Characterization of human monoclonal antibodies selected with a hypervariable loop-deleted recombinant HIV-1IIIB gp120

Immunology Letters 79 (2001) 209– 213 www.elsevier.com/locate/

Characterization of human monoclonal antibodies selected with a hypervariable loop-deleted recombinant HIV-1IIIB gp120 Simon A. Jeffs a,*, Miroslaw K. Gorny b, Constance Williams b, Kathy Revesz c, Barbara Volsky b, Sherri Burda b, Xiao-Hong Wang c, Juan Bandres c, Susan Zolla-Pazner b,c, Harvey Holmes a a

Di6ision of Retro6irology, NIBSC, Blanche Lane, South Mimms, Potters Bar, Herts EN6 3QG, UK Department of Pathology, New York Uni6ersity School of Medicine, New York, NY 10016, USA c Research Center for AIDS and HIV Infection, New York Harbor Veterans Affairs Medical Center, New York, NY 10010, USA b

Received 18 June 2001; accepted 8 August 2001

Abstract Recombinant gp120 of the HIV-1IIIB isolate (BH10 clone) has been mutated to form the PR12 protein with the first 74 C-terminal amino acids and the V1, V2 and V3 hypervariable loops deleted. A variety of studies have shown that the CD4 binding domain (CD4bd) is very well exposed in PR12 in contrast to rgp120LAI. Using PR12 for selection of human monoclonal antibodies (MAbs) from HIV-infected individuals, five MAbs were generated with specificities to the epitopes overlapping the CD4bd (1570A, 1570C, 1570D, 1595 and 1599). The three MAbs, 1570A, C and D, generated from one HIV-infected individual, represent one MAb as determined by sequence analysis of the VH3 region. Since the epitopes overlapping the CD4bd exhibit variability among HIV-1 clades, the specificity of anti-CD4bd MAbs were distinguished by differing patterns of binding to recombinant envelope proteins derived from clade A, B, C, D and E viruses. The PR12-selected MAbs were also compared with a panel of gp120-selected anti-CD4bd MAbs and showed a different range of specificities. MAb 1599 is clade B specific, MAb 1595 reacts with the A, B and D clades, while MAb 1570 recognises the most conserved epitope, as it binds to all proteins. The results show that the exposure of different epitopes in the CD4bd of the PR12 protein allows this protein to serve as an immunogen and to induce anti-CD4bd antibodies. © 2001 Published by Elsevier Science B.V. Keywords: Neutralising antibodies; gp120; Hypervariable loops

1. Introduction Although the correlates of protective immunity remain elusive in both human and simian immunodeficiency virus infections, there is a growing belief that a successful vaccine must stimulate both arms of the immune defence mechanism — humoral and cell-mediated [1,2]. Many of the current vaccines undergoing clinical trials are based on the retroviral envelope glycoprotein (gp120/140/160), which has been shown to be safe and well-tolerated in human volunteers and to induce potentially-protective immune responses [3].

* Corresponding author. Tel.: + 44-1707-654-753; fax: +44-1707649-865. E-mail address: [email protected] (S.A. Jeffs).

However, recombinant gp120 is not a particularly effective immunogen, and a number of approaches have been suggested to enhance its immunogenicity. The goal is to provide an antigen that induces the generation of antibodies with the ability to neutralise primary isolates of HIV-1 from diverse clades. A variety of studies have shown that removal of the hypervariable loops of recombinant HIV-1 gp120 creates a molecule that retains its structural integrity, while providing enhanced accessibility to specific domains of the conserved regions of this glycoprotein, notably the CD4 binding domain (CD4bd) and C1 regions [4–6]. In this study we show that such molecules can be used to select human MAbs using peripheral blood mononuclear cells (PBMC) from HIV-infected individuals, which show a broad repertoire of specificity to the conformation-dependent CD4bd.

0165-2478/01/$ - see front matter © 2001 Published by Elsevier Science B.V. PII: S 0 1 6 5 - 2 4 7 8 ( 0 1 ) 0 0 2 8 9 - 9

S.A. Jeffs et al. / Immunology Letters 79 (2001) 209–213

210

The PR12 protein used for the selection of MAbs is a mutated form of the envelope of the HIV-1IIIB isolate (BH10 clone) from which the first 74 C-terminal amino acids and the V1, V2 and V3 hypervariable loops were deleted [4]. Previous studies have shown that the CD4bd, the C1 and parts of the C2 region are very well exposed in PR12 in contrast to rgp120LAI [4,7]. Initial results suggested that the PR12 protein has different regions exposed for antibody binding than rgp120LAI, indicating that PR12-specific antibodies might recognise epitopes distinct from those seen by rgp120LAI-selected antibodies, particularly in the CD4bd.

2. Materials and methods Human MAbs were generated according to the method previously described [8]. Briefly, Epstein– Barr virus-transformed PBMC, derived from asymptomatic individuals infected with HIV-1 subtype B strains, were screened against the PR12 protein. Cells producing PR12-reactive antibodies were expanded, fused with the human × mouse heteromyeloma cell line SHM-D33 and the resulting hybridomas were cloned at limiting dilution until monoclonality was achieved. The potential identity of the MAbs 1570A, C and D was determined by sequencing their variable domains. Total RNA was isolated from three hybridoma cell lines and then was used to synthesise the first strand cDNA. Amplification of the VH domain was performed by PCR with a different family’s gene-specific primers according to a sequence previously described [9]. The inhibition of sCD4-gp120 binding by these MAbs was determined as previously described [10]. Briefly, the complex of test MAb with rgp120LAI and sCD4 was captured by immobilised sheep anti-C5 antibody. The presence of sCD4 in the complex was determined using mouse anti-CD4 MAb followed by alkaline phosphatase-labelled anti-murine IgG antibody.

The neutralising activity of the MAbs was examined by assay using two types of target cells: GHOST cells expressing CD4 and either of the two coreceptors CXCR4 or CCR5. The assay measures the number of infected cells, which express green fluorescent protein, using flow cytometry. A decrease in the proportion of infected cells in the presence of MAbs is expressed as percentage neutralisation [15].

3. Results Five human MAbs, 1570A, 1570C, 1570D, 1595 and 1599, were generated and characterised (Table 1). Three MAbs, 1570A, C and D, were derived from one HIVinfected individual and display a very similar pattern of binding. The heavy chains of the three 1570 MAbs were classified as members of the VH3 family. Analysis of three VH sequences revealed identity with the exception of five base pairs, confirming that MAbs 1570 A, C and D represent a single MAb derived from three subclones (data not shown). All five MAbs competed with soluble CD4 for binding to rgp120LAI, indicating that they belong to the group of anti-CD4bd antibodies (Table 1). This inhibitory activity, which was measured by the quantity of MAbs needed for 50% of inhibition, was similar among five MAbs tested (Table 1) and comparable to that of other anti-CD4bd MAbs [10] All MAbs displayed similar relative affinity binding to rgp120LAI (Table 1). The amount of MAb required for half-maximal binding to gp120 was determined by an ELISA by titration of the MAbs against antigen as described previously [14]. The anti-CD4bd MAbs are conformation-dependent and recognise discontinuous epitopes, which consist of amino acids located in several regions overlapping the CD4bd [11,12]. As the CD4bd of gp120 is very conserved in terms of binding to CD4, the MAb epitopes overlapping and close to the CD4bd exhibit variability

Table 1 Characteristics of human monoclonal antibodies selected with PR12 protein MAb

Isotype

Family of heavy chain

Specificity

50% inhibition of sCD4–gp120 binding (mg/ml)

50% maximal binding to gp120 (mg/ml)

1570Aa 1570Ca 1570Da 1595 1599

IgG1l IgG1l IgG1l IgG1k IgG1l

VH3 VH3 VH3 nt nt

CD4bdb CD4bd CD4bd CD4bd CD4bd

0.44c 0.27 0.28 0.25 0.93

0.060 90.01d 0.050 9 0.002 0.036 9 0.01 0.043 90.01 0.033 9 0.01

nt, Not tested. a MAbs generated from the same HIV-1-infected individual. b MAb competes for sCD4 binding to rgp120LAI. c Quantity of MAb needed for 50% inhibition. d 50% maximal binding values in mg/ml; means of three experiments 9 SD.

Table 2 Binding of human anti-CD4bd MAbs to recombinant envelope glycoproteins derived from clade A, B, C, D and E viruses MAba

gp160RF

gp120W61D

gp120LAI

gp140HAN2

gp140451

gp120SF2

gp14092UG037

gp14092BR025

gp160ELI/LAI

gp1604020

gp120CM235

gp120CM

gp120CM243

B

B

B

B

B

B

B

A

C

D/B

D

E

E

E

1570Ab 1570Cb 1570Db

14.1d 14.5 14.1

12.8 12.8 12.7

10.5 8.5 8.8

15.6 15.6 15.9

9.0 8.3 8.8

10.6 10.4 10.7

10.9 10.0 10.6

14.0 14.2 14.5

2.0 1.8 1.9

14.5 13.1 13.2

12.8 12.8 12.7

15.0 12.3 12.3

2.7 2.5 2.3

1595b 1599b

14.0 11.7

8.9 1.6

6.2 5.2

15.9 15.4

6.7 4.5

10.5 10.3

5.6 3.6

6.3 1.1

– 1.2

12.4 –

6.0 –

– –

– –

– –

448-D 559/64D 588-D 654-D 729-D 9CL 1008-30D 1027-30D 1202-D 1263 1331E

13.8 13.9 13.8 14.0 13.6 14.1 12.5 14.3 12.6 14.1 13.8

1.7 2.7 9.8 4.3 8.0 8.6 1.8 4.9 5.2 3.6 1.9

2.5 8.5 8.9 6.6 6.4 7.3 8.2 10.5 8.3 8.0 13.0

15.5 15.8 15.8 15.9 15.8 15.5 14.7 15.5 16.9 16.5 15.0

2.2 7.0 11.3 6.4 7.7 13.0 3.2 5.2 11.5 – 7.1

9.6 10.5 10.3 10.4 10.1 10.2 9.8 10.7 10.5 9.7 11.2

4.1 7.4 4.2 6.3 5.8 4.7 7.2 8.0 6.7 6.7 7.4

10.4 13.8 14.6 14.6 14.3 14.1 14.6 15.9 14.0 3.6 15.9

– – 2.5 1.6 1.2 1.4 1.1 1.0 – – 1.4

8.5 12.4 13.0 14.1 12.2 13.2 12.1 14.1 11.8 7.8 13.1

3.5 12.3 12.7 12.7 12.7 12.6 12.5 12.9 12.8 – 12.7

– – 12.6 – – – – – – – 7.9

– – 2.5 – – – – – – – 1.7

– – 10.5 – – – – – – – 5.7

1418c

–

–

–

–

–

–

–

–

–

–

–

–

–

–

3.1 2.7 3.1

Recombinant gp120LAI was purchased from Intracel (Issaquah,WA); gp140451 was purchased from Advanced BioScience Laboratories (Kensington, MD); rgp120SF2 and rgp120CM235 were kindly provided by Chiron (Emeryville, CA); rgp120W61D was kindly provided by SmithKline Beecham (Rixensart, Belgium); rgp160RF (gp120RF–gp41LAI), rgp160ELI/LAI (gp120ELI–gp41LAI), rgp1604020 and rgp120CM243 were kindly provided by Transgene (Strasbourg, France); rgp120CM was obtained through the AIDS Research and Reference Reagent Program (Division of AIDS, NIAID, NIH); rgp140HAN2, rgp14092UG037, rgp14092BR025 and PR12IIIB were produced in CHO cells according to method recently published [4]. All proteins were coated directly onto plastic plates at 1.0 mg/ml and binding of MAbs was detected using alkaline phosphatase-conjugated goat anti-human IgG. a All MAbs tested at 1.0 mg/ml. b MAbs selected with PR12. Others are anti-CD4bd MAbs selected with rgp120LAI or gp140451. c Mab 1418 is specific for human parvovirus B19 and was used as a negative control. d Index value is expressed as the ratio of sample OD to cutoff (mean OD+3 SD with MAb 1418). Values \1.0 are considered reactive. Values B1.0 are indicated by dashes (–).

S.A. Jeffs et al. / Immunology Letters 79 (2001) 209–213

PR12IIIB

211

212

S.A. Jeffs et al. / Immunology Letters 79 (2001) 209–213

among HIV-1 subtypes [13]. In order to distinguish the MAbs specificities for the CD4bd, the binding pattern of the MAbs to various recombinant gp120, gp140 and gp160 molecules derived from clade A, B, C, D and E viruses was examined by ELISA (Table 2). Additionally, the results were compared with binding reactivity of 11 other anti-CD4bd MAbs selected with a ‘fulllength’ gp120LAI or oligomeric gp140451 [10,14] (Table 2). All MAbs bound to all of the clade B envelope glycoproteins tested, but each of the MAbs selected with PR12 (1570A/C/D, 1595 and 1599) shows a different pattern of binding to proteins from other HIV-1 subtypes. MAb 1570 recognises the most conserved epitope as it binds to proteins from the A, B, C, D and E clades. MAb 1595 only recognises the clade A, B and D proteins, while MAb 1599 was predominantly clade-B specific, recognising the A and C clade gp140s poorly and the D and E specimens not at all. The clade A gp140, 92UG037, was strongly recognised by all the CD4bd MAbs except 1599. The clade C gp140 92BR025 was weakly recognised by most of the CD4bd MAbs. In comparison with those MAbs selected with full-length rgp120, PR12-selected MAbs showed a similar but not identical pattern of binding to the CD4bd of envelope glycoproteins from diverse clades. Each of the PR12-selected MAbs reacted with different proteins suggesting different epitopes overlapping the CD4bd. One MAb 1599 displayed a unique pattern of binding, not found with other antibodies. In contrast, the majority of the rgp120-selected MAbs, with the exception of 588-D and 1331E (selected using oligomeric gp140451), showed similar specificity in terms of binding patterns to clade A, B and D proteins. The anti-CD4bd MAbs, whether selected with PR12 or rgp120LAI, displayed no or low levels of neutralising activity against B-clade primary isolates, Bx08, JR-FL and MNp (data not shown). In general, slightly higher activity was noted against the laboratory-adapted strain (IIIB), and 1570A/C/D showed more activity than 1595 or 1599. This low level of activity against primary isolates is in agreement with previous studies concerning CD4bd MAbs, whether selected with fulllength or mutated envelope glycoprotein [7,10]. Indeed, the only anti-CD4bd MAb that is reported to show potent neutralising activity against primary isolates is the recombinant human MAb, IgGb12, which was selected using a phage display technique and recognises a complex epitope related to the CD4bd [16].

4. Discussion The present study has demonstrated the ability of loop-deleted proteins to select MAbs that display a broad range of specificities against conformation-de-

pendent epitopes overlapping the CD4bd. In a separate study, MAbs against CD4bd and C1 have been generated from rats immunized with the same PR12 protein (Jeffs et al., unpublished results). None of these MAbs neutralised HIV-1 although sera from PR12-immunized rats showed cross-clade neutralising activity against a number of chimaeric viruses comprising an HXB2 backbone with the en6 gene from 5 clade B, 2 clade C or 1 clade E or F primary isolate (Jeffs et al., unpublished results). The type of anti-PR12 antibodies that mediate neutralisation remains to be determined. However, the presence of CD4bd, in the V1/V2 and V3 deleted-PR12 protein suggest that this domain could be responsible for serum neutralising activity. The MAb-binding experiments with PR12 showed that hypervariable-loop deletion does not affect the antigenicity of the glycoprotein, while the selection of different anti-CD4bd MAbs from HIV-1-infected individuals using PR12 indicates that various epitopes in the PR12 CD4bd are both antigenic and well exposed for antibody binding. The generation of similar MAbs from PR12-immunized rats confirms the immunogenic potential of the PR12 CD4bd epitopes. The approach of using an immunogen, in the form of a mutated protein, which enhances the exposure of some epitopes in the context of the tertiary and quaternary structure of the gp120 glycoprotein, is to be preferred to peptides that are only able to present linear epitopes. This technique could be extended to expose other regions implicated in viral neutralisation, located within the V1/V2, V3 and gp41 domains. Recently, it was shown that anti-V3 antibodies, which recognise conformation-dependent epitopes, mediate more neutralising activity than those directed to linear epitopes [17,18]. Building upon this and other studies, we can envisage the concept of a ‘superimmunogen’ as a recombinant envelope glycoprotein with selected polypeptide domains and defined N-linked glycosylation sites removed to expose neutralising epitopes and enhance their immunogenicity. Studies to examine these modifications are underway, and will hopefully lead to new directions for the development of HIV vaccine therapy.

Acknowledgements This work was supported in part by grants AI32424 and HL59725 from the National Institutes of Health, by the Research Center for AIDS and HIV Infection and a Merit Review grant funded by the Department of Veterans Affairs and by the European Union through the Programme EVA Centralised Facility for AIDS Reagents (BIOMED2 contract no. BMH4 97/ 2515).

S.A. Jeffs et al. / Immunology Letters 79 (2001) 209–213

References [1] G. Ferrari, W. Humphrey, M.J. McElrath, Proc. Natl. Acad. Sci. 94 (1997) 1396 –1401. [2] S. Zolla-Pazner, S. Xu, S. Burda, A.-M. Duliege, J.-L. Excler, M.L. Clements-Mann, J. Infect. Dis. 178 (1998) 1502 – 1506. [3] D.R. Burton, P.W.H.I. Parren, Nat. Med. 6 (2000) 123 – 125. [4] S.A. Jeffs, J. McKeating, S. Lewis, J. Gen. Virol. 77 (1996) 1403 – 1410. [5] R. Wyatt, N. Sullivan, M. Thali, J. Virol. 67 (1993) 4557 – 4565. [6] P.D. Kwong, R. Wyatt, J. Robinson, R.W. Sweet, J. Sodroski, W.A. Hendrickson, Nature 393 (1998) 648 –659. [7] J. McKeating, C. Shotton, S. Jeffs, Immunol. Lett. 51 (1996) 101 – 105. [8] M.K. Gorny, J.-Y. Xu, V. Gianakakos, Proc. Natl. Acad. Sci. 88 (1991) 3238 – 3242.

213

[9] J.D. Marks, M. Tristem, A. Karpas, G. Winter, Eur. J. Immunol. 21 (1991) 985 – 991. [10] S. Karwowska, M.K. Gorny, A. Buchbinder, AIDS Res. Hum. Retroviruses 8 (1992) 1099 – 1106. [11] U. Olshevsky, E. Helseth, C. Furman, J. Li, W. Haseltine, J. Sodroski, J. Virol. 64 (1990) 5701 – 5707. [12] J.A. McKeating, M. Thali, C. Furman, Virology 190 (1992) 134 – 142. [13] J.P. Moore, F.E. McCutchan, S.-W. Poon, J. Virol. 68 (1994) 8350 – 8364. [14] M.K. Gorny, T.C. VanCott, C. Williams, K. Revesz, S. ZollaPazner, Virology 267 (2000) 220 – 228. [15] D. Cecilia, V.N. KewalRamani, J. O’Leary, J. Virol. 72 (1998) 6988 – 6996. [16] D.R. Burton, J. Pyati, R. Koduri, Science 266 (1994) 1024 –1027. [17] M. Schreiber, C. Wachsmuth, H. Muller, J. Virol. 71 (1997) 9198 – 9205. [18] E.J. Park, M.K. Gorny, S. Zolla-Pazner, G.V. Quinnan Jr., J. Virol. 74 (2000) 4183 – 4191.