Inhibition of Apoptosis by Gamma Interferon in Cells and Mice Infected with Chlamydia muridarum (the Mouse Pneumonitis Strain of Chlamydia trachomatis)

The effect of gamma interferon (IFN- ${gamma}$ ) on apoptosis due to infectionby Chlamydia muridarum (the mouse pneumonitis strain of Chlamydiatrachomatis) was studied in epithelial cells in culture andin the genital tracts of mice. IFN- ${gamma}$ concentrations that inducethe formation of aberrant, persistent chlamydiae inhibit apoptosisdue to C. muridarum infection. In cells treated with an IFN- ${gamma}$ concentration that leads to the development of a heterogenouspopulation of normal and aberrant Chlamydia vacuoles, apoptosiswas inhibited preferentially in cells that contained the aberrantvacuoles. The inhibitory effect of IFN- ${gamma}$ appears to be due inpart to expression of host cell indoleamine 2,3-dioxygenaseactivity, since inhibition of apoptosis could be partially reversedthrough coincubation with exogenous tryptophan. Apoptotic cellswere observed in the genital tracts of wild-type mice infectedwith C. muridarum, and a significantly larger number of apoptoticcells was detected in infected IFN- ${gamma}$ -deficient mice. These resultssuggest that IFN- ${gamma}$ may contribute to pathogenesis of persistentChlamydia infections in vivo by preventing apoptosis of infectedcells.

INTRODUCTION

Top

Introduction

Chlamydia trachomatis strains include a causative agent of ocularinfection in humans and the most common cause of sexually transmittedbacterial infections (20, 54). An inflammatory response is requiredfor the resolution of primary C. trachomatis infection, butchronic inflammation is also responsible for the scarring processobserved in trachoma and chlamydial sexually transmitted disease(2). In addition, the ability of chlamydiae to persist in thehost could be an important factor in exacerbating pathogenesis(4, 6). The immune mechanisms responsible for chronic inflammationare not fully understood, but it is believed that repeated exposureto chlamydial antigens contributes to pathogenesis and thatbacteria in persistently infected cells may serve as a sourceof long-lasting pathology (5).

Chlamydia muridarum (also known as the mouse pneumonitis strainof C. trachomatis) (14, 48) infects the genital tracts of miceand has been used to characterize the host immune response tochlamydial genital tract infection. In C. muridarum infectionsof mice, a major role for CD4⁺ T cells secreting gamma interferon(IFN- ${gamma}$ ) during clearance of the infection has been described(25, 31, 42, 52). Thus, studies with IFN- ${gamma}$ -deficient mice haveshown that IFN- ${gamma}$ is required for prevention of disseminationof genital tract infection (9, 26, 42). Similarly, administrationof anti-IFN- ${gamma}$ antibodies or recombinant IFN- ${gamma}$ prolongs or resolvesthe infection, respectively (45).

Chlamydiae are obligate intracellular bacteria that have a distinctbiphasic infection cycle: the small elementary bodies are infectiousand metabolically inert, and the larger reticulate bodies (RB)are noninfectious and metabolically active (22, 32, 56). A persistentstate of chlamydiae can be induced in vitro by IFN- ${gamma}$ treatment,which leads to altered bacterial forms and an antigen profilethat is different from that observed during active infection(4, 6). Significantly, there is down-regulation of the chlamydialmajor outer membrane protein (MOMP), and expression increasesor decreases for several other proteins and genes, whereas expressionof the immunostimulatory chlamydial antigen, hsp60, remainsat high levels during the persistent state (6, 27, 49). Theintracellular chlamydiae undergo striking morphological changes,becoming much larger than conventional RB (5, 49). In IFN- ${gamma}$ -treatedinfected cells, the chlamydiae are noninfectious, but the bacteriaare viable and can revert to the infectious state followingremoval of IFN- ${gamma}$ (3).

Several Chlamydia spp. have joined a growing number of intracellularmicroorganisms that modulate apoptosis in opposite directions,both promoting and inhibiting apoptosis under different conditions(8, 17, 19, 23, 35, 55). Similar to other pathogens, chlamydiaeprotect infected cells against apoptosis due to external stimuliduring early stages of infection (15) and induce apoptosis ofthe host cell during later stages (21, 38, 40, 41). Chlamydiaemay thus protect infected cells against cytotoxic mechanismsof the immune system, while the apoptosis observed at the endof the infection cycle may contribute to the inflammatory responseor facilitate release of bacteria from infected cells.

It has recently been reported that cells infected with C. trachomatisin the presence of IFN- ${gamma}$ resist apoptosis due to external ligands,via inhibition of caspase activation (11). Inhibition of caspaseactivation has also been shown in the absence of IFN- ${gamma}$ (15),while the apoptosis due to the infection itself is independentof caspases (38, 41). To determine if persistently infectedcells are still susceptible to apoptosis at the end of the infectioncycle, we characterized apoptosis of cells infected with C.muridarum in the presence or absence of IFN- ${gamma}$ in vitro. The studywas then extended to an in vivo infection by measuring the numberof apoptotic cells in the genital tracts of wild-type and IFN- ${gamma}$ -deficientmice infected with the same strain of C. muridarum. In bothcases, IFN- ${gamma}$ inhibited Chlamydia-induced apoptosis, which couldcontribute to limiting propagation of the infection during thestate of persistence.

MATERIALS AND METHODS

Top

Materials and Methods

Cells and animals. HeLa 229 cells (obtained from the American Type Culture Collection,Manassas, Va.) were cultivated in Dulbecco's modified minimalessential medium with Glutamax-1 (Life Technologies, Inc., Rockville,Md.), supplemented with 10% heat-inactivated fetal calf serumand 25 µg of gentamicin per ml. C. muridarum, obtainedoriginally from the American Type Culture Collection, was grownin McCoy cells and purified as previously described (44). Six-week-oldfemale C57BL/6 (C57; H-2^b) and IFN- ${gamma}$ -deficient female C57BL/6-Ifg^tm1(IFN- ${gamma}$ ^-/-) mice were purchased from The Jackson Laboratory (BarHarbor, Maine). Recombinant human IFN- ${gamma}$ was obtained from PeproTech(Rocky Hill, N.J.), and L-tryptophan was from Aldrich (SaintQuentin Fallavier, France). All other reagents were previouslydescribed (40).

Infections of cells with C. muridarum and treatment with IFN- ${gamma}$ . For infections in vitro, HeLa cells were plated at a densityof 2 x 10⁴ cells per well in six-well plates (Costar). After24 h in a humidified incubator at 37°C with 5% CO₂, thecell monolayers were washed and infected at a multiplicity ofinfection (MOI) of 0.5. At 1 h postinfection, the inoculum wasremoved and replaced with fresh culture medium. Infected monolayerswere maintained in the incubator for 2 days or treated 3 h postinfectionwith various concentrations of IFN- ${gamma}$ . In some experiments, 200µg of L-tryptophan per ml was added to the cell culturesat the same time as the IFN- ${gamma}$ (3, 28).

Immunofluorescence. Infected or uninfected cells were fixed in 4% paraformaldehydeand revealed with anti-Chlamydia fluorescein-conjugated antibodies,consisting of a mixture of two monoclonal antibodies specificfor the chlamydial MOMP and chlamydial lipopolysaccharide (1:500dilution; Argene, Varilhes, France), as previously described(37, 38). Host cell and bacterial DNAs were visualized by stainingsaponin-permeabilized cells with 5 µg of Hoechst 33258(Sigma, St. Louis, Mo.) per ml. Samples were examined with anAxiovert 135 TV fluorescence microscope (Zeiss, Oberkochen,Germany) attached to a cooled charge-coupled device camera (Photometrics).

To estimate the number of cells with persistent chlamydiae,infected samples were fixed and incubated with an anti-Chlamydiamonoclonal antibody, as described above. Using the scale of1 to 4 for the level of persistent chlamydial development definedpreviously (3), normal cells correspond to cells with inclusionscontaining typical chlamydia (level 1), while persistently infectedcells included all the cells with inclusions with some evidenceof atypical, enlarged chlamydiae (levels 2 and 3) or exclusivelyatypical chlamydiae (level 4). As the cells were also labeledwith Hoechst, apoptotic cells in the same samples were identifiedby their condensed nuclei.

Analysis of cell death. In vitro, apoptosis was measured by cytofluorimetry using detergent-permeabilizedcells stained with propidium iodide (PI), as previously described(13, 29, 33, 36). Since apoptotic cells were present mainlyin the supernatant (38), adherent cells were detached with 1mM EDTA in phosphate-buffered saline at 37°C, and both adherentcells and cells in suspension were collected and centrifuged.After one washing with phosphate-buffered saline, the cell pelletwas resuspended directly in PI-detergent buffer (62.5 µgof PI per ml, 0.1% Triton X-100, and 1 mg of sodium citrateper ml in water) and transferred into 12- by 75-mm Falcon 2052fluorescence-activated cell sorting tubes (Becton Dickinson,San Jose, Calif.). Data from 10,000 HeLa cells were collectedon a FACScan flow cytometer (Becton Dickinson) with an argonion laser tuned to 488 nm. Apoptosis was measured in the FL-3range, and the marker used to quantify the number of apoptoticcells was set in such a way as to exclude debris and necroticcells in the low FL-3 values. Each sample was performed in triplicate,and each experiment was repeated at least three times on separatedays.

Infection of mice and histological procedures. Wild-type or IFN- ${gamma}$ -deficient mice received 2.5 mg of medroxyprogesteroneacetate (Depo-Provera; Upjohn, Kalamazoo, Mich.) subcutaneously7 days before vaginal infection. The mice were infected by placing30 µl of 250 mM sucrose-10 mM sodium phosphate-5 mM L-glutamicacid (SPG) containing 1.0 x 10⁷ inclusion-forming units of C.muridarum per ml into the vagina of each mouse or 30 µlof SPG without bacteria for negative controls, as describedelsewhere (10). Mice were sacrificed 7 days after vaginal infection.Oviducts and uterine horns were removed, fixed in 4% paraformaldehyde,and embedded in 37°C paraffin. Longitudinal 4-µm sectionswere cut, deparaffinized, and hydrated through xylene and gradedalcohol series.

Apoptotic cells were identified by staining with the terminaldeoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling(TUNEL) method using the apoptosis detection kit from BoehringerMannheim (Meylan, France), following the manufacturer's instructions.Samples were examined with the Zeiss microscope, and imageswere acquired and analyzed with the IPLab spectrum program (SignalAnalytics Corporation, Vienna, Va.). The relative number ofapoptotic cells in each microscope field was quantified withthe IPLab spectrum program, as previously described (40).

RESULTS

Top

Results

Effect of IFN- ${gamma}$ on survival of Chlamydia-infected cells in vitro. Several strains of C. trachomatis and C. muridarum and the guineapig inclusion conjunctivitis strain of Chlamydia psittaci induceapoptosis of host cells towards the end of the infection cycle(21, 38, 40). IFN- ${gamma}$ has been reported to induce a state of persistenceof C. muridarum, human strains of C. trachomatis, and Chlamydiapneumoniae (5, 30, 39), although human strains of C. trachomatisare more sensitive than C. muridarum (30). When incubated 24or 48 h before the start of infection, IFN- ${gamma}$ had a partial effecton growth of C. muridarum at 0.5 ng/ml and a larger effect at2.0 ng/ml. The IFN- ${gamma}$ concentrations used in this work had noeffect on uninfected cells (data not shown), but IFN- ${gamma}$ concentrationsabove 5 ng/ml were toxic to the host (HeLa) cells themselves(30). When cells are infected first and then cultured with IFN- ${gamma}$ ,growth of the chlamydiae is not inhibited, and instead the chlamydiaeassume the biochemical and morphological features of persistence(4). We therefore measured the effect on apoptosis during C.muridarum infection in the presence of IFN- ${gamma}$ at 0, 0.2, and 2ng/ml, which was added 3 h after the start of infection. Apoptosis,defined by DNA content as measured by cytofluorimetry after48 h of infection, was not affected by IFN- ${gamma}$ at 0.2 ng/ml butwas inhibited by approximately half by IFN- ${gamma}$ at 2 ng/ml (Fig.1).

View larger version (20K):
[in this window]
[in a new window]
FIG. 1. Effect of IFN- ${gamma}$ on apoptosis and survival of host cells infected with Chlamydia. HeLa cells were infected with C. muridarum at an MOI of 0.5, and the indicated concentration of IFN- ${gamma}$ was added 3 h after the beginning of infection. After 48 h of infection, specific apoptosis of adherent cells and cells in suspension was measured by cytofluorimetry of PI-treated, detergent-permeabilized cells, as described in Materials and Methods. The marker (M) was used to define the population of apoptotic cells. (A) Cytofluorimetry profile of infected and uninfected cells treated with PI. In the absence of infection or IFN- ${gamma}$ treatment, a spontaneous level of apoptosis of 3 to 5% was always observed after 48 h (filled profile). Over 25% of the cells became apoptotic after a 48 h infection (open profile). (B) Cytofluorimetry profile with expanded y axis to show apoptotic cells in a population of infected cells (dashed line) and infected cells incubated with IFN- ${gamma}$ at 0.2 ng/ml (solid line) or 2 ng/ml (dotted line). The fluorescence intensity (FL-3) is given on the x axis, and the number of cells is on the y axis (A and B). (C) Calculated values of specific apoptosis for infected cells treated with 0, 0.2 or 2 ng of IFN- ${gamma}$ per ml. The values are means and standard deviations for three Chlamydia-infected samples treated with IFN- ${gamma}$ on separate days. P is <0.01 for 0 versus 2 ng of IFN- ${gamma}$ per ml.

Correlation between chlamydial persistence and host cell nuclear condensation. Cells persistently infected with some human C. trachomatis strainsin vitro contain abnormally large noninfectious forms of chlamydiae(5, 6). As for C. trachomatis serovar A (3, 49), the numberof C. muridarum-infected cells containing persistent chlamydiaewas heterogenous for cells treated at the low IFN- ${gamma}$ concentration.At 0.2 ng of IFN- ${gamma}$ per ml, cells with both normal and aberrantinclusions were observed (data not shown), while at 2 ng/ml,most of the inclusions had enlarged chlamydiae, as reflectedby the appearance of patches in the inclusions, observed byimmunofluorescence (Fig. 2). Thus, unlike C. trachomatis L2,whose form does not change significantly in response to IFN- ${gamma}$ (49), the morphological development of C. muridarum in IFN- ${gamma}$ -treatedcells is reminiscent of that of C. trachomatis serovar A.

View larger version (30K):
[in this window]
[in a new window]
FIG. 2. Appearance of Chlamydia inclusions in cells treated with IFN- ${gamma}$ . HeLa cells infected with C. muridarum and treated with IFN- ${gamma}$ at 0 (A) or 2 (B) ng/ml were fixed with paraformaldehyde and prepared for immunofluorescence as described in Materials and Methods, using fluorescein-conjugated anti-Chlamydia antibodies (green) and Hoechst for DNA labeling (red). Chlamydia inclusions were normal in the absence of IFN- ${gamma}$ but were aberrant when infected cells were treated with 2 ng of IFN- ${gamma}$ per ml. Arrows indicate Chlamydia inclusions, and arrowheads show host cell nuclei. Bar, 10 µm.

Approximately 1/10 of the infected cells had persistent chlamydiaeat 0.2 ng of IFN- ${gamma}$ per ml, compared to over 3/4 of the cellsthat had been treated with 2 ng of IFN- ${gamma}$ per ml (Fig. 3A). Atthe MOI of 0.5 used here and in the absence of IFN- ${gamma}$ , half ofthe infected cells had apoptotic nuclei after 48 h of infectionand the other half still had normal nuclei (data not shown).Interestingly, at the lower IFN- ${gamma}$ concentration, inhibition ofnuclear condensation, as assayed by immunofluorescence of Hoechst-stainednuclei, was limited preferentially to cells containing the aberrantchlamydial inclusions (Fig. 3B). Thus, IFN- ${gamma}$ inhibits apoptosismainly of persistently infected cells.

View larger version (18K):
[in this window]
[in a new window]
FIG. 3. Sensitivity to the effects of IFN- ${gamma}$ and correlation between the developmental state of Chlamydia vacuoles and sensitivity to apoptosis. HeLa cells were infected with C. muridarum at an MOI of 0.5 for 48 h in the presence of the indicated concentrations of IFN- ${gamma}$ . (A) The number of cells containing normal inclusions (black bars) or inclusions with persistent chlamydiae (white bars) was determined and defined as described in Materials and Methods. For cells containing normal inclusions, P is <0.01 for 0 versus 0.2 and 0 versus 2 ng of IFN- ${gamma}$ per ml. (B) Host cells were identified as containing normal or persistent chlamydiae, and apoptotic cells were counted based on the condensation of Hoechst-labeled nuclei, as described in Materials and Methods. For cells containing normal inclusions and condensed nuclei, P is <0.001 for 0 versus 2 ng of IFN- ${gamma}$ per ml and <0.01 for 0 versus 0.2 ng of IFN- ${gamma}$ per ml. Each series of experiments was performed at least three times.

Effect of tryptophan on apoptosis of IFN- ${gamma}$ -treated cells. The persistent phenotype of Chlamydia-infected cells can bereversed through incubation of cell cultures with an excessof extracellular tryptophan (3, 28, 53). Similarly, coincubationof Chlamydia-infected cells with IFN- ${gamma}$ and tryptophan partiallyreverses the inhibitory effect of IFN- ${gamma}$ on Chlamydia-inducedapoptosis (Fig. 4), suggesting that IFN- ${gamma}$ exerts its effect mainlyvia host cell indoleamine 2,3-dioxygenase (IDO) activity (5).Thus, IFN- ${gamma}$ inhibits infection-dependent apoptosis by approximately60% in the absence of tryptophan and by approximately 30% inthe presence of tryptophan (Fig. 4). However, tryptophan didnot completely reverse the inhibitory effect of IFN- ${gamma}$ , suggestingthat this cytokine may have some effects on chlamydiae independentof IDO activity. As controls, addition of tryptophan to theextracellular medium in the absence of IFN- ${gamma}$ did not affect Chlamydia-inducedapoptosis, and 2 ng of IFN- ${gamma}$ per ml with or without tryptophanhad no effect on apoptosis in the absence of infection (datanot shown).

View larger version (15K):
[in this window]
[in a new window]
FIG. 4. Reversal by tryptophan of IFN- ${gamma}$ -dependent inhibition of apoptosis. HeLa cells were infected with C. muridarum at an MOI of 0.5 for 48 h. Infected cells were treated with (white bars) or without (black bars) IFN- ${gamma}$ (2 ng/ml) and with or without 200 µg of L-tryptophan per ml. Apoptosis was measured by cytofluorimetry using PI-stained, detergent-permeabilized cells, as described in Materials and Methods. The values were normalized to the highest level of apoptosis (infected cells in the absence of IFN- ${gamma}$ or L-tryptophan), and the values are means and standard deviations for three samples infected with C. muridarum on separate days.

Apoptosis in wild-type and IFN- ${gamma}$ -deficient mice. Both C. muridarum and human strains of C. trachomatis are sensitiveto the inhibitory effects of IFN- ${gamma}$ in mouse models, although,as in the case of the chlamydiae in vitro (30), human strainsare more sensitive than C. muridarum (43). The effect of IFN- ${gamma}$ on apoptosis during infection with C. muridarum in vivo wasevaluated by quantifying the apoptotic cells, revealed by theTUNEL technique, in the upper genital tracts of wild-type andIFN- ${gamma}$ -deficient mice infected for 7 days. There was a large increasein the level of apoptosis in infected wild-type mice comparedto uninfected controls (Fig. 5), as previously reported (40).The extent of apoptosis increased even further, more than twofold,in the IFN- ${gamma}$ -deficient mice that had been infected (Fig. 5),suggesting that secretion of IFN- ${gamma}$ may protect against Chlamydia-inducedapoptosis in vivo.

View larger version (15K):
[in this window]
[in a new window]
FIG. 5. Level of apoptosis in the genital tracts of wild-type and IFN- ${gamma}$ -deficient mice infected with Chlamydia. Mice were infected with C. muridarum for 7 days, histological samples were stained by TUNEL, and the number of apoptotic cells in the uterine horns and oviducts was determined as described in Materials and Methods. All apoptosis values were normalized to the value obtained with infected IFN- ${gamma}$ -deficient (IFN ${gamma}$ KO) mice, which was defined as 100. Histological sections were prepared from three uninfected mice and three infected mice from each group.

DISCUSSION

Top

Discussion

We show here that IFN- ${gamma}$ produces a state of persistence of C.muridarum, as previously reported for human strains of C. trachomatis(5, 6, 39, 49). After treatment with a low concentration ofIFN- ${gamma}$ (0.2 ng/ml), we observed a heterogenous population of Chlamydiainclusions, some being normal and others being aberrant. A homogeneouspopulation of persistent, aberrant chlamydiae arises after treatmentwith a higher concentration (2 ng/ml) of IFN- ${gamma}$ , as reported forthe human strain C. trachomatis serovar A (3). In infected cellsin vitro, IFN- ${gamma}$ treatment also results in inhibition of Chlamydia-inducedapoptosis. As measured by immunofluorescence, there is near-completeinhibition of apoptosis at 2 ng of IFN- ${gamma}$ per ml and partial inhibitionat 0.2 ng/ml. However, at the lower concentration, the inhibitoryeffect is restricted mainly to cells containing inclusions withthe persistent phenotype. The inhibitory effect of IFN- ${gamma}$ on apoptosiswas smaller when assayed by cytofluorimetry, which allowed usto measure apoptosis of both adherent cells and cells in suspension.Thus, the extent of apoptosis is most likely underestimatedwhen measured by immunofluorescence, which is limited to adherentcells. We also quantified the number of apoptotic cells in theupper genital tract of IFN- ${gamma}$ ^-/- mice infected with C. muridarumand found that, consistent with the in vitro results, thereare more apoptotic cells in the IFN- ${gamma}$ ^-/- mice than in wild-typemice.

The concentration dependence and the timing of the effects ofIFN- ${gamma}$ on Chlamydia-induced apoptosis are consistent with previouswork on the persistent state of chlamydiae in vitro. It hasbeen reported that multiplication of C. trachomatis is inhibitedin a dose-dependent manner when IFN- ${gamma}$ treatment begins beforeinfection (50). However, when treatment begins concomitantlywith infection or shortly after infection, as in our experiments,a persistent state develops in which the RB are morphologicallyabnormal and are unable to develop into elementary bodies (4),remaining in an undividing but viable state. In addition, pretreatmentof HeLa cells with IFN- ${gamma}$ before infection with the murine speciesused in this study, C. muridarum, had little effect on the productionof infectious chlamydiae at IFN- ${gamma}$ concentrations of $<=$ 0.2 ng/ml,but there was more than a threefold reduction in the numberof infectious chlamydiae starting at 0.5 ng/ml (30). The concentrationsof IFN- ${gamma}$ that cause persistence in vitro are thus comparableto those found in endocervical secretions from women infectedwith C. trachomatis (1, 5).

Chlamydia spp. have been reported to modulate apoptosis in oppositedirections (19, 55), both protecting acutely infected cellsagainst apoptosis due to external ligands and inducing apoptosistowards the end of the infection cycle (15, 38, 40, 41). Ina recent study, the effect of persistence on the protectiveactivity of C. trachomatis serovar A was evaluated in HeLa cellsincubated with 0.2 and 2.0 ng of IFN- ${gamma}$ per ml (11). During acuteinfection in vitro, infected cells are resistant to apoptosisdue to inhibition of cytochrome c release from mitochondriaand inactivation of host cell caspases (15). Similarly, cytochromec was not translocated from mitochondria to the cytosol in persistentlyinfected cells, suggesting that caspase-9 and consequently caspase-3were not activated (11). Thus, the mechanism of protection inpersistently infected cells appears to be the same as that observedin the absence of IFN- ${gamma}$ (15). The effect of IFN- ${gamma}$ on apoptosisat the end of the infection cycle was not evaluated.

Chlamydia species are endowed with a type III secretion apparatusthat could be used by the bacteria to secrete virulence factorsinto the inclusion or the host cell cytosol (16, 24, 51). Althoughspecific Chlamydia factors that could induce or inhibit apoptosishave not been identified yet, it is tempting to speculate thatproteins secreted by Chlamydia may modulate apoptosis in theinfected cell, as reported for other intracellular bacteria(18, 55). It also remains to be seen whether IFN- ${gamma}$ could alterthe expression of chlamydial proteins that modify survival pathwaysin the infected cell. A recent proteomic analysis of the regulationof chlamydial proteins revealed that the expression of chlamydialMOMP and several other proteins decreases markedly early inthe developmental cycle of C. trachomatis serovar A, while modificationswere also found for other proteins (49). However, no significantdown-regulation was observed for the expression of MOMP or otherproteins in cells infected with C. trachomatis serovar L2. Ifit turns out that chlamydial virulence factors are involvedin IFN- ${gamma}$ -dependent inhibition of host cell apoptosis, then onemay expect that IFN- ${gamma}$ affects Chlamydia-mediated apoptosis differentially,depending on the Chlamydia strain.

In human epithelial cells, IFN- ${gamma}$ treatment results in depletionof tryptophan via IFN- ${gamma}$ -induced expression of IDO. IDO activityleads to tryptophan catabolism, thus decreasing the intracellularconcentration of tryptophan available for chlamydiae (3). C.trachomatis and C. pneumoniae can be rescued from IFN- ${gamma}$ -treatedor tryptophan-depleted cell cultures through the addition ofextracellular tryptophan (3, 28, 53), but IFN- ${gamma}$ may also induceor inhibit apoptosis of eukaryotic cells in the absence of infection(7, 34, 57). In order to distinguish between the effects ofIFN- ${gamma}$ on host cell apoptosis and chlamydial development, apoptosisof IFN- ${gamma}$ -treated infected cells was measured in the presenceand absence of exogenous tryptophan. IFN- ${gamma}$ by itself (at 2 ng/ml)or with tryptophan in the absence of infection had no effecton survival of the host cells. However, reversal by tryptophanof the effect of IFN- ${gamma}$ in infected cells suggests that IDO activityis responsible for the IFN- ${gamma}$ -dependent inhibition of apoptosisobserved during C. trachomatis infection of human host cells.

The increase in apoptosis observed in the genital tracts ofinfected IFN- ${gamma}$ -deficient mice may also have been due to theirinability to induce IDO. Unlike the case for infected humanhost cells (3, 4), the anti-C. muridarum activity of IFN- ${gamma}$ inmurine host cells was not dependent on induction of IDO andtryptophan catabolism in vitro (12). However, in a murine modelof C. pneumoniae infection, IFN- ${gamma}$ was necessary for accumulationof IDO transcripts in the lung (46), and in a murine model ofmalaria, IDO was induced in wild-type mice but not in IFN- ${gamma}$ -deficientmice (47). It is thus likely that IDO activity may also be responsiblefor IFN- ${gamma}$ -dependent inhibition of apoptosis in the murine genitaltract.

ACKNOWLEDGMENTS

This work was supported by the Institut Pasteur, INSERM, UniversitéParis 7, a fellowship from the Fondation pour la Recherche Médicale(to J.L.P.), and National Institutes of Health grant AI43337(to T.D.).

We thank Thomas Jungas for excellent technical assistance, IrisMotta for advice on experiments, and Philippe Kourilsky forsupport and encouragement.

FOOTNOTES

* Corresponding author. Mailing address: Unité de Biologie Moléculaire du Gène, Institut Pasteur, 25 rue du Dr. Roux, 75724 Paris Cedex 15, France. Phone: (33) 1 40 61 36 34. Fax: (33) 1 45 68 85 48. E-mail: ojcius{at}pasteur.fr.

Editor: S. H. E. Kaufmann

REFERENCES

Top