Serotyping Cryptococcus neoformans by immunofluorescence

JOURNALOFCLINICAL MICROBIOLOGY, Sept.1981, p. 313-317 0095-1137/81/090313-05$02.00/0

Serotyping Cryptococcus neoformans by

Immunofluorescence

WILLIAM KAPLAN,l SANDRA L. BRAGG,' STANLEY CRANE,' ANDDONALD G. AHEARN2

CentersforDiseaseControl, Atlanta, Georgia30333,'andDepartment of Biology, Georgia StateUniversity,

Atlanta,

Georgia

303032

Received23 January 1981/Accepted 29 April 1981

Four serotypes of Cryptococcus neoformans designated A, B, C, and D are

currently recognized. Althoughanagglutinationtestismostoften usedtoserotype

C. neofornans in cultures, this testis notappropriate fortyping the fungus in

fixed tissues. A study to prepare fluorescent-antibody reagents for typing C. neofornansin cultures and to determine whetherthey can be used to type this

fungusinfixed tissueswascarriedout.Antiserato onestrainbelongingtoeach of the four serotypes were prepared in rabbits by intravenous injection ofwhole Fornalin-killed cryptococci.Each antiserum waslabeled with fluorescein

isothi-ocyanate and then adsorbed with cells of each of theheterologousserotypes.The adsorbed conjugates were then tested against six serotype A isolates and five

isolates of each of the other threeserotypes. Labeledserotype A or Dantiserum adsorbed with eitherB orCcellsstained theAandD, butnottheB orC,isolates. Labeledserotype Bantiserum adsorbed withAcellsstained theBandC, butnot

theA orD,isolates. Labeled A antiserum absorbed with D cells differentiated A

from D; labeledC antiserum absorbed with B cells differentiated C from B. Of the 21 test isolates, 17 could be serotyped in paraffin sections of tissues of

experimentallyinfected mice.

In1949, Evans dividedisolates of

Cryptococ-cus

neoformans

into three

serological

types,A,

B,and C,onthe basis of theantigenic

makeup

oftheir

capsules (3).

Inalaterstudy,Wilsonet

al. (14)

distinguished

a fourth serotype which

was designated type D. All four serotypes can causedisease; however,they differin their

geo-graphic distribution (1).

Serotype

A has a

cos-mopolitan distribution.

Serotype

D is common

in someparts ofEurope, but it is

infrequently

encountered inthe United States.

Serotypes

B

and Ccause

cryptococcosis

infrequently,

except

in southem California. Recently, isolates of group BorC have been found tobeimportant

causesofthis disease in residents of southeastern Oklahoma

(11).

The four serotypes also differ in

their natural habitats (1).

Serotypes

A and D

are

commonly

foundin association with

pigeon

droppings, whereasthe natural reservoir of se-rotypes BandC is unknown. These differences

indicate that serotyping ofC. neoformans

iso-lateswould be of value from anepidemiological

standpoint.

Agglutinationis thetechniquemostoftenused to serotype C. neoformans. This procedure is

suitablefor use withcultures,but not with fixed tissues or other clinical materials. It would be

desirable to be able to type C.

neoformnans

in

histological specimens, because the results

ob-tained from such analyses would be of epidemio-logicalandecological value. Also, recent findings suggestthat theserotype oftheisolatecausing cryptococcosis might have therapeutic implica-tions (D. K. Henderson, J. E. Edwards, Jr., W.

E. Dismukes, and J. E. Bennett, Abstr. Annu.

Meet. Am. Soc. Microbiol.

1981,

Fll,

p.

315).

The

fluorescent-antibody (FA)

technique

has

beenused

successfully

todetect and

identify

C.

neoformans in fixed tissues (8,9)andincultures (7, 13), suggesting that theFAtechniquemight

also be suitable fortyping C. neoformans in such specimens. The present study was undertaken

todevelopFA reagentsfor serotyping this

fun-gusin culturesand todetermine theirvalue for carryingoutsuchananalysiswith fixedtissues.

MATERLA1S AND METHODS Cultures. Twenty-one isolates of C. neoformans, whoseserotypes had beenestablished bythe

aggluti-nationprocedure, werestudied.Six of these cultures (B551, B3166, B3168, B3169, B3170, and B3270)

be-longed to serotype A, five (B3171, B3172, B3173,

B3268, and

B327J)

belonged to serotype B, five

(B3182, B3183,B3184,B3267, and B3269)belonged to serotypeC,andfive(B3176, B3177, B3178, B3179, and

B3266)belonged to serotype D. With the exception of

isolate B551, these cultureswere obtained from the

collectionmaintainedbythe Department of Biology,

Georgia State University, Atlanta. Isolate B551 was

obtained from the culture collection maintained by

313

on February 7, 2020 by guest

http://jcm.asm.org/

the Mycology Division, Centers for Disease Control, Atlanta, Ga.

Antisera.Theantigens used for the production of

antiserawerepreparedfrom four isolates of C. neofor-mansrepresenting each of the serotypes(B551, B3172, B3182, and B3179). These cultures were selected on the basis of smallness ofcapsule when compared with the otherisolates in theirrespective serologicalgroup when grown onSabourauddextrose agar slants for 72 h at 35°C. This criterion for selection was based on

earlier observationsthathighestserumantibody titers areobtainedwhenlightly encapsulatedC.neoformans

cellsareusedasimmunizationantigens (12). Antigens

used for immunizationwereprepared by growing cul-turesB551, B3172,and B3179 in Sabourauddextrose

broth. The inoculaforthese broth cultures consisted

of two loopfuls of surface growth from Sabouraud

dextrose agar slants incubated for96hat25°C. The

cultureswereshaken at 180rpmfor 72hwhilethey

wereincubatedat25°C.Culture B3182wasgrownon modified Sabourauddextrose agar,pH5.0,containing 2.9% NaCl for 72 h at35°C to reduce the size ofits

capsule (2).The inoculum consisted ofoneloopful of

growthon a Sabouraud dextroseagar slantthat had been incubated for96h at25°C. The cultureswere killed with Formalin (1.0%) and then washed three times with Formalin-treated (0.5%) physiological sa-line. Thewashed cellswereresuspendedin

Formalin-treated (0.3%) 0.01 Mphosphate-buffered saline, pH

7.4.Vaccineswerepreparedby adjustingthe suspen-sions of the B551, B3172, and B3179 cells to the

turbidity ofa no.10and thesuspensionofB3182 cells to the turbidityof a no. 8 McFarland nephelometer standard.

Four rabbits per immunization antigen were used

for theproductionof antiseratostrains B551(type A), B3172 (typeB),and B3179 (type D). However,seven

rabbitswereused for theproductionofantiserumto

strain B3182(typeC).Beforeimmunization,each

rab-bitwasbled and thepreimmunizationserum wasused as a control for its corresponding antiserum. Each

rabbitwasinjectedintravenouslywith 10 3-mldoses

of vaccinespreadoveraperiodof4weeks. Thevaccine wasadministered at3- or4-dayintervals. Starting 1 weekafter the 10th injection, the rabbits were bled onceeachweek for 5 weeks. Doses of1and2mlof the

vaccine were administeredintravenously during this

periodtomaintainthe titer.

Conjugation ofsera. Each preimmunization se-rum andantiserum was labeled with fluorescein

iso-thiocyanateataratioof3mg ofdyeper ml ofserum. The whole-serumconjugateswerepreparedasfollows. The dye wasdissolved in 0.1 M Na2HPO4 (1 mg of fluorescein isothiocyanateper ml ofbuffer).The

flu-oresceinisothiocyanatesolutionwasthen addedtothe serum, and the pH wasadjusted to 9.5 with 0.1 M

Na3PO4.The mixturewasthenallowedtostandfor3 hat roomtemperature. Unreacted fluorescein isothi-ocyanatewasremovedbygelfiltrationwithSephadex

G-25.Conjugateswerediluted with0.01M

phosphate-buffered salinecontaining1.0% bovineserumalbumin and Merthiolate (1:5,000) adjusted to pH 8.5 to 9.0. The titers of these conjugates were based on the volume ofconjugate relative to the volume of

unla-beled serum used to prepare the reagent. The conju-gates werestored at 4°C.

Testing of conjugates. Tests for the sensitivity and specificity of the conjugates were carried out on smears ofsuspensions of Formalin-killed C. neofor-mans cells that had been cultivated on Sabouraud dextrose agarat 37°C for 5 days. The ability of the conjugates to stain their respective immunization strain was the basis for selecting a conjugate of each serotype for detailed study. The conjugates selected were testedfor sensitivity with all strains of the same serotype as the immunization antigen. These conju-gates werethen tested for specificity with all strains belonging to the other serotypes. In all of these tests, reagentswere used at dilutions of 1:10, 1:40, 1:160, and 1:640.

Adsorptionprocedure.The cells used for adsorp-tions were prepared from cultures grown in brain heart

infusionbroth. The broths were inoculated with two loopfuls of growth from Sabouraud dextrose agar slants that had been incubated for 96 h at 25°C. The cultures were shaken for 72 to 96 h while they were incubated at250C.They were treated with Formalin (1.0%),washed twice in Formalin-treated (0.5%)

phys-iologicalsaline, and stored at 4°C in Formalin-treated

(0.5%)saline.Conjugates diluted at 1:10 were adsorbed with half volumes of packed cells at 37°C in a water bathfor2 h. Adsorptions were then continued over-night at 4°C. Each conjugate selected for detailed study was adsorbed with cells of the homologous im-munization strain and with cells of the other three heterologous strains. Adsorptions were repeated until staining of the adsorbing strain had been eliminated.

Staining procedure and visualization of

stainedpreparations. Theprocedure used to stain the preparations was the one described by Kaplan and Ivens (5).Allpreparations were examined with a Leitz

Ortholux II incident-light fluorescence microscope.

Criteriasimilar to those of Goldman (4) and Moody et al. (10) were used to rate the intensity of staining. Results were recorded as -, ± (no reaction), 1+, 2+, 3+,and4+(reaction).

RESULTS

Staining properties of unadsorbed la-beled sera. The lala-beledpreimmunization sera didnotstain theC.neoformans strains. Antisera obtained from all four rabbits immunized with the B551 antigen (serotype A), when labeled, intenselystained cells ofthehomologous strain. One conjugate, prepared from serum collected fromone of the rabbits8 weeks after immuni-zation wasbegun,gave thebrightest staining.It was selected for further study and was desig-nated A-1.

Antisera from onlyone oftherabbits immu-nized with the B3172antigen (serotype B),when labeled,brightly stainedthehomologous organ-ism. The conjugate prepared from serum col-lected 8 weeks after immunization had begun gavethe beststaining reactions.Itwasselected forfurtherstudyandwasdesignatedB-1.

on February 7, 2020 by guest

http://jcm.asm.org/

Antisera obtained from onlyoneof theseven

rabbitsimmunized with theB3182 antigen (se-rotypeC), whenlabeled,stainedthehomologous strain. Theconjugate prepared fromserum

col-lected 5 weeks after immunization had begun appearedtogive the best stainingreactions. It

wasselected for furtherstudyanddesignated

C-7.

Antisera obtainedfrom allof the rabbits im-munized with the B3179 antigen (serotype D), whenlabeled, brightly stained the homologous strain. Theconjugate preparedfromserum col-lected 8 weeks after immunization had begun

gave the brightest staining. It was selected for furtherstudyandwasdesignatedD-4.The stain-ing properties of these fourconjugateswerethen

studied with allof the strainsbelongingtoeach of therespectiveserotypesand withthose ofthe heterologous serotypes. The staining titers ob-tainedarepresentedinTable1.

Staining properties of adsorbed labeled antisera. Portions of each ofthe labeled

anti-sera selected for further study were adsorbed

with cells of the homologous immunization strain andcross adsorbed withcells of each of

the three heterologous immunization strains. Thestainingpropertiesof each of the adsorbed

conjugateswerestudied withall21typedstrains. Thestainingreactions observedaresummarized

in Table2.

Use of adsorbed conjugates to serotype C. neoformans in paraffi sections of For-malin-fixed tissue.Groups of three micewere

experimentally infected with each of the 21

se-rotyped strains. Infections wereinduced by

in-traperitonealinjectionof0.5mlofsaline

suspen-sions of each culture that had beengrown on

Sabourauddextroseagarfor 96 h at25°C.These suspensions were prepared by dispersing the

surface growthfromoneagarslant in 5.0mlof physiological saline. Themiceweresacrificed 1 to2weekslater, and portions ofallorganswith visible lesions were fixed in 10.0% neutral

For-malin and then embedded inparaffin. Sections of these tissues were stained withhematoxylin

andeosinandby the Gomori

methenamine-sil-TABLE 1. Staining titers of selectedunadsorbed

fluorescein-labeled antiseratostrainsof C.

neoformansbelongingto serotypesA, B, C,and D Stainingtiteratoserotype:

Conjugate

A(6)b B (5) C(5) D (5)

A-1 160-640 160-640 40 5640

B-1 >640 5640 5640 5640

C-7 160-5640 >640 160-5640 40-640 D-4 160-640 40-160 40-160 >640

aReciprocalofdilution. 'Numberofstrains.

TABLE 2. Stainingproperties of adsorbed fluorescein-labeled C. neoformans antisera with homologous andheterologousserotypeantigens Staining reactiona of C. neofor-Conju- Adsorbingan- mansserotype:

gate tigen

A(6)b B(5) C(5) D(5)

A-1 SerotypeA - -

-cells

SerotypeB + - - +

cells

Serotype C + - - +

cells

Serotype D + -(4)b

-cells + (1)

B-1 SerotypeA - + +

-cells Serotype B

cells

Serotype C + + - +

cells

Serotype D + + +

-cells

C-7 SerotypeA - +(3) +

-cells -(2)

Serotype B - - +

-cells Serotype C

cells

Serotype D +(3) + +

cells -(2)

D-4 Serotype A

-cells

Serotype B + - - +

cells

Serotype C - - - +

cells Serotype D

cells

a

_,

Negative reaction; +, staining that ranged from

1+to4+.

bNumber of strains.

vernitrate procedure. An examination of these

preparations disclosed the presence of

crypto-cocci in thelesions. Replicatesections were then testedwith selected adsorbedconjugatesto

de-termine whetherthese reagents could be used to serotype the cryptococci present. The

proce-dures usedtopreparethesectionsfor FA

stain-ing and tostain them were those described by KaplanandKraft(6).

The serotyping was carried out in two steps with fouradsorbedconjugates. Inthefirst step, thesections were tested with theA-1 conjugate

that had been adsorbed with B

cells

and also withtheB-1 conjugate that had beenadsorbed

with Acells. The purposeofthis first step was todeterminewhether the cryptococci belonged

toeitherserotype A or D or serotype B or C. In the second step, the sections were tested with theA-1conjugate that had been adsorbed with VOL.14,1981

on February 7, 2020 by guest

http://jcm.asm.org/

316 KAPLAN ET AL.

D cellstoseparate A from D, if they had been foundinthe first step to belong togroupAorD, orwith the C-7 conjugateadsorbed withBcells

toseparate C fromB,ifthey had beenfoundin the first step to belong togroupBorC.

Inthispreliminarystudy, inallcases we were

able to determine whether the cryptococci in the fixedtissues belonged to eithergroup A

orDorgroupBorC.Withtheexceptionofone

serotype C strain,we were abletoseparatethe C strains from the B strains. In tissues, this particulartype Cstrain did not react with the C-7conjugatethathad beenadsorbedwith type B cells. It was also possible in most cases to

differentiatethetype Astrains fromthetypeD strains. In one case, however, a type A strain wasstained equivocallywith the A-1 conjugate that had been adsorbed with D cells; in two

casesinvolvingtype Dstrains,someof thefungal

cellsin the tissue sectionswereobservedto react

equivocallywith theadsorbed A-1conjugate. Use of adsorbed conjugates to serotype

C. neoformans cultures of unknown

sero-type. SixtyC. neoformanscultures ofunknown serotype were typed by tihc 'A technique. Of

these, 47were obtainedfromvarious

investiga-torsin theUnited States and 13wereobtained

fromforeignlaboratories.Thirty-fiveofthe cul-tures fromthe United States had beenisolated from clinical materials, five had been isolated from environmental samples, and thesourceof seven wasunknown. Sixofthe cultures submit-tedbyforeignlaboratorieswerefromIsrael,five

werefrom Colombia,andtwo werefrom Spain.

Theconjugatesusedfortypingwerethoseused with the sections of fixed tissues. As was done

with the tissue sections, thetypingwas carried

outin twosteps.

Thereagents andprocedure usedenabled us

to serotype all 60 cultures. A total of 41 were

foundtobetype A,2weretypeB,5weretype

C,and 12weretypeD.

DISCUSSION

The purposeof thisstudywastodevelop FA

reagentsforserotyping C.neoformans, andthis goalwas met. Byreciprocal cross staining and cross adsorption procedures, FAreagents were

producedthat, when used in appropriate

com-bination, make possible the serotyping of this

fungus. Wefoundthatlabeled antiserum to

se-rotype A orD,whenadsorbedwitheither Bor

C cells,differentiatedtypesA and D fromtypes B and C. Conversely, labeled antiserum to

se-rotypeB,whenadsorbedwithAcells,

differen-tiated B and Cfrom Aand D. Separationof A

from Dwasdone by using labeled antiserumto

A that had been adsorbed withDcells;

separa-tion of C from B was done by using labeled antiserum to C that had beenadsorbed with B

cells.

Serotyping bymeansof these reagentscanbe

carriedoutinone or twosteps.We

prefer

to use atwo-step

procedure

because itsavesreagents. In thefirst step wedetermine whetherthe

iso-late belongstogroup AorDortogroup BorC.

In the second step the group is separatedinto itsrespectivecomponents.

In this study we did notencounter any diffi-culties inserotyping C.neoformansincultures. The staining results wereclear-cut and unequiv-ocal. Inpreliminary tests with sections of fixed tissues, wecould also readily establish whether theC.neoformans

cells

presentbelonged to the A or D or the B or C group. In some cases, however, it was not possible todifferentiate the components in these two groups. Further studies areneededtoresolve this problem.

The agglutination procedure has been

suc-cessfully usedforserotyping C.neoformans

cells

fromcultures. Thus, the question can be raised as tothe needforimmunofluorescence to type

such

cells.

Immunofluorescence hasseveral

im-portant advantages over agglutination. First,

contaminated cultures can be typed by the FA technique, whereas this may not be possible by

agglutination,withouttime-consuming purifica-tion procedures. Also, nonviable cultures with scantgrowth can be typed by the FA technique, whereas this might not be possible by aggluti-nation.

In mostlaboratoriesFAreagents areprepared from globulin fractions of sera. This is done to reducepresumednonspecific staining and to ob-tain more intense specific staining due to the

concentrationofantibodies.Inthepresent study whole-serum conjugates were used. Our decision tousesuch FA reagents wasbased upon the fact that they are easy toprepareand can be made in 1 day. Furthermore, we did not encounter nonspecific staining with these conjugates.

LITERATURE CITED

1. Bennett, J. E., K. J.Kwon-Chung,and D. H.Howard. 1977. Epidemiologic differences among serotypes of Cryptococcusneoformans.Am.J.Epidemiol. 105:582-586.

2.Dykstra,M.A., L. Friedman,and J. W.Murphy.1977. Capsulesize ofCryptococcusneoformans:control and relationshiptovirulence. Infect. Immun. 16:129-135. 3. Evans, E. E. 1949. Animmunologic comparisonoftwelve

strains of Cryptococcus neoformans (Torula histo-lytica).Proc.Soc.Exp.Biol. Med. 71:644-646.

4. Goldman, M. 1957. Staining Toxoplasmagondii with fluorescein-labelled antibody.II.A newserologic test

forantibodiestoToxoplasma baseduponinhibition of specific staining.J.Exp.Med.105:557-573.

5. Kaplan,W.,and M. S. Ivens. 1960. Fluorescentantibody staining ofSporotrichum schenckiiin culturesand

on February 7, 2020 by guest

http://jcm.asm.org/

SEROTYPING C. NEOFORMANS BY 317

ical materials. J. Invest. Dermatol. 35:151-159. 6. Kaplan,W., and D. E. Kraft. 1969. Demonstration of

pathogenic fungiinformalin-fixed tissueby

immunoflu-orescence.Am. J.Clin. Pathol.52:420-421.

7. Kase,A.,and J. D. Marshall. 1960. Astudyof

Crypto-coccusneoformans bymeansof the fluorescent anti-body technique.Am. J.Clin. Pathol.34:52-56. 8. Marshall, J. D., L. Iverson, W. C. Eveland, and A.

Kase.1959.Comparisonoffluorescentantibody stain-ing andspecial histologic stains for the identification of Cryptococcusneoformans.Am.J. Pathol.35:684-685. 9. Marshall, J. D.,LIverson, W.C.Eveland, and A.

Kase.1961.Applicationand limitations of the

fluores-centantibody staininthespecific diagnosisof crypto-coccosis.Lab.Invest.10:719-728.

10. Moody,M.D.,M.Goldman, and B.M. Thomason. 1956.Staining bacterialsmearswithfluorescent

anti-body. I. General methods for Malleomyces pseudo-mallei. J. Bacteriol. 72:357-361.

11.Muchmore, H. G., E. N. Scott, F. G. Felton,andR. A. Fromtling. 1980. Cryptococcus neofonnansserotype

groupsencountered inOklahoma. Am. J. Epidemiol. 112:32-38.

12.Neill, J. M., I. Abrahams, and C. E. Kapros.1950.A

comparisonofthe immunogenicity of weakly

encapsu-lated and of strongly encapsuencapsu-lated strains of

Cryptococ-cusneoformans (Torula histolytica). J. Bacteriol.59:

263-275.

13. Pidcoe,V., and L. Kaufman.1968.Fluorescent-antibody

reagentfortheidentificationofCryptococcus

neofor-mans.Appl. Microbiol.16:271-275.

14. Wilson, D. E., J. E. Bennett, andJ.W.Bailey.1968. Serologic grouping of Cryptococcus neoformans.Proc. Soc.Exp. Biol. Med. 127:820-823.

on February 7, 2020 by guest

http://jcm.asm.org/