Bacterial exotoxins downregulate cathelicidin (hCAP-18/LL-37) and human β-defensin 1 (HBD-1) expression in the intestinal epithelial cells

Authors


*E-mail dasss@icmr.org.in; santasabujdas@yahoo.com; Tel. (+11) 3323 633 851; Fax (+11) 3323 705 066.

Summary

Cathelicidin (hCAP-18/LL-37) and β-defensin 1 (HBD-1) are human antimicrobial peptides (AMPs) with high basal expression levels, which form the first line of host defence against infections over the epithelial surfaces. The antimicrobial functions owe to their direct microbicidal effects as well as the immunomodulatory role. Pathogenic microorganisms have developed multiple modalities including transcriptional repression to combat this arm of the host immune response. The precise mechanisms and the pathogen-derived molecules responsible for transcriptional downregulation remain unknown. Here, we have shown that enteric pathogens suppress LL-37 and HBD-1 expression in the intestinal epithelial cells (IECs) with Vibrio cholerae and enterotoxigenic Escherichia coli (ETEC) exerting the most dramatic effects. Cholera toxin (CT) and labile toxin (LT), the major virulence proteins of V. cholerae and ETEC, respectively, are predominantly responsible for these effects, both in vitro and in vivo. CT transcriptionally downregulates the AMPs by activating several intracellular signalling pathways involving protein kinase A (PKA), ERK MAPKinase and Cox-2 downstream of cAMP accumulation and inducible cAMP early repressor (ICER) may mediate this role of CT, at least in part. This is the first report to show transcriptional repression of the AMPs through the activation of cellular signal transduction pathways by well-known virulence proteins of pathogenic microorganisms.

Introduction

LL-37 and HBD-1 are members of two major families of small cationic peptides, cathelicidins and defensins, respectively. These peptide families have potent antimicrobial activities and form the first line of host defence against pathogenic microorganisms. Their antimicrobial function is demonstrated by in vitro susceptibility of a broad range of bacteria, viruses, fungi and protozoa to concentrations of antimicrobial peptides (AMPs) that are achieved at the mucosal surfaces (Islam et al., 2001; Bergman et al., 2005; Selsted and Ouellette, 2005). LL-37 also has the property to directly bind and neutralize LPS (Larrick et al., 1994). In addition, the AMPs exert multiple immunomodulatory roles (Bals and Wilson, 2003; Dommett et al., 2005). Many researchers believe that the in vivo antimicrobial activities of AMPs, as evidenced by increased susceptibility of the AMP knock-out animals (Nizet et al., 2001; Moser et al., 2002; Morrison et al., 2002; Rosenberger et al., 2004; Chromek et al., 2006) and resistance of the transgenics (Bals et al., 1999) to various infections, are largely due to the direct microbicidal effects. However, others consider the immunomodulatory functions to play the predominant role in vivo (Bowdish et al., 2005).

Despite the wealth of information about the critical role the AMPs play in the immune responses of the skin and the mucosal surfaces, relatively little is known about their regulation under the physiological and pathological conditions. Accumulating evidence suggests that the regulatory mechanisms may be stimulus and tissue specific and operate mainly at the level of mRNA expression (Dommett et al., 2005; Selsted and Ouellette, 2005). Stimulation with several pathogen-associated molecular patterns (PAMPs) like LPS and peptidoglycan that induce various β-defensins illustrates differential regulation of the AMPs (Sorensen et al., 2005). In the human gastrointestinal tract, LL-37 and HBD-1 are ubiquitously expressed (Bals et al., 1998) and their levels may be further modified by multiple stimuli.

Knowledge about the regulation of LL-37 expression remains grossly inadequate. In the intestine, undifferentiated epithelial cells express low levels of LL-37, but cellular differentiation induced by short-chain fatty acids (SCFAs), especially butyrate, significantly upregulates it and has been proposed to be a key regulator of LL-37 expression (Hase et al., 2002; Schauber et al., 2003). However, the precise mechanism behind this regulation remains unclear. Despite both in vitro and in vivo evidence of bacterial infections leading to changes in LL-37 expression in different epithelial cells (Islam et al., 2001; Hase et al., 2002; Schaller-Bals et al., 2002), no pathogen-derived molecule has so far been established to regulate LL-37. It has also been shown to be elevated in allergic and inflammatory disorders of the skin (Frohm et al., 1997), nasal mucosa (Kim et al., 2003) and lung (Beisswenger and Bals, 2005). In addition, LL-37 level was increased in the cultured composite keratinocyte graft in response to IL-1 and IL-6 (Erdag and Morgan, 2002) and the gene encoding LL-37 contains potential binding sites for NF-IL6 (Frohm Nilsson et al., 1999). On the other hand, IL-10 may be responsible for suppression of LL-37 and HBD-2 expression in atopic dermatitis skin (Howell et al., 2005). However, LPS, PMA and pro-inflammatory cytokines (TNF-α, IL-1α, IL-6 and IFN-γ) had little effect on LL-37 expression in the colonic epithelial cells (Hase et al., 2002), suggesting stimulus- and tissue-specific regulation.

Regulation of HBD-1 expression has also remained largely unexplored. Undifferentiated intestinal epithelial cells (IECs) express a high basal level of HBD-1. However, it has never been experimentally demonstrated whether differentiation of the IECs may further modulate its expression. Ca2+-induced differentiation of primary keratinocytes has been reported to upregulate HBD-1 (Harder et al., 2004), while apoptosis signal-regulating kinase-1 (ASK1), an intracellular regulator of keratinocyte differentiation, enhances HBD1–3 and LL-37 gene expression through the activation of p38 MAPK pathway (Sayama et al., 2005). Among the pro-inflammatory cytokines, only IFN-γ was found to augment HBD-1 expression (Sorensen et al., 2005).

An intriguing hypothesis was offered by Islam et al. (2001) who reported marked suppression of LL-37 and HBD-1 expression in the intestinal biopsy-derived epithelial cells of patients suffering from shigellosis as well as acute watery diarrhoea. They further showed that in vitro infection of human colon-derived epithelial cell line HT-29 with Shigella dysenteriae type I significantly downregulates LL-37 mRNA expression and proposed that downregulation of the AMP message may be one of the major immune evasion mechanisms for the mucosal pathogens. Although subsequent investigations carried out by a different group failed to confirm their finding (Hase et al., 2002), this hypothesis was supported by studies showing that pathogenic Neisseria spp. may suppress LL-37 expression in the genitourinary epithelial cells (Bergman et al., 2005) and HBD-1 may be downregulated in the HT-29 cells following Cryptosporidium parvum infection (Zaalouk et al., 2004). A recent report has shown that induction of LL-37 in the gut by SCFAs significantly decreases the severity, duration and bacterial load of experimental shigellosis in rabbits, further strengthening the hypothesis (Raqib et al., 2006). However, the pathogen or the host-derived molecules and the signalling pathways that may regulate LL-37 and HBD-1 expression remain largely obscure.

In the present study, we reported marked upregulation of LL-37, but not of HBD-1 expression in the IECs upon differentiation with sodium butyrate (NaB). However, both the molecules were significantly downregulated by several enteric pathogens, most notably Vibrio cholerae O139. Cholera toxin (CT) was found to be the predominant molecule associated with V. cholerae spp. regulating the AMPs, both in vitro and in vivo. Labile toxin (LT) of enterotoxigenic Escherichia coli (ETEC) exerted a similar effect as CT in the human intestine-derived cell lines as well as in animal studies. We have also shown that multiple signalling pathways, like the protein kinase A (PKA), ERK MAPK and the arachidonic acid (AA) synthesis pathways, activated downstream of intracellular accumulation of cAMP, contribute to CT-mediated suppression of LL-37 and HBD-1 expression in the IECs. Finally, CT-induced pathways function at the level of transcription to downregulate AMP expression and their effects are mediated, at least in part, through the induction of inducible cAMP early repressor (ICER).

Results

Enterocyte differentiation upregulates LL-37 but not human β-defensin expression

We revisited the issue of AMP expression upon NaB-induced differentiation of the IECs. The results showed progressive rise in LL-37 mRNA and protein expression with increasing differentiation of the HT-29 cells, while the expression of the β-defensins remained unchanged (Fig. S1).

Enteric pathogens differentially regulate LL-37 and HBD-1 expression

Controversy exists regarding the influence of enteric bacterial infections on LL-37 expression (Islam et al., 2001; Hase et al., 2002), while HBD-1 regulation in the IECs by bacterial pathogens has not been studied. To investigate the role of enteropathogens in the regulation of AMP expression, we co-cultured differentiated HT-29 cells and several enteric bacteria, both invasive and non-invasive, for 1 h followed by culturing the cells in fresh medium in the absence of extracellular bacteria for an additional 6 h. Greater than 90% of the cells were viable at the end of the experiment, as assayed by trypan blue dye exclusion method. The results showed marked downregulation of LL-37 and HBD-1 mRNA expression by V. cholerae O139 (SG24) and ETEC (4660) and less by S. dysenteriae type I (BCH518), while non-pathogenic Vibrio parahemolyticus (environmental isolate) and E. coli BL21(DE3) strains had no effect. On the other hand, Salmonella typhimurium LT2 only modestly increased LL-37 level. That the decrease in the AMP levels was not due to global repression of gene expression or decreased viability of the cells was proved by the fact that IL-8 expression was significantly upregulated in the same cells infected with the bacterial strains (Fig. 1A). Changes in mRNA expression levels were accompanied by similar changes in LL-37 protein expression (Fig. 1B). As non-invasive pathogens like O139 and ETEC exerted significant influence on AMP expression, we sought to investigate if pathogen-derived secreted molecule(s) may be involved in this regulation. To this end, HT-29 cells were stimulated with bacteria-free culture supernatants of the above pathogens for 6 h. The effects were similar as the live bacteria for O139 and ETEC, suggesting that their secreted products may be involved in AMP regulation (Fig. 1C). That downregulation of LL-37 and HBD-1 expression by the enteric pathogens is functionally relevant was proved by diminished bacterial killing by the culture supernatants of the cells treated with the bacteria that suppressed AMP expression (Fig. 1D). The above results suggest that both secreted and non-secreted products of enteric pathogens regulate AMP expression in the IECs.

Figure 1.

Enteric pathogens, but not the non-pathogenic bacteria regulate LL-37 and HBD-1 expression. NaB (4 mM, 48 h)-differentiated HT-29 cells were stimulated with either live bacteria (A, B, D) or bacteria-free culture supernatants (C).
A. Total RNA was isolated from cells harvested 6 h post stimulation and LL-37, HBD-1 and IL-8 expression was analysed. (Top) Semi-quantitative RT-PCR; one representative agarose gel picture out of three is presented here. (Bottom) qRT-PCR performed with triplicate samples and data from one of the two experiments are shown; bars above and below the x-axis indicate upregulation and downregulation, respectively, compared with the expression in the unstimulated samples.
B. Confocal microscopic images of the cells stimulated with the live bacteria and stained with anti-LL-37 antibody followed by FITC-conjugated secondary antibody.
C. Gene expression analysis by semi-quantitative RT-PCR (top) and qRT-PCR (bottom) performed with RNAs isolated from differentiated HT-29 cells stimulated with bacterial culture supernatants for 6 h. Data from one representative experiment are presented here.
D. Colony-forming unit (cfu) count 16 h after plating of bacteria following exposure to the culture supernatants of differentiated HT-29 cells. X-axis indicates the duration of stimulation of the differentiated HT-29 cells by the respective bacterial strain before the culture supernatants were collected and exposed to the same bacteria. Cfu counts were normalized against the growth potential of individual strains in tissue culture media (DMEM) as well as their killing by the culture supernatant of differentiated but untreated cells during the indicated durations.
*, < 0.01 (highly significant); +, < 0.05 (significant) compared to the undifferentiated cells. VP, Vibrio parahemolyticus; LT2, Salmonella typhimurium LT2; SD, Shigella dysenteriae type I; BL-21, E. coli BL-21(DE3).

CT is the predominant secreted molecule of V. cholerae involved in the downregulation of AMPs

We observed maximum suppression of AMP expression by twofold concentrated culture supernatants of O139 and ETEC (Fig. S2A) and the same concentration was used for all the experiments. To address the question whether only the secreted product(s) of O139 downregulates AMPs or other intracellular, surface or membrane-associated molecule(s) may also contribute to it, we stimulated differentiated HT-29 cells with the bacterial culture supernatant, cell lysate, heat-killed bacteria or the purified membrane fraction. mRNA expression was unaffected by all the cell-derived fractions except that the bacterial lysate suppressed HBD-1. In contrast, remarkable downregulation of both LL-37 and HBD-1 was found following stimulation with the culture supernatant (Fig. 2A), suggesting that the secreted molecules were predominantly involved. As CT is the major virulence factor present in the culture supernatant of O139, we went on to investigate if CT was responsible for the observed effect on AMPs. To this end, we stimulated HT-29 cells with the culture supernatants of several V. cholerae strains that either produced variable amounts of CT or did not produce it at all. While all the strains that avidly produce CT significantly suppressed LL-37 and HBD-1 expression, the supernatants of CT non-producing (ctx−) V. cholerae strains [(non-toxigenic V. cholerae O1 (WO5) and O139 (CO853) and the isogenic CT mutant of O139 (SG24) strains; see Table 1] had no effect. In addition, culture supernatant of the 569B strain (ctx+) grown in nutrient broth where it produces significantly diminished amount of the toxin exerted much less role. These results suggest that CT might be responsible for the observed downregulation of AMP expression by V. cholerae (Fig. 2B). As CT is a heat-labile protein, we treated the culture supernatant of O139 with proteinase K that inactivates all the proteins as well as heat-inactivated it at two different temperatures (at 56°C that inactivated CT-A, but not CT-B and at 100°C that inactivated most proteins including CT-A and CT-B) and stimulated HT-29 cells with them. The results showed that heat inactivation at both temperatures as well as proteinase K treatment entirely abolished the effects of O139 culture supernatant on LL-37 and HBD-1 mRNA expression, suggesting that an enzymatically active CT may be predominantly responsible for this effect (Fig. 2C). Finally, to establish that the heat-labile protein may be CT, we pre-treated HT-29 cells with anti-CT antibody followed by the culture supernatant. This resulted in HBD-1 expression to remain the same as in the unstimulated cells, while LL-37 level was still somewhat lower. Treatment with anti-CT antibody alone had no impact on expression, indicating that CT was the predominant factor present in the supernatant of O139 responsible for suppression of the AMPs (Fig. 2D). The downregulation of LL-37 that persisted after the antibody pre-treatment might be either due to incomplete blocking of CT function or due to a second molecule contributing to this effect. Taken together, these results suggest that secreted heat-labile protein(s) of O139, predominantly CT, may suppress the expression of the AMPs in the cultured IECs.

Figure 2.

Downregulation of LL-37 and HBD-1 by cholera toxin present in the culture supernatant of V. cholerae.
A. LL-37 and HBD-1 expression was analysed by semi-quantitative RT-PCR in NaB-differentiated HT-29 cells stimulated for 6 h with the culture supernatant or different cellular fractions of V. cholerae O139. (Left) Agarose gel (2%) electrophoresis of the amplified PCR products from one of the three independent experiments. NaB-differentiated cells that were not stimulated with any of the V. cholerae fractions were used as controls. (Right) Mean (±SD) relative (GAPDH normalized) expression of LL-37 and HBD-1 transcripts of three experiments.
B. Differentiated HT-29 cells were stimulated with several V. cholerae strains that either produced variable amounts of CT or did not produce it at all and LL-37 and HBD-1 expression was analysed by qRT-PCR. Data from one of the two experiments performed in triplicate are shown.
C. Experiment done as in A with differentially treated culture supernatant of O139 strain and results analysed as in (A).
D. NaB-differentiated HT-29 cells pre-stimulated with anti-CT antibody followed by the culture supernatant of O139 strain. Cells left unstimulated or stimulated with the supernatant or the antibody alone were used as controls. (Left) LL-37 and HBD-1 expression was analysed by semi-quantitative RT-PCR; agarose gel (2%) electrophoresis of one representative experiment out of the three is shown. (Right) qRT-PCR performed in triplicate with differentially treated cells showing relative (GAPDH-normalized) expression of LL-37 and HBD-1.
*, < 0.01 (highly significant) compared to the undifferentiated cells.

Table 1.  Bacterial strains used in the current study.
Strain namePhenotypeGenotypeSourceReference
V. cholerae O1 (WO5)Inabactx(−), tcpA(+)Clinical sampleSaha et al. (1996)
V. cholerae O139 Bengal (SG24) ctx(+), tcpA(+)Clinical sampleMukhopadhyay et al. (1996).
V. cholerae O1 ElTor (CO550)Ogawactx(+), tcpA(+)Clinical sampleMukhopadhyay et al. (1996)
V. cholerae O1 (569B)Ogawactx(+), tcpA(+)CommercialATCC # 25870
V. cholerae O139 (CO853) ctx(−), tcpA(−)Clinical sampleBasu et al., 2000
Enterotoxigenic E. coli # 4660Serogroup O25elt, est, CS4, CS6Clinical sampleGhosal et al. (2007)
Enterotoxigenic E. coli # 6300Serogroup O115est, CS5, CS6Clinical sampleGhosal et al. (2007)
Enterotoxigenic E. coli # 4407Serogroup O6elt, CS3Clinical sampleGhosal et al. (2007)
Salmonella typhimuriumS. enterica, subspecies enterica, serovar typhimurium (LT2) CommercialATCC # 33068
Shigella dysenteriae type I (BCH518)Congo red (+), Sereney test (+) Clinical samplePazhani et al. (2004)
Vibrio parahemolyticus (L-17400)Serovar 01 (K antigen untypable)tdh(−) toxR(+) trh(−)Environmental isolateT. Ramamurthy et al. (unpublished)
E. coli BL21(DE3)  CommercialNovagen cat # 69387
Isogenic mutant of V. cholerae O139 (SG24) ctx(−), tcpA(+)Laboratory generatedS. Das et al. (unpublished)

Purified CT downregulates LL-37 and HBD-1 expression

To further establish the role of CT, we first optimized the dose–response of IECs to CT. Both LL-37 and HBD-1 were suppressed, with 200 ηg ml−1 of CT exerting the maximum effect. Interestingly, a further higher dose diminished the effect which was almost completely abolished with 1 μg ml−1 of CT. Compared to HBD-1, LL-37 downregulation required a lower dose of CT (Fig. S2B). Next, we stimulated HT-29 cells with CT (200 ηg ml−1) for different durations and analysed AMP expression by semi-quantitative RT-PCR, qRT-PCR and immunohistochemistry. The results showed significant suppression of LL-37 mRNA and protein expression as early as 3 h that reached maximum after 6 h of stimulation. In contrast, HBD-1 downregulation followed a delayed kinetics with significant diminution of expression observed only after 6 h of CT treatment. HBD-2 and -3 expressions, on the other hand, remained the same in the unstimulated and CT-stimulated cells (Fig. 3A and B, Fig. S3). To investigate if the downregulation is restricted to the HT-29 cell line or may be observed in other IECs as well, we stimulated other colonic (Caco-2 and T84) and small intestinal (INT407) cell lines with CT. The results showed similar effects on AMP expression as in the HT-29 cells (Fig. 3C). We also studied the expression in response to CT in spontaneously differentiated as well as undifferentiated HT-29 cells and found that CT plays a major role in these cells as well (Fig. S4). However, LL-37 was undetectable in the undifferentiated HT-29 cells. Taken together, the above data suggest that AMP regulation by CT is not cell line specific; rather it is relevant in a broader context and may even be physiologically important.

Figure 3.

Purified CT downregulates LL-37 and HBD-1 expression in HT-29 cells.
A. NaB-differentiated HT-29 cells stimulated with CT for various durations. (Left) Semi-quantitative RT-PCR showing LL-37 and HBD-1 expression; one representative agarose gel electrophoresis out of three is shown. (Middle) qRT-PCR performed in triplicate showing relative (GAPDH-normalized) expression of the above genes in the unstimulated and CT-stimulated cells. (Right) Semi-quantitative RT-PCR of HBD-2 and -3 expressions in the first samples.
B. Confocal microscopic images of LL-37 protein expression in unstimulated and CT-stimulated HT-29 cells, analysed by immunohistochemistry using polyclonal primary and FITC-conjugated secondary antibodies.
C. qRT-PCR analysis of LL-37 and HBD-1 expression in Caco-2, T84 and INT407 cells stimulated with cholera toxin for various durations.
*, < 0.01 (highly significant) compared to the undifferentiated cells.

LT of ETEC exerts a similar role in LL-37 and HDB-1 expression

To investigate whether LT has similar effects on AMP expression as CT, we stimulated HT-29 cells with the clinical isolates of toxin-expressing and -non-expressing strains of ETEC as well as the purified toxins. The results showed significant suppression of LL-37 and HBD-1 mRNA expression by purified LT and the strains that produce LT, but not by stable toxin (ST) (Fig. 4A). A similar trend was observed in LL-37 protein expression when the cells were stimulated with either the live bacteria (Fig. 4B) or purified toxins (data not shown).

Figure 4.

LT but not ST of ETEC downregulates AMP expression in the HT-29 cells.
A. qRT-PCR carried out with RNAs isolated from cells stimulated with either live bacteria (left) or purified toxins (right). The experiments were performed twice and each time with triplicate samples. *, < 0.01 (highly significant) compared to the undifferentiated cells.
B. Confocal microscopic images of HT-29 cells stimulated with live ETEC and immunostained for LL-37 protein expression as in Fig. 3B.

AMP regulation by enteric bacteria is physiologically relevant

To study the physiological relevance of the downregulation of AMPs by enteric pathogens, we performed ileal loop experiments in mice and rabbits. Six-hour treatment of mice ileal loops with O139 and ETEC resulted in significant downregulation of cathelicidin (CRAMP) mRNA (Fig. 5A) and protein (Fig. 5B) expression in the freshly isolated mucosal epithelial cells. Mouse β-defensin 1 (MBD-1) expression was also suppressed in these cells. Diminished viability due to bacterial treatment was ruled out by marked induction of mouse IL-8 (KC) in the same cells (Fig. 5A) as well as by histological examination of the ileal tissues derived from the experimental mice that showed normal architecture of mucosal tissue with intact crypt and villus epithelium (Fig. 5C). Treatment of the ileal loops with purified CT resulted in marked downregulation of both CRAMP and MBD-1 in the primary IECs of mice (Fig. 5D) and cathelicidin (CAP18) in the rabbit intestine (Fig. 5E). The data presented so far suggest that O139 and ETEC play a major role in the regulation of the AMPs in vitro and in vivo.

Figure 5.

Downregulation of antimicrobial peptides by V. cholerae O139/ETEC and their secreted toxins in vivo.
A. Mouse homologs of LL-37 (CRAMP), HBD-1 (MBD-1) and IL-8 (KC) mRNA expressions were analysed by qRT-PCR in the freshly isolated primary IECs following ileal loop experiments with V. cholerae and ETEC. Three loops were created for each bacteria and time point and the same experiment was repeated twice.
B. Confocal microscopic images of mouse enterocytes isolated following ileal loop experiment and stained with polyclonal antibody against CRAMP followed by FITC-conjugated secondary antibody.
C. Light microscopic images (40× magnification) of ileal tissues of experimental mice stained with haematoxylene and eosin. Magnified views of crypt (inset A) and villus (inset B) epithelium are shown.
D and E. Enterocytes isolated from mouse and rabbit following stimulation of the ileal loops with CT were analysed for the expression of CRAMP and MBD-1 (D) or CAP18 (E), respectively, by semi-quantitative RT-PCR. Multiple loops were injected with either PBS or CT in each of the three experiments. (Top) Agarose gel electrophoresis in one representative experiment. GAPDH was used as loading control. (Bottom) Mean (±SD) relative expression of the above-mentioned genes in three independent experiments.
*, < 0.01 (highly significant) compared to the undifferentiated cells.

Multiple signal transduction pathways contribute to CT-induced downregulation of AMPs

Published reports suggest that CT may activate multiple signal transduction pathways like the PKA-CREB, MAPKinases and AA synthesis pathways in the epithelial cells (Stork and Schmitt, 2002; Dumaz and Marais, 2005). Although the principal mode of action of CT is through the accumulation of intracellular cAMP, it is probably capable of activating the signalling pathways independent of cAMP that may contribute to its immunomodulatory role (Dickinson and Clements, 1995; Pizza et al., 2001). To dissect the signalling pathways that regulate AMP expression in the IECs, semi-quantitative and real-time RT-PCR was performed in the cells pre-treated with the signalling pathway inhibitors followed by CT. The results showed that pre-treatment with adenylate cyclase (AC) inhibitor 2′5′DDA completely obliterated the effects of CT, suggesting that CT-induced downregulation of LL-37 and HBD-1 is cAMP dependent (Fig. 6A). This is further supported by stimulating the cells with cell-permeable cAMP agonists (8-br cAMP and dibutyl cAMP) or AC activator (forskolin) as well as endogenous cAMP activator (PGE2), all of which suppressed AMP expression that was comparable to the effect of CT (Fig. 6B). In contrast, inhibition of each of the PKA, ERK MAPK and AA synthesis pathways separately only partially abolished CT-induced modulation of LL-37 and HBD-1 expression (Fig. 6A, C and D). These results indicate that several signal transduction pathways are activated downstream of intracellular cAMP accumulation in response to stimulation with CT and contribute to the regulation of AMP expression.

Figure 6.

Involvement of multiple signal transduction pathways in the cholera toxin-induced regulation of LL-37 and HBD-1 expression.
A. NaB-differentiated HT-29 cells were pre-treated with Rp-cAMP (PKA inhibitor) or 2′5′DDA (AC inhibitor), either alone or in combination, followed by CT. (Left) Semi-quantitative RT-PCR amplification of LL-37 and HBD-1; one representative gel is shown. (Right) qRT-PCR performed with triplicate samples and relative LL-37 and HBD-1 expression is presented as fold changes of the untreated samples in the same experiment; data from one of the two independent experiments are presented here.
B. qRT-PCR analysis of LL-37 and HBD-1 mRNA expression in differentiated HT-29 cells stimulated with cell-permeable cAMP analogs (8-Br cAMP and dibutyl cAMP), adenylate cyclase activator (forskolin) or endogenous cAMP activator (PGE2). The experiment was performed with triplicate samples and the bar diagram represents one of the two independent experiments.
C. Experiments performed as in (A) with cells pre-treated with MEK inhibitor U0126 (MEK1/MEK2 inhibitor) or PD98059 (MEK1 inhibitor) followed by CT. Data presented as in (A).
D. Experiments carried out, as in (A) and (C) above, with cells pre-treated with one of the Cox-2 inhibitors, namely nimesulide, etoricoxib and inhibitor IV (Calbiochem, USA). Data presented as in (A) and (C).
*, < 0.01 (highly significant) compared to the undifferentiated cells. 2′5′DDA, 2′5′ Dideoxyadenosine; db-cAMP, N6,2′-O-dibutyryladenosine-cAMP; PGE2, prostaglandin E2.

Enteric pathogens transcriptionally downregulate LL-37 expression that may be mediated through ICER

To investigate whether suppression of the mRNA expression levels of AMPs by the enteric pathogens and their toxins is due to transcriptional repression or mRNA instability, we performed reporter assays following transfection of HT-29 cells with a luciferase reporter construct containing the upstream regulatory sequence of the LL-37 gene. Butyrate-induced differentiation markedly upregulated the reporter activity that was suppressed by O139 and ETEC as well as by CT and LT, suggesting that the downregulation of LL-37 expression is transcriptional. Lack of changes in the reporter activity following TNF-α stimulation and marked increase in the activity of a reporter containing the IL-8 promoter sequence by all the stimuli that downregulated AMPs indicates specificity of the effect and rules out the possibility of reporter instability or diminished cell viability (Fig. 7A). Because ICER is a well-defined transcriptional repressor of a number of cAMP-regulated genes, we planned to investigate if ICER was responsible for the downregulation of AMPs induced by the enteric pathogens and their secreted toxins. As ICER is transcriptionally regulated, we studied its mRNA expression in the unstimulated and stimulated HT-29 cells. The results showed significant induction of ICER mRNA expression 3 h post stimulation that reached maximum after 6 h, a pattern similar to AMP downregulation by the same stimuli, suggesting that ICER may regulate AMP expression that is observed in the HT-29 cells following infection with O139 and ETEC or stimulation with their purified toxins (Fig. 7B).

Figure 7.

Transcriptional downregulation of LL-37 expression by cholera toxin (CT) that may be mediated through ICER.
A. LL-37 or IL-8 promoter-driven luciferase reporter activity was assayed in HT-29 cells transiently overexpressing the respective reporter construct. Cells were either left undifferentiated or differentiated with NaB (4 mM for 24 h). The differentiated cells were stimulated with live O139 or ETEC or their secreted toxins or TNF-α. The experiment was performed with triplicate samples and repeated three times; data from one representative experiment is presented here.
B. qRT-PCR of ICER in differentiated HT-29 cells stimulated with live bacteria (O139 or ETEC) or purified toxins (CT or LT) for the indicated time points.
*, < 0.01 (highly significant); +, < 0.05 (significant) compared to the undifferentiated cells.

Discussion

We studied the regulation of AMP expression in NaB-differentiated colon epithelial cells, as the maximally differentiated cells line the luminal surface of the intestine and play the most critical role in enteric infections, while dietary fibre-derived butyrate contributes to epithelial differentiation in the gut. In agreement with the published reports (Schauber et al., 2004), we observed that HT-29 cell differentiation and LL-37 expression rise in parallels with increasing dose and duration of NaB. On the other hand, β-defensin levels were independent of cellular differentiation (Fig. S1). In our studies, LL-37 was significantly downregulated by enteric pathogens in the differentiated HT-29 cells, while the relatively low level of expression in the undifferentiated cells (Schauber et al., 2004; Schwab et al., 2007) may have precluded the detection of its suppression by S. dysenteriae type I in a previous study (Hase et al., 2002).

Pathogenicity of Neisseria gonorrhoea, in an earlier study, was found to be related to LL-37 expression in the genitourinary epithelial cell line (Bergman et al., 2005). The authors suggested that bacterial secreted products may downregulate AMPs at the mucosal surfaces. We, for the first time, have shown that the secreted toxins of non-invasive bacteria like V. cholerae and ETEC may significantly suppress LL-37 and HBD-1 expression in the IECs. Non-pathogenic V. parahemolyticus and BL21(DE3) strains failed to regulate AMP expression, further suggesting that this effect may be related to bacterial pathogenicity and may directly result from the virulence proteins. Identical regulation of LL-37 and HBD-1 by enteric bacteria strengthens the hypothesis that their downregulation may significantly aid in immune evasion by the pathogens, as concomitant suppression of both the AMPs with high basal expression levels would be crucial for efficient neutralization of the host innate immune response.

We have shown that CT is the predominant V. cholerae-associated molecule that suppresses the AMPs in the IECs. However, pre-treatment of HT-29 cells with anti-CT antibody only partially obliterated the effect of CT on LL-37 expression, while HBD-1 downregulation was completely abolished (Fig. 2D). This may be explained by either involvement of another secreted protein of O139 in LL-37 regulation or the residual CT effect that persisted after the antibody treatment. The latter explanation seems more probable in view of our subsequent experiments, which showed that LL-37 downregulation, compared with HBD-1, requires a lower dose of CT (Fig. S2). Interestingly, the most remarkable effects on both the AMPs were found with a 200 ηg ml−1 of CT and the response was blunted at a higher dose. This has also been noticed by other investigators (Royaee et al., 2006) and rules out toxicity as the cause for CT-induced downregulation of LL-37 and HBD-1 expression.

Cholera toxin suppressed AMP expression in the IECs of diverse origins. Neither NaB nor differentiation per se possibly had an influence over this role of CT. We found similar results in NaB-induced as well as spontaneously differentiated HT-29 cells (Fig. S4A). However, the basal LL-37 level was still considerably lower after 7 days of spontaneous differentiation, compared with what was observed in the cells differentiated by butyrate (4 mM) for 48 h. Significant downregulation was also seen in freshly isolated small IECs of mice and rabbits (Fig. 5). In addition, we observed identical changes in HBD-1 expression in response to CT in the differentiated and undifferentiated cells (Fig. S4B). Finally, LL-37 upregulation in the differentiated IECs was probably not due to transcriptional activation (Schauber et al., 2003). However, any role butyrate or cellular differentiation might have played in CT-induced regulation of LL-37 was not completely ruled out in our studies.

Similar to the in vitro experiments, AMPs were markedly suppressed in the ileal loop experiments by both live bacteria and their secreted toxins, as early as 3 h post stimulation. This suggests that AMP regulation by enteric pathogens may be physiologically relevant and CT/LT released inside the intestinal lumen following a natural infection may suppress AMPs in the gut epithelium. The latter may help colonization of the gut that is evident within 4 h of infection in an infant mouse model of V. cholerae (Ottemann and Mekalanos, 1995). The biological relevance of the suppression of AMPs is also supported by significantly reduced bacterial killing by the culture supernatants of the cells pre-treated with the pathogenic bacteria compared with the supernatants from non-pathogen treated cells (Fig. 1C).

LL-37 mRNA and protein expression was maximally suppressed after 6 h of stimulation with the live bacteria/secreted toxins, a time-frame that falls well within the range compared to the studies published by other groups who used similar experimental systems (Islam et al., 2001; Hase et al., 2002; Bergman et al., 2005). We and others (Hase et al., 2002; Zaalouk et al., 2004) have found identical changes in the mRNA and protein expression levels of the AMPs in response to different stimuli, suggesting absence of major post-translational regulation. Multiple lines of evidence indeed suggests that the regulatory mechanisms for LL-37 operate mainly at the level of mRNA expression (Schauber et al., 2004; Elloumi and Holland, 2008). We performed reporter assays to experimentally demonstrate, for the first time, that both live bacteria as well as their secreted toxins transcriptionally downregulate LL-37 (Fig. 7). TNF-α did not alter LL-37 reporter activity while IL-8 promoter-driven reporter activity was elevated by many folds following bacterial infections, thus ruling out the possibility of reporter instability or compromised cell viability. In addition, this indicates that AMP regulation by pathogenic bacteria and their virulence proteins was indeed specific and differential and rules out toxicity as the probable cause for the observed downregulation.

Accumulating evidence suggests that CT and LT may regulate, both positively and negatively, a large number of immunomodulatory genes in the IECs (Soriani et al., 2002; Huang et al., 2004). Although both cAMP-dependent and -independent pathways may be responsible for immunomodulation (Pizza et al., 2001), LL-37 and HBD-1 downregulation in our studies was entirely dependent on CT-induced accumulation of cAMP and contributed by several intracellular signalling pathways. Identical regulation of the AMPs by CT in cells of different origins and complete reversal of this effect by an AC inhibitor (2′5′DDA) as well as downregulation by multiple other agents that can activate cAMP signalling pathway (PGE2/forskolin/cell-permeable agonists/LT) underscore the critical role played by the latter in AMP regulation. Intriguingly, downregulation of the AMPs was considerably lower with ETEC-producing LT and ST compared to the bacteria-secreting LT alone. This may have resulted from lower production of LT by the bacteria that expressed both the toxins.

Protein kinase A (PKA) is the best-known cAMP effector, which activates a group of transcription factors (TFs) (CREB, CREM τ, ATF1) as well as repressors (CREM-α, -β and -γ, ICER, E4BP4, CREB2) that bind CRE (cAMP response element) sequences in the promoter regions of a large number of genes and regulate their expression (Brindle and Montminy, 1992; Lalli and Sassone-Corsi, 1994; Sassone-Corsi, 1994; 1995; 1998). Interestingly, a number of PKA-activated molecules are in fact bifunctional TFs, which depending on the context, may either activate or repress gene expression (Ogbourne and Antalis, 1998). cAMP also activates other intracellular signalling pathways like the RAS-RAF-MEK-ERK and AA metabolism pathways (Stork and Schmitt, 2002; Dumaz and Marais, 2005), which extensively cross-talk with the cAMP-CREB pathway. Cox-2, the rate-limiting enzyme in AA metabolism, may also induce transcriptional repressors like ZEB1 and Snail (Dohadwala et al., 2006), while Cox-2 inhibition results in the elevation of HBD-2 mRNA level in response to Actinobacillus actinomycetemcomitans (Noguchi et al., 2003).

LL-37 has several CRE and activator protein-1 (AP-1) consensus sites in its upstream sequence in addition to putative binding sites for SP1, NF-κB, C/EBP-α, C/EBP-β and Oct-1 (Fig. S5). In a recently published report, multiple cis-acting repressor elements have been identified in the upstream regulatory sequence of the LL-37 gene and a number of trans-acting factors binding those elements have been suggested (Elloumi and Holland, 2008). HBD-1 promoter, on the other hand, has multiple binding sites for SP1 and C/EBP (Fig. S6). To fully understand cAMP-induced downregulation of AMPs, the TFs that regulate their basal expression levels must be identified. cAMP-mediated signals may replace the already bound CREB from the promoter sites of the AMPs by an active repressor or may modulate CREB to make it transcriptionally less active (Sun et al., 1994). Alternatively, an activated repressor may interfere with the functions of the basal transcriptional machinery (Gaston and Jayaraman, 2003). CREM-α has recently been found to actively repress transcription by recruiting HDAC1 to the CRE sites of IL-2 and c-Fos promoters (Tenbrock et al., 2006). Finally, cAMP-activated CREB and CREM may inhibit Jun-mediated transcription (Masquilier and Sassone-Corsi, 1992). Here, we have reported the induction of ICER by CT at the time points where LL-37 and HBD-1 were repressed, suggesting that ICER may mediate CT-induced negative regulation of AMP expression. ICER binding has been reported to downregulate the promoter activity of a large number of cAMP-responsive genes that may or may not be regulated by the PKA-CREB pathway (Mioduszewska et al., 2003). Our results suggest involvement of PKA-dependent and -independent pathways in the regulation of AMPs (Fig. 6). However, we could not exclude the possibility of complex regulation involving multiple repressors. Future studies with ICER−/− cells may help to establish the specific role this repressor may play in AMP regulation.

Considering the critical importance of immunomodulation by the AMPs in host defence, transcriptional downregulation would be the most efficient mode of immune evasion by pathogenic microorganisms. High AMP levels in the unstimulated cells and further modulation by multiple stimuli offer us a unique paradigm to study basal and inducible transcriptional regulation. Future studies in this direction would not only expand our knowledge about this fascinating biological process, but also help to identify novel targets to modulate the expression of these endogenous antibiotics to human benefits.

Experimental procedures

Cells and reagents

IL-8 reporter construct was a generous gift from Dr Suresh Tikoo, University of Saskatchewan, Saskatoon, Canada. HT-29 (HTB-38), Caco-2 (HBT-37), T84 (CCL-248) and INT407 (CCL6) cell lines were purchased from American Type Culture Collection (ATCC). NaB, p-nitrophenyl phosphate (NPP), purified CT, anti-CT antibody, ST of E. coli, 8Br-cAMP, N6,2′-O-dibutyryladenosine-cAMP (db-cAMP), prostaglandin, forskolin, Rp-cAMP, nimesulide and etoricoxib were purchased from Sigma. LL-37 antibody and FITC-conjugated anti-goat IgG were procured from Santa-Cruz Biotechnology and Jackson Laboratories respectively. Synthetic inhibitors (U0126, PD98059, Cox-2 inhibitor IV and 2′5′DDA) were purchased from Calbiochem, USA and TNF-α was procured from R&D Systems, USA. Oligonucleotides for the amplification of LL-37 (FP: 5′-CCAAGCCTGTGAGCTTCACAG-3′; RP: 5′-CTTGGCCTTCCCTCTGTAAC-3′), HBD-1 (FP: 5′-CAGGTGGTAACTTTCTCACAG-3′; RP: 5′-CTTGGCCTTCCCTCTGTAAC-3′), HBD-2 (FP: 5′-CCATGAGGGTCTTGTATCTCC-3′; RP: 5′-GAGACCACAGGTGCCAATTTG-3′), HBD-3 (FP: 5′-GTTCCAGTTCATGGAGGAATC-3′; RP: 5′-GTCGAGCACTTGCCGATCTG-3′), MBD-1 (FP: 5′-CTGCTGATATGCTGCCTCCT-3′; RP: 5′-AGGGGTTCTTCTCTGGGAAA-3′), CAP18 (FP: 5′-TACGGTGAAGGAGACGGAGT-3′; RP: 5′-AATCTGTCCTGGGTGGAAGT-3′) CRAMP (FP: 5′-TCCCAAGTCTGTGAGGTTCC-3′; RP: 5′-CCGGCTGAGGTACAAGTTTC-3′), GAPDH (human) (FP: 5′-GAGAACGGGAAGCTTGTCATC-3′; RP: 5′-CATGACGAACATGGGGGCATC-3′), GAPDH (mouse/rabbit) (FP: 5′-GGTCTACATGTTCCAGTATG-3′; RP: 5′-ATCACAAACATGGGGCATC-3′), IL-8 (FP: 5′-CTTGGCAGCCTTCCTGATTTCTGC-3′; RP: 5′-GTGGTCCACTCTCAATCACTCTC-3′), KC (FP: 5′-GCTGGGATTCACCTCAAGAA-3′; RP: 5′-TGGGGACACCTTTTAGCATC-3′), LL-37 upstream regulatory sequences (FP: 5′-TAGATGGAGCAGAGCCTTCG-3′; RP: 5′-AGGGGCGGTAGAGGTTAGC-3′) and ICER (FP: 5′-TGGAGATGAAACAGATGAGGAA-3′; RP: 5′-TCTCTGAGGGCCTTGAGTTC-3′) were custom synthesized from IDT, USA.

Bacterial strains

Salmonella typhimurium LT2 (ATCC # 33068) and S. dysenteriae type I were kindly provided by Professor Maharani Chakraborti, Emeritus Scientist and Dr Swapan Kumar Neogi, Deputy Director of NICED, Kolkata, respectively. All other bacterial strains were provided by the authors T.R. and A.K.M. Isogenic mutant of V. cholerae O139 Bengal (SG24) was generated and characterized in the laboratory of the corresponding author (see later). All the bacterial strains used in the current study are discussed in Table 1.

Generation of isogenic CT mutant of V. cholerae O139 Bengal (SG24)

Isogenic CT mutant of SG24 strain was generated by a technique called recombineering (Datta et al., 2006). Briefly, SG24 strain was electroporated with pSIM6 plasmid carrying the genes for bacteriophage λ recombination. Kanamycin resistance gene was amplified from E. coli Dy411 strain using primers containing partial sequence (50 bp) of CT gene as 5′ overhang (Fp: 5′-aagcagtcaggtggtcttatgccaagaggacagagtgagtactttgaccgtatggacagcaagcgaaccgga-3′; RP: 5′-GATCTTGGAGCATTCCCACAACCCGGCGGTGCATGATGAATCCACGGCTCTCAGAAGAACTCGTCAAGAAG-3′) and electroporated into pSIM6-transformed SG24. Positive clones were selected in Amp/Kan plates grown at 32°C. Deletion mutation of CT was confirmed by PCR amplification of ctx gene and morphological changes in CHO cell assay (data not shown).

Preparation of bacterial supernatants and other cellular fractions

All bacteria were cultured in LB-Miller (DIFCO) broth and the culture was filtered by passing it through 0.22 μm membrane when the OD600 reached 1. Bacteria-free supernatants were concentrated by AmiconTM column (Millipore) followed by the treatment with or without proteinase K (200 μg ml−1), anti-CT antibody (sigma) or heat (56°C and 100°C).

To prepare the crude membrane fraction, cells were pelleted from an overnight culture in LB-Miller. The cell pellet was re-suspended in 100 μl of buffer A (50 mM Tris-Cl, pH 7.5, 10% sucrose) and frozen in liquid nitrogen followed by addition of lysozyme (250 μg ml−1) to the thawed cell suspension. The cells were lysed at 37°C for 30 min, followed by immediate chilling in ice-cold water and centrifuged at 65 000 r.p.m. in an ultracentrifuge (Beckman) for 30 min. The pellet was washed with buffer A and re-suspended in buffer A containing 5 mM MgCl2, 10 μg ml−1 RNase A and 10 μg ml−1 DNase and incubated at 37°C for 45 min.

To prepare heat-killed bacteria, cells grown in LB were harvested by centrifugation. The pellet was re-suspended in PBS and heated at 95°C for 10 min.

Bacterial cell lysate was prepared by repeated cycles of freezing and thawing in liquid N2 and 37°C water bath, respectively.

Purification of LT

Labile toxin of ETEC was purified as described previously (Uesaka et al., 1994). Briefly, bacterial cells were disrupted by sonication and the debris was removed by centrifugation. Crude toxin was prepared by ammonium sulfate fractionation (65%) of the clear lysate followed by dialysis. For further purification, crude toxin was bound to immobilized d-galactose packed in a column and eluted with TEAN buffer (50 mM Tris-HCl, pH 7.4, 0.2 M NaCl, 3 mM NaN3 and 1 mM EDTA). The eluted fractions that gave positive results in an LT-specific bead-ELISA were concentrated and further resolved by HPLC using Superose 12. A major peak, eluted at a retention time of 27 min, gave positive results for both bead-ELISA and morphological changes in CHO cell assay. The protein eluted in this peak was used as purified LT in the subsequent experiments.

Cell culture, differentiation and AP activity assay

All the cell lines except Caco-2 were cultured in DMEM (Gibco) supplemented with 10% FBS, 2 mmol l−1l-glutamine, penicillin (100 U ml−1) and streptomycin (100 μg ml−1) (Gibco). Caco-2 cells were cultured in MEM supplemented with 20% FBS along with l-glutamine and pen/strep. For the purpose of experiments, 2 × 105 cells were seeded in each well of a 24-well plate and allowed to grow. Optimum differentiation was induced by culturing 1-day post-confluent cells in serum-free DMEM supplemented with NaB (4 mmol l−1) for 48 h.

Alkaline phosphatase (AP) activity assay was performed as described previously (Schauber et al., 2003). Briefly, cells were lysed with the lysis buffer (150 mM NaCl, 20 mM Tris-HCl, 1 mM EDTA, 0.1% Triton X-100, pH 7.5), diluted with the AP activity buffer (50 mM Tris-HCl, 1 mM MgCl2, pH 8.0) and incubated with the AP substrate (0.5% NPP) for 10 min at 37°C. Reaction was stopped by adding 50 mM NaOH and AP activity was assayed by measuring the optical density (OD) at 410 nm. Enzyme activity was expressed as mU per mg of protein, where one unit represents the enzyme activity required to hydrolyse 1 μmol of substrate per minute (Bowers and McComb, 1966).

In vitro infection

Bacteria grown in LB were harvested when OD600 reached 1, washed in PBS and re-suspended in DMEM. Differentiated HT-29 cells were infected with bacteria for 1 h at a multiplicity of infection (moi) of 100. Cells were washed five times with PBS and cultured for an additional 6 h in complete DMEM supplemented with gentamicin (200 μg ml−1) to kill the extracellular bacteria. Cell viability was determined by trypan blue dye exclusion at the end of the experiment.

Total RNA extraction and cDNA synthesis

Total RNA was extracted with TrizolTM reagent (Invitrogen) following the manufacturer's instructions. cDNA was prepared from 2 μg of extracted RNA using Superscript II reverse transcriptase (Invitrogen) following the standard protocol.

Semi-quantitative PCR

PCR amplification of the genes of interest and one housekeeping gene (GAPDH) was performed using thermostable Taq polymerase (New England Biolabs). Amplified products were analysed in a 2% agarose gel and the band intensities were measured by Quantity OneTM software (Bio-Rad) in a gel-documentation system.

SYBR-Green® real-time PCR

Real-time PCR was performed using ABI7300 (Applied Biosystems, USA). Relative quantification was performed by the comparative CT method. PCR was performed with SYBR Green™ mastermix™ (Applied Biosystems, USA), where SYBR green is the fluorescent reporter. The internal control gene GAPDH was amplified simultaneously in separate reaction tubes. To eliminate primer dimers, the fluorescent signal was collected at a temperature of 82°C, at which the primer dimers melted, but the PCR products were still in their double-stranded form, thereby emitting fluorescence. The reaction conditions were set as follows: initial heating at 95°C for 5 min, followed by 40 cycles of reactions at 95°C (1 min), followed by 62°C (2.5 min) and finally 82°C (45 s). Final extension was carried out at 62°C for 8 min. Threshold cycle number (CT) of triplicate reactions was determined using the ABI-SDS software and the mean CT of triplicate reactions was determined. The levels of expression of the genes of interest were normalized to GAPDH using the formula 2−ΔΔCT, where −ΔΔCT = ΔCT (sample) − ΔCT (calibrator) and ΔCT is the CT of the target gene subtracted from the CT of the housekeeping gene (GAPDH). The calibrator used in our experiments was the unstimulated HT-29 cells.

Immunohistochemistry and confocal microscopy

HT-29 cells grown in 24-well tissue culture plates were washed with PBS and harvested with trypsin-EDTA (Gibco). Cells were fixed with 2% para-formaldehyde (PFA) and permeabilized by treating with 0.1% Triton X-100 for 30 min at 4°C. Cells were washed again with cold PBS and incubated with anti-LL-37 primary antibody (Santa Crutz) raised in goat at 1:500 dilution for 1 h at 4°C followed by FITC-conjugated anti-goat IgG for 1 h. Cells were washed four times with PBS, mounted on a coverslip and examined under confocal microscope (Zeiss).

Bacterial killing assay

NaB-differentiated (4 mM × 48 h) HT-29 cells were treated with live bacteria (V. cholerae O139, ETEC, S. typhimurium LT2, S. dysenteriae type I, V. parahemolyticus and BL-21) at an moi of 100 for 3 and 6 h duration. Fresh bacterial cultures (104 organisms each) were exposed for 3 h to filtered (0.22 μ) culture supernatants of the cells treated with the same bacteria and plated on LB agar after dilution. Colony-forming units (cfu) were counted for each bacterial strain following overnight growth at 37°C and plotted in a graph after normalization against the growth potential of individual strains and the killing potential by the culture supernatant of NaB-differentiated but unstimulated HT-29 cells.

Ileal loop experiment

The ileal loop experiment was performed in mice and rabbits by a modified Rabbit Ileal Loop (RIL) assay originally described by De and Chatterje (1953). Prior approval for the experiment was taken from the Institutional Animal Ethics Committee. Following gut sterilization, the animals were kept on fasting for 24 h prior to the experiment. Anaesthesia was induced by a cocktail of ketamine (35 mg per kg of body weight) and xylazine (5 mg per kg of body weight). Abdomen was slit open and three (mice) or six (rabbits) loops, each 5 cm long, were made in the ileum by tying with non-absorbable silk. Loops were injected with either live bacteria (108 organisms) or CT (500 ηg) in 0.5 ml of PBS. Loops injected with 0.5 ml of PBS alone were used as controls. Following injection, the loops were inserted back into the abdominal cavity and the animals were kept under observation. Animals were sacrificed after 6 h and each loop was dissected separately for three independent experiments. A small section of each loop was fixed in neutral buffered formalin (10%) for subsequent histological examination. Rest of the loops were washed twice with PBS followed by the addition of isolation buffer (0.9% NaCl, 8 mM KH2PO4, 5.6 mM NaH2PO4, 1.5 mM EDTA pH 7.6, 0.5 mM DTT and 0.25 mM PMSF) and incubated at 37°C for 3–4 min with intermittent tapping to dislodge the cells. The IECs released were collected, washed twice with PBS and treated with gentamicin (200 μg ml−1) for 30 min. Cells were subjected to immunohistochemistry or lysed with TrizolTM reagent for RNA extraction.

Histology

Formalin-fixed ileal loop tissues of mice were dehydrated followed by paraffin embedding. Paraffin-embedded tissue sections (3–4 μm thick) were cut in rotary microtome (Leica 2145, Germany), stained with haematoxylene and eosine (H&E) stain and visualized under light microscope (Leica 2145, Germany) for morphological details.

Generation of LL-37 reporter construct

LL-37 upstream regulatory sequence (upstream 900 bp to downstream 200 bp of the transcription start site of LL-37 gene) was PCR amplified with proof-reading polymerase (Pfu polymerase, Stratagene) using genomic DNA isolated from HT-29 cells as template and cloned into pBlueScript (Stratagene, USA). The sequence was verified using an automated DNA sequencer (Applied Biosystems). LL-37 URS (1100 bp) was subcloned by directional cloning into EcoRI/XhoI-digested luciferase reporter vector pGL3-basic (Promega).

Transfection

Transfection of HT-29 cells was performed using Lipofectamine 2000 (Invitrogen, USA) following the manufacturer's protocol. Briefly, cells (1 × 105 per well) were seeded in 24-well plate 18 h prior to transfection and cultured in DMEM supplemented with 10% FBS, but no antibiotics. Cells were incubated in triplicate wells in the transfection mix containing 0.75 μg of total DNA per well for 6 h followed by addition of complete DMEM and grown for 48 h. Cells were then differentiated with 4 mM of NaB for 24 h.

Reporter assay

Transfected undifferentiated and differentiated HT-29 cells were washed with PBS and lysed with 100 μl of 1× passive lysis buffer (Promega) per well. Reporter activity was analysed with a luminometer (Berthold) using Dual Luciferase Repoter Assay kit (Promega) following the manufacturer's protocol.

Statistical analysis

Semi-quantitative PCR data are presented as mean (±SD) of three independent experiments. qPCR was performed with triplicate samples and the data are presented as mean (±SD) of the values obtained.

Statistical significance was analysed by the Student's t-test. The results were considered significant (+) at P < 0.05 and highly significant (*) at P < 0.01.

Acknowledgements

We thank Dr Simanti Datta, Institute of Postgraduate Medical Education and Research, Kolkata, India for providing us with the pSIM6 plasmid and Dy411 strain. This work is partially funded by Program of Funding Research Center for Emerging and Reemerging Infectious Diseases, Ministry of Education, Culture, Sports, Science and Technology of Japan and extramural funds from Indian Council of Medical Research (IRIS ID 2006-05630).

Ancillary