WO1991004277A1

WO1991004277A1 - Monoclonal antibodies specific to cytomegalovirus glycoprotein

Info

Publication number: WO1991004277A1
Application number: PCT/US1990/005164
Authority: WO
Inventors: Bruce E. Kari; Ronald Goertz; Nancy O. Lussenhop; Richard C. Gehrz
Original assignee: Children's Biomedical Research Institute
Priority date: 1989-09-14
Filing date: 1990-09-12
Publication date: 1991-04-04

Abstract

Disclosed is a set of monoclonal antibodies having reactivity with 93kD glycoproteins of the gCI complex of human cytomegalovirus (HCMV) but not with the 52kD glycoprotein from that same complex. The monoclonal antibodies and serums containing them exhibit neutralizing activity against HCMV without complement.

Description

MONOCLONAL ANTIBODIES SPECIFIC TO CYTOMEGALOVIRUS GLYCOPROTEIN

Background of the Invention

Human cytomegalovirus contains three distinct families of disulfide linked glycoprotein complexes. (D.R. Gretch et al., J. Virol., 62, 875-881 (1988)). One of these families, designated gCI, has been the most extensively studied. (T. Banks et al., J. Gen. Virol., 70, 979-985 (1989); W.J. Britt and D. Auger, J. Virol . , 58, 185-191 ( 1986 ) ) ; W. J. Britt et al . , J . Virol . , 62, 3309-3318 (1988)); D.R. Gretch et al., Anal . Biochem., 163, 270-277 (1987)).

When obtained from extracellular virus, the gCI family consists of disulfide linked glycoprotein

complexes ranging in molecular weight from 150,000 to

250,000 (7, 10, 12). G.H. Farrar and P.J. Greenaway,

Gen. Virol., 67, 1469-1473 (1986). When reduced, these complexes generate three glycoproteins with molecular weights of 130,000 (gpl30), 93-116,000 (gp93) and 52- 58,000 (gp52) (W.J. Britt and L.G. Vugler, J. Virol., 63, 403-410 (1989). All three glycoproteins appear to arise from the same gene product, which exhibits

homology with glycoprotein B of herpes simplex. (D.R. Gretch et al., J. Gen. Virol., 69, 1205-1215 (1988);

M.P. Cranage et al., EMBO J., 5, 3057-3063 (1986)).

Glycoproteins gp52 and gp93 appear to arise from a 138,000 to 160,000 molecular weight precursor by proteolysis. (See W.J. Britt et al., 1988, cited supra; and R.R. Spaete et al.. Virology, 167, 207-225 (1988)). Another pathway may also exist which leads to a mature 130,000 molecular weight glycoprotein which is not subjected to the proteolytic event that generates gp52 and gp93. Several laboratories have generated murine monoclonal antibodies (MoAbs) which recognize gp52, but not gp93 in Western blot analysis or other immune assays. (B. Kari et al., J. Virol., 60, 345-352

(1986)). Many of these MoAbs were found to be

neutralizing in in-vitro assays but the majority

required complement. At least two of these MoAbs to gp52, however, were complement independent (T. Banks, et al., cited supra; (N.O. Lussenhop et al., Virology, 164, 362-372 (1988); L. Rasmussen et al., J. Virol., 55, 274-280 (1985)); L. Rasmussen et al., Virology, 163, 308-318 (1988)).

Consequently, if gp52 and gp93 arise from a single precursor glycoprotein by proteolytic cleavage it should be possible to obtain MoAbs which recognize unique epitopes exposed on either gp52 or gp93 and their precursor.

An object of the invention, therefore, is the development of unique MoAbs to gp93. A further object is the development of complement independent MoAbs to gp93. Further objects include development of an anti-HCMV serum and an HCMV vaccine through the application of purified gp93 and unique MoAbs to it.

Summary of the Invention

These and other objects are achieved by the present invention which is directed to monoclonal antibodies that are specifically immunoreactive with gp93 of gCI but not with gp52 of gCI. These monoclonal antibodies preferably display complement independent neutralizing activity against HCMV. Also included are the hybridomas for production of the gp93-specific MoAbs and a serum containing an effective amount of the gp93-specific MoAbs.

Detailed Description of the Invention It is believed that the development of complement independent MoAbs to the HCMV virus would provide a powerful weapon to cure HCMV infections in humans. In contrast with complement-dependent

neutralization of HCMV, which is believed to involve extracellular lysis of virions, the complement

independent neutralization of HCMV is believed to interfere directly with viral infectivity or viral replication in infected cells, or other biological properties essential for the virus to propagate.

Neutralization at this latter stage has great

therapeutic potential since patients already infected can be treated.

The glycoproteins comprising the gCI family of antigenically related complexes are derived from a single polypeptide precursor, with a molecular weight of 95,000, encoded by a gene in the unique long (U_L) region of the HCMV genome which exhibits homology with

glycoprotein B of herpes simplex virus. Post- translational processing of this polypeptide results in a fully glycosylated stable precursor glycoprotein with a molecular weight ranging from 138,000 to 170,000 depending upon the methods of separation and viral strain according to the various reports cited above.

Proteolytic cleavage of this precursor glycoprotein results in a mature glycoprotein with a

molecular weight of 52,000, representing the carboxyl- terminal half of the glycosylated polypeptide of 907 amino acids in Towne strain HCMV, beginning at amino acid 461, as confirmed by N-terminal sequencing of purified gp52 by Spaete et al., Virology, 167, 207

(1988). A second mature glycoprotein with a molecular weight of 93,000, designated gp93, has also been

obtained from this precursor (see U.S. Application

Serial No. 06/933,789, filed November 24, 1986, the disclosure of which is incorporated herein by

reference). Presumably, this glycoprotein represents the amino terminal end of the glycoprotein precursor following a single proteolytic cleavage event, although the possibility of an alternate cleavage with a small extent of overlapping peptide sequence with gp52 cannot be entirely excluded. It is known that any overlapping peptide sequence of gp93 and gp55 does not include the major B-cell antigenic determinants of either gp93 or gp55, since there is no cross-reactivity of gp52- specific monoclonal antibodies with gp93, or of gp93-specific monoclonal antibodies with gp52. The higher molecular weight of the N-terminal half of the

glycoprotein precursor can readily be explained, based on amino acid and carbohydrate analysis, which suggests that more than 50% by weight of the gp93 molecule is carbohydrate, whereas gp52 has far less glycosylation.

According to the present invention both gp93-specific monoclonal antibodies and gp52-specific

monoclonal antibodies can completely immunoprecipitate radio-labeled, unreduced gCI complexes which indicates that both gp93 and gp55 are contained in these

complexes. However, following reduction. Western blot analysis indicates that gp93-specific monoclonal

antibodies developed according to the present invention react only with gp93 and the precursor glycoprotein, gpl58, whereas gp52-specific monoclonal antibodies react only with gp52 and the precursor glyco-protein, gpl58.

According to the invention, new monoclonal antibodies were derived from subcloning two parent hybridomas, 3C2 and 9F9. The new MoAbs exhibit the IgG₁ isotope, neutralize Towne strain HCMV in a plaque reduction assay in the absence of complement and exhibit specific immunoreactivity with gp93 but not with gp52 (by Western blot analysis).

The gp93-specific hybridomas developed according to the invention include a daughter clone from hybridoma 3C2, designated F15-3B10. It has been

deposited with the In Vitro International and has been assigned IVI- _{1 0207}

The gp93-specific MoAbs are useful in the diagnosis and treatment of HCMV infection whether alone or in combination with other HCMV reactive MoAbs.

Formulation of an effective amount into a serum

containing optional adjuvants, buffers, extenders, physiological salts, solubilizing and stabilizing agents and the like provides an effective regiment for such treatment.

According to the invention, purified gp93 was used as the immunizing agent in mice. After production of the appropriate hybridomas the MoAbs obtained were characterized by immunoprecipitation. Western blot, and neutralization assays.

Towne strain HCMV was used in all experiments. The virus was grown on skin fibroblasts as previously reported, (see B. Kari et al. (1986) cited supra.) Extracellular virus was partially purified from culture media by differential centrifugation. To purify gp93 for the purpose of making MoAbs, gCI complexes were immunoaffinity purified from a non-ionic detergent extract of HCMV using a (gp52 of gCI) gCI-specific biotinylated MoAb and streptavidin agarose. This immunoaffinity method has been previously described, see Gretch et al. cited supra. The gCI complexes were reduced and separated by SDS-PAGE in a 9% polyacrylamide gel using the method of Laemmli. Cleavage of structural proteins during assembly of the head of bacteriophage

T4, U.K. Laemmli, Nature (London), 227, 680-684 (1970). Proteins were detected by Commassie blue staining and the gp93 band was cut from the polyacrylamide gel. The glycoprotein was electroeluted from the gel slice and SDS was removed from the protein with a 0.5 ml column of Extracti-gel D (Pierce Chemical Co.). Mice were

immunized with purified gp93 to make MoAbs. Production and characterization of MoAbs was done as previously described, see B. Kari et al. (1986) cited supra.

Two MoAbs were obtained: 3C2 (IgGl) and 9F9

(IgGl). 3C2 was further subcloned to obtain a daughter clone producing MoAb F15-3B10. Several immune assays were used to compare these MoAbs to MoAbs which

recognized gp52 in Western blot (15). MoAbs were first tested for their ability to immunoprecipitate [³H]GlcN labeled gCI from non-ionic detergent extracts.

The results of these tests were determined by gel electrophoresis, which was conducted as follows. A [³H]GlcN labeled non-ionic detergent extract of Towne strain HCMV was immunoprecipitated with the MoAbs using protein A separose which had been preincubated with goat anti-mouse IgG. The monoclonal antibodies used

including 18F9 which is a MoAb made using whole HCMV and recognizes gp52, and new 3C2 and 9F9 which are MoAbs made using purified gp93. After extensive washing with PBS containing 0.1% NP-40 the proteins were eluted from the beads with SDS-PAGE sample solubization buffer.

Samples were subjected to electrophoresis under non- reducing conditions using a 7% polyacrylamide gel (A) or reducing conditions using a 10% polyacrylamide gel (B and C). SP2 was a negative ascites control. Molecular weights of 28-158 X 10³ were determined. On gel (A) immunoprecipitates were examined under non-reducing conditions. On gel (B) immunoprecipitates were examined under reducing conditions. On gel (C) serial

immunoprecipitations were examined. An extract was first precipitated with MoAb 3C2 and then with 18F9 (C lane 1) or first with 18F9 and then 3C2 (C lane 2).

Serial immunoprecipitations were also done by using 9F9 first followed by 3C2 (C lane 3).

According to the electrophoresis test results, all three MoAbs immunoprecipitated high molecular weight disulfide linked complexes which generated glycoproteins with apparent molecular weights of 52,000, 93,000 and 130,000. Glycoproteins with molecular weights greater than 130,000 were also detected in all

immunoprecipitates. These higher molecular weight glycoproteins are most likely present in the viral preparation due to contaminating cellular membranes.

These can be removed by including a sorbitol gradient purification step, see Gretch et al. (1988) cited supra. However, the presence of these glycoproteins did allow for the detection of MoAb reactivity with precursor glycoproteins. Serial immunoprecipitations of unreduced gCI complexes were also done. When MoAb 18F9 was used first and then 3C2 or visa versa, little additional radioactivity was immunoprecipitated according to the measurements obtained from gel electrophoresis conducted as discussed below. The same result was obtained when 9F9 was used first followed by 3C2.

The MoAbs were also examined for their reactivity with individual gCI glycoproteins in Western blot analysis. The Western blot method used has been previously described. For this experiment unreduced gCI was immunoaffinity purified as described above.

Following reduction glycoproteins from gCI were

separated by SDS-PAGE using a 9% polyacrylamide gel and then electroblotted onto nitrocellulose paper, which was conducted as follows. Complex gCI was immunoaffinity purified from a non-ionic detergent extract of Towne strain HCMV. One hundred and seven micrograms of gCI was reduced and gCI glycoproteins separated by SDS-PAGE in a 9% polyacrylamide gel. Proteins were then electro- blotted onto nitrocellulose paper. The paper was cut into strips and the strips incubated with the indicated negative controls, MoAbs and adult human sera. Lane 1 shows a portion of the gel which was cut off and stained with Commassie blue. T, top of gel and B, bottom of gel. Lane 2 SP2 negative ascites control. Lanes 3 and 4 MoAbs 3C2 and 9F9 made with purified gp93. Lanes 5-7 MoAbs 39E11, 9B7 and 18F9 made using whole HCMV. Lane 8 is a HCMV negative serum (A-). Lanes 9-11 are HCMV positive adult human sera (A+).

The results of the electroblot were determined by cutting the paper into strips and incubating the strips with the MoAbs or adult human sera. Mouse

antibody was detected with phosphatase labeled goat anti-mouse IgG, and human antibody with phosphatase labeled goat anti-human IgG (Kirkegaard and Perry).

After washing the paper was incubated with the substrate 5-bromo-4-chloro-3-indoly phosphate/tetrazolium in 0.1 M Tris buffer (Kirkegaard and Perry) and the reaction stopped by addition of water. Prior to electroblotting a portion of the SDS-PAGE gel was cut off and stained with Commassie blue. Glycoproteins gpl58, gp93 and gp52 appeared much less abundant relative to gp93 when

[³H]GlcN label was used but with Commassie blue staining gp52 appeared much more abundant relative to gp93. This supports previous observations which showed that gp93 was more heavily glycosylated than gp52. The SP2 negative ascites control did not react in Western blot. MoAbs 3C2 and 9F9 made with gp93 reacted with gp93 and gpl58 as well as material at the very top of the paper, but not with gp52. In contrast MoAbs 39E11, 9B7, and 18F9 reacted with gp52 and gpl58 and not gp93. Four human sera were also examined for their reactivity with gCI glycoproteins. A human serum negative for HCMV did not react with gCI glycoproteins.

All three positive human sera reacted with gp52 and gpl58, but only two reacted with gp93. All three positive human sera also reacted with a protein with a molecular weight of 28,000. Very little of this protein was detected by Commassie blue staining. This protein was not associated with gCI by disulfide bonds and is recognized by another of our HCMV MoAbs (35F10) which does not recognize gCI (data not shown). The 28,000 molecular weight protein was most likely a contaminating HCMV protein which co-precipitates with gCI.

The MoAbs recognizing gp93 were also examined for their ability to neutralize Towne HCMV in a plaque reduction assay according to methods previously

described. The MoAbs were purified with a protein G column (Pharmacia) prior to the neutralization assays. MoAbs 3C2 and 9F9 were found to be neutralizing against Towne strain HCMV in the absence of complement (See Table 1 below). In contrast MoAbs which recognized gp52 required complement to neutralize Towne strain HCMV. TABLE 1

MoAb Activity With and Without Complement

MoAb Subtype ug/ml MoAb+C'^* ug/ml MoAb-C'^*

3C2 IgG1 4.4 2.4

9F9 IgG1 4.2 2.4

18F9 IgG2b 3.6 No activity * Amount of MoAb in ug/ml needed to reduce the number of plaques by 50Z with complement (+c') or without complement (-c'). MoAbs were purified with a protein G column. Each MoAb was tested in a range of 1 to 100 ug/ml of antibody with and without guinea pig complement.

Approximately 1000 plaque forming units were used in a 36-mm dish.

Several reports have recently been published which indicate that gp52 and gp93 arise from a common precursor by proteolytic cleavage. In this study, it is shown that two sets of MoAbs can be made which recognize gCI complexes, and the gCI glycoprotein precursor gpl58. One group recognizes gp52 but not gp93, whereas the second group (as disclosed in the present invention) recognizes gp93 but not gp52. This observation supports the hypothesis that gp52 and gp93 arise by proteolytic cleavage of a common high molecular weight precursor. Previous studies have focussed on the importance of neutralizing antibody responses to gp52. However, it is clear that there are also neutralizing sites on gp93.

Until the present invention no murine MoAbs to gp93 had been obtained by using whole virus as the immunizing antigen. Several gCI specific MoAbs have also been generated from different fusions using whole virus as the antigen. Among these MoAbs, those which recognized continuous epitopes reacted with gp52 and not gp93 in Western blot analysis. Glycosylation can be an important modifier of viral antigenicity. One

difference between gp52 and gp93 is the extent to which they are glycosylated. For example, carbohydrate

represents approximately 50% of the molecular weight of gp93 in SDS-PAGE. Carbohydrate represents much less of the molecular weight of gp52. It may be that the

extensive glycosylation of gp93 makes it less immunogenic. It is also possible that the apparent lack of response to gp93 when whole virus is used as the antigen represents a species difference. All three glycoproteins are recognized by human HCMV positive sera. Thus, humans are capable of making antibody to gp93. Because gp93 contains neutralizing sites and is recognized by human HCMV positive sera it is likely to be important in the human immune response to HCMV.

Claims

CLAIMS :

1. A monoclonal antibody having immunoreactivity with the 93kD glycoprotein but not the 52kD glycoprotein derived from the gCI complex of human cytomegalovirus.

2. A monoclonal antibody according to claim 1 which recognizes the 93kD glycoprotein and exhibits neutralizing activity against HCMV.

3. A monoclonal antibody according to claim 1 which exhibits the Ig G1 isotope and neutralizes the human cytomegalovirus without complement.

4. A monoclonal antibody according to claim 1 which is produced by hybridoma F15-3B10.

5. A monoclonal antibody according to claim 1 wherein the 93 kD glycoprotein is substantially pure, contains a peptide sequence different from that of the 52 kD glycoprotein and contains epitopal sites not found upon the 52 kD glycoprotein.

6. A monoclonal antibody according to claim 3 wherein the 93 kD glycoprotein is free of the epitopal sites present in the 52 kD glycoprotein.

7. A monoclonal antibody of claim 1 obtained from murine hybridoma.

8. A serum composition suitable for the diagnosis and treatment of human cytomegalovirus infection comprising an effective amount of a monoclonal antibody of claim 1 in combination with suitable carrier.