalin for histologic analysis. Paraffin sections were stained with hematoxylin and eosin. Four to six colon rings were obtained from each 1-cm colon segment and were thus available for histologic examination. Histologic scoring was performed in a blinded fashion by a pathologist (H. A. L.) as a combined score of inflammatory cell infiltration (0–3) and tissue damage (0–3) as described previously 19.

4.5 Colon organ culture

The culture of whole colon segments was performed as described previously 19.

4.6 Cytokine measurements

Murine IL-18, MIP-1α, MIP-2, IL-1β, IL-10 and TNF-α levels were measured using an electrochemiluminescence (ECL) method as described previously 43, 44. The range of quantification is 20 pg to 10 ng/ml for TNF-α, IL-18, IL-1β, IL-10 or MIP-1α, and 20 pg to 2 ng/ml for MIP-2. IFN-γ, IL-4, and IL-6 were measured using a specific ELISA (BD Pharmingen, San Diego, CA, USA).

4.7 Western blot

Segments (2 cm) of colon adjacent to the section taken for histology were homogenized in radioimmunoprecipitation assay buffer and Western blot analysis was performed. Antibodies included a rabbit anti-COX-2 (Cayman Chemical Company, Ann Arbor, MI, USA), as well as a rabbit anti-iNOS, a rabbit anti-IRF-1, a rabbit anti-total-STAT-3 and a rabbit anti-phosphorylated-STAT-3 (serine and tyrosine) (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

4.8 RT-PCR

Total RNA was extracted from the colon of WT and IRF-1 KO mice. RT-PCR was performed using the following primers: IL-15 forward 5′-CATATGGAATCCAACTGGATAGATTAAGATA-3′, IL-15 reverse 5′-CATATGCTCGAGGGACGTGTTGATGAACAT-3′; IL-18BP forward 5′-ACATCTGCACCTCAGACAACT-3′, IL-18BP reverse 5′-TGGGAGGTGCTCAATGAAGGAACCA-3′; GAPDH forward 5′-ACCACAGTCCATGCCATCAC-3′, GAPDH reverse 5′-TCCACCACCCTGTTGCTGGTA-3′.

4.9 Statistical analysis

Data are expressed as mean ± SEM. The statistical significance of differences between treatment and control groups was determined by factorial ANOVA. Differences in survival were analyzed using a χ² test. Statistical analyses were performed using the XLStat software (Addinsoft, Brooklyn, NY, USA).

Acknowledgements

We wish to thank Antonietta Rastiello and Claudia Braun for excellent technical assistance. This work was supported by grants from the Cystic Fibrosis Foundation, The Broad Medical Research Program of the Eli and Edythe L. Broad Foundation, the Crohn's and Colitis Foundation of America, National Institutes of Health grants DK061483 (to G. F.), AI15614 and HL68743 (to C. A. D.), by the Emmy-Noether program (DFG SI 749/3-1) of the Deutsche Forschungsgemeinschaft and by DFG SI 749/2-1 (to B. S.).

• 1

  WILEY-VCH

• 2

  WILEY-VCH

• 3

  WILEY-VCH

• 4

  WILEY-VCH

• 5

  WILEY-VCH

• 6

  WILEY-VCH

• 7

  WILEY-VCH

• 8

  WILEY-VCH

• 9

  WILEY-VCH