Journal of Medical Virology 71:404-407 (2003)

Cytomegalovirus Glycoprotein B Genotypes andCentral Nervous System Disease in AIDS Patients

Lucy S. Vilas Boas;' Vanda A. Veda F. de Souza,' Augusto C. Penalva de Oliveirar'Ana T. Rodriguez ViSO,I;2Anderson M. Nascimento Filho, 1

Maria C. Nascimento;' and Cláudio S._Pannuti1*lLaboratório de Virolagia (LIMHC), Instituto de Medicina Tropical de São Paulo,Departamento de Doenças Infecciosas e Parasitárias da Faculdade de Medicina,Universidade de São Paulo, São Paulo, Brazil2Instituto de lnfectologia Emilio Ribas, São Paulo, Brazil

To investigate any association between cytome-galovirus glycoprotein B (CMV gB) subtypes andcentral nervous system (CNS) disease in AIOSpatients, proportions of different gB genotypesdetected in AIOS patients with CNS disease werecompared with the gB genotypes detected inAIOS patients with no neurological disorder. Thepatients were matched by C04+ cell counts, CMVwas detected by PCR in cerebrospinal fluid (CSF)samples obtained from AIOS patients with CNSdisease and from urine and saliva samplesobtained from AIDS patients without CNS dis-ease. CMV strains obtained were digested byrestriction enzymes Hinffl and Rsa/to classify thegenotypes. The CMV gB genotype was deter-mined in 26 CSF samples. Ofthese, 11/26 (42.3%)typed as gB group 1, seven (26.9%) as gB2, four(15.4%) as gB3, and four (15.4%) as gB4. The CMVgB genotype frequency distribution in the 42AIOS patients without CNS disease showed that18/42 (42.8%) were classified as gB group 1, 10(23.8%) as gB2, seven (16.6%) as gB3, and seven(16.6%) as gB4. In the present study, no associa-tion was found between CMV gB genotypesand CMV-related central nervous system disease.J. Med. Virol. 71:404-407,2003.© 2003 Wiley-Liss, Inc.

KEY WORDS: cytomegalovirus; PCR; geno-types; gB; CNS disease; AIOS

INTRODVCTION

The glycoproteín B (gB) ofthe cytomegalovirus (CMV)constitutes a major component of the viríon envelope.Experiments in vitro have shown that CMV gB promotesvirion penetration into the cell, transmission ofinfectionfrom cell to cell, and fusion of infected cells [Navarroet al., 1993; Tugizov et al., 1994, 1995; Pietropaolo andCompton, 1997]. The gB protein elicits strong neutraliz-

© 2003 WILEY-LISS, INC.

ing antibody and cytotoxic T-cell immune responses ininfected individuaIs [Rasmussen et al., 1988; Kniesset al., 1991; Liu et al., 1991; Marshall et al., 1992;Schoppel et al., 1997; Speckner et al., 1999]. Levels ofglycoprotein-specific antibodies and detection of viralDNA in peripheral bIood are inversely correlated[Schoppel et al., 1997]. In recipients of allogeneic bonemarrow transplants, high titers of glycoprotein-specificneutralizing antibodies are correlated with the absenceofviral DNA in the blood [Schoppel et al., 1998].

Although strains ofCMV propagated in the laboratoryare genetically distinct from the wild type [Prichardet al., 2001], this variatíon does not affect the gB region.Restriction enzyme analyses have reveaIed geneticvariation in the gB region in clinical strains of theCMV, and clinical CMV isolates have been found toassume one of four gB sequence configurations [Chouand Dennison, 1991]. The gB genotype has been shownto correlate with cell tropism in vivo and may influencethe virulence of CMV [Meyer-Konig et al., 1998b]. Thefinding that CMV envelope proteins activate inflamma-tory cytokine responses via CD14 and Toll-like receptor2 (TLR2) [Compton et al., 2003] provides anotherrationale for studyíng a possible association betweengB and disease olÚcome.

A number of studies have attempted to correlate thedistribution of the four gB CMV genotypes with theclinical outcome of CMV infection [Fries et al., 1994;Bongarts et al., 1996; Shepp et al., 1996; Vogelberget al.,1996; Torok-Storb et al., 1997; Chern et al., 1998; Peeket al., 1998; Rosen et al., 1998; Gilbert et al., 1999;

Grant sponsor: CNPQ; Grant number: 300317/97-2; Grantsponsor: FAPESP; Grant number: 97/00541-7.

*Correspondence to: Dr. Cláudio S. Pannuti, Instituto deMedicina Tropical de São Paulo, Av. Enéas de Carvalho Aguiar,470, São Paulo, 05403-000, Brazil. E-mail: cpannuti@Usp.br

Accepted 30 May 2003DOI 10.1002/jmv.10488Published online in Wiley InterScience(www.interscience.wiley.com)

CMVGenotypes and CNS Disease in AIDS Patients

Aquino and Figueiredo, 2000]. The aim of the presentstudy was to investigate a possible association betweenCMVgB subtypes and central nervous system disease inAIDS patients.

MATERIALSANDMETHODS

PatientsFrom February 1995 to December 2000, cerebrospinal

fluid (CSF) samples from 203 AIDS patients exhibitingvarious central nervous system diseases were tested forCMV-DNA by PCR. As controls, we examined CMVsamples obtained from the urine and saliva of AIDSpatients with asymptomatic CMV infection. Asympto-matic CMVinfection was defined as the isolation ofCMVin a patient with no clinical, neurological, or laboratorysigns ofinfection. AlI control patients were submitted tophysical and neurological examination, fundoscopy anddetermination of CMV antigenemia, and were matchedwith CSF -CMV positive cases by CD4+ cell counts.

The study was approved by the Ethics and ResearchCommittee of the University of São Paulo School ofMedicine (CAPPesq).

Sample PreparationCSF was boiled for 10min and spun in a microfuge for

15 min at 4°C. Urine samples were diluted 1:4 inMILLIQ water, while saliva samples were diluted 1:1in a lysis buffer consisting of 10 mM Tris-HCI (pH 8.3)0.9% NP 40 and 0.9% Tween 20, with proteinase K(100 Ilg/ml). The mixture was then incubated at 55°C for1 hr, followed by denaturation ofthe enzyme at 95°C for10 mino

gB PCR and GenotypinggB gene amplification and subsequent genotyping

were performed directly on patients' specimens usingpreviously described methods [Chou and Dennison,1991]. Primers gB 1319 and gB 1605 amplified a regionof high variability in the gB gene. The gB PCR productswere then cut in two separate reactions withHinffl andRsaI, and were classified into gB types 1-4.

Statistical AnalysisProportions of CMV genotypes detected in CSF and

saliva or urine were compared by chi-square or Fisher'sexact test using Epi-Info 6 (version 6.04d). For contin-uous variables that were normally distributed (CD4+cell counts), Student's t test was used. P values < 0.05were considered significant. -

RESULTSCytomegalovirus Deteetíon in Patíents

With and Without Central N~rv~us .System (CNS)Disease

Cytomegalovirus DNA was successfully amplified in30 samples from 203 AIDS patients exhibiting variousCNS diseases. Of the 129 AIDS patients without CNS

405

disease, CMVwas detected in 39 (26 from urine samplesand 16 from saliva samples). From three patients ofthisgroup, CMV was amplified from both urine and saliva.Of these, two yielded the same gB genotype in urine andsaliva (gB1 or gB2) and the remaining one yielded CMVgB1 in the urine and CMV gB3 in the saliva.

CD4+ Cell CountsThere was no difference between the CD4-t cell counts

in the group of patients exhibiting CMV-related CNSdisease (range, 1-106 cells/mm'', median = 13, mean =25.1 ± 28.0 cells/mm'') and AIDS patients without CNSdisease (range, 5-86 cells/mm", median=37, mean=36.8 ± 24.6 cells/mm'') (P = 0.14).

Distribution of gB Genotypes in CSF SamplesThe gB genotype was determined in 26 out of the 30

CMVsamples. Ofthese, 11/26(42%)typed as gB group 1,7 (26%) as gB2, 4 (15%) as gB3, and 4 (15%) as gB4.

Distribution of gB Genotypes in AIDS PatientsWithout CNS Disease

-~AllCMV strains detected in the clinical specimens ofsaliva and urine could be assigned to gB genotypes 1-4;no additional gB types were found in this study. The gBgenotype frequency distribution in the 42AIDS patientswithout CNS disease showed that 18/42 (42.8%) belong-ed to gB group 1, 10/42 (23.8%) to gB2, 7/42 (16.6%) togB3, and 7 (16.6%) to gB4.

There was no difference between the gB genotypefrequency distribution of CMV samples from the CSF ofAIDS patients exhibiting central nervous system dis-eases and those from the urine or saliva ofAIDS patientswithout CNS disease (Table 1).

There was no difference between the gB1 and gB2genotype frequency both in the CSF of patients mani-festing CMV-related neurologic disease (42.3% vs.26.9%, respectively, P =0.06) and in the urine andsaliva ofAIDS patients without CNS disease (42.8% vs,23.8%, respectively, P = 0.06).

/ DISCUSSIONCorrelation between gB genotype and the risk for

CMV disease has been evaluated in various patientgroups. In bone marrow transplant recipients, patientswho survived CMVinfection more frequently shed CMV

TABLEI. Distribution of GlycoproteinB Types in CSFSamples From AIDSPatients Exhibiting CMV-RelatedCNSDisease, and in Saliva and Urine fromAIDSPatients Without

CNSDisease

Urine or salivagBtype CSf (n=26) (n= 42) P-value

1 11 (42.3%) 18 (42.8%) 0.962 7 (26.9%) 10 (23.8%) 0.773 4 (15.4%) 7 (16.7%) 0.844 4(15.4%) 7 (16.7%) 0.84

406

of the gB type 1 genotype than those who died [Frieset al., 1994]. These findings have been confirmed byother studies showing that disease and death due toCMVare significantly lower in marrow transplant reei-pients infected with the gB1 genotype compared topatients infected with other gB subtypes [Hebart et al.,1997; Torok-Storb et al., 1997; Woo et al., 1997].However, in renal [Vogelberg et al., 1996; Woo et al.,1997; Aquino and Figueiredo, 2000] and liver [Rosenet al., 1998] transplant recipients, gB type did not cor-relate with the development of symptomatic or tissue-invasive CMVdisease.

Although an initial study in AIDSpatients reported ahigher risk ofCMVretinitis in those presenting viremiacaused by gB group 2 CMVstrains [Shepp et al., 1996],subsequent studies found no association ofglycoproteinB genotypes with the viral DNAload or the presence ofretinitis [Chern et al., 1998; Peek et al., 1998; Gilbertet al., 1999]. In the present study, the ratio among gBsubtypes in the CSF of patients manifesting differentCMV-related neurological diseases was similar to that ofsubtypes found in the urine or saliva ofAIDS patientswith asymptomatic CMV infection. In AIDS patients,progression to invasive disease is associated.with a lowCD4+ T-lymphocyte count [Gallant et al., 1992;Monforte et al., 1993; Monforte et al., 1997]. The factthat in the present study both patients with or withoutCNS disease were matched by CD4+ cell countsdecreased possible interference ofthe patients' immunestatus with the invasive capacity of the different CMVstrains.

In previous studies, the gB2genotype was found moreoften in AIDS patients than was gB1 [Fidouh-Houhou'et al., 2001]. In the present study, the gB types 1 and 2were the CMVgenotypes most frequently found, both inthe CSF ofAIDS patients with CMV-related neurologi-cal disease (42.3% and 26.9%, respectively) and in theurine and saliva ofAIDS patients without CNS disease(42.8% and 23.8%, respectively). However, the differ-ence in the prevalence of CMVgB1 and CMVgB2 wasnot statistically significant in our data, neither in theCSF samples (P = 0.24) nor in the urine and salivasamples (P = 0.06). The present study is relativelylimited both in patient numbers and in geographicaldistribution and before this type of information can beextrapolated to the population at large, more samplesare needed, and more patients from a wider geographi-cal distribution are required. In previous reports,geographicaI and demographic differences in patientswere shown to affect the frequency ofCMVgBgenotypes1-4 in immunocompromised patients [Rasmussenet al., 1997;Wada et al., 1997;Wooet aI., 1997;Zipetoet al., 1998]. Further, gB genotypes may exhibit dif-ferent cell tropisms in vivo, as demonstrated by thelower capacity of gB type 1 to infect T-Iymphocytes invivo in BMT and solid organ recipients [Meyer-Koniget al., 1998b],and by the observation that gB4was foundmore frequently in the semen than in leukocytes ofHIV-infected patients with CD4 cell counts <100;"'1.1(Ras-mussen et al., 1997].1t is thus important to consider that

Vilas Boas et al.

the frequency of a specific CMVgB genotype may be afunction of the body compartment under evaluation.Different results could have been obtained in the pre-sent study if CMV plasma/leukocytes samples wereexamined in the control group. Furthermore, prospec-tive analysis of sequential specimens from the samepatient have shown that usually only a single gBgenotype is detected (Wooet al., 1997], although simul-taneous infection with multiple gB genotypes has beendemonstrated in AIDS and allograft recipient patients[Chern et al., 1998; Gilbert et al., 1999; Aquino andFigueiredo, 2000] and in healthy people [Meyer-Koniget al., 1998a;Numazaki et al., 1998].The demonstrationthat extensive recombination takes place betweenstrains of CMV [Haberland et al., 1999; Rasmussenet al., 2003] and the lack of genetic linkage between gBand other Ioci within the CMV genome [Rasmussenet al., 2003] also limit the use of the smallloci in the gBgene to genotype CMVstrains. These findings underlínethat molecular epidemiologic studies must considervaríous clínical, laboratorial, and genetic aspects beforeattempting to establish an association between CMVgBgenotypes and the pathogenesis or virulence of CMV.

REFERENCESAquinoVH, Figueiredo LT. 2000. High prevalence ofrenal transplant

recipients infected with more than one cytomegalovirus glycopro-tein B genotype. J MedViroI61:138-142.

Bongarts AD, von Laer D, Vogelberg e, Ebert K, van Lunzen J,Garweg J, Vaith P, Hufert FT, Haller 0, Meyer-Konig U. 1996.Glycoprotein B genotype of human cytomegalovirus: Distributionin HIV-infected patients. Scand J Infect Dis 28:447-449.

Chern KC, Chandler DB, Martin DF, Kuppermann BD, Wolitz RA,Margolis TP. 1998. Glycoprotein B subtyping of cytomegalovirus(CMV)in the vitreous of patients with AIDS and CMV retinitis.J Infect Dis 178:1149-1153.

Chou SW, Dennison KM. 1991. Analysis of interstrain variation incytomegalovirus glycoproteinBsequences encoding neutralization-related epitopes. J Infect Dis 163:1229-1234.

CornptonT, Kurt-Jones EA, Boehme KW,Belko J, Latz E, GolenbockDT, Finberg RW. 2003. Human cytomegalovirus activates inflam-matory cytokine responses via CD14and Toll-likereceptor 2.JVirol77:4588-4596.

Fidouh-Houhou N, Duval X, Bissuel F, Bourbonneux V,Flandre P, Ecobichon JL, Jordan Me, Vilde JL, Brun-Vezinet F,Leport C. 2001. Salivary cytomegalovirus (CMV)shedding, glyco-protein b genotype / distribution, and CMV disease in humanimmunodeficiency virus-seropositive patients. Clin Infect Dis 33:1406-1411.

Fries BC, Chou S, Boeckh M, Torok-Storb B. 1994. Frequencydistribution of cytomegalovirus envelope glycoprotein genotypesin bone marrow transplant recipients. J Infect Dis 169:769-774.

Gallant JE, Moore RD, Richman DD, Keruly J, Chaisson RE, TheZidovudine Epidemiology Study Group. 1992. Incidence andnatural history of cytomegalovirus disease in patients withadvanced human immunodeficiency virus disease treated withzidovudine. J Infect Dis 166:1223-1227.

Gilbert C, Handfield J, Toma E, Lalonde R, Bergeron MG, Boivin G.1999. Human cytomegalovirus glycoprotein B genotypes inblood of AIDS patients: Lack of association with either the viralDNAload in leukocytes or presence ofretinitis. J MedViroI59:98-103.

Haberland M, Meyer-Konig U, Hufert FT. 1999.Variation within theglycoproteinB gene ofhuman cytomegalovirus is due to homologousrecombination. J Gen ViroI80:1495-1500.

Hebart H, GreifM, Krause H, Kanz L, Jahn G,Muller CA, Einsele H.1997. Interstrain variation ofimmediate early DNAsequences andglycoprotein B genotypes in cytomegalovirus clinical isolates. MedMicrobiol Immunol (Berl) 186:135-138.

CMVGenotypes and CNS Disease in AIDS Patients

Kniess N, Mach M, Fay J, Britt WJ. 1991. Distribution of li~earantigenic sites on glycoprotein gp55 of human cytomegalovirus,J ViroI65:138-146.

LiuYN, KlausA, KariB, StinskiMF, EckhardtJ, Gehrz Re.1991. TheN-terminal 513 amino acids of the envelope glycoprotein gB ofhuman cytomegalovirus stimulates both B- and T-cell immuneresponses in humans. J ViroI65:1644-1648.

Marshall GS,Rabalais GP, Stout GG,Waldeyer SL. 1992.Antibodies torecombinant-derived glycoprotein B after natural human cytome-galovirus infection correlate with neutralizing activity. J Infect Dis165:381-384.

Meyer-Konig U, Ebert K, Schrage B, Pollak S, Hufert FT. 1998a.Simultaneous infection of healthy people with multiple humancytomegalovirus strains. Lancet 352:1280-1281.

Meyer-KonigU,Vogelberg e, Bongarts A,Kampa D,Delbruck R,Wolff-Vorbeck G, Kirste G, Haberland M, Hufert FT, von Laer D. 1998b.Glycoprotein B genotype correlates with cell tropism in vivo ofhuman cytomegalovirus infection. J Med ViroI55:75-81.

Monforte AD, Mainini F, Vago L, Magni C, Musicco M, Antinori S,Bini T, Moroni M. 1993. Factors associated with increased risk ofacquiring cytomegalovirus disease in AIDS patients. J Infect Dis168:1071-1072.

Monforte AD, Mainini F, Testa L, Vago L, Balotta L, Nebuloni M,Antinori S, Bini T, Moroni M. 1997. Predictors of cytomegalovirusdisease, natural history and autopsy findings in a cohort ofpatientswith AIDS. AIDS 11:517-524.

Navarro D, Paz P, Tugizov S, Topp K, La Vail J, Pereira L. 1993.Glycoprotein B of human cytomegalovirus promotes virion pene-tration into cells, transmission of infection from cell to cell, andfusion of infected cells. Virology 197:143-158.

Numazaki K, Ikehata M, Asanuma H, Chiba S. 1998.'Simultaneousinfection of immunocompetent individuals with multiple cytome-galovirus strains. Lancet 352:1710.

Peek R,Verbraak F, Bruinenberg M, van der LelijA,van den Horn G,Kijlstra A. 1998. Cytomegalovirus glycoprotein B genotyping inocular fluids and blood of AIDS patients with cytomegalovirusretinitis. Invest Ophthalmol Vis Sci 39:1183-1187.

Pietropaolo RL, Compton T. 1997. Direct interaction between humancytomegalovirus glycoprotein B and cellular annexin n. J Virol 71:9803-9807.

Prichard MN, Penfold ME, Duke GM,Spaete RR,Kemble GW.2001.Areview of genetic differences between limited and extensivepassaged human cytomegalovirus strains. Rev Med Virolll:191-200.

Rasmussen L, Nelson M,NeffM, Merigan TC, Jr. 1988.Characteriza-tion of two different human cytomegalovirus glycoproteins whichare targets for virus neutralizing antibody. Virology 163:308-318.

Rasmussen L, Hong C, Zipeto D, Morris S, Sherman D, Chou S,MinerR, Drew WL, WolitzR, Dowling A, Warford A, Merigan TC. 1997.Cytomegalovirus gB genotype distribution differs in human

407

immunodeficiency virus-infected patients and immunocompro-mised allograft recipients. J Infect Dis 175:179-184.

Rasmussen L, Geissler A, Winters M. 2003'. Inte~- and intragenicvariations complicate the molecular epidemiology of humancytomegalovirus. J Infect Dis 187:809-819.

Rosen HR, Corless CL, Rabkin J, Chou S. 1998. Ass?ciation ofcytomegalovirus genotype with graft rejection after liver trans-plantation. Transplantation 66:1627-1631.

Schoppel K, KropffB, Schmidt C, Vornhagen R, Mach M. 199y. T~ehumoral immune response against individual cytomagalovirus IS

characterized by a delayed synthesis of glycoprotein-specificantibodies. J Infect Dis 175:533-544.

SchoppelK,Schmidt C,Einsele H, Hebart H, Mach M. 199~.Kinetic~ofthe antibody response against human cytomegalovirus-specificproteins in allogeneic bone marrow transplant recipients. J InfectDis 178:1233-1243.

Shepp DH, Match ME, Ashraf AB, Lipson SM,Millan C, PergolizziR.1996. Cytomegalovirus glycoprotein B groups associated withretinitis in AIDS. J Infect Dis 174:184-187.

Speckner A,GlykofrydesD, Ohlin M,Mach M. 1999.Antigenic domain1of human cytomegalovirus glycoprotein B induces a multitude ofdifferent antibodies which, when combined, results in incompletevirus neutralization. J Gen Viro180:2183-2191.

Torok-Storb B, BoeckhM,Hoy C, Leisenring W,Myerson D, GooleyT.1997.Association ofspecificcytomegalovirus genotypes with deathfrom myelosuppression after marrow transplantation. Blood 90:2097-2102.

Tugizov S, Navarro D, Paz P, Wang Y, Qadri I, Pereira L. 1994.Function of human cytomegalovirus glycoprotein B: Syncytiumformation in cells constitutively expressing gB is blocked by virus-

" neutralizing antibodies. Virology 201:263-276.Tugizov S, Wang Y, Qadri I, Navarro D, Maidji E, Pereira. L. 1995.

Mutated forms of human cytomegalovirus glycoprotem B areimpaired in inducing syncytium formation. Virology 209:580-591.

Vogelberg C, Meyer-Konig U, Hufert FT, Kirste G, von Laer D. 1996.Human cytomegalovirus glycoprotein B genotypes in renal trans-plant recipients. J MedViro150:31-34.

Wada K, Mizuno S, Kato K, Kamiya T, Ozawa K. 1997. Cytomegalo-virus glycoprotein B sequence variation among Japanese bonemarrow transplant recipients. Intervirology 40:215-219.

WooPC,LoCY,LoSK, Siau H, Peiris JS, WongSS, Luk WK,Chan TM,Lim WW, Yuen KY. 1997. Distinct genotypic distributions ofcytomegalovirus (CMV) envelope glycoprotein in bone marrowand renal transplant recipients with CMVdisease. Clin Diagn LabImmuno14:515-518.

Zipeto D, Hong C, Gerna G,Zavattoni M, Katzenstein D, Merigan Te,Rasmussen L. 1998. Geographic and demographic differences inthe frequency of human cytomegalovirus gB genotypes 1-4 inimmunocompromised patients. AIDS Res Hum Retroviruses 14:533-536.

,/