Vol. 12, No. 3 JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1980,p.367-374
0095-1137/80/09-0367/08$02.00/0
Immunoglobulins
M
and
G
to
Varicella-Zoster Virus Measured
by Solid-Phase
Radioimmunoassay: Antibody Responses
to
Varicella and Herpes Zoster Infections
A. M.ARVIN' ANDC. M. KOROPCHAKDepartmentofPediatrics, Divisionof Infectious Diseases, Stanford UniversitySchoolofMedicine,
Stanford,
California
94305Both immunoglobulin M (IgM) and IgG antibodies to varicella-zoster virus (VZV)were detectablein asolid-phase radioimmunoassaywith
'25I-labeled
goat antisera tohumanimmunoglobulins. Primary infection withVZV was associated with early production of IgM and IgG antibodies and rapid development of lymphocyte transformationtoVZVantigen. Amongeight
subjects with varicella tested 1 to 4daysafter onset, sevenpatients hadIgG andsixpatientshad IgM antibodies; all patientshadbothIgGandIgM antibodieswithin 7 days.AnIgMresponse wasdocumented byradioimmunoassayin 18 of 26patientswithherpes
zoster.VZV antibodiescould beassayedby
radioimmunoassay
inunfractionatedserumwith commercialgoatantiseratohumanimmunoglobulin and commercial VZVantigen. VZV-specific IgG bindingwaspresent inallsera from42 subjects withaVZVantibody titer of-1:8 asdeterminedby indirectimmunofluorescence andcellular
immunity
toVZVasdeterminedbylymphocyte
transformation and who had had varicellaatleast20yearsbefore testing. The geometricmeantiterwas 1:6,309, and titers were 21:16,384 in 20 subjects.
Antibody
was present asdeterminedbyradioimmunoassay in14samples negative by complement fixation and in five samples negative by complement fixation and immune adherence hemagglutination.No
specific
bindingwasobserved in21 serafromsubjects whowere not immune to VZV as determined by indirect immunofluorescence or
lymphocyte transformation despitethe presence ofherpes simplexor
cytomega-lovirusantibody indicated by
complement
fixation in15 sera.High titers ofVZVIgM antibody weredetectedinunfractionatedsera despite thepresenceofhigh titers ofVZV
IgG antibody.
The VZVradioimmunoassay
provided a sensitiveand
practical
method formeasuring VZVIgG
andIgM antibodies.Primary infection with varicella-zoster virus (VZV) is common in
early
childhood. Varicellais
usually
mild inhealthy
children butlife-threatening disseminationofthe viruscan occur
in children with
congenital immunodeficiency
disorders or
malignant
diseases (7). A reliable method fordetermining
susceptibility
tovari-cella is
important
because theseverity
of theinfection can be reduced
by
theprophylactic
administration ofzosterimmune
globulin
toim-munocompromised
children (13). The indirectimmunofluorescence methods forVZV
antibody
aresensitive, but thesetechniques
requirelabo-ratoryfacilitiesfor
handling
varicella virus,and theirreliabiity
canbeaffectedby
thesubjective
interpretation involved (26, 27). The standard
serological
methods,
such as the complementfixation technique, have notbeen useful in
de-tectingVZVantibodyforprolonged periodsafter theprimaryinfection (25).None of these meth-ods has been
easily adaptable
to detect VZVimmunoglobulin M (IgM) antibody in
unfrac-tionatedserum.
The reactivation of latent VZV results in
herpes zoster, which is a frequent occurrence among patients with malignancy and organ
transplant
recipients
(18, 19). Inadditionto itsclinical application, a very sensitive assay for
VZVIgG and IgM antibodiesmight detect
defi-cienciesinantibody production that contribute tothe pathogenesis ofsevereprimary varicella and to the reactivation oflatent VZV virus in immunocompromised patients.
The solid-phase radioimmunoassay(RIA) has been asensitivemethod forthe
detection
of IgGantibody to herpesviruses, including VZV,
herpes
simplex,
andcytomegalovirus (9, 10, 16), and has also been used to measure IgMantibody tocytomegalovirus (16). In this study, asimpli-fied solid-phase RIA technique with
commer-ciallyavailable reagents wasemployed to mea-sure both IgG and IgM antibodies to VZV in
367
on February 7, 2020 by guest
http://jcm.asm.org/
368 ARVIN AND KOROPCHAK
unfractionatedserumsamples.TheRIAmethod wascompared with other assaysforVZV
anti-body, including complement fixation, indirect immunofluorescence assay (IFA), and immune
adherancehemagglutinationassay(IAHA).The
VZV IgM and IgGresponsesofnormal subjects
wereevaluated with the RIAtechnique during
primaryvaricella and acuteherpeszoster. MATERIALS AND METHODS Subjects.Serawereobtainedfromhealthy subjects whoseimmunitytoVZVwasestablished by IFA and
bytheVZV-specific lymphocyte transformationassay
(seebelow).The VZV-immune adults had hadprimary varicella 20yearsor morebefore testing. Fiveofthe subjectswhowerenotimmuneby thesecriteria sub-sequently developed varicella.Serawerealso obtained
from otherwisehealthysubjects with acute varicella
orherpeszoster.Thediagnosiswasconfirmedby
VZV-specificdirect fluorescence ofepithelial cellsfromthe skin lesions (24). Serawere tested forantibodies to
herpes simplex and cytomegalovirus by the
comple-mentfixationtechnique (17).
Antigenpreparation.Oneroller bottle(490 cm2) ofconfluenthuman foreskin fibroblastcellswas inoc-ulated with a VZVstrain (passage 44) originally iso-lated from a patient with herpes zoster and kindly provided by LucyRasmussen. When 75% of the cells showedcytopathic effect, theywereremovedand used
toinfect five roller bottlescontainingconfluent fibro-blast monolayers. After 72 h, when 75% cytopathic effectwas observed,theinfectedcellswereremoved with glass beads, washed once with phosphate-bufferedsaline(PBS), pH 7.2,suspendedtoa
concen-tration of 5 X 10" cellsper ml and sonicated. After
centrifugationat400 xgtoremovelargercelldebris, the antigen was dispensed in working volumes and stored at-70°C. Control antigen waspreparedfrom uninfectedcells bythesameprocedure.
Thecommercial antigenwaspreparedfor use asa
VZVcomplementfixationantigen byFlow Laborato-ries, Inc., Rockville,Md.
Varicella-zoster antibody testing by comple-mentfixation,IAHA,andIFAand VZV lympho-cytetransformation.TheVZVcomplementfixation
assay wasperformed byastandard methodusingtwo
units ofantigenandtwounits ofcomplement (17).For the IAHAassay(12, 19),serial dilutions ofsera(0.025
nml) were incubatedwith 0.025 ml ofVZVantigenin
the wells ofamicrotiter"V"plate(Flow Laboratories) for1 hat37°C.The usualoptimaldilution ofantigen
was 1:8 to 1:16.Complementwasaddedatadilution of 1:80to 1:100. Afterincubationat37°C for 40min, dithiothreitolinVeronal-bufferedsaline and ethylene-diaminetetraaceticacid(0.025mlperwell)anda 1.5%
suspension of human "O" cells (0.025 ml per well)
which had been screened forsensitivityinthisassay were added. The plateswere agitated for1 min and left atroom temperature for 1 h. AUl sera werealso tested with control antigen. The antibody titer was
readasthereciprocalof the maximum serial dilution
resultingin 3+hemagglutination.
Ourmodification of the IFA method has been
de-scribed (1). Monolayers of human foreskin fibroblast cells grown in chamberslides (Lab-Tek Division, Miles Laboratories,Inc., Naperville, Ill.)wereinfected with VZV. Whenseveral foci of infectedcellswere visible in each well, but before the cytopathic effect was extensive, the slides were fixed in cold acetone for 5 to 10min. The fixedslides were stored at -70°C. Begin-ning at 1:8, serialdilutions of the serum samples to be tested and known negative and positive sera were coded and added in triplicate to the slide wells. After incubation for 1 h at 37°C, the slides were washed three times with PBS, and fluorescein-conjugated guineapig antihumanIgG (Microbiological Research Co.,Bountiful, Utah) diluted 1:20 was added. After 1 h ofincubation at 37°C, theslideswere washedthree times with PBS and examined under afluorescence microscope. The titer was read as the highest dilution of serum which showedfluorescence ofinfected cells in at least two of the three wells.
VZV cell-mediated immunity was assessed by in vitro lymphocyte transformation to VZV antigen as described previously (1). Ficoll-Hypaque-separated peripheral blood mononuclearcells in30% heat-inac-tivated human serum wereincubated with VZV anti-gen for 5 days. Lymphocyte transformation was de-tected by comparing tritiated thymidineincorporation inantigen-stimulatedcultures(;3H-antigen) with con-trol antigen cultures
(;3H-control).
The results were expressed as astimulation index ;H-antigen/3H-con-trol. Astimulation index of at least threefold to VZV antigen was seen in normalseropositivesubjects (1).RIA for varicella-zoster IgG and IgM anti-bodies. TheRIA for VZVIgG antibodywasa modi-fication of the method described by Friedman et al. (10). Commercial VZV antigen (Flow Laboratories) was diluted to 1:5 with PBS, pH 7.2, and antigens preparedin thislaboratoryweredilutedto 1:8 to 1:10 for use in the assay. Three lots of the commercial antigen were used with equivalentresults. VZV anti-gen (0.025ml)oruninfected controlantigen (0.025 ml) was added to each well of a polyvinyl "U" plate (Dynatech Co., Alexandria, Va.), and the plate was allowedto dry overnight at roomtemperature. After onewashing withPBS,thewells were filled with PBS containing 20% fetal calf serum andincubated for 1 h at37°C. Thewells were then rinsed twice with PBS andallowedto dry. Each serumsample tobe tested wasadded to twowellscontainingantigen and to two control wells. The sera were diluted in PBScontaining 10%fetalcalfserumand0.05%Tween 20(J. T. Baker Chemical Co., Phillipsburg, N.J.) using a microtiter platedilutor(Cooke Co.,Alexandria, Va.). Theplates wereincubated at 37°C for 1 h and washed three times with PBS containing 0.05% Tween 20 before
'2'l-la-beledantihumanglobulinwasadded.
The specificbinding of VZV IgG antibodywas as-sessed using goat antihuman IgG, specific for the Fc fragment (Cappel Laboratories, Downingtown, Pa.) thathad been labeled with '2'Iby amodified chlora-mine T method (20). '25I-labeled monospecific goat antihuman IgM antibody (Tago Co., Burlingame, Calif.) purifiedby affinity chromatography wasused to detect IgM antibody bound to the VZV antigen-coatedwells. Afterlabeling,67to76%of thecountsin J. CLIN. MICROB10L.
on February 7, 2020 by guest
http://jcm.asm.org/
RIA FOR IgM AND IgG ANTIBODIES TOVZV 369 thesepreparationswereprecipitable with 10%
trichlo-roacetic acid. Thespecific activity of the labeled prep-arationsranged from 8to10,iCiof
"25I
permicrogramofprotein.
For use in the RIA, the iodinated globulin was dilutedtocontain2X 10" cpm/ml. Afterpreparation of theplatesasdescribed, '251-labeled antihumanIgG
orIgM (0.025 ml perwell)wasaddedtoeachwell, and theplateswereincubated for 1 hat37°C. Theplates were thenrinsed five times with cold PBScontaining 0.05% Tween20and allowedto dry. The wellswere separated bypassinga"hot wire" below the surface of theplate, transferredto countingvials, and counted for1 mininagammascintillationcounter(Beckman
Instruments,Inc., Fullerton,Calif.). Controls used in each assayincluded knownpositive and negativesera anddiluent alone.Specific antibodybindingwas con-sideredtobepresentateachserumdilution for which theratioof themeancountsper minute (cpm)of the twowells containingserum and VZV antigento the mean cpm of the two wells containing serum and control antigen was -2.5. The maximum cpm seen after the incubation of positive sera with antigen
ranged from 6,000to 15,000 cpm per well;the maxi-mum cpm after the incubation of antibody-positive
serumwithcontrol antigenrangedfrom200to 1,000
cpm per well. The incubation ofnegative serum or diluent with VZV antigenorwith controlantigenalso producedcountsin this range.
Separation ofIgG and IgM serum fractions. The methods used for the removal or separation of IgG andIgM antibody from thetestseraincluded the
adsorptionofIgG by using staphylococcal proteinA (3), the separation of IgG and IgM by usinga gel
filtrationtechnique,andincubation with 2-mercapto-ethanol (2). The gel purification procedure, kindly performed by Gregory Filice, used Bio-Gel A-5M,
which isa6% agarosegel (Bio-Rad Laboratories, Rich-mond,Calif.)toseparate7S and19Sglobulin.
Rheumatoidfactor. Sera thatwerepositivein the VZVIgM assayweretested for rheumatoid factor with a commercial latex agglutination method
(Calbi-ochem,LaJolla,Calif.) (21). VZVIgGandIgM anti-body titerswereassayedinserafrom six adults with hightitersofrheumatoidfactor, rangingfrom1:1,280 to 1:5,120, who had had varicella at least 20 years beforetesting.
RESULTS
BothlaboratoryandcommercialVZVantigen
preparations
were.sufficiently
adherent to thewells of the
polyvinyl
microtiter plates topro-videan
adequate
antigen-coated
surface for thedetection of
specific antibody binding.
The op-timalconcentration of VZV antigen was deter-minedbyevaluatingthespecific binding detect-able in sevensera with VZVantibody titers of 1:8byIFA andnodemonstrable VZVantibody by the complement fixation or IAHA assays.Using a 1:5 dilution of the commercialantigen
and 1:8to 1:10dilutions ofourlaboratory anti-gen, antibodywas detectable byRIA at serum dilutions of 1:256 or higher in sera with low
positive IFA titers. Parallel testing of these sera showed equivalent VZV antibody titers with
commercialantigen diluted 1:5 and our labora-toryantigen diluted 1:8 in PBS.
With the commercial antigen, specific IgG bindingwas detected in all 42 sera from subjects remote from infection who had VZV antibody as determined by IFA (Fig. 1). No specific binding wasobserved in 21 sera from subjects who were not immune to varicella by IFA or in vitro
lymphocyte transformation (Fig. 1). When
25I-labeled goat antihuman IgM was used to detect bound IgM,ail
8 patients with acute varicellaand 18 of 26patients with acute herpes zoster had VZV IgMantibody (Fig. 1 and 2). None of the 15 subjects remote from infection who had VZV IgG antibody had detectable VZV IgM antibody (Fig. 1). No specific binding occurred when sera from nine subjects who were not immune tovaricellawere tested.Binding curves
for
'251-labeled
antihuman IgG and IgM inun-fractionated serum from a subject with acute varicellaareshown in Fig. 3.
Twelve serum samples fromseropositive
sub-jectsweretestedtodetermine the reproducibil-ity of theassay.Of the 12 sera, 10 had IgG titers within a twofold dilution, and two samples showed a greater thanfourfoldvariation intiter.
As is the case with other methods, it was nec-essary to test paired sera in the same assay to
obtain the most accurate comparison of VZV
IgG and IgM antibody titers early and later in
thecourse ofacuteinfection (Fig. 2).
Someassaysforthe
detection
of viralantibodybysolid-phase RIA have been based on a com-parison of the
'25I-labeled
antihuman IgG boundafterincubating the test serum with viral antigen
andthebindingmeasured when a known
nega-tiveserum was incubated with theantigen (15).
SpecificVZVIgGbindingcouldbeassessed with
reference
to astandard negative serum withoutsignificantly altering the antibody titer in 20
seropositive
subjects. However, theelimination
of the incubation of each serum with control
antigen produced false-positive results among
subjects whohad negative VZV IFA titers and nolymphocyte transformation to VZVantigen.
Of 10 nonimmune subjects, 2 showed binding ratios greater than 2.5 at serum dilutions of 1:16 whenbinding was assessed in relation to a stand-ard negative serum. The binding ratios were consistently less than 2.5 when the calculation
wasbased onbinding of the subject's serum to control antigen.
Thesensitivity of the VZV RIA for measuring
IgG antibody wasevaluated in 14 subjects who
had no antibody to
varicella
as determined bycomplement
fixation (Fig. 4). The VZV RIAVOL.12,1980
on February 7, 2020 by guest
http://jcm.asm.org/
370 ARVIN AND KOROPCHAK
AJ
4.2x106in,
1.x10zo
2.6x105CL CL
- Q 65,536 _
. 16,384
-z 4096
-o
D 1024
--J
256
-w 64
-CI,
CO) 16 _ w
C:z <16 k
E
m
n
te
-j,
JIl.
't
I
e0
J,1
e
-* _
-* e
-o- e
-m -_
s
m
1
e~e
4 4 ee
e
O 't 4eSb
VZV IgGANTIBODYTITER 4 VZV IgMANTIBODYTITER
FIG. 1. IgG and IgM antibody titers to VZVmeasured by RIA in nonimmune and immune subjects identified by VZVIFA and VZVlymphocyte transformation. The antibody titers are indicated along the
vertical axis. Theantibody titeris thehighest dilutionoftheserumsample forwhichthe specific binding
ratio(definedasthemean cpmafterincubation intwowellscontaining VZVantigen dividedby themean cpmafterincubationin twowellswithcontrolantigen)is-2.5. Theleftpanelshowsthe VZVIgG antibody titers in nonimmune subjects, immune subjects who had varicellamore than 20years before testing, and subjectswithacuteherpeszostertested 2to14and21 to35daysaftertheonset. The right panelshowsthe VZVIgM antibodytitersinimmunesubjectswhohadvaricellamorethan 20yearsbefore testingandsubjects
withacuteherpeszoster.
titerswerelowerinthisgroup (geometricmean
titer[GMT], 1:380)thanintheunselected
pop-ulation ofseropositive subjects (GMT, 1:6,309),
but the GMTby VZVRIA wasfourfoldhigher
than the GMT measuredbyIFA.TheRIAand IFA were equivalent in detecting seropositive
subjects in this group and in the unselected
population,butantibodywasdetectablebyRIA in more dilute serum samples from the same
individual. TheIAHAfailed todetectantibody
infivesera thatwere positive byRIA and the GMTbyRIAwas12-foldhigherthantheGMT
byIAHA(Fig. 4).
Cross-reactivitywithherpes simplexvirusand
cytomegalovirus antibodies did not appear to affect the VZVantibodyresults byRIA.Despite
thepresenceofantibodytoherpessimplexvirus in sixsubjectsandantibodytocytomegalovirus
in ninesubjects, VZVRIA titerswerenegative
in those individualswhowereVZV-susceptible asdeterminedbyIFAandlymphocyte
transfor-mation. When the assay was used to measure
VZV IgG antibodyinimmunesubjects withno
recent VZV infection, the VZV GMTs were equivalentbetweengroupsthathad antibodyto
herpes simplex virus or cytomegalovirus and
thosethatdid not. The VZVGMT in 17 herpes simplex virus-seropositive subjects was 1:2,454
compared with 1:5,623 in 18herpes simplex vi-rus-seronegative subjects;theVZVGMTs in 15
cytomegalovirus-seropositive and 17
cytomega-lovirus-seronegative subjectswere1:3,238and 1:
5,254,respectively.
Sera from six immunesubjectswithout recent VZVinfectionwhohadVZV IgG antibody titers ranging from 1:1,024to 1:2.6x 105and who also
hadhightitersofrheumatoid factorwere
eval-uated todetermine whetherrheumatoid factor producedafalse-positiveresult in the VZVIgM assay. Rheumatoid factor, which is usually an
IgM antibody against IgG, couldattach to the VZV IgG bound to the VZV antigen. The 125j_
_*
"_
~~*
__
-e
-e
le e
-* l-
--1 ____
J. CLIN. MICROBIOL.
-T
*
a
on February 7, 2020 by guest
http://jcm.asm.org/
RIA FOR IgM AND IgG ANTIBODIES TO VZV
DAY DAY16 DAY DAY16
25
20
x
w
0 15
z
o
0 -J
m 10
ci>
5
DAY DAY16 VZV liG ANTIBDDY TITER VZV IlMANTIBDDYTITER VZVLYMPHOCYTE
TRANSFORMATION
FIG. 2. IgG and IgM antibody titers to VZV measured byRIA and VZV antigen-stimulated in vitro lymphocytestimulation inpatientswith acutevaricella.Theleftand the centerpanelsshowthe VZVIgGand IgM antibody titers, respectively, inpairedseraobtainedfromeight patientstestedI to4and 7 to 16days
aftertheonsetofacute varicella. Theantibodytiter isthehighestdilutionoftheserumsample forwhichthe specificbindingratio(definedasthemeancpmafterincubation in two wellscontainingVZVantigendivided bythemeancpmafterincubationin twowellswith controlantigen)is22.5. Theright panelshowstheVZV antigen-stimulatedlymphocyte transformationresponsesofthesepatients. Thestimulation indexisaratioof cpm measured after addingtritiatedthymidineto VZVantigen-stimulated lymphocyte cultures tocpm in control cultures incubated with uninfectedcell antigen. A stimulation index ofgreater than threefold is positive;the shadedareaindicatesastimulation indexless thanthreefold.
labeled antihumanIgM mightthen bind to the
rheumatoidfactor, resulting in apositive assay forIgM in serumthat did not contain specific VZV IgM. However, five of the six sera tested werenegativeintheVZVIgMassaydespite the presenceof rheumatoid factor. One serumhad
an IgM titerof 1:16. This titerwaslower than
theVZVIgMtiters of mostpatientswithrecent
varicellaorherpeszoster(Fig. 1and2). The IgM antibody detected in the patients with acute infection was considered VZVspecific because
none of thesera from thesepatients contained
rheumatoidfactor.
Both IgG and IgM antibodies to VZV were
presentveryearlyinthecourseofprimary var-icella infection in normal subjects (Fig. 2).
Amongeight patients tested within4days after
theonsetofexanthem, sevenhadIgGantibody
and six had IgM antibody. VZV IgG and IgM antibody titers continued to increase 7 to 16
days after the appearance of the rash in most
patients. The production of VZV IgG and IgM antibody within a few days after the onset of clinicalvaricellawasaccompanied byapositive
invitrolymphocytetransformationresponseto
VZVantigen in fourpatients.Antibody produc-tion precededlymphocyte transformationin the other four patients. The only patient with no
detectable VZV IgG, IgM,orlymphocyte
trans-formationwastestedjust 12 h afterappearance ofthe exanthem.
VZVIgM antibodywasdetectable by RIA in
18of 26patients with herpeszoster,including13 patients tested 3 to 5 weeks after the onsetof the infection (Fig. 1). All of the patients with herpes zosterhad VZV. IgG antibody and posi-tive lymphocyte transformation responses to
VZVantigen.
Thepresenceofhigh titers of VZV IgG
anti-body in patients withacutevaricella and herpes
zoster did not interfere with the detection of VZV IgM antibody in unfractionated serum. >4.2x16
<q
4.2x10Z 0
1.0x106
a CC_ 2.6x10 <ci
211
Cà.,
65,536m 16,384
F--z 4096 o
P 1024
-J
256
Co 64
w
c,
F-.
Co 16
w a <16
s
VOL.12,1980 371
on February 7, 2020 by guest
http://jcm.asm.org/
372 ARVIN AND KOROPCHAK
4096 6.5x104
64 1024 16,384 2.6xl0O
SERUM DILUTION
FIG. 3. Binding of '25I-labeled goat anti-human IgG and IgMantiseraafter the incubation of
unsep-aratedserumandserumfractions fromasubjectwith
acutevaricella with VZVantigen. Thefigureshows thespecific bindingincpmof
"'M-labeled
goatanti-serumtohumanIgGdetectedafterthe incubationof unseparatedserumandserumtreated with 2-mercap-toethanol(ME)with VZVantigen.Thespecific bind-ingin cpm of '251-labeledgoatantiserum to human
IgM after the incubation of unseparated serum, se-rum adsorbed with staphylococcal proteinA (staph A), andserumtreatedwith ME is shown. The shaded areasindicate therangeofcpmmeasured when 1251.
labeledanti-human IgG (S)orIgM (O) wasadded afterthe incubationoftheserawith controlantigen. The horizontal axis indicates the dilutionsatwhich eachserumpreparationwastested.
Testing done after adsorption of serum with
staphylococcalproteinAand aftergelfiltration to remove IgG showed that the antibody de-tected with the VZVIgM procedurewaspresent
in theIgMfractionand thatIgMtiterswerenot increased by the removal ofIgG (Fig. 3). VZV
IgG antibodyremained detectable after
adsorp-tion withstaphylococcal proteinAorgel
filtra-tion in some cases, reflecting thesensitivity of the RIA method.VZVIgGwasmeasurable when
4096 20481
w
1-o
O
z
r.,
N
1024f 5121 256 128 641 321 16 8 <8
CF IAHA IFA RIA
VZV IgG ANTIBODY ASSAY
FIG. 4. Comparison of IgG antibodytiters to VZV measured by IAHA, IFA, and RIA in seropositive
subjects, tested more than 20years after varicella
infection, who had no detectable VZV antibody, as
determined by complement fixation. Twofold serial dilutions are indicated along the vertical axis. The
horizontalaxisshowsthe VZVantibodyassaysthat werecompared.
the totalIgG had been reduced toless than0.9 to 1.8 mg/ml. IgM antibody was not detected
after the treatment of VZV IgM-positive sera
with2-mercaptoethanol, but IgG antibody
per-sisted (Fig. 3).
DISCUSSION
The solid-phase RIA technique provedtobe
asimpleand sensitive method for the
measure-ment of VZV IgG and IgM antibodies. Using commercially available reagents, this method
detected VZV antibody in very dilute serum
samples in which no VZV antibody had been
detectedbyIAHAorIFA. The RIA methodwas
not subject to the false-negative results found with theIAHAtechniqueinolderpersons with-out recentinfection notedby Forghanietal. (8)
and observed in thepresentstudy.The RIAwas as reliable as the IFA in detecting antibody
amongsubjectsremotefrom infection and
mea-suredhigher titersofantibodyin thesesubjects. Among21VZVseronegative subjectswhowere
susceptible by lymphocyte transformation, the IFAmethodgaveone false-positiveresult
com-paredwithnonebyRIA. Thefalse-positiveIFA
ci)
I I I I
a e»
-e»
- en~~~~0q
-_
_
~~*
-_ e_
_ . _
1
J. CLIN. MICROBIOL.
on February 7, 2020 by guest
http://jcm.asm.org/
RIA FOR IgM AND IgG ANTIBODIES TO VZV 373
titerwasduetothedifficulty of accurately judg-ing differences influorescence between
negative
andlow-positive sera.Theendpoint in 'he RIA methodwasclearandeliminatedtheproblem of subjectiveinterpretation. Although theRIA was notcompared withthe VZV enzyme
immunoas-saydescribedby
Forghani
etal.(8), thehighest IgGtiterobserved with thatassay was1:2,048 insubjects remote from infection, whereas the
mean titer forasimilar
population
was 1:6,309by RIA and 20 subjects had titers 1:16,384. Sinceserafrompatientswho weresusceptibleto varicella by lymphocytetransformationand IFA wereconsistently negative for VZV by RIA, and
fiveof thesesera wereobtained fromindividuals
who subsequently developed varicella, we be-lieve that the method can be usedclinically to
identifyVZV-susceptible patients.
Animportant advantage of the RIAtechnique
was itsapplication tothemeasurement ofVZV IgMantibody. Althoughthe IFAprocedurehas beenused tomeasure VZVIgMantibody,
com-petitionbetween IgG andIgM forbindingsites can cause
difficulty
with this method (6). The capacitytomeasureIgMantibody without frac-tionatingserumand thecommercialavailabiity of purifiedmonospecific
goat antihuman IgM antiserum made the RIA a practical assay forVZV IgM
antibody.
Withthemonospecific
an-tiserum,
specific binding
ratioswere ashigh forIgMasfor
IgG
antibody
withoutanyalterationofthe
procedure.
Although
rheumatoid factorcaninterferewith thedetectionofspecificVZV IgM antibody in a few instances, this problem shouldarise
only
withseracontaining veryhightitersofrheumatoid factor. Even in these sera,
false-positive IgM titers can be
expected
tobe lowerthantiters observedduringVZV infection. Thephenomenon is notlikely
tobe importantin mostclinicalcircumstances,sincerheumatoid factorwas not detected in patientswith either
varicella or herpes zoster. Preliminary studies
show thatthe RIA technique can also be used to measureVZV-specific IgA antibody.
Sixty-ninepercentofpatientswithherpes
zos-ter had VZV
IgM antibody
as determined byRIA. These data substantiated theobservations
of Ross and McDaid (22) andBrunelletal.
(4),
who reported that VZV IgM antibodycould be detected by IFAinpatientswithherpes zoster.Fifty percent ofpatients had IgM antibody as
determined by IFA according to Ross and McDaid(22). Schmidt and Lennette foundIgM
antibodyinonly 1 of22 patientstested by VZV neutralization with theIgMserumfraction(23). The VZVIgMdatabyRIA contrast withresults obtained when the IAHA method wasmodified to measureVZVIgM.Usingthe IAHAmethod,
Gernaetal.have suggested thataspecific VZV IgMresponse occurswithvaricella butnotwith herpes zoster (11). The VZV IgMtiters we ob-tained in serumfromsubjects withherpes zoster were as high as those in patients with primary varicella in some cases. The presence of VZV IgM antibody does notdifferentiate primary var-icellafrom generalized herpes zoster.
The studies of kinetics of the immune re-sponse to primary varicella showed that both
IgM and IgG antibody and VZV-specific lym-phocyte transformation develop very early in the course of clinical infection in otherwise
healthysubjects. The simultaneousappearances ofIgG and IgM antibodies within 4 daysafter theonset of varicella demonstrated by RIA con-trastwith the apparent delay in VZV IgM
pro-duction assessed by the IFA method (5). Cra-dock-Watson et al. detected IgM antibody within7days in 31%of patients, when 100% of patientshadhigh titers of IgM antibody byRIA within 7 days. In addition to comparing the
antibodyresponse tovaricella and herpeszoster
innormal subjects with the response in immu-nocompromised patients, the VZV RIA for IgG
and IgM antibody will be useful to determine whether the antibody response tothe varicella vaccine now being evaluated in high-risk
pa-tientsissimilarto that seenafternatural infec-tion (14). The sensitivity of the method will allow further investigation of therole of mater-nally acquired antibody in modifying varicella,
the effect ofreexposure of immune subjects to the virus, and other aspects of the natural his-toryofVZVinfection.
ACKNOWLEDGMENT
We gratefully acknowledge the assistance of Catherine
Albin, Gregory Filice, HalstedHolman, andJanet Kahle. LITERATURE CITED
1.Arvin, A. M., R. B. Rollard, L. E. Rasmussen, andT. C.Merigan.1978.Selective impairment of lymphocyte reactivity tovaricella-zoster virus antigenamong un-treated patientswith lymphoma. J. Infect. Dis. 137: 531-540.
2. Banatvala, J. E., J. M. Best, E. A. Kennedy,E. E.
Smith, and M. E.Spence.1967.Aserological method fordemonstratingrecentinfectionby rubellavirus.Br. Med.J. 3:285-286.
3. Bradley, D.W.,H.A.Fields,K. A.McCaustland,J.
E.Maynard,R.H.Decker,R.Whittington, and L.
R.Overby.1979.Serodiagnosisofviral hepatitisAby
amodifiedcompetitive binding radioimmunoassay for immunoglobulin M anti-hepatitis A virus. J. Clin. Mi-crobiol. 9:120-127.
4. Brunell,P.A.,A. A.Gershon, S.A.Uduman,andS. Steinberg. 1975. Varicella-zoster immunoglobulin during varicella, latencyand zoster. J. Infect. Dis. 132: 49-54.
5. Cradock-Watson,J.E.,M. K.S.Ridehalgh, andM.
S. Bourne. 1979. Specific immunoglobulinresponses after varicella andherpes zoster. J. Hyg. Camb.
82:319-VOL.12,1980
on February 7, 2020 by guest
http://jcm.asm.org/
374 ARVIN AND KOROPCHAK
336.
6. Cohen, I. R., L. C. Norins, and A. J. Julian. 1967. Competition between,andeffectiveness of, IgG and IgM antibodies in indirect fluorescent antibody and other tests. J.Immunol. 98:143-149.
7. Feldman, S.,W.T.Hughes, andC.B. Daniel. 1975. Varicella in children with cancer: seventy-seven cases.
Pediatrics56:388-397.
8. Forghani, B., N. J. Schmidt, and J. Dennis. 1978.
Antibodyassays forvaricella-zoster virus:comparison
of enzyme immunoassay withneutralization, immune adherencehemagglutination,andcomplement fixation. J.Clin.Microbiol.8:545-552.
9. Forghani, B., N.J.Schmidt, andE.H.Lennette.1975.
Solid phase radioimmunoassayfor typing herpes
sim-plexviral antibodies in human sera. J. Clin.Microbiol.
2:410-418.
10. Friedman, M.G., S. Leventon-Kriss, and I. Sarov. 1979. Sensitivesolid-phase radioimmunoassay for
de-tection of humanimmunoglobulinGantibodiesto var-icella-zoster virus. J. Clin. Microbiol. 9:1-10. 11. Gerna, G.,P. M.Cereda,E.Cattaneo, G. Achilli, and
M. T.Gerna.1979.Antibodytoearly antigensof vari-cella-zoster virusduringvaricella and zoster. J. Infect. Dis. 140:33-4 1.
12. Gershon,A.A.,Z. G.Kalter,and S.Steinberg. 1976. Detection ofantibodyto varicella-zoster virusby im-mune adherance hemagglutination. Proc. Soc. Exp. Biol.Med.151:762-765.
13. Gershon, A. A., S.Steinberg, andP.A.Brunell.1974. Zoster immuneglobulin:afurtherassessment. N.Engl. J.Med. 290:243.
14. Izawa,T.,T.Ihara,A.Hattori,T.Iwasa,H.Kamiya,
M.Sakurai, andM.Takahashi. 1977.Applicationof alivevaricellavaccine inchildrenwith acuteleukemia
orothermalignant diseases. Pediatrics6:805. 15. Kalimo,K. O.K.,O.H.Meurman,P.E.Halonen,B.
R.Ziola,J. K.Viljanen,K.Granfors,and P.
Toiv-anen. 1976.Solid-phase radioimmunoassayofrubella
virusimmunoglobulinG andimmunoglobulin M anti-bodies. J.Clin.Microbiol. 4:117-123.
16. Knez, V., J. A. Stewart, and D. W. Zeigler. 1976.
Cytomegalovirus specific IgMandIgGresponse in hu-mansstudiedby radioimmunoassay.J.Immunol. 117:
2006-2013.
17. Lennette,E.H.,J. L. Melnick, and R. L. Magoffin. 1974.Clinical virology: introduction to methods, p. 667-677. In E. H. Lennette,I.H. Spaulding, and J. P. Turant (ed.), Manual of clinical microbiology, 2nd ed. American
SocietyforMicrobiology, Washington,D.C.
18. Luby,J.P., C. Ramirez-Ronda,S.Rinner,A.Hull,
and P. A. Vergne-Marini. 1977. Longitudinal study of varicella-zoster virus infections in renal transplant
re-cipients.J. Infect. Dis.135:659-664.
19. Mazur, M. H. 1978. Immunodiagnosis and
immunopro-phylaxisp. 381-383. In R. Dolin(moderator), Herpes zoster-varicella infections in immunosuppressed pa-tients. Ann. Intern. Med. 89:375-388.
20. Purcell, R. H., D. C.Wong, H. J.Alter, and P. V.
Holland.1973.Microtiter solid-phase radioimmunoas-sayfor hepatitis B antigen. Appl. Microbiol. 26:478-484.
21. Rheins,M. D. 1957. Amodification of thelatexfixation testfor the study of rheumatoid arthritis. J.Lab.Clin. Med. 50:113-118.
22. Ross, C. A. C., and R. McDaid. 1972. Specific IgM antibody in serum of patients with herpes zoster infec-tions. Br. Med. J. 4:522-523.
23. Schmidt, N. J., and E. H. Lennette. 1975. Neutralizing antibody responses to varicella-zoster virus. Infect. Im-mun. 12:606-613.
24. Weller,T. H. 1979. Varicella andherpeszoster, p. 379-398.In E. H. Lennette and N. J.Schmidt (ed.), Diag-nostic proceduresfor viral, rickettsial and chlamydial infections, 5th ed. American Public Health Association, Washington,D.C.
25. Wentworth,B.B., andE.R. Alexander. 1971. Sero-epidemiology of infections due to members of the herpes virus group. Am. J. Epidemiol. 94:496-507.
26. Williams, V.,A. A.Gershon,and P. A. Brunell. 1974.
Serologicresponse to varicella-zoster membrane anti-gensmeasured by indirect immunofluorescence. J. In-fect. Dis. 130:669-672.
27. Zaia, J. A., and M. N. Oxman. 1977. Antibody to vari-cella-zoster induced membrane antigen: immunofluo-rescence assay using monodisperse glutaraldehyde-fixed targetcells. J. Infect. Dis. 136:519-530.
J. CLIN. MICROBIOL.