835
Antibody Response to Human Cytomegalovirus Glycoproteins gB and gH after Natural Infection in Humans Lucy Rasmussen, Chantal Matkin, Richard Spaete, Carol Pachl, and Thomas C. Merigan
Division ofInfectious Diseases. Department of Medicine. Stanford University School of Medicine. Stanford. California; Chiron Corporation. Emeryville. California
Humoral immunity may be important in protection from progressive infection with human cytomegalovirus (CMV), especially in the immunocompromised host [1-4], but our understanding of the CMV proteins that may induce protective immunity is incomplete. Many studies have shown that human sera can react physically with a broad spectrum of CMV proteins from either purified virus or virus-infected cells [5, 6]. However, methods for identification ofthe immunologically important antigens that rely on the detection of proteins of specific molecular sizes from virus-infected cells using antibody-positive human sera are difficult to evaluate. The protein composition of CMV is complex. For example, the polypeptides ofCMV may undergo posttranslational modification or exist as complexes of precursors and products. These modified primary gene products may have antibody-binding properties or epitopes that differ from that of the primary gene product. In addition, some proteins may comigrate on SDS-PAGE gels but have different immunologic potential; for example, the major capsid protein and the basic phosphoprotein are both 150 kDa [7]. The basic phosphoprotein, which is highly antigenic [8], may obscure immunologic reactivity ofantibody with the less immunogenic major capsid protein. The use of immunoaffinity-purified viral proteins or well-characterized proteins produced by recornbi-
Received 19 April 1991; revised 8 July 1991. Written informed consent was obtained from participants: guidelines for human experimentation of the US Department of Health and Human Services were followed; protocols were approved by local institutional review boards. Grant support: National Institutes of Health (AI-05629). Reprints or correspondence: Dr. Lucy Rasmussen, Stanford University School of Medicine, Division of Infectious Diseases, Stanford. CA 94305. The Journal oflnfectious Diseases 1991;164:835-42 © 1991 by The University of Chicago. All rights reserved. 0022-1899/91/6405-000 I$0 1.00
nant DNA technologies should simplify the identification of the immunologically significant viral proteins. The glycoproteins of herpesviruses may be targets for virus-neutralizing antibody [9, 10], mediate viral entry [11, 12], and playa role in the release of the virus from the host cell [13]. They are logical candidates for inclusion in subunit vaccines. Five glycoproteins have so far been detected on the envelope ofCMV. The gB complex, consisting ofa 130-kDa precursor and a 55-kDa product [14-17], the gH homologue, also described as the 86/ 145-kDa complex [18-21], and the 50- to 200-kDa glycoprotein complex II family [22] are being intensively studied. These three glycoproteins are all known to be targets for virus-neutralizing antibody. The other two envelope glycoproteins are a 45-kDa integral membrane glycoprotein [23] and a 48-kDa glycoprotein [24]. Whether these induce virus-neutralizing antibody is unknown. An understanding of the natural history of the immune response to the specific glycoproteins will help to provide a basis for developing new immunotherapeutic inventions for the active or passive prophylaxis of CMV infection. In our experiments, we studied the antibody response to both the un cleaved 110-kDa precursor of the recombinant gB (rgB) family and the 84-kDa recombinant gH (rgH) in different patient groups at variable risk for CMV infection in an effort to understand the nature of the humoral responses to these potentially important antigens.
Materials and Methods Virus. CMV, strain AD169, was used for all experiments. Cells. Human embryonic lung (HEL) cells, grown in MEM
supplemented with 10% fetal calf serum (FCS), were used for the culture and quantitation of CMV [25]. The generation of the rgB producing cell line (67.77) by
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
Antibody to the recombinant gB (rgB) and recombinant gH (rgH) glycoproteins of human cytomegalovirus (CMV) was studied in immunocompetent and immunocompromised humans by immunoprecipitating [35S]methionine-labeled CHO cell lines stably expressing rgB and rgH. Antibody to the rgB precursor was present in >60% of immunocompetent individuals. However rgH antibody was detected in 95% of sera studied, a response to the precursor paralleled
the response to the cleavage products (data not shown). Of sera from immunocompetent seropositive individuals with antibody to CMV, --60% had detectable antibody to gB. In contrast, sera from only one of the nine individuals studied by Western blot had readily detectable levels of antibody to the gH. Because of concerns that the denaturing conditions that are inherent in the immunoblotting procedure might destroy crucial conformational epitopes necessary for binding of human sera to the glycoproteins, the same sera were evaluated by immunoprecipitation of rgB and rgH from CHO cells. The immunoprecipitation assay is done without antigen denaturation and does not perturb conformational epitopes. The data shown in figure 2 (lanes rgB and rgH) represent bands immunoprecipitated by human sera. A ratio of~5% of the intensity of the bands immunoprecipitated by human sera to the bands immunoprecipitated by the MAb control was considered a positive response. The mean (±2 SE) intensity of the reaction, as shown by the bars, was not signifi-
250
200 -
o o
150 0 0
100 -
50 -
I 0
o
0 ..0.
gB
0
0 0
o
0
to 0
.&.
rgB
taB gH
rgH
Figure 2. Comparison of reactivity of sera from immunocompetent cytomegalovirus (CMV)-seropositive individuals with immunoblotted glycoproteinsgBand gH and immunoprecipitated recombinant (r) gB and gH. Reactivity of sera with gB and gH was determined by immunoblot analysis of strips with immunoaffinityproduced glycoprotein isolated from virus-infected cells. Positive control sera were from guinea pigs that were hyperimmune to gB and gH. IgG binding was detected with 1251-labeled protein A and visualized by autoradiography. rgH and rgB were immunoprecipitated from [35S]methionine-labeled CHO cell lines that stably express the designated CMV glycoprotein gene. Positive controls were monoclonal antibodies to gB and gH used at concentrations of 5 J,tg. After SDS-PAGE, immunoprecipitated bands were visualized by autoradiography. Binding ofCMV IgG in human sera was quantified by densitometry of auto radiograph and expressedas percentage of the appropriate control antibody. Bar at left of each group shows mean (±2 SE).
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
LA were provided by the Diagnostic Virology Laboratory, Stanford University Hospital. ISG lots containing 5% gamma globulin were provided by commercial manufacturers.
837
838
Rasmussen et al.
200
rgB o 150
•
symptomatic CMV asymptomatic CMV
100
~
o
~
50
~
Z
o u
0
~
o
~
200
Z
rgH
~
u
~ ~
P...,
150
100
50
o -3 -2 -1
0
1
2
3
4
5
6
>
7
MONTHS Figure 3. Antibody reactivity to cytomegalovirus (CMV) recombinant (r) glycoproteins gB and gH in immunocompetent individuals with symptomatic and asymptomatic CMV infection. Serum samples were studied at various times after infection. For patients with symptomatic infection, time 0 represents period of maximum symptomatology. For asymptomatic seroconverters, 0 point represents time of acquisition of last serum sample before detection of seroconversion. Reactivity of sera with rgB and rgH are expressed as percentage ofband immunoprecipitated by appropriate monoclonal antibody from rSS]methionine-labeled CHO cells stably expressing either rgB or rgH CMV glycoprotein gene.
in virus-neutralizing or LA antibody and by the presence of CMV-specific IgM. Sera from five of eight patients had increases in rgB antibody compared with pretransplant levels. Sera from two of three with no increase in rgB antibody also failed to show increases in virus-neutralizing antibody but
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
candy different in comparisons of gB and gH from virus-infected cells with rgB and rgH. Also, sera from the same individuals who were reactive with gB and gH were reactive with rgB and rgH. Because of the relative ease of the immunoprecipitation assay compared with the immunoaffinity purification of gB and gH from large quantities of virus-infected cells that was required for the immunoblot assay, sera from the patient populations in our study were subsequently tested by immunoprecipitation. Detection ofantibody to rgB and rgH in sera from immunocompetent patients with primary CMV infections. To determine whether an antibody response to the CMV glycoproteins could be detected after acute CMV infection, three normal adults with serologically confirmed CMV disease were studied during the symptomatic (0 time) and convalescent (1-6 months) stages of infection. All three had CMVspecific IgM responses during the symptomatic phase of illness. Acute-phase samples were negative « 1:8 for virusneutralizing antibody), but convalescent serum samples had peak titers of 1:90-1 :450. In these individuals (figure 3, open circles), responses to both rgB and rgH were detectable in the convalescent period. Sera from three individuals (figure 3, closed circles) who converted to CMV seropositivity without symptoms were also studied. In this group preconversion sera had neutralizing titers of < 1:8. Postseroconversion peak titers were 1:50-1: 150. The intensity of the response to the glycoproteins during the postseroconversion period was considerably less than that of sera from the patients with symptomatic disease studied at similar times. Detection of antibody to rgB and rgH in immunocompromised patients. The effect of CMV infection on antibody responses to rgB and rgH was studied in sera from heart and heart-and-lung transplant patients. All had CMV infection documented by seroconversion for primary disease or by a positive IgM response for recurrent infection. None of the patients had received antiviral therapy or passive immunotherapy with ISG before or during the course of their CMV infection. The antibody response to the rgB glycoprotein in sera from patients with either primary or recurrent CMV infection is shown in figure 4. Sera from one of the five patients with primary infection (figure 4A) was slightly reactive with rgB before transplant. The reason for this is unknown, since this patient had neither virus-neutralizing nor LA antibody. One of the five patients had a dramatic increase in rgB antibody between 2 and 4 months, while one had only a moderate increase (0-17%) at 4 months after transplant. Two ofthe three patients tested after 4 months developed significant antibody responses late (~8 months) in the posttransplant period; however, one of five had only moderate antibody (20% of control) in the late posttransplant period. Most patients (six of eight) with recurrent CMV infection (figure 4B) had detectable levels of antibody to rgB before transplant. Only modest increases in rgB antibody were detected after reexposure to the virus, as diagnosed by increases
lID 1991; 164 (November)
1ID 1991;164 (November)
Human Antibody to Cytomegalovirus gB and gH
250
A
rgB
200
150
839
dergoing symptomatic primary CMV infection (figure 5) was slightly depressed in the posttransplant period compared with normal immunocompetent individuals with symptomatic primary CMV infection (figure 3), but all five eventually developed detectable antibody. In contrast to rgB, rgH antibody was detectable before transplant in only one of eight patients with recurrent CMV infection. In those with recurrent CMV, moderate but reproducible increases in rgH anti-
100
250
A 50
rgH
o
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
200
150
250 . . . . . . - - - - - - - - - - - - - - - - - - - - , 100
B 200 50
o
~
150
E-
Z
o
o 250
100
B 50
o
z
200
c,
150
-2
MONTHS Figure 4. Antibody to cytomegalovirus (CMV) recombinant (r) glycoprotein gB in heart and heart-and-lung transplant patients with primary (A) and recurrent (B) CMV infections. Sera were obtained at times shown and used to immunoprecipitate rgB from 35 [ S]methionine-labeled CHO cell line 67.77. Immunoprecipitated IIO-kDa rgB molecule was analyzed by SDS-PAGEandautoradiography. Intensity of bands was quantified by densitometry. Antibody reactivity is expressed as percentage of anti-gB monoclonal antibody control.
did develop a rise in titer for LA antibody. Therefore, in these patients some aspects of the humoral antibody response may be sufficiently suppressed to allow an antibody response only to the most highly immunogenic CMV antigens. Antibody to the rgH in the sera from the five patients un-
100
50
o -2
o
2
4
6
>8
MONTHS Figure 5. Antibody to cytomegalovirus (CMV) recombinant (r) glycoprotein gH in cardiac and heart-and-lung transplant patients with primary (A) and recurrent (B) CMV infections. Sera were obtained at times shown and used to immunoprecipitate rgH from 35 [ S]methio nine-labeled CHO cell line 171. lmmunoprecipitated 84-kDa molecule was visualized by autoradiography after SDSPAGE, and intensity of bands was quantitated by densitometry. Antibody reactivity is expressed as percentage ofanti-gH monoclonal antibody control.
Rasmussen et al.
840
Table 1.
Distribution of antibody to human cytomegalovirus (CMV) recombinant (r) glycoproteins gB and gH in different lots of immune serum globulin (ISG) from manufacturers I and II. % of control antibody* (mean ± 2 SO) ISG
No.
rgB
rgH
3
26 ± 36 48 ± 34
5±O 32 ± 34 5±4
8 6
13 ± 3
* (Densitometry value of immunoprecipitate with ISG/densitometry value of immunoprecipitate with monoclonal antibody) X 100. t ISG lots prepared from donors of unknown CMV serostatus. t ISG lots prepared from donors seropositive for CMV.
the extent of variability from lot to lot within the group the differences were not significant.
Discussion The data in this study show that antibodies to the rgB and rgH envelope glycoproteins ofCMV are detected with different frequencies in individuals who are seropositive to the virus. Antibody to rgB was observed regularly, but antibody to rgH in immunocompetent normal individuals with no history of recent symptomatic infection was rarely detected. Several possibilities may account for this difference. First, the response to each of the glycoproteins may be influenced by the amount of antigen that is present on the immunogen. For example, differences in immunogenicity of the hemagglutinin and neuraminidase of influenza after vaccination have been attributed to the greater concentration of hemagglutinin on the virion surface [29]. The relative quantities of gB and gH on the viral envelope are not known. However using MAbs to the gB glycoprotein that are conjugated to colloidal gold, the viral envelopes are heavily labeled when examined by electron microscopy [30]. Thus, the immunodominance of gB may reflect its heavy concentration on the viral envelope. When virus-infected cells are immunoprecipitated by sera from infants with clinically apparent compared with subclinical infection [31], the proteins of the gB complex are easily detected. This may reflect the level of antigenic stimulus in the children with symptomatic disease who are likely to have a greater viral load than those with asymptomatic infection. Second, it is possible that antibody to rgB may be maintained longer after initial infection than antibody to rgH. After acute infection in immunocompetent individuals, antibody increases to both of the glycoproteins were readily detected. However, < 10% of seropositive patients distant in time from any history of acute CMV-associated symptoms had antibody to rgH, indicating that after seroconversion antibody to rgH may wane with time without reexposure to the virus. A third possibility is that some individuals who seroconvert to CMV without symptoms will fail to produce antibody to rgH. The failure ofthree asymptomatic seroconverters (figure 3) to develop antibody responses of the same magnitude as the symptomatic patients supports this idea. Fourth, the possibility of intrinsic differences in immunogenicity ofrgB and rgH is unlikely. Titers ofvirus-neutralizing antibody are comparable when equivalent amounts of immunoaffinity-isolated gB and gH are used to immunize guinea pigs [27]. Fifth, there may be fundamental differences in sensitivity of our assay for detecting rgB and rgH. The fact that the immunoprecipitation assay was slightly less sensitive for rgB than rgH (five times more anti-gB than -gH MAb was needed to give detectable bands) argues against this possibility.
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
body were detected throughout the posttransplant period. There was a trend toward higher responses in the late transplant period in this group compared with those in patients with primary infection. The elevated responses appeared earlier in patients with recurrent infection than in those with primary CMV. It is of interest that one of the three patients with recurrent CMV infection who failed to show increases in rgB antibody showed a significant increase in antibody to rgH (rising from 0 to 200% of control). Distribution ofantibody to gB and gH in CMV-seropositive leukemic patients. Sera from a group of 18 leukemic patients who were candidates for bone marrow allografts were tested for antibody to rgB and rgH. All were seropositive for CMV with LA titers> 1:8. For rgB, sera from 14 of 18 patients were reactive. However, only one serum sample reacted with the rgH glycoprotein. Lack of relationship between intensity of reaction of sera with rgB or rgH and virus-neutralizing or LA antibody. Seventy serum samples with virus-neutralizing titers> 1:8 were evaluated for a relationship between intensity of the reaction with either rgB or rgH and virus-neutralizing activity by linear regression analysis. No significant relationship was established between virus-neutralizing antibody titer and intensity of reaction with either rgB (r = .32253) or rgH (r = .10069). Serum samples from 17 normal donors with LA antibody titers of 16-1024 were also analyzed by linear regression for a correlation with intensity of reaction with rgB or rgH. No relationship of significance was observed (LA antibody vs. rgB, r = .08306; LA antibody vs. rgH, r = .007875). Antibody to rgB and rgH in ISG preparations. Antibody to rgB and rgH in lots of ISG prepared from unselected donors, represented by groups I and II from two different manufacturers, is shown in table 1. The mean ± 2 SD for both antibody to rgB and rgH in lots ofISG prepared from donors who were of unknown CMV serostatus was comparable to that of the normal seropositive population as shown in figure 2. Antibody to both the rgB and rgH tended to be higher in lots of ISG from hyperimmune donors (table 1, Ib), but because of
110 1991;164 (November)
110 1991; 164 (November)
Human Antibody to Cytomegalovirus gB and gH
ciencies in CMV glycoprotein antibodies may provide information about their role in protective immunity.
Acknowledgments
We thank Linda Logan (Diagnostic Laboratory, Stanford University Hospital) for obtaining serum samples from transplant patients, Marcia Smullen for virus-neutralization assays, and participants from the Day Care Center for Children ofSyntex Employees (Palo Alto, CA), for their cooperation.
References l. Meyers J, Flourney N, Thomas J. Risk factors for cytomegalovirus infection after human marrow transplantation. J Infect Dis 1986; 153:478-8. 2. Pass R, Griffiths P, August A. Antibody response to cytomegalovirus after renal transplantation: comparison of patients with primary and recurrent infections. J Infect Dis 1983; 147:40-6. 3. Yeager A, Grumet F, Halfleigh E, Arvin A, Bradley J, Prober C. Prevention of transfusion-acquired cytomegalovirus infections in newborn infants. J Pediatr 1983;98:281-7. 4. Yow M. Congenital cytomegalovirus disease: a NOW problem. J Infect Dis 1989; 159: 163-7. 5. Landini M, Michelson S. Human cytomegalovirus proteins. Prog Med Virol 1988;35: 152-85. 6. Rasmussen L. Immune response to human cytomegalovirus infection. Curr Top Microbiol ImmunoI1990;154:221-54. 7. Roby C, Gibson W. Characterization of phosphoproteins and protein kinase activity of virions, noninfectious enveloped particles, and dense bodies of human cytomegalovirus. J Virol 1986;59:714-27. 8. Jahn G, Scholl B, Traupe B, Fleckenstein B. The two major structural phosphoproteins (pp65 and pp 150) of human cytomegalovirus and their antigenic properties. J Gen Virol 1987;68: 1327-37. 9. Eing B, Juhn J, Braun R. Neutralizing activity of antibodies against the major herpes simplex virus type I glycoproteins. J Med Virol 1989;27:59-65. 10. Spear P. Glycoproteins specified by herpes simplex viruses. In: Roizman B, ed. The herpesviruses. Vol 3. New York: Plenum Press, 1984:315-56. I I. Fuller A, Santos R, Spear P. Neutralizing antibodies specific for glycoprotein H of herpes simplex virus permit viral attachment to cells but prevent penetration. J Virol 1989;63: 3435-43. 12. Fuller A, Spear P. Anti-glycoprotein D antibodies that permit adsorption but block infection by herpes simplex virus I prevent virion-cell fusion at the cell surface. Proc Nat! Acad Sci USA 1987;84:5454-8. 13. Gompels U, Minson A. The properties and sequence of gH of herpes simplex virus type 1. Virology 1986;153:230-47. 14. Britt W. Neutralizing antibodies detect a disulfide-linked glycoprotein complex within the envelope of human cytomegalovirus. Virology 1984; 135:369-78. 15. Pereira L, Hoffman M, Tatsuno M, Dondero D. Polymorphism of human cytomegalovirus glycoproteins characterized by monoclonal antibodies. Virology 1984; 139:73-86. 16. Rasmussen L, Mullenax J, Nelson R, Merigan T. Viral polypeptides detected by a complement dependent neutralizing murine monoclonal antibody to human cytomegalovirus. J Virol 1985;55:274-80. 17. Cranage M, Kouzarides T, Bankier A, et al. Identification of the human cytomegalovirus glycoprotein B gene and induction of neutralizing antibodies via its expression in recombinant vaccinia virus. EMBO J 1986;5:3057-63. 18. Rasmussen L, Nelson R, Kelsall D, Merigan T. Murine monoclonal
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
Finally, we cannot eliminate the possibility that there may be differences in the ability of the rgH and gH from virus-infected cells to react with human antibody. This, as a reason for lack of regular detection of gH antibody in most human sera, is made less likely by two findings. The patient sera reactive with gH were also reactive with rgH. Also, in patients with CMV mononucleosis, it was not possible to detect an anti-gH response in the convalescent phase with gH antigen (data not shown), but positive responses were detected with rgH (figure 2). This suggests that immunoprecipitation assays for rgH are more sensitive than immunoblot assays using gH for detecting human antibody. A final resolution of this question must await epitope mapping of gH for identification of the sites that are immunogenic for humans. Normal individuals with antibody to rgH may be those who have had extensive exposure to CMV, perhaps through multiple episodes of asymptomatic reactivation. This idea is supported by the finding that the antibody response to rgH was higher in recurrent CMV than in primary infections in the heart transplant population. The replicating virus probably provides sufficient antigen to elicit an rgH response. In serologic studies in children infected with the paramyxovirus parainfluenza virus type 3, the antibody response to the F (fusion) glycoprotein appeared only after repeated infections, whereas antibodies to the hemagglutinin/neuraminidase glycoprotein were detected after the initial infection [32]. This demonstrates that in another viral disease, antibody to at least two glycoproteins can be induced independently. The gB has been shown to be an important inducer of virus-neutralizing antibody in some patients [33, 34]. However, many factors may contribute to the neutralizing antibody response, for example, synergy between [35] or blocking by nonneutralizing antibodies [36]. Viral strain differences may influence the detection of virus-neutralizing antibody. The gene products of all of the open-reading frames for CMV glycoproteins [37] have yet to be identified immunologically. A study of the importance of the glycoproteins should be done using only epitopes known to bind virus-neutralizing antibodies. This would exclude interactions of many other regions on the molecule with nonneutralizing antibodies. There is conflicting evidence for the protective value of passively transferred ISO against CMV-associated interstitial pneumonitis in bone marrow transplant recipients, as recently reviewed by Meyers [38]. A prospective investigation of the clinical efficacy oflSO preparations with known levels ofantibody to the CMV glycoproteins may be useful for evaluating their importance in protection from CMV disease in recipients who are deficient in antibody to either glycoprotein. Progress has also been made in developing human MAbs to the CMV glycoproteins [39]. The evaluation of the efficacy of these MAbs in patients selected for specific defi-
841
842
19.
20.
21.
22.
24.
25.
26.
27.
28.
antibody to a single protein neutralizes the infectivity of human cytomegalovirus. Proc Natl Acad Sci USA 1984;81:876-80. Gretch D, Kari B. Rasmussen L, Gehrz R, Stinski M. Identification and characterization of three distinct families of glycoprotein complexes in the envelopes of human cytomegalovirus. J Virol 1988;62:87581. Cranage M, Smith G, Bell S, et aJ. Identification and expression of a human cytomegalovirus glycoprotein with homology to the EpsteinBarr virus BXLFZ product, varicella zoster gp III and herpes simplex type I glycoprotein H. J Virol 1988;62: 1416-22. Pachl C. Probert W, Hermsen K, et al. The human cytomegalovirus strain Towne glycoprotein gene encodes glycoprotein p86. Virology 1988; 169:418-26. Kari B, Lussenhop N, Goertz R, Wabuke-Bunoti M, Radeke R, Gehrz R. Characterization of monoclonal antibodies reactive to several biochemically distinct human cytomegalovirus glycoprotein complexes. J Virol 1986;60:345-52. Lehner R, Meyer H, Mach M. Identification and characterization of a human cytomegalovirus gene coding for a membrane protein that is conserved among human herpesviruses. J Virol 1989;63:37923800. Chang C. Vesole D, Nelson J, Oldstone M, Stinski M. Identification and expression of a human cytomegalovirus early glycoprotein. J Virol 1989;63:3330-7. Rasmussen L, Kelsall D, Nelson R, et al. Virus specific IgG and IgM antibodies in normal and immunocompromised subjects infected with cytomegalovirus. J Infect Dis 1982;145:191-9. Spaete R, Thayer R, Probert W, et al. Human cytomegalovirus strain Towne glycoprotein B is processed by proteolytic cleavage. Virology 1988; 167:207-25. Rasmussen L, Mullenax J, Nelson R, Merigan T. Human cytomegalovirus polypeptides stimulate neutralizing antibody in vivo. Virology 1985; 145: 186-90. Spaete R, Saxena A, Scott P, et al. Sequence requirements for proteolytic processing of glycoprotein B of human cytomegalovirus strain Towne. J Virol 1990;64:2922-31.
110 1991;164 (November)
29. Johansson B, Bucher D, Kilbourne E. Purified influenza virus hemagglutinin and neuraminidase are equivalent in stimulation ofantibody response but induce contrasting types of immunity to infection. J Virol 1989;63: 1239-46. 30. Stannard L, Rider J, Farrar G. Morphology and distribution ofgp52 on extracellular human cytomegalovirus (HCMV) supports biochemical evidence that it represents the HCMV glycoprotein B. J Gen Virol 1989;70: 1553-60. 31. Britt W. Vugler L. Antiviral antibody responses in mothers and their newborn infants with clinical and subclinical cytomegalovirus infections. J Infect Dis 1990; I 61 :214-9. 32. Kasel J, Frank A. Keitel W, Taber L. Glezen W. Acquisition of serum antibodies to specific viral glycoproteins of parainfluenza virus 3 in children. J Virol 1984;52:828-32. 33. Britt W, Vugler L, Butfilowski E, Stephens E. Cell surface expression of human cytomegalovirus (HCMV) gp55-1 16 (gB): use ofHCMV-vaccinia recombinant virus-infected cells in analysis of the human neutralizing antibody response. J Virol 1990;64: 1079-85. 34. Liu Y, Klaue A. Kari B. et al. The N-terminal 513 amino acids of the envelope glycoprotein gB of human cytomegalovirus stimulates both B- and T-cell immune responses in humans. J Virol 1991; 65: 1644-8. 35. Lussenhop N, Goertz R, Wabuke-Bunoti M, Gehrz R, Kari B. Epitope analysis of human cytomegalovirus glycoprotein complexes using murine monoclonal antibodies. Virology 1988; 164:362-72. 36. Utz .U, Britt W. Vugler L, Mach M. Identification of a neutralizing epitope on glycoprotein 58 of human cytomegalovirus. J Virol 1989;63: 1995-200 I. 37. Chee M, Bankier A, Beck S, et al. Analysis of the protein-coding content of the sequence of human cytomegalovirus strain AD 169. Curr Top Microbiol Immunol 1990; 154: 125-69. 38. Meyers J. Prevention of cytomegalovirus infection after marrow transplantation. Rev Infect Dis 1989;II(suppI7):SI691-1705. 39. Matsumoto Y, Sugano T, Miyamoto C. Masuho Y. Generation of hybridomas producing human monoclonal antibodies against human cytomegalovirus. Biochem Biophys Res Commun 1986; 137:27380.
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
23.
Rasmussen et al.
Antibody Response to Human Cytomegalovirus Glycoproteins gB and gH after Natural Infection in Humans Lucy Rasmussen, Chantal Matkin, Richard Spaete, Carol Pachl, and Thomas C. Merigan
Division ofInfectious Diseases. Department of Medicine. Stanford University School of Medicine. Stanford. California; Chiron Corporation. Emeryville. California
Humoral immunity may be important in protection from progressive infection with human cytomegalovirus (CMV), especially in the immunocompromised host [1-4], but our understanding of the CMV proteins that may induce protective immunity is incomplete. Many studies have shown that human sera can react physically with a broad spectrum of CMV proteins from either purified virus or virus-infected cells [5, 6]. However, methods for identification ofthe immunologically important antigens that rely on the detection of proteins of specific molecular sizes from virus-infected cells using antibody-positive human sera are difficult to evaluate. The protein composition of CMV is complex. For example, the polypeptides ofCMV may undergo posttranslational modification or exist as complexes of precursors and products. These modified primary gene products may have antibody-binding properties or epitopes that differ from that of the primary gene product. In addition, some proteins may comigrate on SDS-PAGE gels but have different immunologic potential; for example, the major capsid protein and the basic phosphoprotein are both 150 kDa [7]. The basic phosphoprotein, which is highly antigenic [8], may obscure immunologic reactivity ofantibody with the less immunogenic major capsid protein. The use of immunoaffinity-purified viral proteins or well-characterized proteins produced by recornbi-
Received 19 April 1991; revised 8 July 1991. Written informed consent was obtained from participants: guidelines for human experimentation of the US Department of Health and Human Services were followed; protocols were approved by local institutional review boards. Grant support: National Institutes of Health (AI-05629). Reprints or correspondence: Dr. Lucy Rasmussen, Stanford University School of Medicine, Division of Infectious Diseases, Stanford. CA 94305. The Journal oflnfectious Diseases 1991;164:835-42 © 1991 by The University of Chicago. All rights reserved. 0022-1899/91/6405-000 I$0 1.00
nant DNA technologies should simplify the identification of the immunologically significant viral proteins. The glycoproteins of herpesviruses may be targets for virus-neutralizing antibody [9, 10], mediate viral entry [11, 12], and playa role in the release of the virus from the host cell [13]. They are logical candidates for inclusion in subunit vaccines. Five glycoproteins have so far been detected on the envelope ofCMV. The gB complex, consisting ofa 130-kDa precursor and a 55-kDa product [14-17], the gH homologue, also described as the 86/ 145-kDa complex [18-21], and the 50- to 200-kDa glycoprotein complex II family [22] are being intensively studied. These three glycoproteins are all known to be targets for virus-neutralizing antibody. The other two envelope glycoproteins are a 45-kDa integral membrane glycoprotein [23] and a 48-kDa glycoprotein [24]. Whether these induce virus-neutralizing antibody is unknown. An understanding of the natural history of the immune response to the specific glycoproteins will help to provide a basis for developing new immunotherapeutic inventions for the active or passive prophylaxis of CMV infection. In our experiments, we studied the antibody response to both the un cleaved 110-kDa precursor of the recombinant gB (rgB) family and the 84-kDa recombinant gH (rgH) in different patient groups at variable risk for CMV infection in an effort to understand the nature of the humoral responses to these potentially important antigens.
Materials and Methods Virus. CMV, strain AD169, was used for all experiments. Cells. Human embryonic lung (HEL) cells, grown in MEM
supplemented with 10% fetal calf serum (FCS), were used for the culture and quantitation of CMV [25]. The generation of the rgB producing cell line (67.77) by
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
Antibody to the recombinant gB (rgB) and recombinant gH (rgH) glycoproteins of human cytomegalovirus (CMV) was studied in immunocompetent and immunocompromised humans by immunoprecipitating [35S]methionine-labeled CHO cell lines stably expressing rgB and rgH. Antibody to the rgB precursor was present in >60% of immunocompetent individuals. However rgH antibody was detected in 95% of sera studied, a response to the precursor paralleled
the response to the cleavage products (data not shown). Of sera from immunocompetent seropositive individuals with antibody to CMV, --60% had detectable antibody to gB. In contrast, sera from only one of the nine individuals studied by Western blot had readily detectable levels of antibody to the gH. Because of concerns that the denaturing conditions that are inherent in the immunoblotting procedure might destroy crucial conformational epitopes necessary for binding of human sera to the glycoproteins, the same sera were evaluated by immunoprecipitation of rgB and rgH from CHO cells. The immunoprecipitation assay is done without antigen denaturation and does not perturb conformational epitopes. The data shown in figure 2 (lanes rgB and rgH) represent bands immunoprecipitated by human sera. A ratio of~5% of the intensity of the bands immunoprecipitated by human sera to the bands immunoprecipitated by the MAb control was considered a positive response. The mean (±2 SE) intensity of the reaction, as shown by the bars, was not signifi-
250
200 -
o o
150 0 0
100 -
50 -
I 0
o
0 ..0.
gB
0
0 0
o
0
to 0
.&.
rgB
taB gH
rgH
Figure 2. Comparison of reactivity of sera from immunocompetent cytomegalovirus (CMV)-seropositive individuals with immunoblotted glycoproteinsgBand gH and immunoprecipitated recombinant (r) gB and gH. Reactivity of sera with gB and gH was determined by immunoblot analysis of strips with immunoaffinityproduced glycoprotein isolated from virus-infected cells. Positive control sera were from guinea pigs that were hyperimmune to gB and gH. IgG binding was detected with 1251-labeled protein A and visualized by autoradiography. rgH and rgB were immunoprecipitated from [35S]methionine-labeled CHO cell lines that stably express the designated CMV glycoprotein gene. Positive controls were monoclonal antibodies to gB and gH used at concentrations of 5 J,tg. After SDS-PAGE, immunoprecipitated bands were visualized by autoradiography. Binding ofCMV IgG in human sera was quantified by densitometry of auto radiograph and expressedas percentage of the appropriate control antibody. Bar at left of each group shows mean (±2 SE).
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
LA were provided by the Diagnostic Virology Laboratory, Stanford University Hospital. ISG lots containing 5% gamma globulin were provided by commercial manufacturers.
837
838
Rasmussen et al.
200
rgB o 150
•
symptomatic CMV asymptomatic CMV
100
~
o
~
50
~
Z
o u
0
~
o
~
200
Z
rgH
~
u
~ ~
P...,
150
100
50
o -3 -2 -1
0
1
2
3
4
5
6
>
7
MONTHS Figure 3. Antibody reactivity to cytomegalovirus (CMV) recombinant (r) glycoproteins gB and gH in immunocompetent individuals with symptomatic and asymptomatic CMV infection. Serum samples were studied at various times after infection. For patients with symptomatic infection, time 0 represents period of maximum symptomatology. For asymptomatic seroconverters, 0 point represents time of acquisition of last serum sample before detection of seroconversion. Reactivity of sera with rgB and rgH are expressed as percentage ofband immunoprecipitated by appropriate monoclonal antibody from rSS]methionine-labeled CHO cells stably expressing either rgB or rgH CMV glycoprotein gene.
in virus-neutralizing or LA antibody and by the presence of CMV-specific IgM. Sera from five of eight patients had increases in rgB antibody compared with pretransplant levels. Sera from two of three with no increase in rgB antibody also failed to show increases in virus-neutralizing antibody but
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
candy different in comparisons of gB and gH from virus-infected cells with rgB and rgH. Also, sera from the same individuals who were reactive with gB and gH were reactive with rgB and rgH. Because of the relative ease of the immunoprecipitation assay compared with the immunoaffinity purification of gB and gH from large quantities of virus-infected cells that was required for the immunoblot assay, sera from the patient populations in our study were subsequently tested by immunoprecipitation. Detection ofantibody to rgB and rgH in sera from immunocompetent patients with primary CMV infections. To determine whether an antibody response to the CMV glycoproteins could be detected after acute CMV infection, three normal adults with serologically confirmed CMV disease were studied during the symptomatic (0 time) and convalescent (1-6 months) stages of infection. All three had CMVspecific IgM responses during the symptomatic phase of illness. Acute-phase samples were negative « 1:8 for virusneutralizing antibody), but convalescent serum samples had peak titers of 1:90-1 :450. In these individuals (figure 3, open circles), responses to both rgB and rgH were detectable in the convalescent period. Sera from three individuals (figure 3, closed circles) who converted to CMV seropositivity without symptoms were also studied. In this group preconversion sera had neutralizing titers of < 1:8. Postseroconversion peak titers were 1:50-1: 150. The intensity of the response to the glycoproteins during the postseroconversion period was considerably less than that of sera from the patients with symptomatic disease studied at similar times. Detection of antibody to rgB and rgH in immunocompromised patients. The effect of CMV infection on antibody responses to rgB and rgH was studied in sera from heart and heart-and-lung transplant patients. All had CMV infection documented by seroconversion for primary disease or by a positive IgM response for recurrent infection. None of the patients had received antiviral therapy or passive immunotherapy with ISG before or during the course of their CMV infection. The antibody response to the rgB glycoprotein in sera from patients with either primary or recurrent CMV infection is shown in figure 4. Sera from one of the five patients with primary infection (figure 4A) was slightly reactive with rgB before transplant. The reason for this is unknown, since this patient had neither virus-neutralizing nor LA antibody. One of the five patients had a dramatic increase in rgB antibody between 2 and 4 months, while one had only a moderate increase (0-17%) at 4 months after transplant. Two ofthe three patients tested after 4 months developed significant antibody responses late (~8 months) in the posttransplant period; however, one of five had only moderate antibody (20% of control) in the late posttransplant period. Most patients (six of eight) with recurrent CMV infection (figure 4B) had detectable levels of antibody to rgB before transplant. Only modest increases in rgB antibody were detected after reexposure to the virus, as diagnosed by increases
lID 1991; 164 (November)
1ID 1991;164 (November)
Human Antibody to Cytomegalovirus gB and gH
250
A
rgB
200
150
839
dergoing symptomatic primary CMV infection (figure 5) was slightly depressed in the posttransplant period compared with normal immunocompetent individuals with symptomatic primary CMV infection (figure 3), but all five eventually developed detectable antibody. In contrast to rgB, rgH antibody was detectable before transplant in only one of eight patients with recurrent CMV infection. In those with recurrent CMV, moderate but reproducible increases in rgH anti-
100
250
A 50
rgH
o
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
200
150
250 . . . . . . - - - - - - - - - - - - - - - - - - - - , 100
B 200 50
o
~
150
E-
Z
o
o 250
100
B 50
o
z
200
c,
150
-2
MONTHS Figure 4. Antibody to cytomegalovirus (CMV) recombinant (r) glycoprotein gB in heart and heart-and-lung transplant patients with primary (A) and recurrent (B) CMV infections. Sera were obtained at times shown and used to immunoprecipitate rgB from 35 [ S]methionine-labeled CHO cell line 67.77. Immunoprecipitated IIO-kDa rgB molecule was analyzed by SDS-PAGEandautoradiography. Intensity of bands was quantified by densitometry. Antibody reactivity is expressed as percentage of anti-gB monoclonal antibody control.
did develop a rise in titer for LA antibody. Therefore, in these patients some aspects of the humoral antibody response may be sufficiently suppressed to allow an antibody response only to the most highly immunogenic CMV antigens. Antibody to the rgH in the sera from the five patients un-
100
50
o -2
o
2
4
6
>8
MONTHS Figure 5. Antibody to cytomegalovirus (CMV) recombinant (r) glycoprotein gH in cardiac and heart-and-lung transplant patients with primary (A) and recurrent (B) CMV infections. Sera were obtained at times shown and used to immunoprecipitate rgH from 35 [ S]methio nine-labeled CHO cell line 171. lmmunoprecipitated 84-kDa molecule was visualized by autoradiography after SDSPAGE, and intensity of bands was quantitated by densitometry. Antibody reactivity is expressed as percentage ofanti-gH monoclonal antibody control.
Rasmussen et al.
840
Table 1.
Distribution of antibody to human cytomegalovirus (CMV) recombinant (r) glycoproteins gB and gH in different lots of immune serum globulin (ISG) from manufacturers I and II. % of control antibody* (mean ± 2 SO) ISG
No.
rgB
rgH
3
26 ± 36 48 ± 34
5±O 32 ± 34 5±4
8 6
13 ± 3
* (Densitometry value of immunoprecipitate with ISG/densitometry value of immunoprecipitate with monoclonal antibody) X 100. t ISG lots prepared from donors of unknown CMV serostatus. t ISG lots prepared from donors seropositive for CMV.
the extent of variability from lot to lot within the group the differences were not significant.
Discussion The data in this study show that antibodies to the rgB and rgH envelope glycoproteins ofCMV are detected with different frequencies in individuals who are seropositive to the virus. Antibody to rgB was observed regularly, but antibody to rgH in immunocompetent normal individuals with no history of recent symptomatic infection was rarely detected. Several possibilities may account for this difference. First, the response to each of the glycoproteins may be influenced by the amount of antigen that is present on the immunogen. For example, differences in immunogenicity of the hemagglutinin and neuraminidase of influenza after vaccination have been attributed to the greater concentration of hemagglutinin on the virion surface [29]. The relative quantities of gB and gH on the viral envelope are not known. However using MAbs to the gB glycoprotein that are conjugated to colloidal gold, the viral envelopes are heavily labeled when examined by electron microscopy [30]. Thus, the immunodominance of gB may reflect its heavy concentration on the viral envelope. When virus-infected cells are immunoprecipitated by sera from infants with clinically apparent compared with subclinical infection [31], the proteins of the gB complex are easily detected. This may reflect the level of antigenic stimulus in the children with symptomatic disease who are likely to have a greater viral load than those with asymptomatic infection. Second, it is possible that antibody to rgB may be maintained longer after initial infection than antibody to rgH. After acute infection in immunocompetent individuals, antibody increases to both of the glycoproteins were readily detected. However, < 10% of seropositive patients distant in time from any history of acute CMV-associated symptoms had antibody to rgH, indicating that after seroconversion antibody to rgH may wane with time without reexposure to the virus. A third possibility is that some individuals who seroconvert to CMV without symptoms will fail to produce antibody to rgH. The failure ofthree asymptomatic seroconverters (figure 3) to develop antibody responses of the same magnitude as the symptomatic patients supports this idea. Fourth, the possibility of intrinsic differences in immunogenicity ofrgB and rgH is unlikely. Titers ofvirus-neutralizing antibody are comparable when equivalent amounts of immunoaffinity-isolated gB and gH are used to immunize guinea pigs [27]. Fifth, there may be fundamental differences in sensitivity of our assay for detecting rgB and rgH. The fact that the immunoprecipitation assay was slightly less sensitive for rgB than rgH (five times more anti-gB than -gH MAb was needed to give detectable bands) argues against this possibility.
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
body were detected throughout the posttransplant period. There was a trend toward higher responses in the late transplant period in this group compared with those in patients with primary infection. The elevated responses appeared earlier in patients with recurrent infection than in those with primary CMV. It is of interest that one of the three patients with recurrent CMV infection who failed to show increases in rgB antibody showed a significant increase in antibody to rgH (rising from 0 to 200% of control). Distribution ofantibody to gB and gH in CMV-seropositive leukemic patients. Sera from a group of 18 leukemic patients who were candidates for bone marrow allografts were tested for antibody to rgB and rgH. All were seropositive for CMV with LA titers> 1:8. For rgB, sera from 14 of 18 patients were reactive. However, only one serum sample reacted with the rgH glycoprotein. Lack of relationship between intensity of reaction of sera with rgB or rgH and virus-neutralizing or LA antibody. Seventy serum samples with virus-neutralizing titers> 1:8 were evaluated for a relationship between intensity of the reaction with either rgB or rgH and virus-neutralizing activity by linear regression analysis. No significant relationship was established between virus-neutralizing antibody titer and intensity of reaction with either rgB (r = .32253) or rgH (r = .10069). Serum samples from 17 normal donors with LA antibody titers of 16-1024 were also analyzed by linear regression for a correlation with intensity of reaction with rgB or rgH. No relationship of significance was observed (LA antibody vs. rgB, r = .08306; LA antibody vs. rgH, r = .007875). Antibody to rgB and rgH in ISG preparations. Antibody to rgB and rgH in lots of ISG prepared from unselected donors, represented by groups I and II from two different manufacturers, is shown in table 1. The mean ± 2 SD for both antibody to rgB and rgH in lots ofISG prepared from donors who were of unknown CMV serostatus was comparable to that of the normal seropositive population as shown in figure 2. Antibody to both the rgB and rgH tended to be higher in lots of ISG from hyperimmune donors (table 1, Ib), but because of
110 1991;164 (November)
110 1991; 164 (November)
Human Antibody to Cytomegalovirus gB and gH
ciencies in CMV glycoprotein antibodies may provide information about their role in protective immunity.
Acknowledgments
We thank Linda Logan (Diagnostic Laboratory, Stanford University Hospital) for obtaining serum samples from transplant patients, Marcia Smullen for virus-neutralization assays, and participants from the Day Care Center for Children ofSyntex Employees (Palo Alto, CA), for their cooperation.
References l. Meyers J, Flourney N, Thomas J. Risk factors for cytomegalovirus infection after human marrow transplantation. J Infect Dis 1986; 153:478-8. 2. Pass R, Griffiths P, August A. Antibody response to cytomegalovirus after renal transplantation: comparison of patients with primary and recurrent infections. J Infect Dis 1983; 147:40-6. 3. Yeager A, Grumet F, Halfleigh E, Arvin A, Bradley J, Prober C. Prevention of transfusion-acquired cytomegalovirus infections in newborn infants. J Pediatr 1983;98:281-7. 4. Yow M. Congenital cytomegalovirus disease: a NOW problem. J Infect Dis 1989; 159: 163-7. 5. Landini M, Michelson S. Human cytomegalovirus proteins. Prog Med Virol 1988;35: 152-85. 6. Rasmussen L. Immune response to human cytomegalovirus infection. Curr Top Microbiol ImmunoI1990;154:221-54. 7. Roby C, Gibson W. Characterization of phosphoproteins and protein kinase activity of virions, noninfectious enveloped particles, and dense bodies of human cytomegalovirus. J Virol 1986;59:714-27. 8. Jahn G, Scholl B, Traupe B, Fleckenstein B. The two major structural phosphoproteins (pp65 and pp 150) of human cytomegalovirus and their antigenic properties. J Gen Virol 1987;68: 1327-37. 9. Eing B, Juhn J, Braun R. Neutralizing activity of antibodies against the major herpes simplex virus type I glycoproteins. J Med Virol 1989;27:59-65. 10. Spear P. Glycoproteins specified by herpes simplex viruses. In: Roizman B, ed. The herpesviruses. Vol 3. New York: Plenum Press, 1984:315-56. I I. Fuller A, Santos R, Spear P. Neutralizing antibodies specific for glycoprotein H of herpes simplex virus permit viral attachment to cells but prevent penetration. J Virol 1989;63: 3435-43. 12. Fuller A, Spear P. Anti-glycoprotein D antibodies that permit adsorption but block infection by herpes simplex virus I prevent virion-cell fusion at the cell surface. Proc Nat! Acad Sci USA 1987;84:5454-8. 13. Gompels U, Minson A. The properties and sequence of gH of herpes simplex virus type 1. Virology 1986;153:230-47. 14. Britt W. Neutralizing antibodies detect a disulfide-linked glycoprotein complex within the envelope of human cytomegalovirus. Virology 1984; 135:369-78. 15. Pereira L, Hoffman M, Tatsuno M, Dondero D. Polymorphism of human cytomegalovirus glycoproteins characterized by monoclonal antibodies. Virology 1984; 139:73-86. 16. Rasmussen L, Mullenax J, Nelson R, Merigan T. Viral polypeptides detected by a complement dependent neutralizing murine monoclonal antibody to human cytomegalovirus. J Virol 1985;55:274-80. 17. Cranage M, Kouzarides T, Bankier A, et al. Identification of the human cytomegalovirus glycoprotein B gene and induction of neutralizing antibodies via its expression in recombinant vaccinia virus. EMBO J 1986;5:3057-63. 18. Rasmussen L, Nelson R, Kelsall D, Merigan T. Murine monoclonal
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
Finally, we cannot eliminate the possibility that there may be differences in the ability of the rgH and gH from virus-infected cells to react with human antibody. This, as a reason for lack of regular detection of gH antibody in most human sera, is made less likely by two findings. The patient sera reactive with gH were also reactive with rgH. Also, in patients with CMV mononucleosis, it was not possible to detect an anti-gH response in the convalescent phase with gH antigen (data not shown), but positive responses were detected with rgH (figure 2). This suggests that immunoprecipitation assays for rgH are more sensitive than immunoblot assays using gH for detecting human antibody. A final resolution of this question must await epitope mapping of gH for identification of the sites that are immunogenic for humans. Normal individuals with antibody to rgH may be those who have had extensive exposure to CMV, perhaps through multiple episodes of asymptomatic reactivation. This idea is supported by the finding that the antibody response to rgH was higher in recurrent CMV than in primary infections in the heart transplant population. The replicating virus probably provides sufficient antigen to elicit an rgH response. In serologic studies in children infected with the paramyxovirus parainfluenza virus type 3, the antibody response to the F (fusion) glycoprotein appeared only after repeated infections, whereas antibodies to the hemagglutinin/neuraminidase glycoprotein were detected after the initial infection [32]. This demonstrates that in another viral disease, antibody to at least two glycoproteins can be induced independently. The gB has been shown to be an important inducer of virus-neutralizing antibody in some patients [33, 34]. However, many factors may contribute to the neutralizing antibody response, for example, synergy between [35] or blocking by nonneutralizing antibodies [36]. Viral strain differences may influence the detection of virus-neutralizing antibody. The gene products of all of the open-reading frames for CMV glycoproteins [37] have yet to be identified immunologically. A study of the importance of the glycoproteins should be done using only epitopes known to bind virus-neutralizing antibodies. This would exclude interactions of many other regions on the molecule with nonneutralizing antibodies. There is conflicting evidence for the protective value of passively transferred ISO against CMV-associated interstitial pneumonitis in bone marrow transplant recipients, as recently reviewed by Meyers [38]. A prospective investigation of the clinical efficacy oflSO preparations with known levels ofantibody to the CMV glycoproteins may be useful for evaluating their importance in protection from CMV disease in recipients who are deficient in antibody to either glycoprotein. Progress has also been made in developing human MAbs to the CMV glycoproteins [39]. The evaluation of the efficacy of these MAbs in patients selected for specific defi-
841
842
19.
20.
21.
22.
24.
25.
26.
27.
28.
antibody to a single protein neutralizes the infectivity of human cytomegalovirus. Proc Natl Acad Sci USA 1984;81:876-80. Gretch D, Kari B. Rasmussen L, Gehrz R, Stinski M. Identification and characterization of three distinct families of glycoprotein complexes in the envelopes of human cytomegalovirus. J Virol 1988;62:87581. Cranage M, Smith G, Bell S, et aJ. Identification and expression of a human cytomegalovirus glycoprotein with homology to the EpsteinBarr virus BXLFZ product, varicella zoster gp III and herpes simplex type I glycoprotein H. J Virol 1988;62: 1416-22. Pachl C. Probert W, Hermsen K, et al. The human cytomegalovirus strain Towne glycoprotein gene encodes glycoprotein p86. Virology 1988; 169:418-26. Kari B, Lussenhop N, Goertz R, Wabuke-Bunoti M, Radeke R, Gehrz R. Characterization of monoclonal antibodies reactive to several biochemically distinct human cytomegalovirus glycoprotein complexes. J Virol 1986;60:345-52. Lehner R, Meyer H, Mach M. Identification and characterization of a human cytomegalovirus gene coding for a membrane protein that is conserved among human herpesviruses. J Virol 1989;63:37923800. Chang C. Vesole D, Nelson J, Oldstone M, Stinski M. Identification and expression of a human cytomegalovirus early glycoprotein. J Virol 1989;63:3330-7. Rasmussen L, Kelsall D, Nelson R, et al. Virus specific IgG and IgM antibodies in normal and immunocompromised subjects infected with cytomegalovirus. J Infect Dis 1982;145:191-9. Spaete R, Thayer R, Probert W, et al. Human cytomegalovirus strain Towne glycoprotein B is processed by proteolytic cleavage. Virology 1988; 167:207-25. Rasmussen L, Mullenax J, Nelson R, Merigan T. Human cytomegalovirus polypeptides stimulate neutralizing antibody in vivo. Virology 1985; 145: 186-90. Spaete R, Saxena A, Scott P, et al. Sequence requirements for proteolytic processing of glycoprotein B of human cytomegalovirus strain Towne. J Virol 1990;64:2922-31.
110 1991;164 (November)
29. Johansson B, Bucher D, Kilbourne E. Purified influenza virus hemagglutinin and neuraminidase are equivalent in stimulation ofantibody response but induce contrasting types of immunity to infection. J Virol 1989;63: 1239-46. 30. Stannard L, Rider J, Farrar G. Morphology and distribution ofgp52 on extracellular human cytomegalovirus (HCMV) supports biochemical evidence that it represents the HCMV glycoprotein B. J Gen Virol 1989;70: 1553-60. 31. Britt W. Vugler L. Antiviral antibody responses in mothers and their newborn infants with clinical and subclinical cytomegalovirus infections. J Infect Dis 1990; I 61 :214-9. 32. Kasel J, Frank A. Keitel W, Taber L. Glezen W. Acquisition of serum antibodies to specific viral glycoproteins of parainfluenza virus 3 in children. J Virol 1984;52:828-32. 33. Britt W, Vugler L, Butfilowski E, Stephens E. Cell surface expression of human cytomegalovirus (HCMV) gp55-1 16 (gB): use ofHCMV-vaccinia recombinant virus-infected cells in analysis of the human neutralizing antibody response. J Virol 1990;64: 1079-85. 34. Liu Y, Klaue A. Kari B. et al. The N-terminal 513 amino acids of the envelope glycoprotein gB of human cytomegalovirus stimulates both B- and T-cell immune responses in humans. J Virol 1991; 65: 1644-8. 35. Lussenhop N, Goertz R, Wabuke-Bunoti M, Gehrz R, Kari B. Epitope analysis of human cytomegalovirus glycoprotein complexes using murine monoclonal antibodies. Virology 1988; 164:362-72. 36. Utz .U, Britt W. Vugler L, Mach M. Identification of a neutralizing epitope on glycoprotein 58 of human cytomegalovirus. J Virol 1989;63: 1995-200 I. 37. Chee M, Bankier A, Beck S, et al. Analysis of the protein-coding content of the sequence of human cytomegalovirus strain AD 169. Curr Top Microbiol Immunol 1990; 154: 125-69. 38. Meyers J. Prevention of cytomegalovirus infection after marrow transplantation. Rev Infect Dis 1989;II(suppI7):SI691-1705. 39. Matsumoto Y, Sugano T, Miyamoto C. Masuho Y. Generation of hybridomas producing human monoclonal antibodies against human cytomegalovirus. Biochem Biophys Res Commun 1986; 137:27380.
Downloaded from http://jid.oxfordjournals.org/ at Sheffield University on September 13, 2015
23.
Rasmussen et al.
