characteristics of invasive candidiasis in gamma interferon- and ... · nosocomial infections...

INFECTION AND IMMUNITY,0019-9567/98/$04.0010

Apr. 1998, p. 1708–1717 Vol. 66, No. 4

Copyright © 1998, American Society for Microbiology

Characteristics of Invasive Candidiasis in Gamma Interferon-and Interleukin-4-Deficient Mice: Role of Macrophages

in Host Defense against Candida albicansRITA KAPOSZTA,1 PETER TREE,2 LASZLO MARODI,1 AND SIAMON GORDON2*

Department of Pediatrics, University School of Medicine Debrecen, H-4012 Debrecen, Hungary,1 andSir William Dunn School of Pathology, University of Oxford, Oxford OX1 3RE, United Kingdom2

Received 12 September 1997/Returned for modification 3 November 1997/Accepted 20 January 1998

Murine models of invasive candidiasis were used to study the in vivo importance of gamma interferon(IFN-g) and interleukin-4 (IL-4) in host defense against Candida albicans and to characterize the tissueinflammatory reactions, with special reference to macrophages (Mf). Knockout (KO) IFN-g-deficient (GKO)and IL-4-deficient (IL-4 KO) and C57BL/6 parental mouse strains were challenged intraperitoneally with 108

C. albicans blastoconidia. Survival of GKO mice was significantly lower (16.7%) than that of C57BL/6 control(55.5%) and IL-4 KO (61.1%) animals, but was not correlated with the extent of organ colonization. Immu-nohistological analysis with a panel of myeloid and lymphoid markers revealed multiple renal abscesses,myocarditis, hepatitis, meningoencephalitis, and pneumonia in each strain, with a dominant presence of Mf.In the absence of IFN-g, C. albicans induced striking changes in the phenotype of alveolar Mf and extensiveperivascular lymphoid infiltrates in the lung. Impairment in nitric oxide production by peritoneal Mf wasshown only in GKO mice, and they produced Candida-specific immunoglobulin G (IgG), IgM, IgA, and IgGsubclasses in lower titers. Our in vivo studies with KO mice elucidate a critical role for IFN-g, but not IL-4,in host defense against C. albicans.

Candida albicans is a common commensal organism in hu-mans, and its importance as an opportunistic pathogen, par-ticularly in immunocompromised patients, has continued toincrease over the last two decades. According to the NationalNosocomial Infections Surveillance System, the ratio of C.albicans isolates among nosocomial fungal infections increasedfrom 52% to 63% in the 1980s (4). Phagocytic cell defectsgenerally predispose to disseminated candidiasis; candidemiawas calculated to result in 38% excess mortality and extendhospitalization by approximately 30 days (40). Besides the ef-forts to develop more effective and safer antifungal agents, anew therapeutic approach to augment the antifungal capacityof the host’s immune system should be investigated.

The mechanisms of host defense and pathogenesis of can-didiasis are not completely understood. Optimal phagocytosisof C. albicans requires opsonization; however, unopsonizedyeast can be internalized by macrophages (Mf) through themannose receptor (21). Efficient killing of C. albicans by mono-nuclear phagocytes requires respiratory burst-associated toxiccompounds (22), and recent data suggest that nitric oxide(NO) may also be involved in anticandidal functions of Mf (5).Experimental evidence suggests that mononuclear phagocytescould play an important role in eradication of this pathogen,and their anticandidal activity can be augmented in vitro withgranulocyte-Mf and Mf colony-stimulating factors and cyto-kines (no significant change could be measured in the level ofspecific immunoglobulin A [IgA] in serum or among the levelsof interleukin-3 [IL-3] and gamma interferon [IFN-g]) in bothhuman and murine systems (23, 25, 28, 39).

The in vivo benefit of cytokine treatment in disseminatedcandidiasis has not been established, and data from different

murine models are controversial. Administration of IFN-g hasbeen reported to be associated with improved survival of miceafter lethal challenge with C. albicans, which correlated withthe anticandidal activity of peritoneal Mf (28); another studyshowed a reduction in tissue fungal burden in IFN-g-treatedmice (19). However, in a different murine model, in vivo ad-ministration of IFN-g resulted in increased susceptibility andorgan colonization of four infected inbred strains (13). In vivoadministration of IL-12, which has been reported to primenaive T cells for high IFN-g expression and skew cytokineproduction toward a Th1-type response (38), did not modifythe course of systemic candidiasis (32). In contrast, Th2-typecytokines IL-4 and IL-10 have been reported to exacerbateinfection, and neutralization of IL-4 by specific antibody orsoluble IL-4 receptor resulted in an enhanced production ofTh1 cytokines, associated with increased resistance to systemicmurine candidiasis (26, 30, 37). The controversial results of invivo cytokine treatment may be the consequence of geneticdifferences among the infected strains and also the variation inprotocols; the kinetics of cytokine production are influenced byseveral host and pathogen factors, and the effect of exogenouscytokine might depend on the condition of the infected hostand stage of infection.

Cytokine and receptor gene disruption strategies make itpossible to examine the role of cytokines in host response todifferent pathogens directly. Recent studies showed an in-creased susceptibility of IFN-g–receptor knockout (KO) miceto Mycobacterium bovis or Mycobacterium tuberculosis, but notto Schistosoma mansoni (1, 7, 8). Another study reported thatdisruption of the IFN-g receptor gene was associated withhigher susceptibility to Leishmania major and that IL-4 defi-ciency resulted in increased resistance, but only in certaininbred strains (17).

Our study was undertaken to investigate the in vivo role ofIFN-g and IL-4 in disseminated C. albicans infection and char-acterize the tissue inflammatory cells by immunohistochemis-

* Corresponding author. Mailing address: Sir William Dunn Schoolof Pathology, South Parks Rd., Oxford OX1 3RE, United Kingdom.Phone: 1865-275-534. Fax: 1865-275-515. E-mail: [email protected].

1708

on March 29, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

try and by functional assays ex vivo. We demonstrate thatIFN-g, but not IL-4, is essential for survival in invasive candi-diasis and show the dominant participation of Mf in the in-flammatory lesions of different tissues in KO as well as wild-type mice. In the absence of IFN-g, a striking local immuneregulatory alteration was observed in the lungs.

MATERIALS AND METHODS

Mice. IFN-g-deficient KO (GKO) mice were generated on the C57BL/6 back-ground by D. Dalton et al. (8), and IL-4 KO mice were generated on the(129Sv 3 C57BL/6) background by M. Kopf et al. (16). They were backcrossedfor 14 generations onto C57BL/6 mice at the Sir William Dunn School ofPathology, Oxford, United Kingdom.

Experimental infection and semiquantitative organ culture. C. albicans(ATCC 18804) was cultured as described previously (21). Specific-pathogen-freeinbred C57BL/6, IL-4 KO, and GKO mice of both sexes, 6 to 8 weeks old, werechallenged intraperitoneally (i.p.) with 108 C. albicans blastoconidia in two sep-arate experiments. The half-lethal dose of pathogen in a month had been deter-mined previously by titration in C57BL/6 mice. At 7, 14, 21, and 28 days after theC. albicans challenge, four to five mice from each strain were sacrificed by CO2asphyxiation. Quantitation of viable C. albicans within the various organs ofinfected mice was made by colony counting. The brain, lungs, liver, spleen, andkidneys were removed aseptically, weighed, and homogenized in 0.1% TritonX-100 (Sigma Chemical Co., St. Louis, Mo.). Samples were plated onto Sab-ouraud dextrose agar (Difco Laboratories, Detroit, Mich.) in duplicate serialdilutions and incubated for 24 to 48 h at 37°C. Data were recorded as the meanlog10 CFU per gram of organ. Animals were observed over 28 days.

Antibodies for immunohistology. The following rat monoclonal antibodies(MAbs) were prepared in our laboratory and used at optimal concentrations forimmunohistology: FA/11, which is specific for macrosialin and is a pan-Mfmurine homolog of CD68 (27, 33); and 7/4, which defines a polymorphic differ-entiation antigen on mouse neutrophils (15) and on immunologically activatedmurine Mf (37a). Other antibodies used were B220 (PharMingen, San Diego,Calif.); CD19 (Serotec Ltd., Oxford, United Kingdom); and MAbs recognizingT-cell markers, which included CD3 (KT3.1.1), CD4 (YTS 191.1.1.2), and CD8(YTS 169.4.2.1), a gift from S. P. Cobbold (Sir William Dunn School of Pathol-ogy). A hybridoma producing a MAb against major histocompatibility complex II(MHC II) (TIB120) was obtained from the American Type Culture Collection(Rockville, Md.).

Immunohistology. Organs were excised, immersed in Tissue-Tek OCT com-pound (BDH-Merck, Poole, Dorset, United Kingdom), and rapidly frozen inisopentane-dry ice. Frozen sections were cut at a depth of 5 mm onto glass slidesand stored at 220°C. Shortly before staining, sections were thawed at roomtemperature for 30 min and fixed for 10 min in 2% paraformaldehyde in HEPES-buffered isotonic saline on ice. Fixed sections were washed in 0.1% Triton X-100in phosphate-buffered saline (PBS) and then incubated with 10 mM glucose–1mM NaN3–0.4 U of glucose oxidase per ml (Sigma) in PBS for 15 min at 37°C toblock endogenous peroxidase activity. Slides were treated with 5% normal rabbitserum for 30 min and then with the primary MAb (FA/11, 7/4, TIB120, B220,CD3, CD4, or CD8) or PBS as a control for 60 min at room temperature.Affinity-purified biotinylated rabbit anti-rat IgG (Vector Laboratories, Peterbor-ough, United Kingdom) was used as a secondary antibody at 1% for 30 min,followed by avidin-biotin-peroxidase complex (ABC elite standard; Vector Lab-oratories). The presence of antigens was revealed by incubation with 0.5 mg ofdiaminobenzidine (Polysciences, Inc., Northampton, United Kingdom) per ml–0.024% H2O2 in 10 mM imidazole in PBS (pH 7.4). Sections were counterstainedin Cresyl fast violet acetate and mounted in DPX (BDH-Merck).

For detection of fungi in tissues, sections were fixed in 10% neutral bufferedformalin for 10 min and then stained with Gomori’s methenamine silver (Sigma).In brief, sections were treated with 10% periodic acid solution for 5 min tooxidize the polysaccharides to aldehydes and then with 0.11% silver methena-mine in 4% borax solution for 20 min at 62°C. The aldehyde group at alkaline pHreduced silver ion to metallic silver. Sections were rinsed in 2% gold chloride toform a more stable gold complex. Finally, excess silver was removed by washingin 20% sodium thiosulfate. Fungi appeared brown to black; sections were coun-terstained with light green (BDH-Merck).

Ex vivo nitrite assay. Peritoneal cells were harvested by peritoneal lavage 5days after i.p. challenge with a lower dose of C. albicans (105 CFU/mouse). Thedose of pathogen was reduced to avoid the ex vivo infection and killing of Mf byextracellular Candida in the peritoneal fluid. Peritoneal Mf (5 3 105/well) werecultured in 96-well plates (Falcon) in Optimem 1 (GIBCO BRL, Paisley, UnitedKingdom) supplemented with 2 mM L-glutamine (Sigma) and 50 IU of penicillinper ml–50 mg of streptomycin (Sigma) per ml. Culture supernatants were col-lected after 48 h of culture, and the nitrite concentration was assayed by theGriess reaction adapted for microplates (14). Briefly, equal volumes of 2%sulfanilamide (Sigma) in 10% phosphoric acid and 0.2% naphthylethylene dia-mine dihydrochloride (Sigma) were mixed to prepare the Griess reagent. Re-agent (100 ml) was added to equal volumes of test supernatants, and then thesemixtures were incubated for 30 min in the dark. The A550 of the formed chro-mophore was measured by means of a plate reader. The nitrite content of the

samples was calculated by using sodium nitrite as a standard and was used as arelative measure of NO synthesis. The viability of the adherent cells was .92%by Trypan blue assay. The proportion of peritoneal Mf was .90%, as assayed byFA/11 antibody staining.

Candida-specific antibody detection. A standard enzyme-linked immunosor-bent assay (ELISA) was done to quantify specific antibodies in mice sera. Briefly,polystyrene microtiter plates (Falcon, Becton Dickinson, Paramus, N.J.) werecoated overnight at 4°C with 107 heat-killed candida per well in 0.1 M bicarbon-ate buffer (pH 9.6). Wells were then saturated with 1% bovine serum albumin(Sigma) in PBS for 1 h at 37°C. Appropriate serial dilutions of the samples wereincubated in the plates for 1 h at 37°C. After extensive washing, bound antibodieswere revealed by the addition of alkaline phosphatase-conjugated rabbit anti-mouse IgG1, IgG2a, IgG2b, IgG3 (Zymed Laboratories, San Francisco, Calif.),goat anti-mouse IgG, IgM, IgA (Sigma), or peroxidase-conjugated rat anti-mouse IgE (Serotec) for 1 h at 37°C. The colorimetric change was measured bymeans of a plate reader with 405- and 492-nm filters. The antibody titers wereexpressed as the reciprocal of the dilution giving an absorbance of 0.1 above thatof the control (no serum added).

Statistical analysis. The significance of differences between the mean survivalof infected animals, organ colony counts, candida-specific antibody titers indifferent strains, and nitrite production in different strains was determined byStudent’s t test. A difference was considered statistically significant at P , 0.05.

RESULTS

Mortality of different infected strains. Survival of C57BL/6,IL-4 KO, and GKO strains was followed over 28 days, after i.p.challenge with 108 C. albicans (Fig. 1). No significant differ-ences could be observed in the first 2 weeks among the differ-ent infected strains; the mortality of each group reached 22%by the end of the second week. During the third week, thesurvival of infected GKO mice decreased rapidly, and by day28, their mortality rate was significantly higher (83.3%) thanthat of the infected IL-4 KO (38.9%) and C57BL/6 (44.5%)mice. The rates of survival of the IL-4 KO and C57BL/6 groupswere not significantly different at the P , 0.05 level during thefirst month of the Candida infection.

Semiquantitative organ cultures. Semiquantitative organcultures of the brain, lungs, spleen, liver, and kidneys wereanalyzed 7, 14, 21, and 28 days after i.p. C. albicans infection ofmice (Fig. 2); data from at least five animals were pooled at

FIG. 1. Survival of GKO, IL-4 KO, and C57BL/6 control mice challenged i.p.with 108 C. albicans blastoconidia. Survival of GKO mice was significantly lowerthan that of IL-4 KO and C57BL/6 control animals from day 14 (P , 0.05 at days21 and 28 for both). Each group contained 40 animals initially. The figurerepresents pooled data of two separate sets of experiments, which gave results inclose agreement.

VOL. 66, 1998 INVASIVE CANDIDIASIS IN IFN-g- AND IL-4-DEFICIENT MICE 1709


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

each time. The highest numbers of CFU were found in thekidneys of both KO and control mice at each time point, exceptat day 28 in IL-4 KO mice. The lowest numbers of CFU wereobserved in the brain and lungs in each strain during the first2 weeks of candidiasis. Comparing the different strains, theoverall tissue fungal burden was moderately higher in most ofthe organs of GKO mice, but the differences were not signif-icant at days 7, 14, and 28 (P , 0.05). At day 21, the numbers

of CFU in each organ were significantly higher in both GKOand IL-4 KO mice than in the parental strain (P , 0.05). Theresults of two separate experiments showing significant differ-ences in tissue fungal burden at day 21 were in close agree-ment; pooled data of these two sets of experiments are pre-sented. No correlation, however, was found between the fungalburden of organs and the mortality of the different groups.

Histology. Characteristics of the kidney, heart, liver, brain,spleen, and lungs of C57BL/6, GKO, and IL-4 KO mice werestudied by immunohistology 3, 14, and 28 days after i.p. C.albicans injection (Table 1). The pathogen was detected indifferent tissues by Gomori’s silver methenamine staining.

In the kidney, extensive necrotic abscesses containing fila-mentous and yeast forms of the fungus were present in eachstrain from day 3. Evidence of local, capsular invasion of C.albicans could also be detected besides lymphohematogenousspread (Fig. 3). The vast majority of the inflammatory cellswere FA/111 Mf and stained weakly with MAb 7/4; few ofthem were MHC II1. By day 14, lymphoid infiltrates withCD31, CD41, and CD81 T cells and B2201 B cells appearedin the lesions beside FA/111 giant Mf, and the proportion ofMHC II1 cells increased. Lesions were smaller, but still com-plex, localized both to the cortex and medulla at day 28, andcontained fewer inflammatory cells; there were MHC II1 cells,mostly FA/111 Mf.

In the myocardium, similar to the kidney, extensive necroticabscesses were observed at day 3 with filamentous and yeastforms of fungus both extracellularly and inside Mf (Fig. 3).Again the vast majority of inflammatory cells were FA/111 Mfand stained only weakly with 7/4 MAb, and few were MHC II1.At day 14, no focal lesions but diffuse mononuclear cell reac-tions were detected in the myocardium of each strain; theinflammatory cells were MHC II1, mostly FA/111 Mf, andthere were fewer lymphoid cells. Diffuse mononuclear infil-trates were still observed at day 28.

In the liver, no necrotic lesions could be observed at day 3.At day 14, multifocal inflammatory reactions appeared mostlyperivascularly. The inflammatory cells were MHC II1 (Fig.4A), the vast majority of which were FA/111 Mf (Fig. 4C) andwere weakly stained with MAb 7/41. There were fewer Tcells—mostly CD41 (Fig. 4D) and fewer CD81 (4F)—and veryfew B2201 B cells (Fig. 4E), and debris of C. albicans wasassociated with the inflammatory lesions. At day 28 in thesurvivors, multifocal hepatitis could still be observed, withmononuclear infiltrates.

In the brain (not shown), multifocal inflammatory reactionscould be observed from day 14, which were associated with theblood vessels, and the cerebrospinal fluid-brain barrier. Themononuclear cells were MHC II1, mostly FA/111 microglia orMf, and there were fewer T cells, mostly CD41 cells. Debris of

FIG. 2. CFU of C. albicans from various organs of GKO, IL-4 KO, andC57BL/6 control mice after i.p. challenge with 108 Candida blastoconidia. Eachbar represents the mean 6 standard error. Experiments were performed induplicate with five or more mice.

TABLE 1. Kinetics and cell contents of tissue inflammatory reactions in murine candidiasisa

Candidiasiseffect

Result in tissue at postinfection day:

Kidney Heart Liver Brain

3 14 28 3 14 28 3 14 28 3 14 28

Lesions 11 1 1/2 11 1 1/2 1/2 11 1 2 11 1/2Macrophages 11 11 11 11 11 11 11 11 11 2 11 11T cells 1/2 1 1/2 1/2 1 1/2 1/2 1 1/2 2 1 1/2B cells 1/2 1 1/2 1/2 1 1/2 1/2 1/2 1/2 2 2 2Neutrophils 1/2 2 2 1/2 2 2 1 1/2 2 2 2 2

a The extension or multiplicity of lesions and the ratio of different inflammatory cells were scored between 2 (none) and 11 (dominant). Note that Mf were thedominant cell type at each time of infection, the ratio of lymphocytes was increased at 14 days, and small numbers of neutrophil granulocytes could only be observedat an early stage. Most of the T cells were CD41, and Mf became MHC II1 at a later stage of the infection.

1710 KAPOSZTA ET AL. INFECT. IMMUN.


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

fungus could also be seen inside the blood vessels. At day 28 inthe survivors, fewer lesions were present with MHC II1, FA/111 microglia or Mf.

In the spleen (data not shown), there were no necrotic le-sions resembling those in the kidneys and heart at any timestudied; the structure and cell content of white and red pulpwere similar in infected KO and control mice and showed

diffuse lymphoid hyperplasia with no disorganization of organmicroarchitecture.

The histopathologies of the kidneys, heart, liver, brain, andspleen were not significantly different in the control and im-munocompromised mice at any time. However, the lungs ofGKO mice presented a striking, distinct phenotype by day 3 ofcandidiasis, having extensive perivascular lymphoid infiltrates

FIG. 3. Histological appearance of lesions in the kidneys (A to D) and heart (E and F) at day 3 of murine invasive candidiasis. Inflammatory lesions in both thekidneys and heart consisted of mostly FA/111 macrophages (A, C, and E). Yeast and mycelial forms of fungus were shown in the lesions by Gomori’s silvermethenamine staining and were both extra- and intracellular (B, D, and F). The histopathologies of the kidneys and heart in GKO, IL-4 KO, and C57BL/6 control micewere similar. Representative photographs were taken of sections from IL-4 KO (A and B) and GKO mice (C to F). Bar, 50 mm.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

with B2201 and CD191 B cells (arrowheads) (Fig. 5F) andCD31 (Fig. 5C), CD41 (Fig. 5D), and CD81 (Fig. 5D) T cells.Mf (FA/111) (arrowheads in Fig. 5G and H) appeared in twodistinct phenotypes: disseminated, large alveolar Mf and elon-gated cells, peripherally associated with the perivascular lym-phoid infiltrates and beneath the endothelium. Neutrophilgranulocytes (7/41) could not be observed among the inflam-matory cells in the infiltrates. By day 14, the extension of theinfiltrates was reduced, became more focal, and contained

MHC II1 cells (arrowheads) (Fig. 5O): FA/111 Mf (notshown) and mainly CD31 (Fig. 5P), CD41 T cells, and fewerB2201 B cells (not shown). The size of alveolar Mf decreasedby that time. Diffuse pneumonia, but no perivascular infil-trates, could be seen in the survivors at day 28. Extensivenecrotic lesions with filamentous or yeast forms of C. albicans,similar to those in the kidneys and heart, could not be observedin the lungs of GKO mice by Gomori’s staining at any time,which revealed only debris of fungal origin (Fig. 5A). In

FIG. 4. Histopathology of the liver in invasive candidiasis at day 14. Perivascular infiltrates contained MHC II1 cells (A) and included CD31 (B), CD41 (D), andCD81 (F) T cells and FA/111 Mf (C); only a few B2201 B cells (E) were associated with the lesions. The histopathologies of the liver in GKO, IL-4 KO, and C57BL/6control mice were similar; representative photographs of liver in C57BL/6 mice are shown. Bar, 50 mm.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

FIG. 5. Histopathology of the lungs in invasive murine candidiasis at day 3 in GKO (A to H) and C57BL/6 mice (I to L). Inflammatory lesions in GKO mice atday 2 (M and N) and day 14 (O and P) are also shown. Perivascular infiltrates appeared only in GKO mice and contained MHC II1 (TIB1201) mononuclear cells (B),CD31 T cells (C), mostly CD41 (D) and fewer CD81 (E) cells and B2201 B cells (F), which were also CD191 (not shown). Mf FA/111 (G) were not a dominantcomponent of the infiltrates, but were present in two different phenotypes in the lungs: elongated, perivascular cells and engorged, diffusely distributed alveolar Mf(H). Perivascular infiltrates contained only debris of C. albicans, as shown by Gomori’s staining (A). Histopathology of the lungs in IL-4 KO and C57BL/6 mice wasmarkedly different from that of GKO mice; a similar inflammatory reaction did not develop in these strains even at later stages of infection. Lung sections obtainedat day 3 from C57BL/6 mice were stained with TIB120 (I), FA/11 (J), B220 (K), and anti-CD3 (L) MAbs and showed only minimal inflammatory responses. At day2, perivascular MHC II1 cell aggregates (M) appeared in GKO mice, most of which were CD31 T cells (N), and by day 14, the extension of perivascular infiltrateswas reduced compared with that of day 3, containing MHC II1 cells (O), which were partly CD31 T cells (P). Bars, 50 mm.

1713


http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

marked contrast to the results presented above, only diffusemononuclear infiltrates could be observed from day 14 in thelungs of C57BL/6 and IL-4 KO mice (Fig. 5I to L), without theperivascular infiltrates characteristic of lesions in GKO mice.

In order to determine the onset of perivascular infiltrate for-mation, GKO and control mice were also examined at 1 and 2days after i.p. challenge. At day 1, the histopathologies of lungsin GKO and control mice were similar: Mf and CD31 CD41

T cells appeared perivascularly and formed more prominentaggregates by day 2 in GKO mice (Fig. 5M and N).

Nitrite production by peritoneal Mf. The nitrite concentra-tions in the culture supernatant of peritoneal Mf from non-infected mice were comparably low in each strain. Increasednitrite production was detected by Mf isolated from infectedanimals after in vivo challenge. The increase was significantlylower in GKO mice (16.1-fold) and higher in IL-4 KO mice(55.3-fold) than in the C57BL/6 control mice (24.6-fold) com-pared to the baseline (Fig. 6). Impaired NO production by Mffrom GKO mice could be restored in vitro by addition of IFN-g(100 U/ml) to the culture suspension (Fig. 6).

C. albicans-specific serum antibody responses. In controlC57BL/6 mice, an increase in Candida-specific IgM titer wasobserved from day 7, with a peak at day 14 after the challenge;the specific IgG titer reached its maximum 1 week later at day21 (Fig. 7A), whereas among different IgG subclasses, the IgG1response was most prominent in C57BL/6 mice, followed bythe IgG2b and IgG3 responses, and only a moderate increasecould be detected in the level of IgG2a in serum during can-didiasis (Fig. 7B [IgG1.IgG2b$IgG3.IgG2a]).

In GKO mice, the increase in Candida-specific IgM, IgG,and IgG subclass levels was lower than that measured in con-trol mice from day 14 (Fig. 7A and B). The patterns of the IgM,IgG, and IgG subclass responses were similar to those of thecontrol group (IgG1.IgG2b$IgG3.IgG2a). The level of spe-cific IgA in serum in GKO mice was low and remained un-changed during the infection.

FIG. 6. Nitrite concentration in 48-h culture supernatants of peritoneal Mfisolated from C57BL/6, IL-4 KO and GKO infected mice at day 5 of infection.The nitrite production of peritoneal macrophages from IL-4 KO mice was sig-nificantly higher than that of the control. The impaired function of Mf fromGKO animals could be restored by ex vivo IFN-g treatment for 48 h (100 U/ml).The height of each bar represents the mean 6 standard error of three experi-ments performed in duplicate.

FIG. 7. (A) C. albicans-specific IgM, IgG, and IgA titers in sera of mice infected i.p. with 108 Candida blastoconidia. Pooled sera of at least four mice/group wereanalyzed for specific antibodies by standard ELISA. The height of each bar represents the mean 6 standard error of three experiments performed in duplicate. (B)C. albicans-specific IgG1, IgG3, IgG2a, and IgG2b titers in sera of mice infected i.p. with 108 Candida blastoconidia. Pooled serum samples of at least four mice pergroup were analyzed for specific antibodies by standard ELISA. The height of each bar represents the mean 6 standard error of three experiments performed induplicate.



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

The Candida-specific titers of IgM, IgG, and IgA in the seraof IL-4 KO mice did not differ significantly from those ofcontrols (P , 0.05). Titers of specific IgG1 were significantlylower, whereas production of IgG3 and IgG2b was significantlyhigher in IL-4 KO mice than in controls (P , 0.05) (Fig. 7B[IgG2b$IgG3.IgG1.IgG2a]).

DISCUSSION

This study is the first to report disseminated candidiasis inGKO and IL-4 KO mice and to analyze the kinetics and cellcontent of the different tissue inflammatory reactions. In situanalysis establishes that Mf play a key role in the host re-sponse to Candida. Our results indicate that IFN-g, but notIL-4, is essential for survival in invasive candidiasis. In contrastto previous reports, which suggested that neutralization of IL-4or IL-4 receptor was associated with an increased resistance toC. albicans (26, 30), in our model, the susceptibility of IL-4 KOmice was similar to that of control mice. This difference couldbe explained by unrelated genetic factors, which influence theimmune response and susceptibility of other inbred strains tothe growing pathogen. Previous studies reported that C57BL/6mice are relatively resistant to C. albicans infection, with apredominance of Th1-type anticandidal response, while BALB/cand especially DBA/2 strains are more susceptible to this fun-gal pathogen and produce Th2 cytokines in higher titer (30, 31,36); we hypothesize that anti-IL-4 treatment can improve theresistance only of certain strains, which exhibit a predomi-nantly Th2-type immune response to the pathogen, skewingthe immune response towards a Th1 type.

In our model, higher mortality of GKO mice could be ob-served only from day 14 of candidiasis, although the infectionbecame disseminated within 24 h of the i.p. C. albicans chal-lenge. Pyelonephritis and myocarditis developed at an earlystage (day 3), while hepatitis and meningoencephalitis could beobserved only later in the infection (day 14) in each strain. Thetissue fungal burden was marginally higher in GKO mice, andit peaked at day 21 in each strain associating with increasedmortality in GKO mice alone, which suggests that Candidaovergrowth was not the only factor responsible for higher mor-tality. The increased mortality of GKO mice after 14 days ofcandidiasis could be explained by a combination of differentfactors, including an altered inflammatory reaction and im-mune responses to C. albicans.

Immunohistology indicated that Mf were the dominant celltype in the inflammatory reaction of several organs, both inKO and control mice. The kinetics and phenotype of tissueinflammatory lesions were organ rather than mouse strain spe-cific, except in the lungs. Lesions appeared first in the kidneyand heart and were more extensive and necrotic. Kidneys areoften reported to be one of the main targets in different can-didiasis models, although their phagocytic system can elimi-nate C. albicans as effectively as those of the spleen and liver(2). The proteoglycan- and glycoprotein-rich basement mem-brane of glomeruli contains ligands for the CR2/CR3 receptoron the Candida surface, and the acidic condition in renal tu-bules may favor fungal growth (6). The rapid development ofnecrotic lesions in the heart might be explained by limited localinnate immune defenses. The tissue inflammatory reactions inboth the liver and brain were most prominent at later stages ofinfection (day 14), but were less extensive and necrotic, whichcould be explained by the efficiency of the phagocytic system inthe liver and by anatomical segregation in the brain. The dom-inance of Mf rather than neutrophils in the tissue infiltratesfrom the early stages of infection and in immunocompetent aswell as KO mice differs from our unpublished observations

with M. bovis BCG infection, in which Mf are poorly recruitedto lesions in GKO mice, which also contain abundant neutro-phils. The appearance of lymphoid, especially CD41, T cells atlater stages (day 14) was similar in most organs of immuno-competent and KO mice, except for the lungs, in which thehost cellular response in the absence of IFN-g was strikinglydifferent in invasive candidiasis.

In the lungs, activated Mf (but few neutrophil granulocytes)and CD41 T cells appeared initially (day 1) at perivascularsites in all strains, but lymphoid infiltrates developed only inGKO mice (day 3). These infiltrates were dominated by B cellsand were associated peripherally with FA/111 elongated, den-dritic cells. The phenotype of alveolar Mf was also distinctiveand unusual in GKO mice; from an early stage of candidiasis,they were engorged and scattered throughout the lungs. In thepresent study, the FA/11 MAb could detect not only Mf, butalso dendritic cells, which are constitutively MHC II1, as wellas Mf that had been induced to express MHC II, and we havenot distinguished between the different potential antigen-pre-senting cells.

The mechanism of the development of perivascular infil-trates in the lungs of GKO mice could be due to relative excessof Th2-type cytokines. Our model illustrates the role of IFN-gin regulating the local inflammatory response, and we suggestthat it may act by (i) influencing the dose and persistence ofantigens, (ii) down-regulating directly or indirectly the localproduction of certain chemokines and cytokines, and/or (iii)terminating the proliferation or survival of activated lymphoidcells. Formation of extensive lymphoid infiltrates might be theconsequence of persistent recruitment, in situ proliferation,and impaired apoptosis of lymphoid cells. The role of thedifferent Mf populations in this process is also undefined: thedendritic cells at the periphery of the infiltrates might providean important inductive function, but these require further in-vestigations.

In our model, C. albicans was disseminated to the lungs viathe bloodstream, adhered to the endothelial cells, and pene-trated through them. Germinated C. albicans has been re-ported to stimulate endothelial cells to produce proinflamma-tory cytokines in vitro, such as IL-6, IL-8, and monocytechemoattractant protein (MCP-1), and to induce adhesionmolecules (ICAM-1 and VCAM-1) (12), resulting in the re-cruitment of phagocytes and initiation of a local inflammatoryreaction. Mf and, possibly, dendritic cells are able to ingestCandida even in the absence of opsonins (21, 29) and thenproduce proinflammatory cytokines, such as IL-1b, IL-6, Mfinflammatory proteins (MIP-1b and MIP-2), and granulo-cyte-Mf colony-stimulating factor (35, 41). In transgenic mice,which overexpress IL-6 in airway epithelial cells, the phenotypeof peribronchial inflammatory reactions resembles that wehave observed in GKO mice (10), indicating that overproduc-tion of IL-6 might contribute to the development of the B-cell-rich lymphoid infiltrates. IL-12 produced by Mf, as well asdendritic cells, acts both as a proinflammatory cytokine and animmunomodulator, priming naive T cells for high IFN-g pro-duction and therefore bridging the innate and adaptive im-mune responses (38). The importance of innate immune mech-anisms (phagocytes and/or NK cells) in host defense to C.albicans (3, 20) has also been demonstrated in candidiasismodels with SCID mice.

NO production by peritoneal Mf, collected after i.p. injec-tion with C. albicans, reflected the function and activation ofMf in the different strains. In GKO mice, NO production wasimpaired, but could be restored by in vitro IFN-g treatment,excluding an intrinsic deficiency of Mf, while in IL-4 KO mice,the increased NO production can be explained by the absence



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

of inhibition by IL-4 and/or an increased stimulatory effect byIFN-g. The decreased NO production by GKO peritoneal Mfcorrelated with enhanced susceptibility to C. albicans, but en-hanced NO in IL-4 KO Mf did not increase in vivo resistanceto infection, compared with that in the wild-type mice.

In our disseminated candidiasis model, the levels of MHC IIexpression on Mf were similar in GKO and other strains, incontrast to a previous study, which established that IFN-g playsa crucial role in MHC II expression on activated Mf in murineM. bovis BCG infection (8). Our finding indicates the existenceof an IFN-g-independent mechanism which regulates MHC IIexpression, possibly via IL-4 and IL-13 (11).

The pattern of Candida-specific Ig subclass production wasaltered in IL-4 KO mice. The specific IgG1 response was re-duced, and IgG2b and IgG3 were increased, which can beexplained by the absence of a stimulatory effect of IL-4 onexpression of IgG1 and inhibition of IgG2a and IgG3 produc-tion (18, 24, 34). IFN-g also influences isotype switching; it caninhibit the expression of IgG1, IgG2b, and IgG3 and augmentthe production of IgG2 (9, 34). However, in our model, theproduction of Candida-specific IgG1, IgG2b, and IgG3 was notupregulated in GKO mice compared with that in the controlstrain.

In conclusion, the findings reported here illustrate the invivo importance of IFN-g for survival in disseminated candi-diasis and indicate the role of Mf in the tissue inflammatoryreactions in both immunocompetent and immunodeficient an-imals. We have also demonstrated localized immune regula-tory disturbances in organs such as lungs, which might contrib-ute to higher susceptibility to this increasingly importantpathogen.

ACKNOWLEDGMENTS

This work was supported by grants from the Medical ResearchCouncil to S. Gordon (06BI) and the National Science Foundation ofHungary (OTKA T 017 100) and European Commission (PECO no.CIPD CT 9400303) to L. Marodi. The first author is a participant in theUK-Hungary Health Scientist Exchange Program jointly funded by theBritish Council, Hungarian Ministry of Education, and Soros Founda-tion.

We thank Rosangela da Silva (Sir William Dunn School of Pathol-ogy, Oxford) for helpful discussion and Elizabeth Darley, Lance Tom-linson, Stephen Clark, and Jeremy Sanderson (Sir William DunnSchool of Pathology, Oxford) for technical assistance.

REFERENCES

1. Akhiani, A. A., N. Lycke, L.-Å. Nilsson, S. Olling, and O. Ouchterlony. 1996.Lack of interferon-g receptor does not influence the outcome of infection inmurine Schistosomiasis mansoni. Scand. J. Immunol. 43:257–262.

2. Baghian, A., and K. W. Lee. 1991. Elimination of Candida albicans fromkidneys of mice during short-term systemic infections. Kidney Int. 40:400–405.

3. Balish, E., J. Jensen, T. Warner, J. Brekke, and B. Leonard. 1993. Mucosaland disseminated candidiasis in gnotobiotic SCID mice. J. Med. Vet. Mycol.31:143–154.

4. Beck-Sague, C. M., W. R. Jarvis, and the National Nosocomial InfectionsSurveillance System. 1993. Secular trends in the epidemiology of nosocomialfungal infections in the United States, 1980–1990. J. Infect. Dis. 167:1247–1251.

5. Blasi, E., L. Pitzurra, M. Puliti, A. R. Chimienti, R. Mazzolla, R. Barluzzi,and F. Bistoni. 1995. Differential susceptibility of yeast and hyphal forms ofCandida albicans to macrophage-derived nitrogen-containing compounds.Infect. Immun. 63:1806–1809.

6. Calderone, R., R. Diamond, J.-M. Seneet, J. Warmington, S. Filler, and J. E.Edwards. 1994. Host cell-fungal cell interactions. J. Med. Vet. Mycol.32(Suppl. 1):151–168.

7. Cooper, A. M., D. K. Dalton, T. A. Stewart, J. P. Griffin, D. G. Russel, andI. M. Orme. 1993. Disseminated tuberculosis in interferon g gene-disruptedmice. J. Exp. Med. 178:2243–2247.

8. Dalton, D. K., S. Pitts-Meek, S. Keshav, I. S. Figari, A. Bradley, and T. A.Stewart. 1993. Multiple defects of immune cell function in mice with dis-

rupted interferon-g genes. Science 259:1739–1742.9. Damore, M. A., S. A. Omori, and R. Wall. 1996. IFN-g induces the k intron

enhancer via an IFN-stimulated response element. J. Immunol. 156:2451–2457.

10. DiCosmo, B. F., G. P. Geba, D. Picarella, J. A. Elias, J. A. Rankin, B. R.Stripp, J. A. Whitsett, and R. A. Flavell. 1994. Airway epithelial cell expres-sion of interleukin-6 in transgenic mice. Uncoupling of airway inflammationand bronchial hyperreactivity. J. Clin. Invest. 94:2028–2035.

11. Doyle, A. G., G. Herbein, L. J. Montaner, A. J. Minty, D. Caput, P. Ferrara,and S. Gordon. 1994. Interleukin-13 alters the activation state of murinemacrophages in vitro: comparison with interleukin-4 and interferon-g. Eur.J. Immunol. 24:1441–1445.

12. Filler, S. G., A. S. Pfunder, B. J. Spellberg, J. P. Spellberg, and J. E.Edwards, Jr. 1996. Candida albicans stimulates cytokine production andleukocyte adhesion molecule expression by endothelial cells. Infect. Immun.64:2609–2617.

13. Garner, R. E., U. Kuruganti, C. W. Czarniecki, H. H. Chiu, and J. E. Domer.1989. In vivo immune responses to Candida albicans modified by treatmentwith recombinant murine gamma interferon. Infect. Immun. 57:1800–1808.

14. Green, L. C., D. A. Wagner, J. Glogowski, P. L. Skipper, J. S. Wishnok, andS. R. Tannenbaum. 1982. Analysis of nitrate, nitrite and [15N]nitrite inbiological fluids. Anal. Biochem. 126:131–138.

15. Hirsch, S., and S. Gordon. 1983. Polymorphic expression of a neutrophildifferentiation antigen revealed by monoclonal antibody 7/4. Immunogenet-ics 18:229–239.

16. Kopf, M., G. Le Gros, M. Bachmann, M. C. Lamers, H. Bluethmann, and G.Kohler. 1993. Disruption of the murine IL-4 gene blocks Th2 cytokineresponses. Nature 362:245–247.

17. Kopf, M., F. Brombacher, G. Kohler, G. Kienzle, K.-H. Widmann, K. Le-frang, B. Ledermann, and W. Solbach. 1996. IL-4-deficient Balb/c mice resistinfection with Leishmania major. J. Exp. Med. 184:1127–1136.

18. Kuhn, R., K. Rajewsky, and W. Muller. 1991. Generation and analysis ofinterleukin-4 deficient mice. Science 254:707–710.

19. Kulberg, B. J., J. W. van’t Wout, C. Hoogstraten, and R. van Furth. 1993.Recombinant interferon-g enhances resistance to acute disseminated Can-dida albicans infection in mice. J. Infect. Dis. 168:436–443.

20. Mahanty, S., R. A. Greenfield, W. A. Joyce, and P. W. Kincade. 1988.Inoculation candidiasis in a murine model of severe combined immunode-ficiency syndrome. Infect. Immun. 56:3162–3166.

21. Marodi, L., H. M. Korchak, and R. B. Johnston, Jr. 1991. Mechanism of hostdefense against Candida species. I. Phagocytosis by monocytes and mono-cyte-derived macrophages. J. Immunol. 146:2783–2789.

22. Marodi, L., J. R. Forehand, and R. B. Johnston, Jr. 1991. Mechanisms ofhost defense against Candida species. II. Biochemical basis for the killing ofCandida by mononuclear phagocytes. J. Immunol. 146:2790–2794.

23. Marodi, L., S. Schreiber, D. C. Anderson, R. P. Macdermott, H. M. Korchak,and R. B. Johnston, Jr. 1993. Enhancement of macrophage candidacidalactivity by interferon-g. Increased phagocytosis, killing, and calcium signalmediated by a decreased number of mannose receptors. J. Clin. Invest.91:2596–2601.

24. Mongini, P. K. A., P. F. Highet, and J. K. Inman. 1995. Human B cellactivation. Effect of T cell cytokines on the physicochemical binding require-ments for achieving cell cycle progression via the membrane IgM signalingpathway. J. Immunol. 155:3385–3400.

25. Perfect, J. R., D. L. Granger, and D. T. Durack. 1987. Effects of antifungalagents and g-interferon on macrophage cytotoxicity for fungi and tumorcells. J. Infect. Dis. 156:316–323.

26. Pucetti, P., A. Mencacci, E. Cenci, R. Spaccapelo, P. Mosci, K.-H. Enssle, L.Romani, and F. Bistoni. 1994. Cure of murine candidiasis by recombinantsoluble interleukin-4 receptor. J. Infect. Dis. 169:1325–1331.

27. Rabinowitz, S. S., and S. Gordon. 1991. Macrosialin, a macrophage-re-stricted membrane sialoprotein differentially glycosylated in response toinflammatory stimuli. J. Exp. Med. 174:827–836.

28. Redmond, H. P., J. Shou, H. J. Gallagher, C. J. Kelly, and J. M. Daly. 1993.Macrophage-dependent candidacidal mechanisms in the murine system.Comparison of murine Kupffer cell and peritoneal macrophage candidacidalmechanisms. J. Immunol. 150:3427–3433.

29. Reis-E-Sousa, C., P. D. Stahl, and J. M. Austyn. 1993. Phagocytosis ofantigens by Langerhans cells in vitro. J. Exp. Med. 178:509–519.

30. Romani, L., A. Mencacci, U. Grohmann, S. Mocci, P. Mosci, P. Pucetti, andF. Bistoni. 1992. Neutralizing antibody to interleukin 4 induces systemicprotection and T helper type 1-associated immunity in murine candidiasis. J.Exp. Med. 176:19–25.

31. Romani, L., A. Mencacci, E. Cenci, R. Spaccapelo, E. Schiaffella, L. Tonetti,P. Puccetti, and F. Bistoni. 1993. Natural killer cells do not play a dominantrole in CD41 subset differentiation in Candida albicans-infected mice. Infect.Immun. 61:3769–3774.

32. Romani, L., A. Mencacci, L. Tonetti, R. Spaccapelo, E. Cenci, P. Pucetti,S. F. Wolf, and F. Bistoni. 1994. IL-12 is both required and prognostic in vivofor T helper type differentiation in murine candidiasis. J. Immunol. 152:5167–5175.

33. Smith, M. J., and G. L. E. Koch. 1987. Differential expression of murine



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

macrophage surface glycoprotein antigens in intracellular membranes.J. Cell Sci. 87:113–119.

34. Snapper, C. M., and W. E. Paul. 1987. Interferon-g and B cell stimulatoryfactor-1 reciprocally regulate Ig isotype production. Science 236:944–947.

35. Taub, D. D., K. Conlon, A. R. Lloyd, J. J. Oppenheim, and D. J. Klein. 1993.Preferential migration of activated CD41 and CD81 cells in response toMIP-1a and MIP-1b. Science 260:355–358.

36. Tavares, D., P. Ferreira, M. Vilanova, A. Videira, and M. Arala-Chaves.1995. Immunoprotection against systemic candidiasis in mice. Int. Immunol.7:785–796.

37. Tonetti, L., R. Spaccapello, E. Cenci, A. Mencacci, P. Pucetti, R. L. Coff-mann, F. Bistoni, and L. Romani. 1995. Interleukin-4 and -10 exacerbatecandidiasis in mice. Eur. J. Immunol. 25:1559–1565.

37a.Tree, P., and S. Gordon. Unpublished observations.

38. Trinchieri, G., and F. Gerosa. 1996. Immunoregulation by interleukin-12.J. Leukocyte Biol. 59:505–511.

39. Wang, M., H. Friedman, and J. Y. Kjeu. 1989. Enhancement of humanmonocyte function against Candida albicans by the colony-stimulating fac-tors (CSF): IL-3, granulocyte-macrophage-CSF, and macrophage-CSF. J.Immunol. 143:671–677.

40. Wenzel, R. P. 1995. Nosocomial candidemia: risk factors and attributablemortality. Clin. Infect. Dis. 20:1531–1534.

41. Yamamoto, Y., T. W. Klein, and H. Friedman. 1997. Involvement of mannosereceptor in cytokine interleukin-1b (IL-1b), IL-6, and granulocyte-macro-phage colony-stimulating factor responses, but not in chemokine macro-phage inflammatory protein 1b (MIP-b), MIP-2, and KC responses, causedby attachment of Candida albicans to macrophages. Infect. Immun. 65:1077–1082.

Editor: S. H. E. Kaufmann



http://iai.asm.org/

Dow

nloaded from

http://iai.asm.org/

characteristics of invasive candidiasis in gamma interferon- and ... · nosocomial infections...

Documents