GASTROENTEROLOGY 2001;120:99 –107
Interferon Gamma Induces Enterocyte Resistance Against Infection by the Intracellular Pathogen Cryptosporidium parvum RICHARD C. G. POLLOK,* MICHAEL J. G. FARTHING,* MONA BAJAJ–ELLIOTT,* IAN R. SANDERSON,‡ and VINCENT MCDONALD* *Digestive Diseases Research Centre and ‡Department of Paediatric Gastroenterology, St. Bartholomew’s and The Royal London School of Medicine and Dentistry, London, England
Background & Aims: Interferon (IFN)-␥ plays an important role in the immunologic control of infection by the protozoan enteropathogen Cryptosporidium parvum. We tested the hypothesis that IFN-␥ may directly inhibit infection of enterocytes by this pathogen. Methods: HT29, Caco-2, and H4 human enterocyte cell lines were grown in monolayers and incubated with IFN-␥ before exposure with C. parvum. IFN-␥ receptor expression in the cell lines was determined by Western blot analysis. Results: IFN-␥ inhibited C. parvum infection of both HT-29 and Caco-2 cells but not H4 cells. Response to IFN-␥ was related to the expression of the IFN-␥ receptor in the respective cell lines. The effect of IFN-␥ was partially reversed by inhibition of the JAK/STAT signaling pathway. IFN-␥ mediated its action by at least 2 mechanisms: (1) inhibition of parasite invasion and (2) by modification of intracellular Fe2ⴙ concentration. No role for tryptophan starvation or nitric oxide synthase activity was found. TNF-␣ and IL-1 also had anti–C. parvum activity but had no synergistic effect with IFN-␥. Conclusions: IFN-␥ directly induces enterocyte resistance against C. parvum infection; this observation may have important consequences for our understanding of the mucosal immune response to invasive pathogens.
he intestinal epithelium does not function solely as a physical barrier but also has an important role in the regulation of the mucosal immune response that is critical in the control of intestinal infection. Enterocytes are able to respond to extrinsic stimuli including infection and cytokines derived from immune cells in the mucosa by, for example, producing a range of cytokines and chemokines.1 In addition, specific cytokines including interferon (IFN)-␥ and tumor necrosis factor (TNF)-␣ may alter enterocyte phenotype, such as alteration of transepithelial resistance or up-regulation of inducible nitric oxide synthase (iNOS).1–3 Furthermore, a range of enterocyte responses may be synergistically enhanced by the combination of IFN-␥ with other proinflammatory cytokines.3–5 Recent evidence indicates mucosal proinflammatory cytokines may also induce an im-
T
mune effector response to enteroinvasive pathogens, as well as the cytokine-induced changes in enterocyte phenotype.6 –9 The coccidian parasite Cryptosporidium parvum, the infectious agent of cryptosporidiosis, infects the intestinal epithelium. Oocysts excyst in the small bowel, releasing sporozoites that invade enterocytes and form an extracytoplasmic parasitophorous vacuole in the apical portion of the cell. Meront development takes 8 –12 hours, and after a number of cycles of merogony sexual development and production of oocysts occurs. Cryptosporidiosis is characterized by watery diarrhea and cramping abdominal pain and, in immunocompromised individuals, symptoms may be persistent and associated with mortality.10 –12 Studies with murine models of infection indicate that IFN-␥ plays a key role in both innate and CD4⫹ T cell–mediated immune responses.13–16 Also, IFN-␥ production by CD4⫹ intraepithelial lymphocytes (IELs) has been shown to be an important protective mechanism in the control of infection.17,18 Humans with a primary deficiency of IFN-␥ are susceptible to persistent cryptosporidiosis which may respond to IFN-␥ therapy.19,20 Because human intestinal epithelial cells express a range of cytokine receptors including IFN-␥ receptor (IFN␥R),21,22 we hypothesized that IFN-␥ could act directly on human enterocytes in the immunologic control of C. parvum infection. Following the binding of IFN-␥ to its receptor and subsequent activation of the JAK/STAT (Janus kinase/signal transducer and activator of translaAbbreviations used in this paper: BSA, bovine serum albumin; DMEM, Dulbecco’s modified Eagle medium; IEL, intraepithelial lymphocyte; IFA, immunofluorescence assay; IFN, interferon; IFN-␥R, interferon gamma receptor; IL, interleukin; INDO, indoleamine-2,3-dioxygenase; JAK/STAT, Janus kinase/signal transducer and activator of translation; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; NOS, nitric oxide synthase; PBS-T, phosphate-buffered saline and 0.5% Tween 20; TNF, tumor necrosis factor. © 2001 by the American Gastroenterological Association 0016-5085/01/$10.00 doi:10.1053/gast.2001.20907
100
POLLOK ET AL.
tion) pathway, nuclear signaling results in the translation of a range of specific genes,4,23,24 thereby leading to the control of infection. In other intracellular infections, IFN-␥ has a direct inhibitory action on pathogen development,6 – 8 but the mechanisms by which IFN-␥ mediates its action on cryptosporidial infection remain to be determined. Nitric oxide (NO)-mediated killing is known to be an important mechanism in the control of intracellular pathogens that invade monocytes, following activation of inducible nitric oxide synthase (iNOS).25,26 In enterocytes, iNOS may be activated by a number of cytokines including IFN-␥, although the general significance of this enzyme in the control of pathogens that invade these cells is unclear. IFN-␥ may exert its inhibitory effect in other intracellular infections by induction of indoleamine-2,3-dioxygenase (INDO) and consequent depletion of cellular tryptophan, and this inhibitory effect may be reversed by the introduction of exogenous tryptophan.27 IFN-␥ may also mediate its inhibitory action on invasive pathogens by limiting intracellular Fe2⫹ availability6,26 or by inhibition of host-cell invasion by the pathogen.7 In this study, we aimed to determine possible mechanisms by which IFN-␥ mediated its putative inhibitory action on C. parvum infection in a previously described in vitro infection model.28,29 Specifically, we determined the role of NOS, tryptophan depletion, cellular Fe2⫹ concentration and inhibition of parasite invasion as possible mechanisms by which IFN-␥ mediates its action against parasite infection. The results suggested that IFN-␥ inhibited C. parvum development by at least 2 mechanisms requiring the direct action of the cytokine with enterocytes.
Materials and Methods Reagents All cytokines were purchased in lyophilized form and reconstituted according to supplier’s instructions. Recombinant human IFN-␥ was supplied by Sigma Ltd. (Poole, England), and TNF-␣, IL-1, and anti–IFN-␥R ␣-chain antibody were obtained from R&D systems (Abingdon, England). The JAK 2 inhibitor, tyrphostin B42 (N-benzyl-3,4-dihydroxybenzylidenecyanoacetamide), and the selective iNOS inhibitor, 1400 W (N-[3-(aminomethyl)benzyl]acetamadine), were supplied by Calbiochem (Nottingham, England). Dulbecco’s modified Eagle’s medium (DMEM) and fetal calf serum were supplied by Life Technologies (Inchinnan, England), and all other chemicals were supplied by Sigma-Aldrich Ltd.
Parasite Preparation C. parvum oocysts were prepared from infected neonatal lambs, purified by salt flotation, stored at 4oC in phosphate-
GASTROENTEROLOGY Vol. 120, No. 1
buffered saline (PBS), and used within 3 months of preparation (Moredun Scientific Ltd., Penicuik, UK). Before use, oocysts were suspended in 10% commercial bleach solution (0.55% sodium hypochlorite) for 10 minutes and washed in DMEM before enumeration in a Neubauer hemocytometer. Nonviable oocysts were prepared by heat-inactivation at 70°C for 45 minutes. Sporozoites were prepared by excysting oocysts in DMEM for 45 minutes at 37°C and purified by passage through a polycarbonate filter with 5-m pore size (Millipore, Bedford, England).
Cell Culture and Parasite Infection HT-29 and Caco-2 cells were obtained from the European Collection of Cell Culture (Salisbury, England) and used between passage numbers 17 and 35. H4 cells, a recently derived primary human fetal intestinal cell line,1 were used between passage numbers 19 and 32. Cells were grown at 37°C in 5% CO2 in DMEM supplemented with 10% fetal calf serum, 4 mmol/L glutamine, 100 U/mL penicillin, 100 g/mL streptomycin, and 1% nonessential amino acids; medium for H4 cells also contained 4 g/mL insulin. Trypsinized cells were seeded on to glass coverslips in 24-well tissue culture plates (Corning Costar, High Wycombe, England) and grown to confluence over 5 days. In experiments involving cytokines alone or in combination, unless otherwise stated, cells were cultured in the presence of the cytokine(s) at the stated concentrations for 24 hours before inoculation of C. parvum, and for a further 24 hours after inoculation. To determine if the inhibitory action of IFN-␥ on C. parvum infection was mediated through activation of the JAK/STAT pathway, HT-29 cells were treated with IFN-␥ in the presence or absence of the selective JAK-2 inhibitor, tyrphostin B42 (1–200 U/mL). To evaluate the role of NOS activity in the IFN-␥–mediated inhibition of C. parvum, HT-29 cells were cultured in medium containing IFN-␥ in the presence or absence of 3 different NOS inhibitors at a high and a low dose: NG-monomethyl-L-arginine (NG-MMA; 200 or 500 mol/L); N-nitro-L-arginine (L-NNA; 20 or 200 mol/ L); or the selective iNOS inhibitor, 1400 W (10 or 100 mol/L), starting 24 hours before infection. To determine the role of tryptophan depletion in the IFN-␥–mediated inhibition of C. parvum, HT-29 cells were cultured in medium containing IFN-␥ in the presence or absence of exogenous tryptophan or with exogenous tryptophan alone (50 –250 g/ mL) for 24 hours before infection. To examine modification of cellular Fe2⫹ in the IFN-␥–mediated control of infection, HT-29 cells were treated with IFN-␥ for 24 hours before infection in the presence or absence of FeSO4 or with FeSO4 alone (12.5–200 mol/L).
Infection of Intestinal Cell Lines With C. parvum Cell monolayers were infected by the addition of 2 ⫻ 105 viable or heat-inactivated oocysts or 4 ⫻ 105 purified sporozoites suspended in 200 L DMEM. After addition of parasites, cells were incubated for 2 hours at 37°C unless
January 2001
otherwise stated, to allow parasite attachment and invasion to occur. Following this period, unattached parasites were removed by repeated washing with DMEM. Wells were filled with fresh medium, and incubated at 37°C for 24 hours unless otherwise stated. Infection was quantified by both immunofluorescence assay (IFA) and Giemsa staining unless otherwise stated, as previously described.28,29 For IFA, infected monolayers were washed thoroughly with DMEM and fixed with ice-cold 100% methanol for 5 minutes. The cells were either immediately stained for IFA or stored at ⫺20°C for up to 2 weeks before staining. Fixed monolayers were blocked for 1 hour at 37°C by incubation with PBS (pH 7.2) containing 2% bovine serum albumin (BSA) and 2% normal sheep serum. After washing, monolayers were incubated at 37°C for 45 minutes with a 1:1000 dilution of anti–C. parvum rabbit serum in PBS containing 1% normal sheep serum and 0.1% BSA. Anti–C. parvum antibodies were raised by subcutaneous injection of a New Zealand white rabbit with 1 ⫻ 107 purified sporozoites in Freund’s complete adjuvant 3 times at intervals of 14 days, and serum was collected 10 days after the last immunization. After washing with PBS, monolayers were incubated with 1:150 dilution of fluorescein isothiocyanate– conjugated affinity purified sheep anti-rabbit immunoglobulin (Ig) G antibody (Binding Site Ltd., Birmingham, England) in PBS containing 1% normal sheep serum, 0.1% BSA, and 0.001% Evan’s blue for 30 minutes at 37°C. Washed coverslips were inverted and mounted on glass slides with Vector Shield antifade mountant (Vector Ltd., Peterborough, England), and monolayers were examined by epifluorescence using a Zeiss Axioskop 2 microscope (Carl Zeiss Ltd., Welwyn Garden City, England). To quantify infection, parasites were counted in 20 random fields across the diameter of the coverslip at 400⫻ magnification. Monolayers inoculated with heat-inactivated oocysts had no detectable evidence of infection as determined by IFA or Giemsa staining. Parasites were enumerated by Giemsa staining, following monolayer fixation by treatment with 10% Giemsa in PBS (pH 7.2) for 1 hour. Intracellular parasitic stages were counted in 50 random fields at 1000⫻ magnification. In preliminary experiments, 2 independent observers masked to the treatment conditions of each coverslip quantified infection, and found interobserver and intraobserver variation to be ⬍10% for both IFA and Giemsa staining. Experiments were performed with 4 – 6 replicates, infection was expressed as mean percentage of inhibition of untreated control wells ⫾ SD unless otherwise stated, and experiments were repeated on at least 3 separate occasions. Data from experiments in which parasites were enumerated by Giemsa staining are shown in Results unless otherwise stated; similar results were also obtained by IFA but for brevity are not shown.
Attachment and Early Invasion of C. parvum Experiments were performed to distinguish the action of IFN-␥ on parasite attachment and early invasion. Attach-
IFN-␥–MEDIATED INHIBITION OF C. parvum INFECTION
101
ment of C. parvum sporozoites to paraformaldhyde-fixed HT-29 cells was assessed, using a modification of a method previously described by other investigators.30,31 Briefly, HT-29 cells grown on coverslips to confluence in the presence or absence of IFN-␥ were treated with 4% paraformaldhyde in PBS (pH 7.2), or PBS alone for 5 minutes at room temperature, and then washed 3 times with DMEM before the addition of sporozoites and incubation for 90 minutes at 37°C. Wells were then washed 3 times with DMEM to remove unattached sporozoites before methanol fixation. In another approach, attachment and invasion were quantified by epifluorescence in viable HT-29 cells either at 37°C or 4°C. Because parasite invasion at 4°C is blocked,32 respective inhibitory action of IFN-␥ on attachment and invasion could therefore be determined.
Viability Assay The cytotoxic action of IFN-␥, IL-1, and TNF-␣ on Caco-2, HT-29, and H4 cell lines was determined by microscopic inspection for morphology and trypan blue exclusion. In addition, reduction of MTT (3-[4,5dimethylthiazol-2-yl]-2,5diphenyltetrazolium bromide) to formazan was assayed in HT-29 and Caco-2 cell lines grown in 96-well tissue culture plates treated with the maximal working concentration of each cytokine in supplemented medium. After cytokine exposure, medium was removed and replaced with medium containing 0.2 mg/mL MTT and incubated at 37°C for 30 minutes. Cells were then solubilized with dimethyl sulfoxide, and MTT conversion to formazan was quantified by measurement of optical density at 550 nm. Results are expressed as a percentage of untreated controls.
Western Blot Analysis for IFN-␥R Caco-2, HT-29, and H4 cells were grown to confluence in 75-cm2 flasks and exposed to lysis buffer (25 mmol/L Tris [pH 8.0], 1 mmol/L EDTA [pH 8.0], 150 mmol/L NaCl, 1% Triton, 0.5% NP-40, and the protease inhibitors 10 g/mL phenylmethylsulfonyl fluoride, 1 g/mL aprotonin, 1 g/mL leupeptin). Clear supernatants were collected by centrifugation, and the protein concentration was determined by the Bradford protein assay (Bio-Rad Ltd., Hemel Hempstead, England). Total protein (100 –200 g) from the 3 cell lines was resolved by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. Nonspecific binding was blocked by overnight incubation in PBS and 0.5% Tween 20 (PBS-T) with 2% BSA. After a 2-hour incubation with the primary antibody (2 g/mL rabbit anti–IFN-␥R ␣ chain; Santa Cruz Biotechnology, Santa Cruz, CA), the membrane was washed 3 times in PBS-T, followed by incubation with horseradish peroxidase– conjugated goat anti-rabbit IgG (1:2000 dilution; Dako Ltd., Cambridge, England) for 2 hours. Following 3 washes in Tris-buffered saline, the reaction was developed using diaminobenzene as substrate.
102
POLLOK ET AL.
GASTROENTEROLOGY Vol. 120, No. 1
Table 1. Effect of Various Cytokines on C. parvum Infection of HT-29 Cells Cytokine IFN-␥ IL-1 TNF-␣ IFN-␥ ⫹ IL-1 IFN-␥ ⫹ TNF-␣
Concentration
% Inhibition (⫾SD)
95% CI
100 U/mL 0.1 ng/mL 1 ng/mL 100 U/mL ⫹ 0.1 ng/mL 100 U/mL ⫹ 1 ng/mL
4.8a
73.8 ⫾ 33.1 ⫾ 2.1a 61.0 ⫾ 5.2a 72.8 ⫾ 4.1a
64.2–83.4 28.9–37.3 50.6–71.4 64.6–81.0
81.2 ⫾ 7.1a
67.0–95.4
NOTE. Cells were cultured in medium containing cytokine(s) for 24 hours before parasite exposure. aInfection in cytokine treated replicates was significantly different from controls, P ⬍ 0.01.
Analysis of Data Results were analyzed by comparison of means using Student t test or analysis of variance (ANOVA) and are expressed as mean ⫾ SD or 95% confidence interval where stated.
Results IFN-␥ Inhibits C. parvum Development IFN-␥ was found to have a marked inhibitory effect on infection of HT-29 cell; for example, in cells exposed to 100 U/mL IFN-␥ for 24 hours before parasite inoculation, there was a 73.8% ⫾ 4.8% inhibition of parasite development (P ⬍ 0.001; Table 1). The level of inhibition was dependent on the exposure time of HT-29 cells to IFN-␥, with highest levels of inhibition being obtained after 24 –72 hours of exposure of cells to IFN-␥ before parasite inoculation (data not shown). Experiments were performed to determine the effect of IFN-␥ when added 3 hours before, or 2–19 hours after parasite inoculation (Figure 1). Infection of cells exposed to
Figure 2. Effect of concentration of IFN-␥ on inhibition of C. parvum infection in 3 different cell lines: HT-29 ( ), Caco-2 (䊐), and H4 cells (■). Each cell line was incubated with medium containing 100 U/mL IFN-␥ for 24 hours before parasite inoculation.
IFN-␥ starting 3 hours before, or 2 hours after, parasite inoculation was significantly reduced (P ⬍ 0.0005) but to a lesser extent compared with exposure to the cytokine starting 24 hours before inoculation. In contrast, cells exposed to IFN-␥ starting 19 hours after parasite inoculation showed no significant resistance to infection. No significant difference between IFA and Giemsa quantification was found (Figure 1). The effect of varying the concentration of IFN-␥ on C. parvum development in 3 different cell lines is shown in Figure 2. A marked difference in the effect of this cytokine on infection was observed among the 3 cell lines, with the greatest inhibition occurring in HT-29 cells (ANOVA, P ⬍ 0.001) and a more moderate effect being observed in Caco-2 cells (ANOVA, P ⬍ 0.01). In contrast, no significant effect of IFN-␥ on C. parvum infection of H4 cells was found. IFN-␥–mediated inhibition of infection increased in a dose-dependent manner in both HT-29 and Caco-2 cells, with maximal effect being obtained with the largest IFN-␥ dose tested, 1000 U/mL, in both HT-29 (ANOVA, P ⬍ 0.001) and Caco-2 cells (ANOVA, P ⬍ 0.01). Action of IFN-␥ Requires Expression of IFN-␥ Receptor by Epithelial Cells
Figure 1. Effect of duration of IFN-␥ treatment of HT-29 cells on C. parvum infection. Cells were incubated with medium containing 100 U/mL IFN-␥ from between 24 hours before parasite inoculation to 19 hours after parasite inoculation. Parasites were enumerated by Giemsa staining ( ) and IFA ( ).
To determine whether variation in IFN-␥–mediated inhibition of C. parvum infection among the cell lines tested reflected differences in the level of expression of IFN-␥R, Western blot analysis was used to examine expression of the heteromeric IFN-␥R in each cell line grown under similar conditions as those used for parasite infection experiments (Figure 3). Variation was observed in the level of expression of the IFN-␥R among HT-29, Caco-2, and H4 cells, which reflected the relative sus-
January 2001
IFN-␥–MEDIATED INHIBITION OF C. parvum INFECTION
103
Table 2. Effect of IFN-␥ on C. parvum Attachment and Invasion of HT-29 Cells Monolayer
Control (SD)
IFN-␥ treated (SD)
Inhibition of control (%)
Fixed (37°C) Viable (37°C) Viable (4°C)
296.2 (⫾45.6) 1005.5 (⫾45.4) 474.4 (⫾38.6)
296.0 (⫾26.8) 609 (⫾39.4) 454.6 (⫾54.6)
0.1 (⫾9.1) 39.4 (⫾4.8)a ⫺4.4 (8.5)
aInvasion in IFN-␥–treated replicates at 37°C was significantly different from untreated controls, P ⬍ 0.001.
Figure 3. Differences in expression of IFN-␥ R (␣-chain) in H4, Caco-2, and HT-29 cell lines determined by Western blotting. Molecular weight markers (lane 1) are 43, 80, 116, and 200 kilodaltons. The IFN-␥R is located at approximately 80 kilodaltons, indicated by the arrow ([a] 200 g/mL or [b] 100 g/mL).
ceptibility of each cell line to the inhibitory action of IFN-␥ on infection. IFN-␥R expression was greatest in HT-29 cells, lesser expression was obtained in Caco-2 cells, whereas there was no detectable IFN-␥R expression in H4 cells under the conditions described. Thus, IFN-␥R expression was critical in mediating the inhibitory actions of IFN-␥. To evaluate the role of the JAK/STAT pathway in the cellular signaling mechanism mediating the IFN-␥ inhibitory effect already described, the action of the selective JAK-2 inhibitor tyrphostin B42 was determined. IFN-␥–mediated inhibition of infection was decreased by tyrphostin B42 in a dose-dependent manner (Figure 4; ANOVA, P ⬍ 0.01); tyrphostin B42 alone had no significant effect on parasite development.
Figure 4. The effect of blockade of the JAK/STAT pathway by the selective JAK-2 inhibitor, tyrphostin B42, on IFN-␥–mediated inhibition of C. parvum infection. HT-29 cells were incubated with medium containing 100 U/mL IFN-␥. Parasites were enumerated by IFA.
No Synergistic Action of IFN-␥ With IL-1 or TNF-␣ The effect of treatment with IL-1 or TNF-␣ on C. parvum infection is shown in Table 1. The concentration of each cytokine that produced optimal inhibition of parasite development was determined in preliminary experiments. All 3 cytokines, when used alone, had significantly greater inhibitory effects on C. parvum infection compared with untreated controls, but IFN-␥ and TNF-␣ reduced parasite development to a significantly greater extent than IL-1. No significant additive effect was noted when IFN-␥ was combined with either TNF-␣ or IL-1. None of the cytokines tested had a cytotoxic effect on HT-29 cells as assessed by gross morphology, trypan blue exclusion, and MTT reduction assay. Mechanism of Action of IFN-␥ Action of IFN-␥ is dependent on parasite invasion but not attachment. Experiments were performed to
examine whether IFN-␥ exerted its effect through events involving host-parasite attachment or invasion (Table 2). No significant inhibition of attachment of sporozoites was observed in paraformaldhyde-fixed HT-29 cells that had been pretreated with IFN-␥ compared with untreated controls. At 4°C, at which temperature parasite invasion is blocked,33 no significant difference in the numbers of parasites attached to viable cells was found with or without IFN-␥ pretreatment of cells. These results suggest parasite attachment is not affected by IFN-␥. At 37°C, a temperature which allows invasion to occur, significantly fewer parasites were observed in viable cells pretreated with IFN-␥ compared with untreated controls (inhibition of penetration, 40.4% ⫾ 3.4%; P ⬍ 0.01). This suggests that an important mechanism of action of IFN-␥ was to prevent penetration of the host cell by the parasite. The inhibitory action of IFN-␥ on early intracellular development of C. parvum at 4 hours after infection was quantified using Giemsa staining (Figure 5). At 4 hours after infection, fewer developing parasites in IFN-␥–
104
POLLOK ET AL.
Figure 5. Inhibition of early development of C. parvum by IFN-␥. Intracellular development was quantified 4 and 24 hours after infection in HT-29 cells cultured in the presence ( ) or absence (䊐) of IFN-␥.
treated cells were observed, with an inhibition of 74.7% ⫾ 6.6% compared with untreated controls (P ⬍ 0.001). Similarly, at 24 hours, infection in IFN-␥–treated cells was inhibited 72.5% ⫾ 9.0% compared with untreated cells (P ⬍ 0.001). However, the absolute parasite count had increased in both IFN-␥–treated and untreated control wells at 24 hours compared with 4 hours after infection. This result suggests that the higher level of inhibition consistently obtained in these experiments compared with that found in the study of invasion indicates that IFN-␥ retards intracellular development, as well as invasion. Action of IFN-␥ is independent of NO synthase activity. The role of NOS activity in mediating the
inhibitory action of IFN-␥ on infection was evaluated using the NOS inhibitors NG-MMA, L-NNA, and the selective iNOS inhibitor 1400 W. NG-MMA was used at a dose previously found to block IFN-␥–mediated inhibition of Toxoplasma gondii replication in rodent enterocytes,6 whereas L-NNA and 1400 W were used at doses previously established to inhibit NOS activity in enterocytes.4 No significant increase in infection was noted in wells treated with both IFN-␥ and any of the NOS inhibitors compared with wells treated with IFN-␥ alone (Figure 6), and none of the NOS inhibitors alone had a significant effect on infection. These findings indicate that in this in vitro model, the IFN-␥–mediated inhibition of C. parvum development in enterocytes is independent of NOS activity.
GASTROENTEROLOGY Vol. 120, No. 1
Figure 6. Effect of the NOS inhibition on IFN-␥–mediated inhibition of C. parvum infection. ■, IFN-␥ alone (100 U/mL); 䊐, IFN-␥ and highconcentration NOS inhibitor; z, IFN-␥ and low-concentration NOS inhibitor; , high-concentration NOS inhibitor alone.
shown to increase T. gondii infection of IFN-␥–treated fibroblasts,27 had no significant effect on the IFN-␥– induced inhibition of infection of HT-29 cells (results not shown). However, IFN-␥–mediated inhibition of infection was moderately decreased by exogenous FeSO4 in a dose-dependent fashion (ANOVA, P ⬍ 0.001) (Figure 7). FeSO4 alone (12.5–200 mol/L) had no significant effect on infection.
Discussion Our findings indicate for the first time mechanisms by which IFN-␥ mediates its inhibitory action in the control of C. parvum infection in intestinal epithelial cells. Using an in vitro model of infection we have demonstrated a direct inhibitory action of IFN-␥ on C. parvum–infected human enterocytes independent of cell-
Mechanism of IFN-␥–mediated inhibition of infection is dependent on cellular Fe2ⴙ but not tryptophan.
To evaluate the role of cellular Fe2⫹ concentration and tryptophan depletion in mediating the action of IFN-␥, the effect of exogenous tryptophan or Fe2⫹ on the IFN␥–induced inhibition of C. parvum was determined. Addition of exogenous tryptophan, at a dose previously
Figure 7. Effect of FeSO4 on IFN-␥–mediated inhibition of C. parvum infection. HT-29 cells were incubated with medium containing 100 U/mL IFN-␥.
January 2001
to-cell interactions with other mucosal immune effector cells. The inhibitory effect of IFN-␥ against C. parvum infection is dependent on both the concentration of cytokine present and the duration of exposure of enterocytes to the cytokine. Compared with cells treated with IFN-␥ starting 24 hours before parasite inoculation, a smaller but significant inhibition of infection was noted when cells were treated with IFN-␥ around the time of parasite inoculation. However, no inhibitory action was noted when cells were exposed to IFN-␥ starting 19 hours after parasite inoculation, i.e., 5 hours before the end of the experiment. This latter observation reflects the minimum period of time required for IFN-␥ to mediate phenotypic changes in the enterocyte (6 –12 hours) reported by others.3 IFN-␥ also inhibits the intracellular development of viral (e.g., rotavirus) and bacterial pathogens (e.g., Shigella dysenteriae) in epithelial cells.7,8 Development of the related parasites Eimeria vermiformis and T. gondii in rodent epithelial cell lines may also be inhibited by IFN-␥.6,9 In contrast with results obtained with HT-29 and Caco-2 cells, C. parvum infection of the human primary fetal intestinal cell line H4 was refractory to the action of IFN-␥. The lack of expression of IFN-␥R by H4 cells probably explains the inability of IFN-␥ to inhibit parasite development in these cells. Coincidentally, the failure of IFN-␥ to inhibit C. parvum infection in H4 cells effectively excludes a direct inhibitory effect of the cytokine on the parasite. Previous work indicating a differential expression of the IFN-␥R in HT-29 and Caco-2 cell lines was confirmed in the present study and may explain why IFN-␥–mediated inhibition of parasite development was more effective in HT-29 cells.34 IFN-␥ is known to mediate its action at the cellular level by binding IFN-␥R, resulting in the activation of STAT1␣, by JAK2.35 Further evidence for the specific action of IFN-␥R in the control of C. parvum infection of enterocytes is indicated by the partial reversal of the IFN-␥ inhibitory action on infection by a specific JAK 2 inhibitor, tyrphostin B42.36 IFN-␥ may up-regulate iNOS gene expression and NO synthesis in enterocytes, but inhibition of NOS had no effect on the action of IFN-␥ on C. parvum infection in our in vitro model. In addition, although a synergistic effect of IL-1 or TNF-␣ with IFN-␥ in the induction of iNOS by enterocytes has been reported,3,4,37 no additive effect of these cytokines was found in the control of C. parvum infection in our study. Our findings contrast with recent work indicating IFN-␥–mediated inhibition of Listeria monocytogenes infection in Caco-2 cells is dependent on cellular NOS activity.38 The reason for this
IFN-␥–MEDIATED INHIBITION OF C. parvum INFECTION
105
difference is unclear but may reflect the extracytosolic location of C. parvum within the cell. Findings on the role for reactive nitrogen species in the control of C. parvum infection in animal models are contradictory.39,40 A recent report found neonatal iNOS⫺/⫺ mice had a moderately increased susceptibility to infection compared with control animals,41 but it is unclear if the persistence of infection in these mice was related to a lack of iNOS expression in mouse enterocytes or another cell type. We examined other possible mechanisms for the antiparasitic action of IFN-␥ observed in our model of C. parvum infection. In the early stages of infection, IFN-␥ had no effect on attachment of C. parvum sporozoites to enterocytes, but was found to significantly reduce the level of host cell invasion (Table 2). Studies of host cell invasion by rotavirus have also indicated a role for IFN-␥ in inhibiting epithelial cell membrane penetration.7 Both IFN-␥ and IFN-␣ have been shown to inhibit Shigella infection of epithelial cells8,42; IFN-␣ exerts its effect by inhibition of the Src-dependent signaling cascade normally triggered by Shigella that leads to host cell cytoskeletal rearrangement and bacterial uptake.42 Invasion and the formation of the parasitophorous vacuole result from membrane remodeling through actin polymerization, and inhibitors of cytoskeletal rearrangement may also inhibit C. parvum invasion of host cells.43 In other coccidial infections, inhibition by IFN-␥ of invasion of nonenterocyte rodent epithelial cells by E. vermiformis has been observed, but in T. gondii infection of rodent enterocytes, no effect of IFN-␥ on host cell invasion was reported.6,9 IFN-␥ consistently had a greater inhibitory effect on development up to 4 hours after infection than invasion alone, which indicated that another important effect of the cytokine was to retard development of intracellular parasites. Inhibition of T. gondii infection of fibroblasts by IFN-␥ results from depletion of cellular tryptophan; this effect may be reversed by the addition of exogenous tryptophan.24 However, this mechanism did not play a role in T. gondii infection of rodent enterocytes6 and, in the present investigation, did not appear to be involved in the inhibition of C. parvum infection in human enterocytes. In contrast, reduction of the availability of cellular Fe2⫹ by IFN-␥ seemed to be a protective mechanism because Fe2⫹ supplementation of culture medium was found to partially reverse the action of IFN-␥ on C. parvum reproduction. However, this effect was smaller than that observed with T. gondii infection of rodent enterocytes,6 which might reflect the host species of enterocytes or the variable requirements for endogenous Fe2⫹ by the respective parasites.
106
POLLOK ET AL.
Evidence from our recent studies suggests IFN-␥ production by intraepithelial lymphocytes (IELs) is crucial in the control of cryptosporidial infection in mice17,18; furthermore, other investigators have shown that recovery from infection correlates with the production of IFN-␥ in the intestinal mucosa.44,45 Taken together with these previous observations, our present findings therefore suggest that immune IELs acting at the site of infection may exert control of cryptosporidial infection through the production of IFN-␥, which acts directly on enterocytes to inhibit parasite invasion and intracellular development. The identification of an IFN-␥–regulated immune effector role for the enterocyte explains, at least in part, the requirement for IFN-␥ production in the immunologic control of cryptosporidiosis and may have potential immunotherapeutic applications in the future.
GASTROENTEROLOGY Vol. 120, No. 1
15.
16.
17.
18.
19.
20.
References 1. Jung HC, Eckmann L, Yang SK, Panja A, Fierer J, MorzyckaWroblewska, Kagnoff MF. A distinct array of proinflammatory cytokines is expressed in human colon epithelial cells in response to bacterial invasion. J Clin Invest 1995;95:55– 65. 2. Madara JL, Stafford J. Interferon-gamma directly affects barrier function of cultured intestinal epithelial monolayers. J Clin Invest 1989;83:724 –727. 3. Salzman AL, Eaves-Pyles T, Linn SC, Denenberg AG, Szabo C. Bacterial induction of inducible nitric oxide synthase in cultured human intestinal epithelial cells. Gastroenterology 1998;114: 93–102. 4. Cavicchi M, Whittle BJR. Potentiation of cytokine induced iNOS expression in the human intestinal epithelial cell line, DLD-1, by cyclic AMP. Gut 1999;45:367–374. 5. Fish SM, Proujansky R, Reenstra WW. Synergistic effects of interferon gamma and tumour necrosis factor alpha on T84 cell function. Gut 1999;45:191–198. 6. Dimier IH, Bout DT. Interferon-gamma-activated primary enterocytes inhibit Toxoplasma gondii replication: a role for intracellular iron. Immunology 1998;94:488 – 495. 7. Bass DM. Interferon gamma and interleukin 1, but not interferon alfa, inhibit rotavirus entry into human intestinal cell lines. Gastroenterology 1997;113:81– 89. 8. Niesel DW, Hess CB, Cho YJ, Klimpel KD, Klimpel GR. Natural and recombinant interferons inhibit epithelial cell invasion by Shigella spp. Infect Immun 1986;52:828 – 833. 9. Rose ME, Smith AL, Wakelin D. Gamma interferon-mediated inhibition of Eimeria vermiformis growth in cultured fibroblasts and epithelial cells. Infect Immun 1991;59:580 –586. 10. Molbak K, Hojlyng N, Gottschau A, Sa JC, Ingholt L, da Silva AP. Cryptosporidiosis in infancy and childhood mortality in Guinea Bissau, West Africa. BMJ 1993;307:417– 420. 11. Blanshard C, Jackson AM, Shanson DC, Francis N, Gazzard BG. Cryptosporidiosis in HIV-seropositive patients. Q J Med 1992;85: 813– 823. 12. Flanigan T, Whalen C, Turner J, Soave R, Toerner J, Havlir D, Kotler D. Cryptosporidium infection and CD4 counts. Ann Intern Med 1992;116:840 – 842. 13. McDonald V, Bancroft GJ. Immunological control of Cryptosporidium infection. Chem Immunol 1998;70:103–123. 14. Theodos CM, Sullivan KL, Griffiths JK, Tzipori S. Profiles of healing and nonhealing Cryptosporidium parvum infection in C57BL/6 mice with functional B and T lymphocytes: the extent of
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
gamma interferon modulation determines the outcome of infection. Infect Immun 1997;65:4761– 4769. Ungar BL, Kao T, Burris JA, Finkelman FD. Cryptosporidium infection in an adult mouse model. Independent role for interferongamma and CD4⫹ T-lymphocytes in protective immunity. J Immunol 1991;147:1014 –1022. Perryman LE, Mason PH, Chrisp CE. Effect of spleen cell populations in resolution of Cryptosporidium parvum infection in SCID mice. Infect Immun 1994;62:1474 –1477. McDonald V, Robinson HA, Kelly JP, Bancroft GJ. Immunity to Cryptosporidium muris infection in mice is expressed through gut CD4⫹ intraepithelial lymphocytes. Infect Immun 1996;64: 2556 –2562. Culshaw RJ, Bancroft GJ, McDonald V. Gut intraepithelial lymphocytes induce immunity against Cryptosporidium infection through a mechanism involving gamma interferon production. Infect Immun 1997;65:3074 –3079. Gomez Morales MA, Ausiello CM, Guarino A, Urbani F, Spagnuolo MI, Pignata C, Pozio E. Severe, protracted intestinal cryptosporidiosis associated with interferon gamma deficiency: pediatric case report. Clin Infect Dis 1996;22:848 – 850. Gooi HC. Gastrointestinal and liver involvement in primary immunodeficiency. In: Heatley RV, ed. Gastrointestinal and hepatic immunology. Cambridge Review Clinic Immunology. Cambridge, England: Cambridge University Press, 1994:169 –177. Reinecker HC, Podolsky DK. Human intestinal epithelial cells express functional cytokine receptors sharing the common gamma chain of the interleukin 2 receptor. Proc Natl Acad Sci U S A 1995;92:8353– 8357. Raqib R, Lindberg AA, Bjork L, Bardham PK, Wretlind B, Andersson U, Andersson J. Down-regulation of gamma interferon, tumor necrosis factor type I, interleukin 1 (IL-1) type I, IL-3, IL-4, and transforming growth factor beta type I receptors at the local site during the acute phase of Shigella infection. Infect Immun 1995; 63:3079 –3087. Reinecker HC, MacDermott RP, Mirau S, Dignass A, Podolsky DK. Intestinal epithelial cells both express and respond to interleukin 15. Gastroenterology 1996;111:1706 –1713. Ceravolo IP, Chaves AC, Bonjardim CA, Sibley D, Romanha AJ, Gazzinelli RT. Replication of Toxoplasma gondii, but not Trypanosoma cruzi, is regulated in human fibroblasts activated with gamma interferon: requirement of a functional JAK/STAT pathway. Infect Immun 1999;67:2233–2240. Proudfoot L, Nikolaev AV, Feng GJ, et al. Regulation of the expression of nitric oxide synthase and leishmanicidal activity by glycoconjugates of Leishmania lipophosphoglycan in murine macrophages. Proc Natl Acad Sci U S A 1996;93:10984 –10989. Lane TE, Wu-Hsieh BA, Howard DH. Antihistoplasma effect of activated mouse splenic macrophages involves production of reactive nitrogen intermediates. Infect Immun 1994;62:1940 – 1945. Gupta SL, Carlin JM, Pyati P, Dai W, Pfefferkorn ER, Murphy MJ Jr. Antiparasitic and antiproliferative effects of indoleamine 2,3dioxygenase enzyme expression in human fibroblasts. Infect Immun 1994;62:2277–2284. Griffiths JK, Moore R, Dooley S, Keusch GT, Tzipori S. Cryptosporidium parvum infection of Caco-2 cell monolayers induces an apical monolayer defect, selectively increases transmonolayer permeability, and causes epithelial cell death. Infect Immun 1994;62:4506 – 4514. Pollok RC, MacDonald V, Farthing MJ. Direct action of gammainterferon in inhibiting infection of intestinal epithelial cells by Cryptosporidium parvum (abstr). Gastroenterology 1999;116: G3090. Joe A, Verdon R, Tzipori S, Keusch GT, Ward HD. Attachment of Cryptosporidium parvum sporozoites to human intestinal epithelial cells. Infect Immun 1998;66:3429 –3432.
IFN-␥–MEDIATED INHIBITION OF C. parvum INFECTION
January 2001
31. Hamer DH, Ward H, Tzipori S, Pereira ME, Alroy JP, Keusch GT. Attachment of Cryptosporidium parvum sporozoites to MDCK cells in vitro. Infect Immun 1994;62:2208 –2213. 32. Chen XM, LaRusso NF. Mechanisms of attachment and internalization of Cryptosporidium parvum to biliary and intestinal epithelial cells. Gastroenterology 2000;118:368 –379. 33. Chen XM, Levine SA, Tietz P, et al. Cryptosporidium parvum is cytopathic for cultured human biliary epithelia via an apoptotic mechanism. Hepatology 1998;28:906 –913. 34. Chavez AM, Morin MJ, Unno N, Fink MP, Hodin RA. Acquired interferon gamma responsiveness during Caco-2 cell differentiation: effects on iNOS gene expression. Gut 1999;44:659 – 665. 35. Weber-Nordt RM, Mertelsmann R, Finke J. The JAK-STAT pathway: signal transduction involved in proliferation, differentiation and transformation. Leuk Lymphoma 1998;28:459 – 467. 36. Meydan N, Grunberger T, Dadi H, et al. Inhibition of acute lymphoblastic leukemia by a Jak-2 inhibitor. Nature 1996;379:645– 648. 37. Kolios G, Rooney N, Murphy CT, Robertson DA, Westwick J. Expression of inducible nitric oxide synthase activity in human colon epithelial cells: modulation by T lymphocyte derived cytokines. Gut 1998;43:56 – 63. 38. Ouadrhiri Y, Sibille Y, Tulkens PM. Modulation of intracellular growth of Listeria monocytogenes in human enterocyte Caco-2 cells by interferon-gamma and interleukin-6: role of nitric oxide and cooperation with antibiotics. Infect Immun 1999;180:1195– 1204. 39. Kuhls TL, Mosier DA, Abrams VL, Crawford DL, Greenfield RA. Inability of interferon-gamma and aminoguanidine to alter Cryptosporidium parvum infection in mice with severe combined immunodeficiency. J Parasitol 1994;80:480 – 485. 40. Leitch GJ, He Q. Arginine-derived nitric oxide reduces fecal oocyst
41.
42.
43.
44.
45.
107
shedding in nude mice infected with Cryptosporidium parvum. Infect Immun 1994;62:5173–5176. Leitch GJ, He Q. Reactive nitrogen and oxygen species ameliorate experimental cryptosporidiosis in the neonatal BALB/c mouse model. Infect Immun 1999;67:5885–5891. Dumenil G, Olivo JC, Pellegrini S, et al. Interferon alpha inhibits a Src-mediated pathway necessary for Shigella-induced cytoskeletal rearrangements in epithelial cells. J Cell Biol 1998;143: 1003–1012. Forney JR, DeWald DB, Yang S, Speer CA, Healey MC. A role for host phosphoinositide 3-kinase and cytoskeletal remodeling during Cryptosporidium parvum infection. Infect Immun 1999;67: 844 – 852. Kapel N, Benhamou Y, Buraud M, Magne D, Opolon P, Gobert JG. Kinetics of mucosal ileal gamma-interferon response during cryptosporidiosis in immunocompetent neonatal mice. Parasitol Res 1996;82:664 – 667. White AC Jr, Robinson HA, Okhuysen PC, et al. Interferon-gamma expression in jejunal biopsies in experimental human cryptosporidiosis correlates with prior sensitization and control of oocyst expression. J Infect Dis 2000;181:701–709.
Received January 24, 2000. Accepted August 23, 2000. Address requests for reprints to: Richard C. G. Pollok, M. D., Digestive Diseases Research Centre and Department of Paediatric Gastroenterology, St. Bartholomew’s and The Royal London School of Medicine and Dentistry, Turner Street, London E1 2AD, England. e-mail:
[email protected]; fax: (44) 20-7882-7192. Supported by a grant from the Wellcome Trust Research Training Fellow (to R.C.G.P.) and partly funded by a Wellcome Trust project grant (to M.B.-E. and V.M).
