Download as a PDF

Journal of General Virology (1994), 75, 301-307. Printed#l Great Britain
301
Molecular analysis of the immune response to human cytomegalovirus
glycoprotein B (gB). II. Low gB-specific T and B cell responses are
associated with expression of certain HLA-DR alleles
Julie M. Curtsinger, 1. Yung-Nan C. Liu, 1 Robert Radeke, 1 Mary Kay Bryon, 3 Susan Fuad, 3
Fritz H. Bach 3,4 and Richard C. Gehrz L2
1Biomedical Research Center, St Paul Children's Hospital, 345 North Smith Avenue, St Paul, Minnesota 55102 and
Departments of ~'Pediatrics and 3Laboratory Medicine and Pathology, University of M#mesota, St Paul, Minnesota
and 4Lutheran Deaconess Hospital and Harvard University, Cambridge, Massachusetts, U.S.A.
Human cytomegalovirus (HCMV) is a common cause of
congenital infection leading to birth defects and a
leading cause of serious illness in patients with immunodeficiencies. Studies in this laboratory have focused on a
molecular analysis of the immune response to glycoprotein B (gB) of HCMV. This protein has been
shown to elicit B cell, helper T cell (Th), and cytotoxic T
cell responses, suggesting that it may be useful as a
subunit HCMV vaccine. However, previous studies
showed that although peripheral blood mononuclear
cells (PBMC) from all HCMV-seropositive donors
proliferate in response to stimulation with whole
HCMV, not all donors respond to purified recombinant
gB. In the present study, PBMC from HCMVseropositive donors homozygous for HLA-DR were
tested for proliferative responses to whole HCMV and to
purified gB expressed in vaccinia virus. PBMC from all
donors proliferated in response to HCMV, but those
from multiple donors expressing the HLA-DR3Dw3 and
-DR4Dw4 specificities, and single donors expressing the
-DR15Dw2, -DR13Dwl9 and-DR14Dw9 specificities,
failed to respond to gB. These results suggested a
possible HLA-DR association with low proliferative
responses to gB. In further studies, PBMC from donors
expressing both putative gB-high responder and low
responder HLA-DR alleles were stimulated multiple
times with gB to generate gB-specific T cell lines. These
cells were then tested for proliferative responses to gB
presented by irradiated PBMC sharing only one D R
allele with the responder cells. Cells from the gB-specific
lines proliferated only when antigen was presented in the
context of a responder DR allele but not when presented
in the context of a low responder DR allele. Analysis of
immune sera revealed that those from donors with
PBMC proliferative responses always contained antibodies reactive with B cell epitopes on both the Nterminal gp93 and C-terminal gp55 portions of gB. In
contrast, many of the sera from donors with low gBspecific proliferative responses had gp55-specific antibodies but lacked antibodies to gp93. These results
suggest that immunogenetic differences in T h responsiveness to gB may lead to lack of antigen-specific help for
antibody responses to gp93 in some cases. The prevalence of these low responder HLA alleles in the
population, and the central importance of the T cell
response to the generation of antibodies suggest that
native gB alone may not be an attractive candidate for an
HCMV subunit vaccine.
Introduction
effector functions develop only in the presence of a
strong helper T cell (T 0 response which results in the
production of various cytokines which are essential to
the differentiation of B cells and To. The T h response to
HCMV has been shown to be stimulated primarily by
envelope glycoproteins forming the glycoprotein complexes gcI and gclI (Liu et al., 1988a, b), the matrixtegument protein pp65 (Forman et al., 1985; Liu et al.,
1988b), an early 72K protein (Rodgers et al., 1987) and
the major immediate-early protein (Alp et al., 1991).
One focus of this and other laboratories has been the
characterization of humoral and cell-mediated immune
Infection with human cytomegalovirus (HCMV) stimulates both humoral and cellular immune responses in
man (Rasmussen, 1990; Gehrz, 1991). Antibodies reactive with regulatory and structural viral proteins can
be detected in convalescent sera (Pereira et al., 1982a;
Landini et al., 1985 ; Hayes et al., 1987; Gold et al., 1988 ;
Kari & Gehrz, 1990). In addition, a number of HCMV
proteins have been shown to elicit cytotoxic T cell (To)
responses (Charpentier et al., 1986; Borysiewiez et al.,
1988; Liu et al., 1991a; Riddell et al., 1991). These
0001-1916 © 1994SGM
302
J. M. Curtsinger and others
responses to the envelope glycoprotein B (gB), which is
a component of glycoprotein complex gcI (Gretch et al.,
1988; Kari et al., 1990). Our laboratory and others have
derived murine monoclonal antibodies (MAbs) which
react with either the N-terminal gp93 subunit or the Cterminal gp55 subunit of gB (Pereira et al., 1982b; Britt,
1984; Rasmussen et al., 1985; Kari et al., 1986, 1990b;
Lussenhop et al., 1988) and human antibodies to both
proteins have been detected in sera from HCMVseropositive donors (Kari & Gehrz, 1990). T~ lines
generated in response to HCMV-infected cells have
specificity for gB (Liu et al., 1991a). In addition,
numerous T h clones which are specific for gB have been
isolated in our laboratory (Liu et al., 1993). We have
previously reported that the majority of HCMV-specific
T h clones are restricted by HLA-DR molecules (Gehrz et
al., 1987). We have now analysed a large number of
clones which are specific for gB to show that DR is
immunodominant in the T h response to this glycoprotein
(P. R. Donahue et al., unpublished). In addition, the
epitopes that stimulate many of these clones have been
mapped (Liu et al., 1993). Each D R allele associates with
a limited number of epitopes (1 to 3) on gB, suggesting
the possibility that certain DR alleles might not associate
productively with any epitopes of gB.
In this paper we demonstrate that the magnitude of the
in vitro T cell proliferative response to gB is determined
by the HLA-DR haplotype of the donor. In addition, we
report that the absence of antibodies to the N-terminal
gp93 subunit of gB correlates with the lack of an in vitro
proliferative response to gB. The implications of these
results for development of a subunit vaccine for HCMV
are discussed.
Methods
H C M V antigen preparation. HCMV antigen was prepared by
purification of Towne strain HCMV virions and dense bodies as
previously described (Gehrz et al., 1980).
Gl.veoproteins expressed #z vaccinia virus. Recombinant vaccinia
viruses containing the entire HCMV gB gene (vac-gB) or the nucleotides
coding for the N-terminal 163 amino acids (aa) of gB (vac-gBm163)
were generated by Tn5 transposon-induced C-terminal deletion as
previously described (Liu et al., 1991 a). Recombinant glycoproteins gB
and gBm 163 were immunoaffinity-purified from lysates of HuTK cells
infected with the corresponding vaccinia virus recombinants using a
biotinylated MAb, 3C2 (which reacts with an epitope in the region ofaa
50 to 77), according to methods previously described (Gretch et al.,
1987). Recombinant proteins of the appropriate molecular masses were
identified by Western blot analysis (data not shown).
Peripheral blood mononuclear cells (PBMC) and gB-specific T cell
lines. PBMC were obtained from heparinized peripheral blood of
normal adult donors expressing known HLA haplotypes; the cells were
cryopreserved in 10% DMSO in liquid nitrogen prior to use as
described previously (Liu et al., 1988b). gB-specific T cell lines were
generated by initial stimulation of PBMC at 1 x 108/ml with
immunoaffinity-purified recombinant gB at the optimal stimulatory
concentration in RPMI 1640 (Celox) containing 15 % pooled human
serum from HCMV-seronegative normal male donors. After 7 days,
Lympbocult T (Biotest) was added as a source of interleukin (IL)-2 to
a final concentration of 15%. On days 14 and 21 the cells were
restimulated with recombinant gB antigen in the presence of autologous
irradiated PBMC, as antigen-presenting cells (APC) and Lymphocult
T as described previously for HCMV-specific T cell lines (Liu et al.,
1988b). These lines were tested for the expression of the T cell marker
CD3 by flow cytometry and were all greater than 90 % CD3 + (data not
shown).
Antigen-presenth~g cells. Autologous or allogeneic PBMC expressing
shared or disparate HLA class II alleles with those of the T cell donor
were irradiated with 5000 rad prior to use as APC.
Donor H L A haplotypes. The HLA class II haplotypes of donors used
as a source of PBMC for proliferation assays, T cell lines and APCs
were determined in the HLA Typing Laboratory at the University of
Minnesota using allotype-specific antisera and mixed lymphocyte
culture responses to homozygous typing cells of defined D-region
specificity.
Lymphocyte proliferation assay. PBMC proliferation assays were
performed by stimulation of 105 PBMC/well in 96-well round bottom
microtitre plates for 7 days with optimal concentrations of HCMV
antigen or recombinant gB, or in culture medium alone as a negative
control, as previously reported (Liu et al., 1988 b). Proliferation assays
of gB-specific T cell lines were performed by incubating 2 x 104 T cells
for 72 h with 105 irradiated autologous or allogeneic PBMC as APC in
the presence or absence (negative control) of antigen. Radiolabelling,
harvesting and counting were the same as for PBMC. Results are
expressed as the mean counts per minute (c.p.m.) from triplicate wells.
The S.E.M. was typically < 10% of the mean proliferative response.
Positive responses were considered to be those greater than five times
the response of negative controls, although antigen-specific responses
were typically > 10-fold higher than those of the controls.
Antibody detection. HCMV antibody titres were determined as
described previously (Gehrz et aI., 1980). Complement fixation and
indirect immunofluorescence titres of < 4 to < 10, respectively, were
considered negative. Antibody responses to gB were determined by
Western blot analysis as described previously (Lussenhop et al., 1988);
antigens used were gB glycoproteins containing the entire gB sequence
(aa 1 to 907) or the N-terminal 163 amino acids (aa 1 to 163) expressed
in recombinant vaccinia viruses. The N-terminal 163 amino acid
glycopeptide contains the major complement-independentneutralizing
B cell epitope on gp93 (aa 68 to 77) which has been reported previously
(Meyer et al., 1992; Ohlin et al., 1993) and which is recognized by our
murine gp93-specific MAbs, 3C2 and 9F9. Positive and negative
responses in immunoblots were determined by the presence or absence
of a band comigrating with the positive bands formed by reaction of
MAbs 3C2 and 9F9 with the antigens used.
Results
Analysis of proliferative responses by P B M C from
H L A - D R homozygous donors
Previous studies from this laboratory demonstrated that
the T cell proliferative response to HCMV is primarily
mediated by CD4 + T cells which are restricted by HLADR (Gehrz et al., 1987). In earlier studies (Liu et at.,
1988a; Gehrz et al., 1991), we found that PBMC from
some HCMV-seropositive donors exhibited high pro-
HLA-DR regulation of T cell responses to HCMV-gB
Analysis of proliferative responses by gB-specific T cell
lines
Table 1. Proliferative responses to H C M V gB by
PBMCs from donors homozygous,for HLA-DR
Proliferative responses to antigen
(c.p.m.)t
PBMC
donor
HLA-DR/Dw
specificity* No antigen HCMV:~
KKL-2
MS-7
MMC-3
LS-3
DH-3
LL-4
MJ-4
AL-10
VG-19
SSEK
ART-7
DR15/Dw2
DR15/Dw2
DR3/Dw3
DR3/Dw3
DR3/Dw3
DR4/Dw4
DR4/Dw4
DR4/Dwl0
DR13/DwI9
DR14/Dw9
DR7/Dw7
6801
1132
4684
425
161
1174
693
397
254
661
3474
70538
81221
36658
70176
15974
63 384
174895
124987
55697
181818
138043
303
gB:~
4527
25930
1588
345
88
1158
586
8148
638
2303
53461
* HLA haplotype analysis was performed using allotype-specific
antisera and mixed lymphocyte culture responses to homozygous
typing cells of defined HLA specificity.
t Proliferation assays were set up using 1 x l0s PBMC and HCMV
antigen or purified recombinantgB. The cellswere cultured for 7 days,
and [3H]thymidinewas included in the cultures during the last 16 h.
Results are expressed as mean c.p.m, of triplicate wells; positive
proliferativeresponses are underlined. PBMC from some donors (MJ4, AL-10, VG-19, SSEK) were assayed only once. PBMC from other
donors were assayed twice(MS~7,MMC-3, LS-3, DH-3) or three times
(KKL-2, LL-4, ART-7) with comparable results.
:~ Sucrose gradient-purified Towne strain HCMV was used at
1 lag/ml; affinity-purifiedrecombinant gB expressed in vaccinia virus
was used at multiple concentrations; the maximum response is
reported.
liferative responses to both whole H C M V antigen and
purified gB, whereas other donors responded to whole
H C M V but gave a low proliferative response after
stimulation with gB. These donors were all heterozygous
for H L A - D R . Therefore, to assess the possible role of
H L A - D R in determining the ability to respond to gB,
PBMC from HCMV-seropositive donors who are homozygous at the H L A - D R locus were tested in a proliferation assay.
As seen in Table 1, PBMC from one donor expressing
the D R 4 D w l 0 allele, one donor expressing the DR7Dw7
allele and one of two donors expressing the DR15Dw2
allele exhibited high proliferative responses to stimulation with both whole H C M V antigen and purified gB.
These data suggest that these D R alleles can associate
with epitopes on gB and are therefore gB-high responder
alleles. Proliferative responses of PBMC from three
donors expressing DR3Dw3, two donors expressing
DR4Dw4, one donor expressing D R 1 3 D w l 9 and one
donor expressing DR14Dw9 were positive when the
antigen was whole HCMV, but clearly negative when the
antigen was gB. These data suggest that certain D R
alleles may not recognize any T h epitopes on gB and are
therefore gB-low responder alleles.
The possibility always exists that lack of a PBMC
proliferative response to gB is due to a lack of crossreactivity of T cells induced by wild H C M V strains with
the Towne strain gB sequence rather than gB nonresponsiveness primarily determined by expression of
certain D R alleles. In fact, we have determined that even
among the limited number of T cell epitopes on gB that
have been defined to date, there is sequence variation
among H C M V strains which affects T cell responsiveness
(Liu et al., 1991 b). Therefore, we wished to assess the
response to gB by PBMC from additional donors
expressing putative low responder alleles. We chose to
focus on DR3Dw3 and DR4Dw4. As we did not have
access to additional HCMV-seropositive individuals who
were homozygous for DR3Dw3 or DR4Dw4, we used a
different approach to confirm the correlation between
these H L A - D R alleles and a low proliferative response
to gB. Donors were selected who were heterozygous for
H L A - D R and expressed one of the putative gB-low
responder alleles and one of the gB-high responder
alleles. T cell lines specific for gB were generated from
these donors by incubation of their PBMC with
recombinant gB, followed by restimulation an additional
one or two times with purified gB in the presence of
irradiated autologous PBMC as APC and conditioned
medium containing IL-2. These cells were greater than
90 % CD3 + by flow cytometry analysis (data not shown).
The gB-specific T cell lines were then tested for their
proliferative responses to gB in the presence of a panel of
APC, including those which were matched for only one
of the H L A - D R alleles expressed by the donor of the
responding cells.
Fig. 1 shows the results of studies with two lines from
donors expressing the putative gB-low responder
DR4Dw4 allele and a gB-high responder allele, either
D R 4 D w l 0 or DR7Dw7. The gB-specific T cells from
donor CG (Fig. l a) proliferate well in response to
antigen in the presence of autologous APC or APC from
two donors sharing the D R 4 D w l 0 allele with the
responding cells. However, APC from three donors
sharing only DR4Dw4 with the responder fail to support
any proliferation in response to gB or HCMV. Also,
unmatched APC do not present antigen to the T cell line.
Similarly, the gB-specific T cells from donor VM (Fig.
1 b) proliferate well in response to antigen presented by
autologous APC or APC from donors sharing the
DR7Dw7 allele with the responding cells. Again, APC
from three donors sharing only DR4Dw4 with the
responding cells fail to promote any proliferation in
response to gB or HCMV. As for responder cells from
donor CG, cells from donor VM do not respond to
J. M. Curtsinger and others
304
APC
'
co IoR4Dw,,
DR,DQw3)
Dw~0~~
(DQw3.
'
I
.................111111111
(a)
VM (DR4Dw4. DR7Dw7) [1
(DQw2. DQw3)
SC (DR,Dw4. DR5Dw5 )
(DQw3, DQw3)
• N o antigen
1131H C M V
• gB
(a)
~[I
• No antigee
[] HCMV
• gB
B
i
GM {DR3Dw3, DR4Dw?)
(DQw2, DQw3)
VM (DR4Dw4, DR7Dw7)
(DQw2, DQw3)
CS (DR5DwS. DR7DwT)
(DQw2. DQw3)
SJK IDRIDwI, DR2Dw2)
(DQwl. DQwl )
JA (DRIDw?, DR4DwI0) ~WTiTOTTTiTWTTiTWT~
( DQwl, DQw3)
I
AO (DRIDw?, DR4D',uI£q h ~ l u i ~ m ~ m m m m m l m m ~
(DQwl, DQw3)
RCG IDR]Dw?, DR3Dw3)
(DQwl, DQw2) [1
(DQw2,DQw2)~
RCG (DRIDwl, DR3Dw3)
(DQwl, DQw2)
i
CED IDR3Dw3, DR5Dw?)
(DQw2, DQw3)
JD (DR2DwL DR4Dw4)
(DQw3, DQ?)
DC (DR3Dw3. DR7Dw7)
APC
I
i
I
DC'DR3Dw3,
OR7Dw7) Illlll............ IIIIIHI......................................
tDQw2, DQw2)
,
I
10 000
i
I
20 000
SC (DR4Dw4. DR5Dw5)
(DQw2, DQw3}
I
i
IJlllillJllHflllllllll[llfllllllllllllllll
iliiiiilliiililllllltlllllllllllllltlillilllll
I
I
I
/T]]I
l
I
I
I
i
'0
0
30 000
50 0
I
10000
I
15000
I
I
I
(b)
J P H (DR3Dw3, DR7Dw7)
(DQw2, DQw2)
CED (DR3Dw3, DR5Dw?)
(DQw2, DQw3)
[ ] N o antigen
[I] H C M Y
igB
GM (DR3Dw3, DR4Dw?)
(DQw2, DQw3)
VM (DR4Dw4. DR7Dw7)
(DQw2, DQw3)
CS (DRSDwS, DR7D'w7)
(DQw2, DQw3)
RI
SJK (DRIDwl, DR2Dw2)
(DQwl, DQwl)
J
SC (DR4Dw4, DR5Dw5)
(DQw2. DQw3)
0
10000
20000
30000
Proliferation (c.p.m)
40000
Fig. 1. Responses by gB-specific T cell lines from donors expressing
DR4Dw4. T cell lines with specificityfor gB were generated by repeated
stimulation of PBMC from donors CG (a) and VM (b) with purified
recombinant gB in the presence of autologous irradiated PBMC as
APC and a source of IL-2, as described in Methods. Proliferation
assays were set up using 2 x 104 T cells, 10~irradiated PBMC from the
donors indicated as APC, and either HCMV at 1 gg/ml or recombinant
gB at several concentrations; the optimal response to gB is reported
here. Cells were cultured for 3 days, with [aH]thymidine added for the
final 16 h. Results are expressed as mean c.p.m, of triplicate wells.
antigen presented by u n m a t c h e d APC. In this experiment, several o f the A P C that do not present antigen to
one line are able effectively to present antigen to the
other, ruling out a general defect in the antigenpresenting ability o f these cells, These results demonstrate that the majority, if not all, o f the gB-specific T
cells in these two lines are restricted by one o f the H L A D R gB-high responder alleles, either D R 4 D w l 0 or
D R 7 D w 7 , with little or no detectable response restricted
by D R 4 D w 4 . These results, together with the results with
two d o n o r s h o m o z y g o u s for D R 4 D w 4 (Table 1), support
the conclusion that D R 4 D w 4 is a gB-low responder
allele.
Fig. 2 shows the results o f studies with two gB-specific
lines f r o m d o n o r s expressing D R 3 D w 3 and D R 7 D w 7 ,
which are putative gB-low responder and high responder
alleles, respectively. As seen in Fig. 2(a), gB-specific T
cells f r o m d o n o r D C proliferated in response to antigen
presented by autologous A P C and A P C expressing
D R 7 D w 7 , but did not proliferate in response to antigen
presented by A P C expressing D R 3 D w 3 or by A P C
~lllrllllllllllllltlllrll|
II I II)111HIIIIII Illllllllllllll Illlllllll Illlllllllllllr Jill II
H!
m
0
[
2000
m
I
,
I
~
L
i
I
J
4000 6000 8000 10000 12000
Proliferation (c.p.m)
Fig. 2. Responses by gB-specific T cell lines from donors expressing
DR3Dw3. T cell lines with specificity for gB were generated from
donors DC (a) and JPH (b), and analysed for proliferative responses to
HCMV and recombinant gB as described in Fig. 1 and Methods.
u n m a t c h e d at D R . The gB-specific T cells f r o m the
second donor, J P H (Fig. 2b), also gave the strongest
proliferative response when A P C expressed D R 7 D w 7 ,
although a u t o l o g o u s A P C did not stimulate well in this
experiment. A P C f r o m two d o n o r s sharing the D R 3 D w 3
allele with the responder and f r o m two u n m a t c h e d
d o n o r s failed to support a proliferative response to gB or
H C M V . The results in Fig. 2, together with the results
with three d o n o r s h o m o z y g o u s for D R 3 D w 3 (Table 1),
support the conclusion that D R 3 D w 3 is a gB-low
responder allele.
The results in Fig. 1 and 2 also confirm that D R 4 D w l 0
and D R 7 D w 7 are gB-high responder alleles, as was
suggested by the data in Table 1. In addition, these
results support previous studies f r o m this l a b o r a t o r y
(Gehrz et al., 1987) showing that the DR-restricted
response to H C M V is i m m u n o d o m i n a n t . We anticipated
that we might see at least a low level o f proliferation
stimulated by u n m a t c h e d A P C or A P C m a t c h e d for the
low responder D R allele, since some o f these A P C
d o n o r s are m a t c h e d with the responder for DQ. Also,
D P p o l y m o r p h i s m is m o r e limited than that o f D R or
DQ, so it is likely that some o f these A P C d o n o r s were
matched with the responder for DP. However, as noted
above, only matching for the high responder D R allele
led to significant proliferation in response to gB.
HLA-DR regulation of T cell responses to HCMV-gB
Table 2. HCMV-specific B and T cell responses of
normal donors to gB glycoproteins
HCMV-specific
HLA*
Donor
DR
Dw
SJK
WRC
JDD
SRR
CED
KM
CG
DH
JDe
SC
DC
VM
CS
1, 2
2, 4
2, 4
2, 4
3, 5
4,?
4, 4
4, 4
4, 5
4,5
3,7
4,7
5,7
1, 2
2, 4
2 4
21, 13
3, ?
14 ?
4, 10
15, 15
?, 5
4,5
3,7
4,7
5,7
KKL-2
MS-7
LS-3
DH-3
LL-4
ART-7
2,
2,
3,
3,
4,
7,
SBu
RCG
4, 7
1, 3
2
2
3
3
4
7
proliferation:~
Whole
H C M V gB gp93
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
2
2
3
3
4
7
+
+
+
+
+
+
+
+
+
+
+
+
+
+
4, 7
I, 3
.
.
2,
2,
3,
3,
4,
7,
HCMV-specific
Abt
.
.
Whole
HCMV Vac-gB
+
+
+
+
+
+
+
.
.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
.
.
* H L A h a p l o t y p e analysis was p e r f o r m e d using allotype-specific
antisera and mixed lymphocyte culture responses to homozygous
typing cells of defined HLA specificity.
~ Antibody responsesto wholeHCMV were assessedby complement
fixation and indirect immunofluorescence. Antibody responses to gB
were determined by Western blot analysis; antigens used were gB
glycoproteins containing the entire gB sequence (aa 1 to 907) or the Nterminal 163 amino acids (aa 1 to 163) expressed in recombinant
vaccinia viruses. Positive and negative responses in immunoblots were
determined by the presence or absence of a band comigrating with the
positive bands formed by reaction of our murine gp93-specificMAbs
3C2 and 9F9 with the antigens used.
:~ Proliferation assays were set up using 1 x 105 PBMC and HCMV
antigen or purified recombinant gB. The cells were cultured for 7 days,
and [3Hlthymidine was included in the cultures during the last 16 h.
Mean c.p.m, of triplicate wellswas determined. Results werescored as
positive when mean c.p.m, in the presence of antigen was > fivefold
the mean c.p.m, in the absence of antigen. Sucrose gradient-purified
Towne strain HCMV was used at 1 gg/ml; affinity-purified recombinant gB was used at multiple concentrations.
Analysis of antibody and proliferative responses to gB
by HLA-DR heterozygous and homozygous donors
Because the T cell proliferative response is dependent
upon T h, and because the T h response is central to the
development of both humoral and cytotoxic responses,
we reasoned that the effect of the H L A - D R phenotype
on the gB-specific T cell proliferative response might
secondarily influence the specificity o f the B cell response
to gB. To test this possibility, antibody and proliferative
responses of a panel of H L A - t y p e d HCMV-seropositive
donors were determined.
As seen in Table 2, sera from five HCMV-seropositive
donors did not contain detectable levels of antibodies
305
reactive with gp93. O f particular interest, we noted that
all of the five donors who lacked gp93-specific antibodies
also lacked P B M C proliferative responses to gB. These
results suggest that the absence of gB-specific Th which
can provide cognate help to B cells could lead to failure
to generate gp93-specific antibodies. The two homozygous donors who lacked both gp93-specific antibodies
and P B M C proliferative responses to gB expressed
D R 3 D w 3 and DR4Dw4, the two low responder D R
alleles defined in this paper.
Discussion
Previous studies from this laboratory have suggested
that the T~ response to gB o f H C M V is limited to a small
number of epitopes for each restricting class II allele (Liu
et al., 1993). This observation led to the hypothesis that
certain alleles would fail to associate with any epitope
on gB, resulting in a low T cell proliferative response to
gB. We have demonstrated in this paper that this is
indeed the case. First, P B M C from H L A - D R homozygous donors expressing certain D R alleles exhibited
high proliferative responses to gB, whereas those from
donors expressing other alleles had low responses to gB.
Analysis of polyclonal gB-specific T cell lines from H L A D R heterozygous donors confirmed that the two alleles
chosen for study, D R 4 D w 4 and DR3Dw3, were in fact
associated with a low proliferative response to gB. The
data also support the conclusion that D R 4 D w l 0 and
DRTDw7 are gB-responder alleles.
In our studies using H L A - D R homozygous donors,
we noted that the gB-specific response by P B M C from
one of the DR15Dw2 homozygous donors (KKL-2) was
low, whereas the gB response of a second donor (MS-7)
was high. One possible explanation for this observation
is that the DR15Dw2-restricted T h epitope in Towne
strain H C M V may not be conserved in all wild strains.
Consequently, the infecting strain that elicited the
HCMV-specific T cell response in K K L - 2 m a y either not
contain an epitope on gB which associates with
DR15Dw2, or the DR15Dw2-restricted T cells elicited
by this strain m a y not cross-react with the Towne
DR15Dw2-restricted gB epitope. We have previously
shown that sequence variation in gB can occur in defined
T cell epitopes, and that this variation affects T cell
responses (Liu et al., 1991b).
Class II-restricted T h can have a direct role in host
defence against persistent viruses such as H C M V by
production of antiviral cytokines such as interferon
g a m m a and, in some cases, by cell-mediated cytotoxicity.
In addition, T h provide help to specific B and H L A class
I-restricted T c precursors. Thus, deficits in the T h
response to gB could have larger effects on the immune
response to gB. In this paper, we examined the possible
306
J. M . Curtsinger and others
consequences of low T h responses to gB on the gBspecific antibody response. We found that five of 19
HCMV-seropositive donors tested did not have antibodies specific for epitopes on the gp93 portion of the gB
polypeptide. These five donors did have antibodies that
reacted with the whole gB complex, presumably with
specificity for gp55. The failure to detect gp93-specific
antibodies in these five is unlikely to be due to lack of
cross-reactivity of antibodies in our test sera with the
gB(aa 1 to 163) polypeptide derived from the Towne
strain which was used in our Western blot experiments,
because we have previously shown that gp93-specific
MAbs react with all HCMV strains (Gehrz et al., 1992).
Furthermore, sequence analysis of gB from a large
number of HCMV strains has shown that the major
neutralizing B cell epitope on gp93 in the region of aa 68
to 77 is highly conserved (Chou, 1992). Interestingly, the
two D R homozygous donors who lacked gp93-specific
antibodies expressed HLA-DR alleles which were consistent with a gB-low responder phenotype and in fact
exhibited low proliferative responses to gB. The remaining three donors also expressed at least one allele
which our observations suggest may be a low responder
allele for gB (DR2Dw21, DR4Dwl3 and DR4Dwl5,
unpublished observations; DR3Dw3, this paper) and did
not express D R alleles which we have shown to be
associated with gB responsiveness.
Our results suggest that the absence of gB-specific T~
that can provide antigen-specific help to B cells could
lead to a failure to generate gp93-specific antibodies in
some cases. The ability of the gp93 antibody-negative
donors to generate gp55-specific antibodies suggests that
the gp55 antibody response somehow bypasses the need
for antigen-specific cognate help or, alternatively, there
may be a DQ- or DP-restricted T h cell response to gB
that contributes little to the proliferative response but is
sufficient to provide T cell help. These explanations may
also account for the fact that sera from two donors
whose PBMC do not proliferate in response to gB do
contain gp93-specific antibodies.
All routine and human MAbs that react with the
major antigenic site on gp93 which have been obtained to
date neutralize HCMV in the absence of complement
(Masuho et al., 1987; Karl et al., 1990; Meyer et al.,
1992; Ohlin et al., 1993), whereas most gp55-specific
MAbs require complement for neutralization (Pereira et
al., 1984; Britt, 1984; Rasmussen et al., 1985; Karl et al.,
1986; Banks et at., 1989; Ohlin et al., 1993). Human
immune sera also contain gB-specific neutralizing antibodies, including a significant proportion which reacts
with epitope(s) on the N-terminal portion of the gB
polypeptide (Britt et al., 1990; Liu et al., 1991 a). This
suggests that antigenic sites on gp93 recognized by
complement-independent antibodies may reside in func-
tional domains involved in virus attachment or fusion
with the plasma membrane. Consequently, antibodies to
gp93 may be particularly important in limiting HCMV
infection and cell-to-cell spread.
We conclude from these studies that certain HLA-DR
alleles are associated with a low gB-specific T cell
proliferative response, suggesting a diminished Th response. This low T cell response can be associated with
a deficiency in the antibody response to gp93. During
natural infection with HCMV, HLA-related deficits in
the T h response to gB may not compromise the host since
compensatory T h responses to other viral proteins are
likely to be sufficient to provide help for protective
neutralizing antibody and To responses to gB and other
HCMV proteins. However, HLA regulation of the T h
response to gB, and probably other HCMV proteins as
well, may present a major obstacle to development of a
subunit vaccine consisting of a single recombinant viral
protein. Individuals expressing two low responder HLA
haplotypes would not be expected to develop T h
responses to the vaccine protein, nor would compensatory T~ responses to other viral proteins be elicited to
provide cognate help for B cells and T c specific for the
vaccine proteins. Consequently, a substantial portion of
the vaccinated population would still be susceptible to
natural infection with HCMV.
Supported by grant no. PO1-HD19937 from the National Institute
of Child Health and Human Development, and the Biomedical
Research Center, St Paul Children's Hospital.
References
ALP, N. J., ALLPORT, T. D., VAN ZANTEN, J., RODGERS,B., SISSONS,
J. P. G. & BORYSIEWlCZ,L.K. (1991). Fine specificity of cellular
immune responses in humans to human cytomegalovirus immediateearly 1 protein. Journal of Virology 65, 4812--4820.
BANKS,T., Huo, B., KOUSOULAS,K., SPAETE,R., PACHL,C. & PEREIRA,
L. (1989). A major neutralizing domain maps within the carboxyterminal half of the cleaved cytomegalovirus B glycoprotein. Journal
of General Virology 70, 979-985.
BORYSIEWICZ, L.K., HICKLING, J.K., GRAHAM, S, SINCLAIR, J.,
CRANAGE, M. P., SMITH, G. L. & SISSONS,J. G. P. (1988). Human
cytomegalovirus-specific cytotoxic T cells: relative frequency of
stage-specific CTL recognizing the 72KD immediate early protein
and glycoprotein B expressed by recombinant vaccinia viruses.
Journal of Experimental Medicine 168, 919-931.
BRITT, W. J. (1984). Neutralizing antibodies detect a disulfide-linked
glycoprotein complex within the envelope of human cytomegalovirus. Virology 135, 369-378.
BRITT, W. J., VUGLER,L., BUTFILOSKI,E. J. & STEPHENS,E. B. (1990).
Cell surface expression of human cytomegalovirus (HCMV) gp55116 (gB): use of recombinant vaccinia virus-infected cells in analysis
of the human neutralizing antibody response. Journal of Virology 64,
1079-1085.
CHARPENTIER,B., MICHELSON,S. 8¢. MARTIN, B. (1986). Definition of
HCMV specific target antigens recognized by cytotoxic T cells
generated in vitro by using an autologous lymphocyte system.
Journal of Immunology 137, 330-336.
CHOU, S. (1992). Comparative analysis of sequence variation in gp 116
and gp55 components of human cytomegalovirus. Virology 188,
388-390.
H L A - D R regulation o f T cell responses to H C M V - g B
FORMAN,S. J., ZAIA,J. A., CLARK,B. R., WRIGHT,C. L., MILLS, B, J.,
POTTATHIL, R., RACKLIN,B. C., GALLAGHER,M. T., WELTE, K. &
BLUME, K.G. (1985). A 64,000 dalton matrix protein of human
cytomegalovirus induces in vitro immune responses similar to those
of whole viral antigen. Journal oflmmunology 134, 3391-3395.
GEHRZ, R. C. (1991). Human cytomegalovirus. Biological and clinical
perspectives. Advances in Pediatrics 38, 190-219.
GEHRZ, R. C., CHRISTIANSON,W. R., LINNER,K. M., GROTH, K. E. &
BALFOUR, H.H., JR (1980). Cytomegalovirus vaccine: specific
humoral and cellular immune responses in human volunteers.
Archives of Internal Medicine 140, 93(~939.
GEHRZ, R, C., FUAD, S., LIE, Y.-N. C. & BACH, F. H. (1987). HLA
class II restriction of T helper cell response to cytomegalovirus
(CMV). I. lmmunogenetic control of restriction. Journal of Immunology 138, 3145-3151.
GEHRZ, R.C., LIE, Y.-N. C., ECKHARDT, J. &, KLAUS, A. (1991).
Relevance of immune responses to pathogenesis of cytomegalovirusassociated diseases. Transplantation Proceedings 3 (Suppl. 3), 75-84.
GEHRZ, R. C., NELSON, C. M. & KARl, B. (1992). A combination of
human cytomegalovirus (HCMV)-specific murine monoclonal antibodies exhibits synergistic antiviral activity in vitro. Antiviral
Research 17, 115-131.
GOLD, D., ASHLEY,R., HANDSFJELD,H. H., VERDON, M., LEACH, L.,
MILLO, J., DREW, L. & COREY,L. (1988). Immunoblot analysis of the
humoral immune response in primary cytomegalovirus infection.
Journal of Infectious Diseases 157, 319-326.
GRETCH, D.R., SUTER, M. & STINSKI, M.F. (1987). The use of
biotinylated monoclonal antibodies and streptavidin affinity chromatography to isolate herpes virus hydrophobic proteins and
glycoproteins. Analytical Biochemistry 163, 270-277.
GRETCH, D. R., GE~4RZ, R.C. & S'rlNSKI, M. F. (1988). Characterization of a human cytomegalovirus glycoprotein complex (gcI).
Journal of General Virology 69, 1205-1215.
HAYES, K., ALFORD, C. & BRITT, W. (1987). Antibody response to
virus-encoded proteins after cytomegalovirus mononucleosis.
Journal of Infectious Diseases 156, 615 621.
KARl, B. & GErmZ, R, (1990). Analysis of human antibody responses
to human cytomegalovirus envelop glycoproteins found in two
families of disulfide linked glycoprotein complexes designated gC-I
and gC-II. Archives of Virology 114, 213-228.
KARl, B., LUSSENHOP,N., GOERTZ, R., WABUKE-BUNOTI,M. & GEHRZ,
R. C. (1986). Characterization of monoclonal antibodies reactive to
three biochemically distinct human cytomegalovirus glycoprotein
complexes. Journal of Virology 60, 345 352.
KARL B., LIu, Y.-N. C., GOERTZ,R., LUSSENHOV,N., STINSKI,M. F. &
Gl?rmz, R. (1990). Structure. and composition of a family of human
cytomegalovirus glycoprotein complexes designated gC-I (gB).
Journal of General Virology 71, 2673 2680.
LANDINLM. P., RE, M. C., MIROLO,G., BALDASS~RI,B. & LAPLACA,
M. (1985). Human immune response to cytomegalovirus structural
polypeptides studied by immunoblotting. Journal of Medical Virology 17, 303-311.
LIU, Y.-N. C., KARI, B. & GEHRZ, R. C. (1988a). Immune responses to
major human cytomegalovirus glycoprotein complexes. Journal of
Virology 62, 1066~1070.
LIu, Y.-N. C., ECKH~DT, J., KARL B. & GEHRZ, R.C. (1988b).
General characterization of human cytomegalovirus-specific proliferative CO4 + T cell clones. International Journal of Cell Cloning 2,
352-364.
LIU, Y.-N. C., KLAUS, A., KARl, B., STINSKI, M_ F., ECKHARDT,J. &
307
GEHRZ, R.C. (1991 a). The N-terminal 513 amino acids of the
envelope glycoprotein gB of human cytomegalovirus stimulates both
B and T cell immune responses in humans. Journal of Virology 65,
1644-1648.
LIu, Y.-N. C., OvIrz, G., KLAUS, A., COOPER, J. & GEHRZ, R. C.
(1991 b). Helper T cell (Th) epitope mapping of gp93 in gcI complexes
of human cytomegalovirus (HCMV). In Progress 01 Cytomegalovirus
Research, pp. 187-190. Edited by M. P. Landini.
LIu, Y.-N. C., CURTSINGER,J., DONAHUE,P. R., KLAUS,A., OPITZ, G.,
COOPER, J., KARR, R.W., BACH, F.H. & GEmtZ, R.C. (1993).
Molecular analysis of the immune response to human cytomegalovirus glycoprotein B. I. Mapping of HLA-restricted helper T cell
epitopes on gp93. Journal of General Virology 74, 2207-2214.
LUSSENHOP, N. O., GOERTZ, R., WABUKE-BUNOTI, M., GEHRZ, R. &
KARI, B. (1988). Epitope analysis of human cytomegalovirus
glycoprotein complexes using mnrine monoclonal antibodies. Virology 164, 362 372.
MASUHO, Y., MATSUMOTO, Y.-I., SUGANO, T., FUJINAGA, S. &
MINAMISHIMA,Y. (1987). Human monoclonal antibodies neutralizing
human cytomegalovirus. Journal of General Virology 68, 1457-1461.
MEYER, H., StrNDQVlST, V.-A., PEREIRA, L. & MACH, M. (1992).
Glycoprotein gp 1l 6 of human cytomegalovirus contains epitopes for
strain-common and strain-specific antibodies. Journal 03" General
Virology 73, 2375-2383.
OHLIN, M., SUNDQVIST,V.-A., MACH, M., WAHREN,B. & BORREBAECK,
C. A. K. (1993). Fine specificity of the human immune response to
the major neutralization epitopes expressed on cytomegalovirus
gp58/116 (gB), as determined with human monoelonal antibodies.
Journal of Virology 67, 703-710.
PEmSIRA, L., HOFFMAN, M. & CROMER, N. (1982a). Electrophoretic
analysis of polypeptides immune precipitated from cytomegalovirusinfected cell extracts by human sera. Infection and Immunity 36,
933 942.
PEREIRA, L., HOFFMAN, M., GALLO, D. & CREMER, N. (1982b).
Monoclonal antibodies to human cytomegalovirus: three surface
membrane proteins with unique immunological and electrophoretic
properties specify cross-reactive determinants. Infection and ImmuniO' 36, 924-932.
PEREIRA, L., HOFFMAN, M., TATSUNO, M. & DONDERO, J. (1984).
Polymorphism of human cytomegalovirus glycoproteins characterized by monoclonal antibodies. Virology 139, 73-86.
RASMUSSEN,L., MULLENAX,J., NELSON, R. & MERIGAN,T. C. (1985).
Viral polypeptides detected by a complement-dependent neutralizing
murine monoclonal antibody to human cytomegalovirus. Journal of
Virology 55, 274-280.
RASMUSSEN,L. (1990). Immune response to human cytomegalovirus
infection. Current Topics hi Microbiology and Immunology 154,
221 254.
RIDDELL,S. R., RABIN,M., GEBALLE,A. P., BRITT,W. J. & GREENBERG,
P.D. (1991). Class I MHC-restricted cytotoxic T lymphocyte
recognition of ceils infected with human cytomegalovirus does not
require endogenous viral gene expression. Journal of bnmunology
146, 2795-2804.
RODGFRS,B., BORYSlEWICZ,L., MUNDIN,J., GRAHAM,S. & SISSONS,P.
(t987). lmmunoaffinity purification of a 72K early antigen of human
cytomegalovirus: analysis of humoral and cell-mediated immunity to
the purified polypeptide. Journal of General Virology 68, 2371-2378.
(Received 23 June 1993: Accepted 17 September 1993)