induced by adenovirus and vaccinia virus - CiteSeerX

Journal of General Virology (1993), 74, 25022512. Printed in Great Britain

2507

Murine cytotoxic T cell response specific for human cytomegalovirus glycoprotein B (gB) induced by adenovirus and vaccinia virus recombinants expressing gB Klara Berencsi,' Robert F. Rando,2~ Charles deTaisne, 3 E n z o P a o l e t t i , 4 S t a n l e y A. P l o t k i n 3 and E v a G o n c z o l ' * 1 The Wistar Institute of Anatomy and Biology, 3601 Spruce Street, Philadelphia, Pennsylvania 19104-4268, 2National Institute of Dental Research, 9000 Rockville Pike, Bethesda, Maryland 20892, U.S.A., ~Institut Pasteur, Merieux, 1541 Avenue Marcel-Merieux, 69280 Marcy-rEtoile, France and 4 Virogenetics Corporation, Rensselaer Technology Park, 465 Jordan Road, Troy, New York 12180, U.S.A.

A murine model of the cytotoxic T lymphocyte (CTL) response to glycoprotein B (gB) of human cytomegalovirus (HCMV) was developed based on the use of adenovirus (Ad) and vaccinia virus (Vac) recombinants expressing gB. Mice of different major histocompatibility haplotypes [CBA (H-2k), BALB/k (H-2 k) and BALB/c (H-2d)] infected with the Ad-gB recombinant developed an Ad-specific CTL response. However, only the H-2 k mice developed a significant HCMV gBspecific CTL response, as indicated by the major

histocompatibility complex class I-restricted lysis of Vac strain Copenhagen (VacC)-gB recombinant-infected target cells by H-2 k mouse immune spleen cells. The VacC-gB recombinant elicited only a weak gB-specific CTL response in these mice, indicating that the observed gB-specific C T L response in mice is dependent on the expression vector used for immunization. The gBspecific cytotoxicity observed in H-2 k mice was mediated by the CD8 lymphocyte subset.

Human cytomegalovirus (HCMV) is ubiquitous in humans, with usually mild or inapparent acute infection followed by persistence or latency in the host. However, HCMV is a significant cause of morbidity and mortality in infants infected in utero and in immunocompromised hosts (Weller, 1971; Glenn, 1981). Although the role of individual HCMV proteins in protective immunity is not completely understood, glycoprotein B (gB) is of interest because evidence suggests that it induces neutralizing antibody following natural infection (Britt et al., 1990; G6nczS1 et al., 1991) and following immunization with purified gB (G6ncz61 et al., 1990). Moreover, human anti-gB monoclonal antibodies (MAbs) neutralize HCMV in vitro (Masuho et al., 1987). There is also evidence for activation of helper T cells and cytotoxic T lymphocytes (CTLs) by gB in naturally seropositive humans (Borysiewicz et al., 1988; Liu et al., 199I ; Riddell et aI., 1991). Finally, immunization of humans and experimental animals with purified gB induces lymphocytes that proliferate in the presence of gB in vitro (GSnczS1 et al., 1986b, 1990). In other

experiments, it has been demonstrated that animals infected with heterogeneous viral recombinants that express gB develop neutralizing antibody to HCMV (Britt et al., 1988; Cranage et al., 1986; G6nczS1 et al., 1991; Marshall et al., 1990). In addition, adenovirus (Ad)-gB recombinant virus induces gB-specific helper T cells in BALB/c and CBA mice (Berencsi et al., 1991). Recently recombinant viruses have been considered for use as vaccines for the prevention of viral diseases (Tartaglia et al., 1990). In the present study, we asked whether recombinant viruses carrying the gB gene induce gB-specific CTL responses. As HCMV does not infect mice in vivo or murine tissue culture cells, we developed a model system suitable for the evaluation of CTL responses to individual HCMV proteins in mice. This model system is based on two types of viral expression vectors, vaccinia (Vac) and Ad virus, carrying the same HCMV gB gene and used alternately for immunization of mice or for infection of target cells to show that HCMV gB induces specific CTLs. Ad and Vac recombinant viruses expressing the Towne strain gB were constructed previously. The Ad-gB recombinant expresses the gB gene under the control of the E3 promoter (Marshall et al., 1990). In the Vac

t Present address: Triplex Pharmaceutical Corporation, 9391 Grogen's Mill Road, The Woodlands, Texas 77380, U.S.A. 0001-1860 © 1993SGM

2508

Short communication

01 (a)

(b)

IL

(c)

50

~

20

50

25

12.5

6

50

25

12.5

6 50 E : T ratio

25

12.5

6

(d) _

'-

50

25

12.5

6

Fig. 1. H C M V gB-specific CTL response in Ad-gB-immunized mice. Mice were immunized i.p. with 1 x 108 p.f.u, virus. Three weeks later, splenocytes were restimulated in vitro with Ad gB-infected autologous spleen cells and tested for their ability to lyse target cells: II, uninfected; 0 , Ad-5AE3-infected, O, VacC-gB-infected; A, wt Vac-infected. (a) A&gB-immunized CBA mice; (b) wt Adimmunized CBA mice; (c) Ad-gB-immunized B A L B / k mice; (d) Ad-gB-immunized BALB/c mice. Solid line ( ), M H C class Imatched L-929 cells (a, b and c) or P815 cells (d); dotted line (---), mismatched MC57 cells (a, b).

Copenhagen strain gB recombinant (VacC-gB) the H C M V gB gene is under the control of the Vac H6 early/late promoter (G6ncz61 et al., 1991). For use in immunization, recombinant or wild-type (wt) Ad or Vac were purified by CsC1 or sucrose gradient centrifugation, respectively. To test the ability of the Ad-gB recombinant to induce an H C M V gB-specific CTL response, T cell cytoxicity assays were performed, as described previously (Rawle et al., 1989). Briefly, spleens of CBA mice immunized intraperitoneally (i.p.) with 108 p.f.u, of the Ad-gB recombinant virus or wt Ad were aseptically removed and cell suspensions were prepared from them. Cells were suspended at 2.5 × 10~ viable cells/ml in R P M I 1640 medium containing 5% fetal bovine serum, 2 × 10-5 M-2-mercaptoethanol, 10 mM-HEPES buffer, 2 mM-glutamine and 50 ~tg/ml gentamicin. Spleen cell cultures were restimulated in vitro with Ad-gB-infected (m.o.i. 10) autologous spleen cells (effector:stimulator ratio 2: 1) for 5 days in 24-well plates. Cytolytic activity of non-adherent spleen cells was tested in a 51Cr release assay. Target cells (L-929, P815 and MC57) were infected with the E3-deleted Ad-5 mutant virus lacking the XbaI D fragment of virus D N A (Ad-5AE3) (m.o.i. 40 to 80, for 40 h) or with VacC-gB or wt VacC (m.o.i. 5 to 10, for 4 h). The Ad-5AE3 virus was constructed by overlap recombination, using plasmid pAd-5 containing map units (m.u.) 59.5 to 100 deleted in E3 sequences (m.u. 78-5 to 84), and pAd-5, m.u. 0 to 75-9, as described (Marshall et al., 1990). Cells were labelled with 100 ~tCi of [51Cr]Na2CrO~ (Amersham, specific activity 250 to 500 mCi/mg) for 1 h. Labelled target cells were mixed with effector cells at various effector:target (E:T) ratios in triplicate using 96-well U-bottomed microtitre plates and incubated for 4 h. Percentage specific 5~Cr release was calculated as [(c.p.m. experimental release-c.p.m.

spontaneous release)/(c.p.m, maximal release-c.p.m. spontaneous release)] × 100. S.D. of the mean of triplicate culture was less than 10 %, and spontaneous release was always less than 25 %. As shown in Fig. 1 (a), in vitro restimulated spleen cells from CBA mice immunized with Ad-gB induced no significant lysis of uninfected or wt Vac-infected major histocompatibility complex (MHC) class I-matched L-929 (H-2 k) target cells but did lyse Ad5AE3-infected L-929 cells. H C M V gB-specific lysis of VacC-gB-infected L-929 target cells was determined to be 37, 36, 26 and 22 % at E : T ratios of 50: 1, 25 : 1, 12.5: 1 and 6: 1, respectively (Fig. 1 a). The gB-specific lysis was calculated as percentage of specific lysis of VacC-gBinfected target cells-percentage of specific lysis of wt Vac-infected target cells. Mice immunized with wt Ad virus exhibited no gB-specific lysis of these target cells (Fig. 1 b). VacC-gB-infected MHC-mismatched (H-2 b) MC57 target cells were not lysed by the splenocytes of Ad-gB-immunized mice (Fig. 1 a). Spleen cells of nonimmunized mice restimulated in vitro with Ad-gB showed neither Ad- nor gB-specific cytotoxicity (data not shown). Thus Ad-gB-infected CBA mice generated not only Ad-specific CTLs, but also CTLs that specifically recognize gB expressed by VacC-gB in an M H C class I-restricted manner. To determine whether the CBA mouse CTL response to gB was specific to the H-2 k haplotype, congenic BALB/c (H-2 d) and B A L B / k (H-2 k) mice, which differ only in M H C haplotype, were inoculated i.p. with Ad-gB, and their spleen cells were tested for gB-specific lytic activity using P815 (H-2 d) and L929 (H-2 k) target cells. Ad-specific lysis was elicited by the spleen cells of both mouse strains (Fig. 1 c, d). However, an H C M V gBspecific response was observed only in Ad-gB-immunized B A L B / k mice, i.e. VacC-gB-infected target cells,

Short communication i

60

i

i

i

1 week

i

t

i

2 weeks

i

i

4 weeks

i

i

i

12 weeks

50

~- 40

20

10

--~

•

I

M gB

wt

~

I

I

"i'-

I

I

I

M gB wt M gB Infected target cells

I

wt

I

•a

I

M gB

wt

Fig. 2. Kinetics of H C M V gB-specific C T L response after A ~ g B immunization. C B A mice were immunized i.p. with 108 to 2 x 108 p.f.u. of Ad-gB. One to 12 weeks later, splenocytes were restimulated in vitro with Ad-gB-infected autologous spleen cells and tested on L-929 target cells. Black circles represent results of individual mice; horizontal bars represent m e a n values from individual mice. Percent specific lysis of target cells is shown at an E : T of 50: 1. M, mock-infected; gB, VacC-gB-infected; wt, wt Vac-infected target cells.

but not wt Vac-infected L-929 cells, were specifically lysed (Fig. l c). No gB-specific CTL response was observed in BALB/c mice (Fig. 1d), demonstrating that H-2 k but not H-2 a mice recognize the gB antigen. The mouse strain-related difference in CTL response to gB is consistent with findings that immunogenicity of different viral proteins for the mouse CTL response varies with MHC haplotype (Tanaka et al., 1991). Kinetic analysis of the gB-specific CTL response (Fig. 2) indicated that HCMV-gB-specific CTLs were present 1 week after immunization and remained detectable throughout the 12-week test period. To characterize the cells responsible for anti-gB cytotoxicity, 3 x 106 in vitro restimulated spleen cells of Ad-gB-immunized mice were incubated with anti-mouse CD4 MAb (Pharmingen; cat. no. 3:01061D; 20 lag/3 x 106 cells) or CD8 MAb (Accurate; cat. no. CL-8921; diluted 1 : 4) for 60 min at 4 °C, and further incubated in the presence of rabbit complement (Accurate; Low-Tox M; diluted 1 : 10) for 30 min at 37 °C. Cells were washed twice and used as effector cells in a ~lCr release test. Almost no anti-gB activity was detected using CD8depleted spleen cells, whereas CD4-depleted cells continued to mediate this activity (Table 1). Thus the CD8 lymphocyte subset mediates the gB-specific lysis of the target cells.

2509

The above results show that the Ad-gB recombinant can elicit an MHC class I-restricted CD8 lymphocytemediated HCMV gB-specific CTL response in mice and that by using our approach the CTL response to an individual HCMV antigen is readily detectable. However, the recognition of a CTL epitope by the immune system depends on the T cell repertoire of the individual and the structure of its MHC molecules. Some T cell epitopes are recognized by both human and mouse T cells (Hosmalin et al., 1990). Although human CTLs recognize the gB protein at a low precursor frequency (Borysiewicz et al., 1988; Riddell et al., 1991), it remains to be demonstrated that human and mouse CTLs recognize the same epitope on this protein. To determine whether the CTL mouse model functions in the opposite direction, e.g. a gB-specific CTL response is induced in CBA mice immunized i.p. with VacC-gB, spleen cell cultures derived from these mice were tested for their gB-specific CTL activity. As restimulation of recombinant Vac-immune spleen cells with recombinant Vac may lead to a predominant response to Vac and difficulty detecting weaker CTL reponses to epitopes on the recombinant gene product, spleen cells from VacCgB-immunized CBA mice (108 p.f.u.) were pooled (cells from two mice from each group), divided in half and restimulated in vitro with VacC-gB (m.o.i. 0.5) or Ad-gB (m.o.i. 10) before testing for lyric ability on Ad-gB-, Ad5AE3-, VacC-gB- and wt Vac-infected MHC class Imatched L-929 and mismatched MC57 target cells (Table 2). In this assay, spleen cells of Ad-gB-immunized mice (108 p.f.u.) served as controls. It was observed that Ad-gB-immunized mice exhibited significant HCMV gB-specific cytolytic activity, regardless of the virus used for in vitro restimulation and for target cell infection, whereas the gB-specific CTL response observed after VacC-gB immunization was very low (< 10% in any combination of in vitro restimulation and target cell infection). The low CTL response associated with VacC gB does not reflect suboptimal experimental conditions since Vac-infected mice showed a high Vacspecific CTL response, and Ad-gB-immune spleen cells restimulated either with Ad-gB or VacC-gB elicited an identical gB-specific target cell lysis pattern. The studies on the comparative ability of CTL responses induced by an i.p. dose of 10Sp.f.u. of Ad-gB or VacGgB recombinant were performed five times with similar results. The gB-specific lysis obtained with VacC-gBimmune spleen cells was always significantly below the potency of the Ad-gB immune cells (Student's t-test, P ~< 0.01). It seems unlikely that the failure of the VacC-gB recombinant to elicit a significant CTL response rests in the lack of presentation of gB epitopes by the MHC molecules, since VacC-gB-infected target cells were readily lysed by Ad gB-primed mouse CTLs. Analysis

2510

Short communication

T a b l e 1. Phenotype o f H C M V gB-specific C T L s Specific lysis of target cells (%) Immunization* Ad-gB Ad-gB Ad gB Ad-gB Ad gB Ad-gB

Restimulation VacC-gB VacC-gB VacC-gB VacC-gB VacC gB VacC-gB

Treatment of spleen cellst VacC gB wt Vac C' Anti-CD4 Anti-CD4 + C" Anti-CD8 Anti-CD8 + C'

40 35 36 37 36 6

3 4 4 4 4 3

* CBA mice were immunized i.p. with 1 x 108 p.f.u, of A~gB, sacrificed 2 weeks later, and spleen cells restimulated in vitro with VacC-gB. t Spleen cells were depleted of CD4 or CD8 lymphocytes and used as effector cells in a 4 h 51Cr release assay. C', Complement.

with a m o n o s p e c i f i c a n t i b o d y (4A g u i n e a - p i g s e r u m ; G6ncz61 et al., 1986b) to gB s h o w e d t h a t in V a c C - g B infected H e L a cells gB was expressed i n t r a c e l l u l a r l y in 6 5 % o f cells at 5 h p o s t - i n f e c t i o n (p.i.) rising to a m a x i m u m o f 80 % at 8 h p.i., a n d surface expression o f gB o c c u r r e d in 50 to 70 % o f infected cells (G6ncz61 et al., 1991). T h e expression o f g B in V a c C - g B - or A d - g B infected t a r g e t L-929 a n d M C 5 7 cells was also tested. W e s t e r n b l o t analysis using the 4 A g u i n e a - p i g s e r u m s h o w e d similar gB expression in L-929 cells infected with V a c C - g B (m.o.i. 5 to 10 at 8 h p.i.) or A d - g B (m.o.i. o f 40 to 80 at 40 h p.i.) ( d a t a n o t shown). T h e in vitro

expression o f gB m a y n o t reflect identical gB expression in vivo in A d gB- o r V a c C - g B - i n f e c t e d mice a n d the l a c k o f gB-specific C T L m a y s i m p l y be the result o f i m p a i r e d gB expression in V a c C - g B - i m m u n e C B A mice. Since gB is n o t a s t r u c t u r a l p r o t e i n o f the r e c o m b i n a n t viruses, the f o r m a t i o n o f gB-specific a n t i b o d y is a g o o d i n d i c a t o r o f gB expression in vivo. Therefore, the presence o f H C M V n e u t r a l i z i n g a n t i b o d y in sera o f V a c C - g B - o r A d - g B i m m u n i z e d mice was tested in a m i c r o n e u t r a l i z a t i o n assay (G6ncz61 et al., 1986a). T h e results s h o w e d t h a t b o t h r e c o m b i n a n t viruses i n d u c e d n e u t r a l i z i n g antib o d i e s in C B A mice ( 1 : 5 4 o r 1 : 124, m e a n titre o f 10 mice i m m u n i z e d w i t h V a c C - g B o r A d - g B , respectively), i n d i c a t i n g t h a t gB is expressed in vivo in a m o u n t s sufficient to s t i m u l a t e the i m m u n e system. I n s u b s e q u e n t experiments, doses o f 1 x 107, 5 x 107 o r 1 x 109 p.f.u, o f V a c C - g B were used for i m m u n i z a t i o n to exclude the p o s s i b i l i t y t h a t 1 x 108 p.f.u, was n o t the o p t i m a l dose for C T L i n d u c t i o n . Results s h o w e d the absence o f C T L responses in mice i m m u n i z e d w i t h l o w e r t h a n 1 x 108 p.f.u. I m m u n i z a t i o n w i t h 1 x 109 p.f.u, o r b o o s t e r i n o c u l a t i o n with 1 x 108 p.f.u., however, i m p r o v e d the gB-specific C T L response (14 to 2 1 % o r 15 to 17 % lysis, respectively at a 50: 1 E : T ratio), b u t r e m a i n e d b e l o w the gB-specific r e s p o n s e i n d u c e d b y 1 x 108 p.f.u, o f A d - g B r e c o m b i n a n t . Thus, we c o n c l u d e t h a t V a c C - g B , a l t h o u g h i n d u c i n g a gB-specific n e u t r a l i z i n g a n t i b o d y response, is a p o o r C T L i n d u c e r for this i m m u n o g e n . V a r i o u s V a c r e c o m b i n a n t s , i n c l u d i n g C o p e n h a g e n strain r e c o m b i -

T a b l e 2. Ability of A d - g B and VacC-gB to induce H C M V gB-specific C T L responses in

CBA mice In vitro-

Mice immunized with*

restimulated with

Target ceils

E:T ratio

A~gB

Ad E3

VacC-gB

wt Vac

Ad-gB Ad-gB Ad gB Ad~B Ad~B Ad-gB Ad gB Ad-gB Ad gB Ad gB VacC-gB VacC gB VacC-gB VacC gB VacC-gB VacC gB VacC-gB VacC-gB VacC gB VacC gB

Ad-gB Ad-gB Ad-gB Ad-gB VacC-gB VacC-gB VacC gB VacC-gB Ad-gB VacC-gB Ad-gB Ad-gB Ad-gB Ad-gB VacC-gB VacC-gB VacC gB VacC-gB Ad-gB VacC gB

L-929 L-929 L-929 L-929 L-929 L-929 L-929 L-929 MC57 MC57 L-929 L-929 L-929 L-929 L-929 L-929 L-929 L-929 MC57 MC57

40:1 20:1 10:1 5:1 40:1 20:1 10:1 5:1 40:1 40:1 40:1 20:1 10: 1 5: 1 40: 1 20:1 10:1 5:1 40: 1 40:1

75 56 32 19 69 47 25 15 3 1 15 8 4 3 9 7 4 1 1 3

56 36 20 10 5 3 1 1 2 1 9 6 3 1 5 3 1 1 1 1

34 24 14 8 53 35 20 12 4 1 10 5 4 3 74 62 43 28 2 1

8 6 3 1 5 4 3 1 5 3 8 6 3 2 88 72 53 33 3 4

Specific lysis of target cells (%)

* CBA mice were immunized i.p. with 1 x 108 p.f.u, virus on the same day. Two weeks later spleen cells of two mice from each group were pooled and restimulated in vitro either with VacC-gB or Ad-gB for 5 days.

Short communication

nants expressing viral genes under the H6 early/late Vac promoter (as in our construct), have been used successfully to induce CTL responses specific for the expressed foreign antigen (Tartaglia et al., 1990; W. Cox, personal communication); however, there is other evidence of CTL responses to antigen when expressed by the natural but not a recombinant Vac (Tanaka et al., 1991). Our results with VacC-gB provide an example of the latter. The difference between the Ad-gB and VacC-gB recombinants in CTL induction may rest in the influence of the vector itself on immunogenicity. Adenoviruses are B cell mitogens in experimental animals (Bakay et aI., 1992; Gibson et al., 1982) and, because B cells are important antigen-presenting cells in immune responses to viruses, activation of these cells may favour CTL responses to simultaneouslyintroduced antigens. Alternatively, the CTL epitope on gB might successfully compete with the Ad CTL epitopes but may not be immunodominant in Vac-infected cells; T cell responsiveness to antigenic determinants can be influenced by competition among different peptides derived from the same protein or other proteins for binding to MHC molecules (Adorini et al., 1988). Early proteins are usually considered to be CTL inducers and to confer CTL-mediated protection in a natural virus infection (Del Val et al., 1991). In naturally seropositive humans, the percentage of gB-specific CTL precursors is low (Borysiewicz et al., 1988; Riddell et al., 1991). Early viral proteins may be more effectively presented to the immune system than late proteins (Coupar et al., 1986; Townsend et al., 1988; Zhou et al., 1991), and also wt viruses may have evolved strategies to evade immune surveillance. For example, HCMV has been shown to evade immune recognition by interfering with class I-mediated presentation (Browne et al., 1990). Therefore, it is possible that an antigen expressed as an early protein by a recombinant virus elicits a more effective immune response than that elicited by natural virus infection. Determination of a strong gB-specific CTL response when gB is expressed by an Ad E3 replacement vector, as compared with a weak CTL response when gB is expressed by natural HCMV infection (Borysiewicz et al., 1988; Riddell et al., 1991) or by the VacC-gB recombinant, provides important information on the choice of a delivery vector in immunization efforts where, in addition to an insertspecific neutralizing antibody response, a strong insertspecific CTL response is a favourable goal. The authors thank Leopoldo Luistro, Weichi Lin and Joseph Marton for expert technical assistance, and Marina Hoffman for editorial work. This study was supported by grant HD18957 from the National Institute of Child Health and Development and by the Institut Merieux, France.

2511

References ADORIN1, L., APPELLA, E., DORIA, G. & NAGY, Z.A. (1988). Mechanisms influencing the immunodominance of T cell determinants. Journal of Experimental Medicine 168, 2091-2099. BAKAY,M., BELADI,I., BERENCSI,K., SIDOROVA,E., AGADZANYAN,M., FACHET,J. & ERDEI,J. (1992). Immunoenhancement and suppression induced by adenovirus in chickens. Acta virologica 36, 26%276. BERENCSI, K., GONCZOL, E., VALYI-NAGY,T., URI, A., RANDO, R., DETAISNE, C. & PLOT~ZlN, S.A. (1991). Recombinant human cytomegalovirus (HCMV)-gB glycoprotein induces gB-specific cellmediated immunity in mice. In Proceedings of the 3rd International Cytomegalovirus Workshop, Bologna, pp. 191-194. Edited by M. P. Landini. Netherlands: Elsevier. BORYSIEWICZ,L., HICKLING,J., GRAHAM,S., SINCLAIR,J., CRANAGE, M., SMITH,G. & SISSONS,J. (1988). Human cytomegalovirus specific cytotoxic T cells- relative frequency of stage specific CTL recognizing the 72 kDa immediate early protein and glycoprotein B expressed by recombinant vaccinia viruses. Journal of Experimental Medicine 168, 919-931. BRITT, W.J., VUGLER, L. & STEPHENS, E.B. (1988). Induction of complement dependent and independent neutralizing antibodies by recombinant-derived human cytomegalovirus gp 55-116 (gB). Journal of Virology 62, 3309-3318. BRIT, W. J., VUGLER, L., BUTFILOSKI,E. J. & STEPHENS,E. B. (1990). Cell surface expression of human cytomegalovirus (HCMV) gp55116 (gB): use of HCMV-recombinant vaccinia virus-infected cells in analysis of the human neutralizing antibody response. Journal of Virology 64, 1079-1085. BROWNE, H., SMITH, G., BECK, S. & MINSON, T. (1990). A complex between the MHC class I homologue encoded by human cytomegalovirus and t2 microglobulin. Nature, London 347, 770-772. COUPAR, E. B. H., ANDREW,M. E., BOTH, G. W. & BOYLE,V. (1986). Temporal regulation of influenza hemagglutinin expression in vaccinia virus recombinants and effects on the immune response. European Journal of Immunology 16, 147%1487. CRANAGE, M.P., KOUZARIDES, T., BANKIER, A., SATCHWELL,S., WESTON,K., TOMLINSON,P. & BARRELL,B. (1986). Identification of the human cytomegalovirus glycoprotein B gene and induction of neutralizing antibodies via its expression in recombinant vaccinia virus. EMBO Journal 5, 3057 3063. DEL VAL, M., SCHLICHT, H.-J., VOLKMER, H., MESSERLE, M., REDDEHASE, M.J. & KOSZlNOWSKI,U. (1991). Protection against lethal cytomegalovirus infection by a recombinant vaccine containing a single nonameric T-cell epitope. Journal of Virology 65, 3641-3646. GIBSON, M., TIENSIWAKUL,P. & KHOOBYARIAN,V. (1982). Adenovirus fiber protein (FP) functions as a mitogen and an adjuvant. Cellular Immunology 73, 397-403. GLENN, J. (1981). Cytomegalovirus infections following renal transplantation. Reviews of Infectious Diseases 3, 1151 1178. G6NCZ6L, E., FURLINI,G., IANACONE,J. & PLOTKIN, S. A. (1986a). A rapid microneutralization assay for cytomegalovirus. Journal of Virological Methods 14, 37-41. GONCZ6L, E., HUDECZ, F., IANACONE,J., DIETZSCHOLD,B., STARR,S. & PLOTK1N, S.A. (1986b). Immune responses to isolated human cytomegalovirus envelope proteins. Journal of Virology 58, 661 664. G6NCZ6L, E., IANACONE,J., HO, W., STARk, S., MEIGNIER, B. & PLOTKIN, S.A. (1990). Isolated gA/gB glycoprotein complex of human cytomegalovirus envelope induces humoral and cellular immune-responses in human volunteers. Vaccine 8, 130-136. GONCZ6L, E., DETAISNE, C., HIRKA, G., BERENCSI, K., LIN, W., PAOLETTi, E. & PLOTKIN, S.A. (1991). High expression of human cytomegalovirus (HCMV)-gB protein in cells infected with a vaccinia-gB recombinant: the importance of the gB protein in HCMV immunity. Vaccine 9, 631-637. HOSMALIN, A., CLERICI, M., HOUGHTEN, R., PENDLETON, C.D., FLEXNER, C., LUCEY, D. R., MOSS, B., GERMAIN, R. N., SHEARER, G. M. & BERZOFSKY,J. A. (1990). An epitope in human immunodeficiency virus 1 reverse transcriptase recognized by both mouse and human cytotoxic T lymphocytes. Proceedings of the National Academy of Sciences, U.S.A. 87, 2344-2348. LIu, Y.-N. C., KLAUS, A., KARI, B., STINSKI, M. F., ECKHARDT,J. &

2512

Short communication

GEHRZ, R.C. (1991). The N-terminal 513 amino acids of the envelope glycoprotein gB of human cytomegalovirus stimulate both B- and T-cell immune responses in humans. Journal of Virology 65, 1644-1648. MARSHALL, G. S., RICCIARDI, R.P., RANDO, R., PUCK, J., GE, R., PLOTKIN, S. A. &: G6NCZOL, E. (1990). An adenovirus recombinant that expresses the human cytomegalovirus major envelope glycoprotein and induces neutralizing antibodies. Journal of Infectious Diseases 162, 1177-1181. MASUHO, Y., MATSUMOTO, Y.-I., SUGANO, T., FUJINAGA, S. & MINAMISmMA,Y. (1987). Human monoclonal antibodies neutralizing human cytomegalovirus. Journal of General Virology 68, 1457-1461. RAWLE, F. C., TOLLEFSON,A. E., WOLD, W. S. M. & GOODING, L. R. (1989). Mouse anti-adenovirus cytotoxic T lymphocytes. Inhibition of lysis by E3 gpl9K but not E3 17.7K. Journal of Immunology 143, 2031~037. RIDDELL,S. R., RABIN,M., GEBALLE,A. P., BRITT,W. J. & GREENBERG, P.D. (1991). Class I MHC-restricted cytotoxic T lymphocyte recognition of cells infected with human cytomegalovirus does not require endogenous viral gene expression. Journal of Immunology 146, 2797-2803.

TANAKA, Y., ISOBE, A., MASUDA,M., TOZAWA, H., KOYANAGY,Y., YAMAMOTO,N. 8z SHIDA,H. (1991). Immunogenicity of human T cell leukemia virus type-I (HTLV-I) antigens for cytotoxic T lymphocytes in the rat system. Journal of Immunology 147, 3646-3652. TARTAGLIA, J., PINCUS, S. & PAOLETTI, E. (1990). Poxvirus-based vectors as vaccine candidates. Critical Reviews in Immunology 10, 13 30. TOWNSEND,A., BASTIN,J., GOULD, K., BROWNLEE,G., ANDREW,M., COUPAR, B., BOYLE, D., CHAN, S. & SMITH, G. (1988). Defective presentation to class I-restricted cytotoxic T lymphocytes in vacciniainfected ceils is overcome by enhanced degradation of antigen. Journal of Experimental Medicine 168, 1211-1224. WELLER, T. H. (1971). The cytomegalovirus: ubiquitous agents with protean clinical manifestations. New England Journal of Medicine 285, 203-214. ZHou, J., MCINDOE, A., DAVIES, H., SUN, X.-Y. & CRAWFORD, L. (1991). The induction of cytotoxic T-lymphocyte precursor cells by recombinant vaccinia virus expressing human papillomavirus type 16 L1. Virology 181, 203 210.

(Received 1 June 1993; Accepted 20 July 1993)