Inhibitory Effect of Human NaturalYeast Killer Toxin-like CandidacidalAntibodies on Pneumocystis carinii

Nathalie Seguy,l Jean-Charles Cailliez,"2 Philippe Delcourt,'Stefania Conti,' Daniel Camus,"4 Eduardo Dei-Cas,45and Luciano Polonelli3'Unite 42 INSERM, Villeneuve d'Ascq, France2Faculte' Libre des Sciences, Universite Catholique de Lille, Lille, France3Istituto di Microbiologia, Universita degli Studi di Parma, Parma, Italy4Laboratoire de Parasitologie, Centre Hospitalier Retgional Universitaire,Lille, France5Institut Pasteur de Lille, Lille, France

ABSTRACT

Background: Human natural antibodies have beenfound that owe their candidacidal action to the mimicryof a yeast killer toxin produced by the yeast Pichiaanomala (PaKT). Candidacidal human natural antibodies(KTAb) are elicited by and bind to a KT receptor (PaKTR)present on the cell surface of infectious PaKT-sensitivemicroorganisms. Because of the recognized susceptibilityof Pneumocystis carinii organisms to PaKT upon the occur-rence of specific PaKTR, we examined whether humannatural KTAb could also bind to and inhibit P. carinii.Materials and Methods: Immunoaffinity-purifiedKTAb from the vaginal fluid of patients affected by can-didiasis were tested and compared with PaKT for theirability to inhibit rat-derived P. carinii attachment to epi-thelial lung cells as well as infectivity to nude rats. Im-munofluorescence studies were also performed by bio-tinylated PaKT in competition with human KTAb toestablish their specific binding to PaKTR on the surface ofrat-derived and human P. carinii organisms.

Results: Human natural candidacidal KTAb exerted astrong, specific inhibitory activity against rat-derived P.carinii organisms that are susceptible to PaKT itself. Theantimicrobial activity of human KTAb was abolished byadsorption with a specific PaKT-neutralizing mAb KT4.Immunofluorescence studies of competition with PaKTshowed that human KTAb efficiently bind to the specificPaKTR on the surface of rat-derived and human P. cariniiorganisms.Conclusions: The results strongly suggest that humanKTAb, elicited by a common transphyletic receptor ofdifferent pathogenic microorganisms during infection,may play a role in antibody-mediated cross-immunityand, if properly engineered, as functionally equivalentrecombinant antibodies they could exert a therapeuticactivity against pneumocystosis in vivo.

INTRODUCTIONAlthough the development of new therapeuticstrategies against pneumocystosis remains a pri-ority, few drugs capable of inhibiting specifically

Address correspondence and reprint requests to: Prof. Lu-ciano Polonelli, Istituto di Microbiologia, Facolta di Medi-cina e Chirurgia, Universita degli Studi di Parma, Via A.Gramsci 14, 43100 Parma, Italy. Phone: 011-39-521-988877; fax: 011-39-521-993620

544

the growth of Pneumocystis carinii without sec-ondary effects are known (1). During the last fewyears, the susceptibility of mouse- or rat-derivedP. carinii to a Pichia anomala candidacidal killertoxin (PaKT) has been demonstrated by using invitro attachment tests or infectivity assays (2-4).However, the PaKT, which is toxic and antigenic,other than inactive under human physiologicalconditions, could not be used as a systemic anti-biotic (5,6). For this reason, a new strategy to

© 1997, THE PICOWER INSTITUTE PRESS. All rights reserved.MolecularMedicine, Volume 3, Number 8, August 1997 544-552

N. Seguy et al.: Killer Toxin-like Antibodies against P. carinii 545

control candidiasis or pneumocystosis was devel-oped by using yeast killer toxin-like anti-idiotypicantibodies (KT-IdAb) (6,7). An immunoglobulinGI (IgGI) monoclonal antibody (MAbKT4) thatneutralized in vitro the fungicidal activity of PaKTwas used as an immunogen (idiotypic vaccina-tion) to elicit KT-IdAb capable of inhibiting invitro the growth of PaKT-sensitive Candida albi-cans (8-11). KT-IdAb raised in sera and vaginalsecretions of mice and rats through parenteraland intravaginal immunization with MAbKT4were able to confer immunoprotection againstsystemic or mucosal experimental candidiasis(9,10). Presumably, the paratope of MAbKT4 hasregions that are similar to those of the yeast killertoxin receptor (PaKTR) at the surface of sensitivetarget microbial cells. Thus, KT-IdAb reactivewith the MAbKT4 idiotype could also react withthe PaKTR at the surface of C. albicans or P. cariniicells (2,12). The immune system thereforeshould recognize the PaKTR of infectious micro-organisms such as the MAbKT4 idiotype, andconsequently, anti-receptor yeast killer toxin-like antibodies (KTAb) should be part of the an-tibody repertoire of animals experimentally in-fected with PaKTR-bearing microorganisms or ofpatients affected by candidiasis. Anti-receptorKTAb have been elicited and detected in animalsnever immunized with MAbKT4 but repeatedlyinfected with PaKT-sensitive C. albicans cells ( 13).Likewise, human natural KTAb prevalently ofthe IgA isotype were found in the vaginal secre-tions of patients affected by vaginal candidiasis(13). MAbKT4-affinity chromatography-purifiedhuman KTAb proved to be as candidacidal invitro as those KT-IdAb raised in rat vaginas byidiotypic vaccination with MAbKT4 (10). Theantifungal activity of human KTAb was totallyneutralized by previous reaction with MAbKT4,confirming the high specificity of their killer ef-fect. Human natural purified KTAb were alsocapable of transferring passive immunity to naiveanimals experimentally infected with PaKTR-bearing C. albicans cells (13). Human naturalKTAb were shown to interact with specific Pa-KTR of C. albicans cells by competition experi-ments with PaKT in immunofluorescence assays.(14).

In this study, the anti-P. carinii activity ofhuman natural candidacidal KTAb was evalu-ated. They were tested for their ability to inhibitthe in vitro attachment of rat-derived P. carinii tocultured cells as well as the infectivity of thesame organism to nude rats. PaKT was used tocontrol the sensitivity of P. carinii cells to the

killer effect and to test its binding to PaKTR ofrat-derived and human P. carinii in competitionwith human KTAb.

MATERIALS AND METHODSProduction of PaKTA standard suspension of 5 X 108 P. anomalaATCC 96603 (formerly defined as UCSC 25F)yeast killer cells, grown for 24 hr at 250C onSabouraud agar (Diagnostic Pasteur, Paris,France), was inoculated into 100 ml flasks ofSabouraud broth, buffered at pH 4.6. After 48 hrat 25°C, the yeast suspension was centrifugedand the supernatant containing the PaKT wasfiltered through 0.22-,um filters as previously de-scribed (4).

Monoclonal AntibodyAn IgGI monoclonal antibody (MAbKT4), whichneutralized the activity of PaKT against recog-nized PaKT-sensitive strains of C. albicans, wasproduced according to the standard procedurespreviously described (1 5). MAbKT4 was pro-duced as ascitic fluid, purified by ammoniumsulfate precipitation, dialyzed against Dulbecco'sphosphate-buffered saline (PBS), and stored at-200C.

Human Natural Killer Toxin-likeCandidacidal Antibodies (KTAb)Human natural KTAb were obtained and puri-fied as described recently (13). Briefly, naturalKTAb present in the vaginal secretions of womenafflicted by candidal vaginitis were purified byMAbKT4-affinity chromatography and dialyzedagainst PBS. The candidacidal activity of thesepurified KTAb was controlled by using a colony-forming unit (CFU) assay against C. albicans cellssensitive to the activity of PaKT according to aprocedure previously described (1 3). In thisstudy, candidacidal KTAb from two patients (Aand B) at the approximate concentration of 175jig/ml were used separately. Patients A and Bcorresponded to patient 1 (KTAb of prevalentisotype M) and patient 8 (KTAb of prevalentisotype G), respectively, as described by Polonelliet al. (1 3). Vaginal secretions from patients A andB that were not fixed by the MAbKT4 affinitychromatography (KTAb-depleted) were used asnegative controls.

546 Molecular Medicine, Volume 3, Number 8, August 1997

Source and Quantification of P. cariniiOrganismsCorticosteroid-treated outbred Wistar rats (IffaCredo, Lyon, France) were used as the source ofthe parasite. They were administered dexameth-asone (Merck Sharp and Dohme, Qualimed,France) in drinking water (2 mg/i) for 9-12weeks. Animals were housed in a conventionalroom and were given standard food (UAR,France) and water ad libitum. Parasite extractionand quantification were performed as previouslydescribed (2).

In Vitro Attachment AssayW138 VA13 subline 2RA (ECACC 85062512)cells were used as target cell populations. Theseepithelial lung cells were selected to develop anin vitro assay having similarities with in vivoinfection models. W138 VA13 cells were grownin Dulbecco's modified eagle's medium (DMEM;Biowhittaker, Verviers, Belgium) supplementedwith 10% heat-inactivated fetal calf serum (FCS)on glass coverslips placed in the wells of 24-wellflat bottom plates (Costar, Brumath, France).When cell monolayer cultures were subconflu-ent, the medium was removed and 1 ml of asuspension of 4 X 106 parasites in DMEM, pH7.2, was placed in each well. Before being incu-bated with cultured cells, P. carinii organismswere treated for 90 min at 370C in 500 ,ul of thefollowing reagents: DMEM, pH 7.2; KTAb immu-nopurified from patients A or B; vaginal secre-tions from patients A and B that were not fixedby the MAbKT4 chromatography (KTAb-deplet-ed); and KTAb (from patient B) mixed withMAbKT4. The last two reagents were used tocontrol the specificity of human KTAb activity.The plates were incubated for 24 hr in an atmo-sphere of 5% CO2 at 37°C. After incubation, thecoverslips were washed three times with phos-phate buffer at 370C to eliminate unattached anddead parasites. Each experiment was performedin triplicate. Attached P. carinii organisms weremicroscopically detected using methanol-Giemsaand counted by a previously described method(2,16).

Nude Rat ModelMale and female nude (nu/nu) rats 8 weeks ofage were used as recipient animals. These ani-mals came from a colony (Pasteur Institute ofLille, France) previously used in several P. carinii

transmission experiments which had been foundto be free of latent P. carinii infection. They wereimmunosuppressed for 15 days with corticoste-roid (dexamethasone; Merck Sharp and Dohme,France) in their drinking water (1 mg/i). Undercorticosteroid administration, the parasite num-ber showed low interindividual or intersexualvariation. Nude rats were anesthetized with adrug cocktail given intraperitoneally (17). Theywere inoculated by the intratracheal route with10 X 107 parasites/ 100 ,lI of medium along with0.3 ml of air. In these inocula, the cysts' percent-age by comparison with trophozoites was 4-7%.The tracheal wound was closed with nonresorb-able sutures (3/8 Seracap, Serquighy, France).All manipulations were carried out under sterileconditions and the hooded rat cages were placedin sterile boxes maintained under sterile air-flow(Esi Flufrance, Wissous, France). The cages, food,water, and bedding were sterilized before beingused.

Evaluation of P. carinii Infectivity inNude Rats

The experimental animal model of nude rats wasused at first to control that P. carinii infectivitywas affected by the antibiotic activity of PaKT.For this evaluation, five independent experi-ments were performed with four groups of fouranimals each. Samples of freshly extracted P.carinii organisms were incubated for 90 min at25°C in 500 ,ul of the following reagents beforebeing inoculated into the animals: DMEM, pH7.2; PaKT, pH 4.6 (active killer toxin); PaKT, pH7.2 (inactive killer toxin); and PaKT that hadbeen neutralized by mixing with mAbKT4 for 90min at 25°C. Another control group was repre-sented by rats inoculated with parasite-freeDMEM.

For the evaluation of human natural KTAbantimicrobial activity, before being inoculated inanimals, samples of freshly extracted parasiteswere incubated for 90 min at 37°C in 500 ,ul ofthe following preparations in PBS (pH 7.2):KTAb-depleted vaginal secretions from patientsA and B; KTAb (patient A); and KTAb (patient A)neutralized by previous mixing with MAbKT4.

The treated parasite samples were then cen-trifuged before being resuspended in completefresh DMEM, pH 7.2, and inoculated to immu-nosuppressed nude rats (4 rats/group) as de-scribed above.

All rats were given dexamethasone in theirdrinking water for 40 days. They were then sac-

N. Seguy et al.: Killer Toxin-like Antibodies against P. carinii 547

rificed for P. carinii organism extraction andcount.

xce

00m

le.

00

cn4.),I

co

Immunofluorescence Assay

The immunofluorescence studies were carriedout with rat-derived and human P. carinii organ-isms obtained from experimental animals andbronchial washings of patients affected by pneu-mocystosis by using biotinylated PaKT in compe-tition with human natural KTAb according to aprocedure previously described (18,19). Briefly,concentrated PaKT (8 mg/ml) was dialyzedagainst 500 ml of labeling buffer (0.1 M NaHCO3,0.1 M NaCl adjusted to pH 7.4 with concentratedHCI) at 40C with three changes over 2 days. Afterdialysis, 30 ,A of 10 mg/ml biotin (Calbiochem,La Jolla, CA) in dimethyl sulfoxide (SigmaChemical Co., St. Louis, MO) were added to eachmilligram of PaKT. The mixture was incubatedfor 3 hr at room temperature in the dark and atslow agitation. The unbound biotin was removedby dialysis against 500 ml of PBS at 40C withthree changes over 2 days. The biotinylated PaKTwas maintained at 4°C in the dark until used.Twenty microliters of the biotinylated PaKT (di-luted 1:2 in PBS) was added to each well of animmunofluorescence slide containing fixed P. ca-rinii organisms of rat and human origin andmaintained for 1 hr at 37°C in a humid chamber.The slide was successively washed three times(10 min each) with shaking in PBS. The slide wasallowed to dry at room temperature and 20 gl ofstreptavidin-fluorescein (Amersham Interna-tional, Little Chalfont, England) diluted 1:100 inEvans blue was added to each well and allowedto react for 20 min under the same conditions asabove. The slide was then rinsed with distilledwater and mounted with a coverslip using amounting fluid (Syva Microtrak, Palo Alto, CA).The slide was observed with a fluorescence mi-croscope (Zeiss Axiophot, Jena, Germany). As anegative control, the immunofluorescence assaywas carried out by using, in the same procedure,PBS in place of biotinylated PaKT. For competi-tion experiments, the same procedure was car-ried out by using biotinylated PaKT diluted 1:2 inhuman natural KTAb.

Statistical AnalysisResults were expressed as mean values + stan-dard deviation. Differences in parasite numbers

200

150

100

50

0

T- T I

T

3 4 5

FIG. 1. Inhibitory effect of human naturalyeast killer toxin-like antibodies (KTAb) onthe attachment of rat-derived P. carinii to W138VA13 cellsBefore the attachment assay, the parasites were pre-incubated in different reagents: DMEM, pH 7.2 (1);KTAb immunopurified from patient A (2); KTAb im-munopurified from patient B (3); KTAb-depletedvaginal secretions (4); and KTAb (from patient B)neutralized by MAbKT4 (5). *p < 0.001, the parasitenumber compared with that of parasites preincu-bated in KTAb-depleted vaginal secretions.

were evaluated by using the Student t-test (p <0.05 was considered significant).

RESULTSEffect of Human KTAb on P. carinii inVitro Attachment

Human natural KTAb immunopurified from thevaginal secretions of patients A and B induced asignificant and similar inhibition of rat-derived P.carinii attachment to W138 VA13 cells (81 +13.3 X 104 and 78 ± 20.1 x 104 attached para-sites, respectively). In contrast, this inhibitoryeffect was not observed when parasites weretreated either with DMEM (pH 7.2) or KTAb-depleted vaginal secretions from the same pa-tients used as controls (190 ± 23.6 x 104 and176 ± 28.5 x 104 attached parasites, respec-tively) (Fig. 1). The inhibition level obtainedwith KTAb (54% for patient A and 55.5% forpatient B) was evaluated in comparison to thatobtained with KTAb-depleted vaginal secretions.

250J

548 Molecular Medicine, Volume 3, Number 8, August 1997

700

600

x

In

;3

0wI)

.0a.3)

500

400

300

200

100

0

T

FIG. 2. Inhibitory effect of PaKT on the infec-tivity of rat-derived P. carinii to nude rats

In this experiment, a control group consisted of ani-mals inoculated with DMEM without parasites (1).Parasites were incubated for 90 min at 25°C in thefollowing reagents before being inoculated in ani-mals: DMEM, pH 7.2 (2); PaKT, pH 4.6 (active killertoxin) (3); PaKT, pH 7.2 (inactive killer toxin) (4);and PaKT neutralized for 90 min at 250C byMAbKT4 (5). *p < 0.001, the parasite number com-pared with that of parasites preincubated in DMEM,pH 7.2.

Moreover, when rat-derived P. carinii organismswere treated with KTAb from patient B neutral-ized by MAbKT4, the number of parasites at-tached to W138 VA13 cells was similar to thatobtained in the two controls (162 ± 28.4 X 104).

Effect of PaKT on P. carinii Infectivity in

Nude Rats

Five experiments were carried out consecutivelywithin 1 year. The corticosteroid-treated nuderats were nonlatently infected by P. carinii be-cause animals not inoculated with parasites didnot develop pneumocystosis (control withoutparasite). Figure 2 shows that rat-derived para-

sites preincubated in DMEM, pH 7.2, inducedsevere pneumocystosis in nude rats (439 ±

155 X 106 parasites/animal). The number of par-

asites in rats inoculated with inocula treated withactive PaKT (pH 4.6) was significantly lower(57 ± 10 X 106 parasites/animal) than that re-

corded in animals inoculated with parasitestreated with either inactive PaKT (pH 7.2)(431 120 X 106 parasites/animal) or parasites

treated with active PaKT neutralized by MAbKT4(472 ± 113 X 106 parasites/animal). An effectdue to the acidity of active PaKT was also ex-cluded on the basis of previous data showing thatparasites preincubated in DMEM, pH 4.6, re-mained as viable and infective as parasites pre-incubated in DMEM, pH 7.2 (2). Histological ex-amination showed that lungs of nude ratsinfected with parasites pretreated with inactivePaKT (pH 7.2) or active PaKT (pH 4.6) neutral-ized by MAbKT4 were characterized by the pres-ence of an eosinophilic exsudate typical of pneu-mocystosis which was absent in lungs of animalsinfected with parasites that had been pretreatedwith PaKT, pH 4.6 (data not shown).

Effect of KTAb on P. carinii Infectivity inNude Rats

Rat-derived parasites preincubated in KTAb-de-pleted vaginal secretions from patients A and Binduced severe pneumocystosis in nude rats(246 ± 45 X 106 parasites/animal). On the con-trary, when P. carinii organisms were preincu-bated in purified human KTAb from the samepatients, their infectivity was inhibited (63.5 ±17 x 106 parasites/animal). In comparison withthe appropriate control, the percentage of inhi-bition was 74.2% when the P. carinii organismswere pretreated with KTAb. MAbKT4 proved toneutralize the inhibitory activity of KTAb (frompatient A) as attested by the recorded number ofparasites (206 ± 6.8 X 106 parasites/animal),which was comparable to that obtained in ratsinoculated with parasites pretreated with KTAb-depleted vaginal secretions (Fig. 3).

Visualization of PaKTR in P. cariniiImmunofluorescence studies carried out by usingbiotinylated PaKT permitted visualization of theputative KTR in rat-derived and human P. cariniiorganisms (Fig. 4). The immunofluorescence re-activity was lost by the competition of humannatural KTAb against PaKT for binding to thesurface of P. carinii organisms of both animal andhuman origin.

DISCUSSIONP. carinii is an opportunistic microorganism re-cently reclassified in the kingdom Fungi on thebasis of nucleic acid analysis (20-23). P. cariniipneumonia became an increasingly important

---

II

ile

1%e-

N. Seguy et al.: Killer Toxin-like Antibodies against P. carinii

FIG. 4. Immunofluorescence visualization ofputative yeast killer toxin cell wail receptorson human P. carinii by biotinylated PaKT

0

FIG. 3. Inhibitory effect of the human naturalyeast killer toxin-like antibodies (KTAb) onthe infectivity of rat-derived P. carinii to nuderats

The preincubation media were: KTAb-depleted vagi-nal secretions (1); KTAb from patient A (2); andKTAb (from patient A) neutralized by MAbKT4 (3).*p < 0.001, the parasite number compared with thatof parasites preincubated in KTAb-depleted vaginalsecretions.

problem in terms of morbidity and mortalityamong immunocompromised patients, such as

those receiving transplantations and immuno-suppressive drugs. The advent of AIDS broughtabout an explosive increase in its prevalence. P.carinii became the most common life-threateningopportunistic pathogen in HIVinfected people(24). In immunocompetent hosts, P. carinii infec-tion generally stimulates cell-mediated and anti-body immunity. There is strong evidence thatboth of these types of immune responses are

important for the control of infections. The pre-cise role of humoral immunity, however, re-

mains to be clarified (25-27). An improved def-inition of the mechanisms of host immuneresponse would contribute to the developmentof new vaccination strategies for the control of P.carinii infections. The identification of new cellu-lar targets would also constitute the basis for thedevelopment of new drugs against pneumocys-tosis. Interestingly, new lipopeptidic cyclic anti-fungal compounds that inhibit the synthesis of ,31-3 glucan, such as echinocandin and papulo-candin, have proven to be highly effective in vivoagainst P. carinii (28). In the cell wall of P. carinii,

the presence of yeast glucan has been revealed(29). This represents one the major constituentsof cell wall receptors for some yeast killer toxinssuch as that of P. anomala, which displays a largespectrum of antimicrobial activity against manypathogenic microorganisms, including C. albi-cans, a common opportunistic pathogen in HIVpatients (30-32). In previous studies, we dem-onstrated that a candidacidal P. anomala killertoxin (PaKT) was also able to inhibit the in vitroattachment of rat-derived P. carinii (2) or its invivo infectivity (3). These antimicrobial activitieshave proven to be highly specific because theycan be abolished by prior adsorption of PaKT witha neutralizing monoclonal antibody (MAbKT4)that also allows the immunofluorescence detectionof specific PaKTR on the cell wall of P. cariniiorganisms treated with PaKT (2). MAbKT4 hasbeen extensively used to vaccinate mice and ratsby parenteral or mucosal administration againstexperimental intravenous and intravaginal infec-tions with PaKT-sensitive C. albicans cells (idio-typic vaccination) (9,10). Immunoprotectionwas related to the elicitation in the serum or

vaginal secretions of vaccinated animals of hightiters of PaKT-like antidiotypic antibodies (KT-IdAb) that were shown to interact with putativeC. albicans PaKTR and to kill in vitro the C. albi-cans cells used for experimental challenge. Theyalso conferred passive immunity when trans-ferred to unvaccinated animals, thus confirmingthat they represented the internal image ofPaKT. Inherent studies based on the speculationthat the dual steric homology between PaKT andKT-IdAb should conversely imply one betweenthe paratope of MAbKT4 used as a vaccine inidiotypic vaccination and the PaKTR of sensitiveC. albicans cells, led to the detection of anti-receptor PaKT-like candidacidal antibodies (KTAb)in the vaginal secretions of rats experimentally in-

T250

200

150

100

x

bJa2

5-,

.0U*0z

so

T

549

la An.J v

550 Molecular Medicine, Volume 3, Number 8, August 1997

fected with C. albicans cells and in women withvaginal candidiasis (13). The goal of this study wasto verify if such natural candidacidal antibodiesrepresenting the internal image of PaKT could alsodisplay activity against recognized PaKT-sensitiveP. carinii organisms, even though the study was notdesigned to draw definitive conclusions about therelationship of antibody concentration to killer ef-fect. We have shown that human natural candi-dacidal KCTAb were able to inhibit the in vitro at-tachment of rat-derived P. carinii organisms toepithelial lung cells as well as parasite infectivity inthe nude rat model (16). Data obtained from the invitro attachment assay showed that 55.5% of thehuman natural KTAb-treated parasites were notattached to WI38 VA13 cells. The same effecthad been observed by using PaKT-treated par-asites in other cell line systems (2,3). Humannatural KTAb were shown to display an inhib-itory activity comparable to that of PaKT (16).Data obtained with the nude rat model con-firmed this finding by showing inhibition ofthe infectivity of P. carinii organisms treatedwith KTAb as well as with PaKT. Moreover,when human KTAb were adsorbed withmAbKT4, the inhibitory effect on the attach-ment of parasite and infectivity was signifi-cantly neutralized (p < 0.001). These datasuggest that KTAb, representing the internalimage of PaKT, were able to interact with theidiotype of PaKT-neutralizing MAbKT4, thusattesting to the specificity of their antibioticactivity.

As marked differences related to host specieswere reported among P. carinii strains (33,34), itcould be argued that the results are only appli-cable to rat-derived P. carinii isolates. However,we have previously shown that PaKT is also ac-tive against mouse-derived P. carinii (4). More-over, PaKT or KTAb are equally active against C.albicans, which suggests that the killer activity ofthese compounds is dependent on transphyleticreceptors (14). The putative PaKTR on C. albicansand P. carinii, which were cytochemically locatedby using polyclonal rabbit KT-IdAb or PaKT andMAbKT4 with indirect immunofluorescence as-says (2,12), were also visualized either in rat-derived or human P. carinii organisms by usingbiotinylated PaKT in the direct immunofluores-cence procedure. The competition exerted byKTAb to PaKT, particularly in the human clinicalsamples, proved that despite the animal origin ofparasite, these antibodies bind to the same Pa-KTR of P. carinii as PaKT.

Although our experiments demonstrated theability of KTAb to inhibit rat-derived parasite invitro attachment and in vivo infectivity to thenude rat, these observations did not indicatewhether trophozoites were less or more affectedthan the cystic forms by the killer effect. Furtherexperiments are needed to evaluate the sensitiv-ity of the different life stages of P. carinii to PaKTor PaKT-like antibodies.

Our findings suggest that humoral immunitycould play a potential role in the control of P.carinii infection because PaKT-like anti-receptorantibodies are elicited through immunization byspecific PaKTR of infectious microorganisms act-ing as antigens (14). The validation of KT-likeantibodies, moreover, could lead to the cloningof genes from human lymphocytes of selecteddonors that encode for KTAb-variable regions,thus obtaining Fab characterized by antibioticactivity that could be most properly used as im-munotherapeutants in the treatment of pneumo-cystosis. The potential therapeutic effect of murineyeast killer toxin-like monoclonal and single-chainrecombinant antibodies previously described in anexperimental model of candidal vaginitis stronglysupports this expectation (18,19).

Further studies using properly engineeredMAbKT4 idiotype-like vaccines will be requiredto extend the prophylactic value of idiotypic vac-cination against experimental pneumocystosis toascertain if it could be used in individuals at riskof developing the disease.

ACKNOWLEDGMENTSThe authors thank Ms. Chantal Mullet from U42INSERM (Villeneuve d'Ascq) for the numerationof parasites, and Mr. Emile Fleurbaix and PierreBilliaut from the Institut Pasteur (Lille) for pro-viding the nude rats. This work was developed aspart of a European BIOMED- 1 Concerted Actionentitled "Pneumocystis and Pneumocystosis. Im-pact of the Biodiversity of Pneumocystis carinii onEpidemiology, Pathology, Diagnosis, Monitoringand Prevention of Pneumocystosis. New Thera-peutic Approaches" (BMHI-CT94-1 118). Thisproject was also supported by the Agence Natio-nale de Recherches sur le SIDA (ANRS). N.S. wasawarded a grant from ANRS (94635). L.P. andS.C. were supported by the Ministero dellaSanita-Istituto Superiore di Sanita, IX ProgettoAIDS Roma, Italy (9404/28).

N. Seguy et al.: Killer Toxin-like Antibodies against P. carinii 551

REFERENCES1. Sattler FR, Jelliffe RW. (1994) Pharmacoki-

netic and pharmacodynamic considerationsfor drug dosing in the treatment of Pneumo-cystis carinii pneumonia. In: Walzer PD (ed).Pneumocystis carinii Pneumonia, 2nd ed.,Vol. 69. Marcel Dekker, New York, pp. 467-485.

2. Aliouat EM, Cailliez JC, Seguy N, et al.(1993) Inhibitory effect of a yeast killer toxinto the in vitro Pneumocystis carinii attachment.Serodiagn. Immunother. Infect. Dis. 5: 102-106.

3. Seguy N, Aliouat EM, Dei-Cas E, Polonelli L,Camus D, Cailliez JC. (1994) Susceptibility ofPneumocystis carinii to a Pichia anomala killertoxin. J. Eukaryot. Microbiol. 41: 109S.

4. Seguy N, Cailliez JC, Polonelli L, Dei-Cas E,Camus D. (1996) Inhibitory effect of a Pichiaanomala killer toxin on Pneumocystis cariniiinfectivity to the SCID mouse. Parasitol. Res.82: 114-116.

5. Conti S, Cantelli C, Gerloni M, et al. (1996)Killer factor interference in mixed opportu-nistic yeast cultures. Mycopathologia 135: 1-8.

6. Polonelli L, Conti S, Gerloni M, Magliani W,Chezzi C, Morace G. (1991) Interfaces of theyeast killer phenomenon. Crit. Rev. Microbiol.8: 47-87.

7. Cailliez JC, Seguy N, Aliouat EM, Polonelli L,Camus D, Dei-Cas E. (1994) The yeast killerphenomenon: A hypothetical way to controlPneumocystis carinii pneumonia. Med. Hypoth-eses 43: 167-171.

8. Polonelli L, Morace G. (1988) Yeast killertoxin-like anti-idiotypic antibodies. J. Clin.Microbiol. 26: 602-604.

9. Polonelli L, Lorenzini F, DeBernardis F, et al.(1993) Idiotypic vaccination: Immunopro-tection mediated by anti-idiotypic antibodieswith antibiotic activity. Scand. J. Immunol. 37:105-110.

10. Polonelli L, De Bernardis F, Conti S, et al.(1994) Idiotypic intravaginal vaccination toprotect against candidal vaginitis by secre-tory, yeast killer toxin-like anti-idiotypic an-tibodies. J. Immunol. 152: 3175-3182.

11. Polonelli L, Conti S, Gerloni M, et al. (1991)'Antibiobodies': Antibiotic-like antiidiotypicantibodies. J. Med. Vet. Mycol. 29: 235-242.

12. Polonelli L, Fanti F, Conti S, et al. (1990)Detection by immunofluorescent anti-idio-typic antibodies of yeast killer toxin cell wallreceptors of Candida albicans. J. Immunol.Methods 132: 205-209.

13. Polonelli L, De Bemardis F, Conti S, et al.(1996) Human natural yeast killer toxin-likecandidacidal antibodies. J. Immunol. 156:1880-1885.

14. Cassone A, Conti S, De Bernardis F, PolonelliL. (1997) Antibodies, killer toxins and anti-fungal immunoprotection: A lesson from na-ture? Immunol. Today 18: 164-169.

15. Polonelli L, Morace G. (1987) Production andcharacterization of yeast killer toxin monoclo-nal antibodies. J. Clin. Microbiol. 25: 460-462.

16. Seguy N, Polonelli L, Conti S, Dei-Cas E,Camus D, Cailliez JC. (1996) In vitro decreaseof rat-derived Pneumocystis carinii attachmentinduced by human natural yeast killer toxin-like antiidiotypic candidacidal antibodies. J.Eukaryot. Microbiol. 43: 27S.

17. Aliouat EM, Dei-Cas E, Billaut P, Dujardin L,Camus D. (1995) Pneumocystis carinii organismsfrom in vitro culture are highly infectious to thenude rat. Parasitol. Res. 81: 82-85.

18. Magliani W, Conti S, De Bernardis F, et al.(1997) Therapeutic potential of antiidiotypicsingle chain antibodies with yeast killertoxin activity. Nat. Biotechnol. 15: 155-158.

19. Polonelli L, Seguy N, Conti S, et al. (1997)Monoclonal yeast killer toxin-like candi-dacidal antiidiotypic antibodies. Clin. Diagn.Lab. Immunol. 4: 142-146.

20. Cailliez JC, Seguy N, Denis CM, et al. (1996)Pneumocystis carinii: An atypical fungal mi-croorganism. J. Med. Vet. Mycol. 34: 227-239.

21. Edman JC, Sogin ML. (1994) Molecularphylogenic of Pneumocystis carinii. In: Pneu-mocystis carinii Pneumonia, 2nd ed., NewYork, Marcel Dekker, Vol. 69. pp. 91-106.

22. Stringer JR, Edman JC, Cushion MT, Rich-ards FF, Watanabe J. (1992) The fungal na-ture of Pneumocystis carinii. J. Med. Vet. Mycol.30: 271-278.

23. Wakefield AE, Peters SE, Banerji S, BridgePD, Hopkin JM. (1992) Pneumocystis cariniishows DNA homology with the Ustomycet-ous Red yeast fungi. Mol. Microbiol. 6: 1903-1911.

24. Walzer PD. (1991) Pneumocystis carinii-Newclinical spectrum? New Engl. J. Med. 324:263-265.

25. Beck JM, Liggitt HD, Brunette EN, Fuchs HJ,Shellito JE, Debs RJ. (1991) Reduction inintensity of Pneumocystis carinii pneumoniain mice by aerosol administration of gammainterferon. Infect. Immunol. 59: 3859-3862.

26. Ezekowitz RAB, Williams DJ, Koziel H, et al.(1991) Uptake of Pneumocystis carinii medi-

552 Molecular Medicine, Volume 3, Number 8, August 1997

ated by the macrophage mannose receptor.Nature 351: 155-158.

27. Su TH, Martin WJ. (1994) Pathogenesis andhost response in Pneumocystis carinii pneu-monia. Annu. Rev. Med. 45: 261-272.

28. Schmatz D, Romancheck M, Pittarelli L, etal. (1990) Treatment of Pneumocystis cariniiwith 1, 3-glucan synthesis inhibitors. Proc.Natl. Acad. Sci. USA 87: 5950-5954.

29. Matsumoto Y, Matsuda S, Tegashi T. (1989)Yeast glucan in the cyst wall of Pneumocystiscarinii. J. Protozool. 36: 21S-24S.

30. Sawant AD, Ahearn DG. (1990) Involve-ment of a cell wall receptor in the mode ofaction of an anti-Candida toxin of Pichiaanomala. Antmicrob. Agents Chemother. 34:1331-1335.

31. Polonelli L, Morace G. (1986) Reevaluationof the yeast killer phenomenon. J. Clin. Mi-crobiol. 24: 866-869.

32. Magliani W, Conti S, Gerloni M, BertolottiD, Polonelli L. (1997) Yeast killer systems.Clin. Microbiol. Rev. 10: 369-400.

33. Kovacs JA, Halpern JL, Lungren B, Swan JC,Parrillo JE, Masur H. (1989) Monoclonal an-tibodies to Pneumocystis carinii: Identificationof specific antigens and characterization ofantigenic differences between rat and hu-man isolates. J. Infect. Dis. 159: 60-70.

34. Mazars E, Odberg-Ferragut C, Dei-Cas E, etal. (1995) Polymorphism of the thymidylatesynthase gene of Pneumocystis carinii fromdifferent host species. J. Eukaryot. Microbiol.42: 26-32.

Communicated by A. Martinez-Palomo. Accepted June 18, 1997.