Potential of In Situ Hybridization for Early Diagnosis of Productive Cytomegalovirus Infection

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1988, p /88/ $02.00/0 Copyright 1988, Americn Society for Microbiology Vol. 26, No. 12 Potentil of In Situ Hybridiztion for Erly Dignosis of Productive Cytomeglovirus Infection ELISABETH STOCKL,1 THERESE POPOW-KRAUPP,' FRANZ X. HEINZ,1* FERDINAND MUHLBACHER,2 PETER BALCKE,3 AND CHRISTIAN KUNZ' Institute of Virology,l First Surgicl Clinic,2 nd First Clinic of Internl Medicine,' University of Vienn, A-1095 Vienn, Austri Received 3 My 1988/Accepted 23 August 1988 In situ hybridiztion with probe specific for immedite-erly genes ws used for detection of cytomeglovirus (CMV) trnscripts in peripherl blood mononucler cells, nd the potentil use of this technique s dignostic tool ws ssessed. The results were compred with those obtined with conventionl ssy systems. In 8 of 18 continully observed ptients who developed productive CMV infection, high number of hybridiztion-positive cells were observed 1 to 2 weeks before the conventionl tests yielded positive results. Thus, quntittive evlution of hybridiztion results provided n erly nd specific mrker for beginning CMV infection or rectivtion. In three cses, quntittive in situ hybridiztion ssys provided the only lbortory mrker indicting CMV infection or rectivtion. It ws lso found tht probe specific for immedite-erly genes ws superior to probe specific for lte genes for dignosis of productive infections. Cytomeglovirus (CMV) is known to cuse symptomtic nd ltent infections in helthy individuls (8). In immunosuppressed hosts such s trnsplnt recipients, ptients suffering from cquired immune deficiency syndrome, or neontes, however, it cn cuse severe clinicl mnifesttions (1, 7, 9). The conventionl dignosis of CMV infection is bsed on serodignosis; isoltion of virus from urine smples, throt wshings, nd peripherl blood mononucler cells (PBMs); nd use of specific monoclonl ntibodies for detection of CMV-induced immedite-erly (IE) protein in infected tissue culture cells (6, 12, 13, 18). PBMs supposedly ply n importnt role in CMV infection or rectivtion (17). Atypicl lymphocytes re found in cutely infected ptients, nd depression of lymphocyte functions in these ptients hs lso been shown (12, 15). Detection of CMV IE mrna in PBMs by use of n in situ hybridiztion ssy ws recently described (16). In this study, the in situ hybridiztion ssy ws used for dignostic purposes. A CMV frgment tht encodes IE proteins ws used s probe. Since IE sequences re the first sequences trnscribed t the beginning of CMV infection, use of this ssy seemed the most efficient wy to obtin n erly dignostic mrker. Since it is ssumed tht only productive repliction necessittes the expression of lte genes (14), we lso nlyzed limited number of smples with probe specific for lte genes. MATERIALS AND METHODS Ptients. For the hybridiztion ssys, 163 heprin blood smples were obtined from 38 ptients fter kidney trnsplnttion nd from 15 ptients fter liver trnsplnttion. Forty-one heprin blood smples were tken from 33 humn immunodeficiency virus-infected ptients, nd eight smples were obtined from eight helthy controls. For evlution of CMV sttus, ech ptient ws tested for the following indictors t the time the in situ hybridiztion smple ws obtined: presence of infectious virus prticles in urine, throt wshings, nd PBMs by virus isoltion methods; nd presence of virus-specific ntibodies by complement fix- * Corresponding uthor. tion test nd solid-phse enzyme-linked immunosorbent ssy. Virus isoltion. Virus isoltion ws crried out by the stndrd method (13). In brief, 2 ml ech of urine nd throt wshings nd 2 ml of PBMs, which were isolted from 10 ml of heprin blood by Ficoll-Hypque grdient nd suspended in 3 ml of phosphte-buffered sline, were inoculted into 50-ml Roux bottles with primry humn foreskin fibroblsts. Infected cultures were kept t 37 C in 5% C02 tmosphere for 6 weeks nd exmined twice weekly for cytopthic effects. In ddition, 2 ml of the clinicl smple ws inoculted into tube (Sterilin 129 AX/I; Sterilin Ltd., Felthm, Englnd) contining round cover slip with humn foreskin fibroblst monolyer. The tubes were centrifuged t 37 C nd 2,700 x g for 1 h, decnted, filled with medium, nd incubted t 37 C. After 16 h of incubtion, the cover slip ws wshed in phosphte-buffered sline, fixed in cetone, nd stined by the indirect immunofluorescence technique with monoclonl ntibody ginst the 72,000- dlton erly nucler protein of CMV AD169 (Biotech Reserch Lbortories Inc., Rockville, Md.) nd fluorescein isothiocynte-lbeled got nti-mouse immunoglobulin (Jckson Immunoreserch Lbortory, Avondle, P.) (6). CMV serology. Complement fixtion tests were performed by stndrd method with 2 U of ntigen (13). CMV-specific immunoglobulin M (IgM) nd IgG ntibodies were determined by solid-phse enzyme-linked immunosorbent ssy, with isolted nuclei of infected cells s the ntigen (18). Nuclei of uninfected cells served s the control ntigen. Conventionl tests were considered to indicte productive CMV infection when one or both of the following criteri were met: presence of infectious virus prticles in clinicl mteril, presence of CMV-specific IgM ntibodies, or t lest fourfold increse in complement-fixing ntibody titers. In situ hybridiztion of PBMs. (i) Probes. Two frgments of the CMV genome were used s hybridiztion probes: the CMV EcoRI J frgment of strin AD169 (10.6 kilobses), which encodes the IE trnscripts (cloned in pacyc184, host bcteril strin Escherichi coli HB101), nd the CMV AD169 HindIII L frgment (11.7 kilobses), which is within 2536

2 VOL. 26, 1988 IN SITU HYBRIDIZATION FOR DIAGNOSIS OF CMV 2537 the regions encoding the lte structurl virus proteins (cloned in pbr322, host bcteril strin E. coli 490A). Both clones were kindly provided by B. Fleckenstein, University of Erlngen, Erlngen, Federl Republic of Germny (5). For use s probes, the vector sequences were removed by gel electrophoresis nd electroelution. The probe DNA ws lbeled by nick trnsltion (10) with [35S]dCTP (Amershm Corp., Arlington Heights, 111.) to specific ctivity oft lest 108 cpm/,ug. (ài) Procedure. Cell preprtion nd in situ hybridiztion were performed s described by Brhic nd Hse (2) nd Schrier et l. (16). PBMs were isolted from heprin blood smples by Ficoll-Hypque grdients nd wshed in phosphte-buffered sline. A totl of 105 cells per smple were centrifuged onto glss slides coted with polylysine (Boehringer GmbH, Mnnheim, Federl Republic of Germny). The cells were then fixed with 4% freshly prepred prformldehyde for 20 min. Fixed cells were incubted with 0.2 M HCl for 10 min, wshed, nd treted with 1% Triton X-100 for 2 min. Cells were fixed gin with 4% prformldehyde nd wshed in phosphte-buffered sline-glycine nd in phosphte-buffered sline. The PBMs were then dehydrted with grded rtios of ethnol to wter. Hybridiztion ws crried out by stndrd protocols (4, 10) in 50% formmide nd 10% dextrn sulfte. 35S-lbeled probes were dded to the hybridiztion mix to concentrtion of 0.2,ug/ ml (108 cpm/,ug). The solution ws boiled for 5 min nd chilled on ice. Dithiothreitol ws dded to concentrtion of 10 mm. Portions (10 pil) of the mixture were plced on ech smple; the smples were then covered with cover slips. After hybridiztion overnight t 37 C in humidified chmber, the cover slips were removed nd the slides were wshed under conditions described by Schrier et l. (16). For utordiogrphy, Kodk NTB-2 emulsion (Estmn Kodk Co., Rochester, N.Y.) ws used. The slides were exposed for 3 dys t 4 C nd developed with Kodk D19. The smples were exmined under microscope, nd strict criteri for selection of positive signls were used for identifiction of infected cells. We confirmed tht virl mrna ws detected with the EcoRI-J probe in the PBM smples by showing tht highly positive smples gve no hybridiztion signls fter RNse digestion. It ws not cler, however, whether RNA or DNA ws detected with the HindIII L frgment, since only few positive signls were obtined with this probe nd since negtive hybridiztion results fter RNse or DNse digestion my be cused by vrying distribution of the positive cells. RESULTS Comprison of in situ hybridiztion with conventionl tests. In situ hybridiztion of PBMs ws performed by using the EcoRI J frgment of the CMV genome, which encodes IE trnscripts, s probe. A totl of 212 PBM preprtions from 86 ptients t risk (e.g., ptients who hd undergone trnsplnts or were suffering from cquired immune deficiency syndrome) s well s from eight helthy controls were screened for the presence of CMV-specific nucleic cids by in situ hybridiztion. Hybridiztion results of individul PBM smples were compred with those obtined by testing simultneously collected smples of urine, throt wshings, heprin blood, nd serum in conventionl test systems. For 84 of the 212 smples, evidence of productive infection ws provided by one or more conventionl tests. In 57 of these smples the in situ hybridiztion ssy lso showed positive signls; negtive results, however, were obtined with 27 smples. In 2 ofthe 84 smples viremi ws detected by virus isoltion from PBMs; in both smples the PBMs lso gve positive in situ hybridiztion results. Virus excretion in urine or throt wshings or both ws observed in 50 smples. In the in situ hybridiztion ssy, 39 of the simultneously tken PBM smples lso gve positive results but 11 showed no hybridiztion signls. In 32 smples productive CMV infection ws confirmed only by serologicl findings, such s presence of CMV-specific IgM ntibodies or significnt increse in complement-fixing ntibody titers. Positive hybridiztion results were obtined with 16 of these smples, wheres the remining 16 yielded negtive results. Conventionl tests did not provide evidence of productive infection in 128 smples. Of these, 94 showed no hybridiztion signls nd 34 yielded positive results by in situ hybridiztion. Considering the existence of CMV-specific IgG ntibodies in the individuls from whom these smples were tken, it is noteworthy tht ll of the 34 smples positive by in situ hybridiztion were tken from IgG-positive individuls, one ofwhom ws helthy control. In none of the 15 smples tken from seronegtive individuls were positive in situ hybridiztion results observed. It is pprent from these results tht not ll smples obtined from productively infected ptients yielded positive results in the in situ hybridiztion ssy. In other cses, however, positive in situ hybridiztion results were lso obtined in the bsence of ny other indiction of productive CMV infection t tht time. The discrepncy in these results my be ttributble to the fct tht smples were nlyzed only t single time points, without considertion of the time course of ppernce nd disppernce of the vrious indictors of infection. Therefore, smples collected t constnt intervls from 40 trnsplnt ptients were nlyzed by ech of the ssy systems. All of the 40 ptients hd CMV-specific IgG ntibodies, nd none showed evidence of productive CMV infection t the time of trnsplnttion. During the observtion time, 18 of the ptients developed productive CMV infection s indicted by conventionl tests; the in situ hybridiztion ssy yielded positive results for 17 of the ptients. In 14 ptients, positive results were obtined 6 weeks to 1 week erlier by the hybridiztion ssy thn by conventionl tests. In one cse the first positive results were obtined by hybridiztion nd conventionl tests t the sme time, nd in two cses the in situ hybridiztion ssy yielded positive results lter thn did conventionl tests. For 22 of the ptients, conventionl tests produced no signs of CMV infection or rectivtion. In six of these ptients, however, CMV-positive PBMs were found by in situ hybridiztion. These results showed tht during the course of most productive infections, in situ hybridiztion yielded positive signls erlier thn did conventionl tests. Detection of CMV sequences in PBM smples from individuls with no cute infections s well s in smples obtined more thn 3 weeks before the productive infection ws indicted by conventionl tests prompted us to perform quntittive nlysis of in situ hybridiztion results during the course of infection. Quntittive nlysis of PBM smples positive by in situ hybridiztion. All positive PBM smples obtined from the 40 continully observed trnsplnt ptients were nlyzed for number of positive cells per smple, which two observers counted independently by microscopy. Counting ws done under stringent conditions, nd only those cells tht unmbiguously showed specific hybridiztion signls were included (Fig. 1).

3 2538 STOCKL ET AL. A B J. CLIN. MICROBIOL., 4 r T..4 FIG. 1. In situ hybridiztion of PBMs, with the 35S-lbeled EcoRI J frgment s probe. (A) Smple contining positive cells; (B) negtive smple. Mgnifiction, x320. The results of the quntittive nlysis re shown in Fig. 2. Although positive results were obtined from productively s well s from nonproductively infected ptients (Fig. 2), more thn five positive cells per smple were observed exclusively in those individuls who were developing productive CMV infection (Fig. 2A). By this quntittive criterion, the beginning of productive CMV infection ws detected 1 to 2 weeks erlier thn by conventionl tests in 8 of 18 ptients. Quntittive in situ hybridiztion yielded positive results t the sme time s did conventionl ssys for two ptients; in eight ptients no more thn five positive cells were detected t ny time. In ny of the smples nlyzed, only one of the seril smples tken fter trnsplnttion reveled more thn five positive cells. Hybridiztion dt for ll PBM smples obtined from trnsplnt nd humn immunodeficiency virus-infected ptients were quntittively evluted. Of these smples, 22 contined more thn five positive cells; in 19 of these smples productive CMV infection ws confirmed by conventionl tests. In three smples from humn immunodeficiency virus-infected ptients, however, no other ssy provided evidence of CMV infection or rectivtion; two of these smples were obtined from ptient who ws suffering from severe pneumoni for which no other explntion could be found. In the third smple, no clinicl indiction of CMV infection ws observed. In situ hybridiztion, using the HindIII L frgment of the CMV genome s probe. As shown bove, quntittive evlution of in situ hybridiztion results by using probe specific for IE trnscripts represented n excellent mrker for productive infection nd llowed differentition of IE trnscripts from genomes present in ltently infected cells. As n lterntive method, we ssessed the use of probe specific for sequences encoding lte structurl proteins, which re thought to be expressed only during the course of productive infection. A totl of 68 smples from 32 continully observed trnsplnt ptients were nlyzed by both methods. As expected, none of the smples from 15 nonproductively infected ptients gve positive results. In 17 ptients with productive infections, hybridiztion with the HindIII-L probe proved inferior to quntittive hybridiztion with the IE probe, especilly with respect to erly detection of productive infection. In two ptients the lte probe gve positive results 1 week erlier thn did the conventionl tests, wheres with the IE probe more thn five infected cells were observed 1 to 2 weeks erlier in seven ptients. Among smples for which conventionl tests hd lredy yielded positive results, the lte probe showed positive signls for five nd the IE probe did so for only one. In only one cse did both ssys yield positive results t the sme time, nd in three cses * z 15 z 10 5, A 501 B 2._ r ' 'i u 30- r 25 o o 20-. E 15* I e.@**e : e * ** -s * * * e se e * tee X We k * e * e e * e 0 X 9~~~~ Weeks FIG. 2. Quntittive evlution of in situ hybridiztion results for ptients who hd undergone orgn trnsplnttion. (A) Results for 18 ptients developing productive CMV infections; (B) results for 22 ptients with no evidence of productive CMV infections. X, Time of trnsplnttion. e

4 VOL. 26, 1988 productive infection ws detected by none of the hybridiztion ssys. IN SITU HYBRIDIZATION FOR DIAGNOSIS OF CMV 2539 virus repliction, this ssy provides n dditionl vluble tool for the rpid dignosis of productive CMV infection. DISCUSSION Rpid dignosis of CMV infection is essentil for the mngement of immunosuppressed ptients, since these individuls re especilly prone to severe clinicl mnifesttions of CMV. Conventionlly, CMV is dignosed by serologicl tests s well s by virus isoltion nd detection of CMV erly ntigen in inoculted tissue culture cells. There is incresing evidence tht PBMs represent mjor pool of ltent CMV nd consequently site for virus repliction during rectivtion (14). In situ hybridiztion hs been found to be sensitive mens of detecting CMVspecific nucleic cids in PBMs obtined from ltently infected persons (16). We investigted whether this method provides useful tool for the recognition of CMV infection or rectivtion t n erlier stge thn is possible by conventionl methods. For hybridiztion we used the EcoRI J frgment, which encodes IE proteins, since it provided n erlier mrker for CMV repliction thn did HindIII L frgment encoding lte proteins. Positive hybridiztion results were obtined for smples for which conventionl methods showed evidence of productive infection s well s for smples lcking such evidence. The ltter result most likely rose from the detection of ltent CMV genomes (11, 16), n ssumption confirmed by the fct tht in situ hybridiztion signls were observed in none of the individuls with no IgG ntibodies. On the other hnd, negtive hybridiztion results were obtined in PBMs from ptients with productive CMV infections s indicted by conventionl tests. The reson for these prtilly discrepnt results ws ssessed by comprison of the time courses of the hybridiztion ssy nd conventionl tests nd by quntittive evlution of hybridiztion dt. We found tht most of the discrepnt results in productively infected persons were cused by differences in the time course of ppernce nd disppernce of the vrious indictors of infection. In most cses the discrepncies resulted from the fct tht positive results were obtined erlier by hybridiztion thn by conventionl tests. However, it ws necessry to evlute hybridiztion on quntittive bsis becuse low number of hybridiztion-positive PBMs were lso detectble in individuls who did not develop productive infection t ll or did so only more thn 3 weeks fter positive hybridiztion result. Under the conditions of evlution tht we used, detection of more thn five positive cells in given smple ws lwys correlted with productive infection nd in mny cses provided n erlier mrker thn did other methods for the onset of CMV repliction. In few humn immunodeficiency virus-infected ptients this quntittive ssy provided the only lbortory evidence for CMV infection, which in some cses lso correlted well with clinicl symptoms. It ppers tht this pek of ongoing virus repliction occurs only once during the course of infection or rectivtion nd for limited period of time. This result correltes well with the findings tht IE trnscripts nd proteins re preferentilly synthesized within the first few hours fter infection nd tht therefter the rte of synthesis decreses (3, 19, 20). Therefore, filure to detect such pek in only one smple does not preclude the possibility of productive infection. However, since hybridiztion result of t lest five positive cells lwys represented mrker for ongoing ACKNOWLEDGMENTS We re grteful to B. Fleckenstein, Institute of Virology, University of Erlngen, Erlngen, Federl Republic of Germny, for providing us with the CMV DNA-contining plsmids. We thnk Sbine Lnger, Heide Dippe, nd Krin Festl for excellent technicl ssistnce nd Susnne Pfuser for typing the mnuscript. LITERATURE CITED 1. Ahifors, K., S. A. Ivrsson, T. Johnsson, nd I. Svensson Congenitl nd cquired cytomeglovirus infections. Virologicl nd clinicl studies on Swedish infnt popultion. Act Peditr. Scnd. 67: Brhic, M., nd A. T. Hse Detection of virl sequences of low reitertion frequency by in situ hybridiztion. Proc. Ntl. Acd. Sci. USA 75: DeMrchis, J. M., C. A. Schmidt, nd A. S. Kpln Ptterns of trnscription of humn cytomeglovirus in permissively infected cells. J. Virol. 35: Denhrdt, D. T A membrne filter technique for the detection of complementry DNA. Biochem. Biophys. Res. Commun. 23: Fleckenstein, B., I. Muller, nd J. Collins Cloning the complete humn cytomeglovirus genome in cosmids. Gene 38: Gleves, C. A., T. F. Smith, E. A. Shuster, nd G. R. Person Rpid detection of cytomeglovirus in MRC-5 cells inoculted with urine specimens by using low-speed centrifugtion nd monoclonl ntibody to n erly ntigen. J. Clin. Microbiol. 19: Glenn, J Cytomeglovirus infection following renl trnsplnttion. Rev. Infect. Dis. 3: Gold, E., nd G. A. Nnkervis Cytomeglovirus, p In S. Evns (ed.), Virl infections of humns: epidemiology nd control, 2nd ed. Plenum Publishing Corp., New York. 9. Mcher, A. M., C. M. Reichert, nd S. E. Struss Deth in the AIDS ptients: role of CMV. N. Engl. J. Med. 309: Mnitis, T., E. Fritsch, nd J. Smbrook Moleculr cloning: lbortory mnul. Cold Spring Hrbor Lbortory, Cold Spring Hrbor, N.Y. 11. Mocrski, E. S., nd M. F. Stinski Persistence of the cytomeglovirus genome in humn cells. J. Virol. 31: Montplisir, S., nd B. Mrtineu Persistent cytomeglovirus infections in mn nd mice: role of lymphocytes, p In E. Kurstk nd K. Mrmorosch (ed.), Viruses nd environment, Acdemic Press, Inc., New York. 12.Popow-Krupp, T., nd C. Kunz Detection of cytomeglovirus in clinicl specimens by virus isoltion nd by use of monoclonl ntibody ginst the erly nucler ntigen: comprtive study. J. Med. Virol. 24: Reynolds, D. W., S. Stgno, nd C. A. Alford Lbortory dignosis of cytomeglovirus infections, p In E. M. Lennette nd N. J. Schmidt (ed.), Dignostic procedures for virl, ricketsil nd chlmydil infections, 5th ed. Americn Public Helth Assocition, Wshington, D.C. 14. Rice, G. P. A., R. D. Schrier, nd M. B. A. Oldstone Cytomeglovirus infects humn lymphocytes nd monocytes: virus expression is restricted to immedite-erly gene products. Proc. Ntl. Acd. Sci. USA 81: Rinldo, C., P. Blck, nd M. Hirsch Interction of cytomeglovirus with leukocytes from ptients with mononucleosis due to cytomeglovirus. J. Infect. Dis. 136: Schrier, R., J. Nelson, nd M. Oldstone Detection of humn cytomeglovirus in peripherl blood lymphocytes in nturl infection. Science 230: Spector, S. A., J. A. Ru, D. H. Spector, nd R. McMilln Detection of humn cytomeglovirus in clinicl specimens by DNA-DNA hybridiztion. J. Infect. Dis. 150: Stgno, S., D. W. Reynolds, nd R. J. Smith Use of

5 2540 STOCKL ET AL. isolted nuclei in the indirect fluorescent-ntibody test for humn cytomeglovirus infection: comprison with microneutrliztion, nticomplement, nd conventionl indirect fluorescent-ntibody ssys. J. Clin. Microbiol. 7: Wthen, M. W., nd M. F. Stinski Temporl ptterns of humn cytomeglovirus trnscription: mpping the virl RNAs J. CLIN. MICROBIOL. synthesized t immedite erly, erly, nd lte times fter infection. J. Virol. 41: Wthen, M. W., D. R. Thomsen, nd M. F. Stinski Temporl regultion of humn cytomeglovirus trnscription t immedite erly nd erly times fter infection. J. Virol. 38: