CA2091253A1 - Method for inhibiting the infectivity of human immunodeficiency virus - Google Patents

Abstract

Disclosed herein are methods for inhibiting cell fusion between human T-lymphocytes infected by human immunodeficiency virus (HIV) or free human immunodeficiency virus and uninfected human T-lymphocytes comprising administering inhibitory effective amounts of HIV protein gp120 and CD4-Immunoadhesion and immunogenic composiitions for use in the methods.

Description

.
.

WO 92/05799 ~, PCI~/US91/0{iO79 ~ 1 2 ~ 2 ~ 3 10M13T~IOD FOR IN~IBITING THE INFECTIVITY
OF HUM~N IMMUNODEFICIEMCY VIRUS
Field of the Invention This application is a continuation-in part application 15of co-pending U.S. Patent Application Serial No. 287,270 filed December 20, 1988.
The government has rights to this invention by virtue of funding from U.S. Pu~lic Service Grant No. ROl-AI28194.
One aspect of this invention relates to methods and compositions for influencing the immunogenicity of human im-munodeficiency virus ~HIV) antigens and more specifically to methods and compositions for raising antibodies that inhibit the propagation of HIV infection.
Another aspect of this invention relates to antibodies that have the ability to inhibit such propagation.

Back~round of the_Invention Acquired Immunodeficiency Syndrome (AIDS) is believed to be caused by a retrovirus called human immunodeficiency virus t~IV), also known as ~TLV III or hAV. This syndrome is considered responsible for a variety of immunologic abnor-malities including but not limited to the depletion and/or selective infection by this vixus of helper/inducer (T4) lymphocytes, which results in impairment of the helper/inducer T cell function in affected individuals and in eventual inhibition of normal immune response in such individuals.
The viral envelope (~IV env) includes a population of glycoproteins tcalled gp 160) anchored in the viral cell .. .; . ; . ,.
.
.

W092/05799 PCT/US9l/06079 2 ~ 2 !~
membrane bilayer via their C-terminal region. Each glycoprotein contains two segments: the N-terminal segment, called gpl20, which protrudes from the membrane into the surrounding medium; and the C-terminal segment, gp 41, which spans the membrane.
It has been reported that ~IV infects CD4+ T lym-phocytes by a sequence of events beginning with attachment of gpl20 to its cellular receptor CD4, a nonpolymorphic surface glycoprotein described in Maddon, P.G. et al, Cell, 42:93-104, 1985; Clark, S., et al, P.N.A.S. (USA), 84:1649, 1987 and Germain, R.N., Cell, 54:441-444, l9R8. It is believed that the binding of gpl20 to CD4 then tri~gers anchorage of gp 41 to the lymphocyte membrane, an event in turn followed by cell fusion between the virion and the target lymphocyte; Kowalski, M. et al, Science, 237:1351-1355, 1987; Gallaher, W.R., Cell, 50:327-328, 1987 and Gonzalez~Scarano, F. et al, AIDS Research & Human Retroviruses, 3(3):245-252, 1987.
In addition, HIV infection is propagated by direct lymphocyte-lymphocyte fusion between virus-infected cells (which have been shown to express gpl20 and gp 41 on their surface) and uninfected CD4~ cells. This fusion takes place even in the absence of free HIV in the surrounding medium.
Lifson, J.D. et al, Science, 232:1123-1127, 1986; Sodroski, J.
et al, Nature, 322:470-474, 1986; and Lifson, J.D. et al, Nature, 323:725-728, 1986.
The properties and isolation of gpl20 from HIV par-ticles and its sequencing from different HI~ isolates are well-known and have been extensively described e.g. in the foregoing references. The preparation of gpl20 via reco~binant DNA
techniques has been described in Lasky, L.A. et al, Science, 233:209, 1986 and in published European patent application of Genentech, Inc. published on August 24, 1988, Serial No.
279,688 (based on U.S.S.N. 155,336 ~iled 02/12~88) naming ~erman, P.W.; Gregory, T.J.; Lasky, L.A.; Nakamura, G.R; et al.
as inventors, as well as in Landau, N.R., et al., Nakure, 334:159-162, 1988; and Lasky, L.A., et al., Cell, 30:975-985, 1987. Finally, isolation of gpl20 (whether native or recom-:~ . , , - ' ' ' - , ~, . .
.

binant) has been described in Essex, U.S. Patent NoO 4,725,669 (2/16/88).
CD4, the cellular receptor of gpl20l has been isolated from lymphocytes. Synthetic (soluble) and recombinant CD4 have been described in Smith, D.H. et al, Science, 238:1704-1707, 1937. Other methods as well as properties of CD4 have been descri~ed in Jameson, B.A. et al, Science, 240:1335-1339, 1988;
Fisher, R.A. et al, Nature, 331:76-78, 1988; ~ussey, R.E. et al, Nature, 331:78-81, l9B8; Deen, K.C. et al, Nature, 331:82-84, 1988; Traunecker, A. et al., Nature, 331:84-86, 1988; and Lison/ J.D. et al, Science, 241:712-716, 1988.
Two regions of HIV env (gp 160) exhibit the highest immunogenicity that has been observed against this protein:
The first region has been placed between residues 307 and 330 of gpl20 and represents an immunodominant epitope since animals immunized with whole gp 160 or gpl20 (or with fragments of gpl20 containing the epitope) produce high titers of HIV-neutralizing antibodies (i.e., antibodies that inhibit virion-lymphocyte fusion). This immunodominant epitope is situated in a highly variable segment of gpl20 that varies from isolate to isolate and, as a result, the antibodies are also isolate-specific. This limits their utility in immunological studies and in therapy against (or prevention of) HIV infection. Also, sera from HIV-infected humans do not contain high titers of these antibodies.
Instead, sera from infected humans contain HIV-neutra-lizing antibodie~ (specifically antibodies that inhibit the binding of viral gpl20 to lymphocytic CD4) directed to other epitopes of gpl20 which have not been precisely identified (though it is thought to be proximal to the CD4-binding site of gpl20); Lasky, L.A. et al, C~ll, 50:975-935, 1987. The antibodies are group- and not isolate-specific, which indicates that this second epitope is located in a more conserved domain of gpl20. However, animals immunized with gpl20, gp 160 or various fragments of gpl20 have not produced HIV-neutralizing antibodies. Thus, these other epitopes of gpl2a appears to be less immunogenic.

., .. ~ . . :

W092/057~9 PCT/US91/06~79 ~2 a ~ 4 Accordingly, one object of this invention is to provide significant amounts of antibodies that neutralize the infec-tivity of HIV virus ti.e./ inhibit its ability to invade T4 lymphocytes).
Another object of this invention is to provide an-tibodies that prevent HIV-induced cell fuc;ion between healthy T4 lymphocytes and lymphocytes infected with HIV.
Yet another object of this invention is to use such antibodies to improve the understanding of the pathogenesis of HIV and to inhibit propagation of ~IV infection in human T4 lymphocytes.
Still another object is to provide compositions and methods for raising such antibodies.
Further objects of this invention include use of such antibodies in research to elucidate the structure and function of HIV components and the mechanism of HIV infectivity and use of such antibodies in the passive or active immunization of humans for propylactic or therapeutic purposes.
These and other objects of the invention will be apparent to persons skilled in the art in light of the present specification, accompanying claims and appended drawings in which:

B ef Descrl~tion of the Drawinqs Figure l A-C is a series of plots comparing the magnitude and cell fusion-blocking ability of various antibody titers in successive bleedings of animals immunized with CD4, gpl20 and a combination of CD4 and gpl20.
Figure 2 is a series of graphs illustrating the time course of an experiment with three groups of mice respectively injected with CD4 (A,D,G), CD4-gpl20 complex (B,E.H) and gpl20 (G,H,I). The weekly serum samples were assayed individually for gpl20-binding antibodies (A,B,C), CD4-binding antibodies (D,E,F), and syncytia-blocking capacity (G,H,I). The mice immunized with the complex showed a somewhat higher anti-gpl20 response than those immunized with gpl20 alone (panels B ~ersus C); a markedly lower titer of CD4 binding as compared to those ~, :
, . . ~ -W092/0~799 PCT/USg1/~079 ~ 2~2~3 , 5 receiving CD4 alone (panel E versus D); and a significantly higher syncytia-blockin~ response (panel El versus G).
Figure 3 is a graph showing titration of 11-week serum from 4 mice injected with CD4-gpl20 complex for IIIB and RF
syncytia-blocking capacity. The parallel behavior of in-dividual sera in the two tests suggests that the antibodies are directed at group-specific determinants.
Figure 4 is a graph showing the effect of antibodies on rCD4-phosphatase binding to solid phase rgpl20. In panel A the antibody is OKT4A; in panel B, 94; in panel C, OK~4, in panel D, 55. CD4-phosphatase concentrations: closed circles = 10 micrograms/ml; open circles - 3 micrograms/ml~ Ordinate:
OD405/60 min. Abscissa: antibody concentration, from left to right, 0, 0.3, 1, 3, 10 micrograms/ml.
Figure 5 is a graph showing effect of rgpl20 on rCD4-phosphatase binding to various antibodies captured on solid phase goat anti-mouse Ig. In panel A the antibody is OKT4A; in panel B, 94; in panel C, OKT4; in panel D, 55. Concentration of rCD4-phosphatase: closed circles = 10 micrograms/ml; open circles - 1 microgram/ml; squares = 0.1 microgram/ml. Or-dinate: OD40s/60 min (A,B) or OD40s/120 min (C,D)- Abscissa:
concentration of gpl20 (0, 0.4, 4, 40 micrograms/ml).
Figure 6 is a graph showing the effect of rgpl20 on phosphatase-labeled antibody binding to solid phase rCD4. In panel A, the antibody is 94; in panel B, 55. Open circles = no rgpl20; closed circles = rgpl20 at 0.1 microgram/ml; squares =
rgpl20 at 1.0 microgram/ml. Ordinate: OD40s/60 min. Abscis-sa: antibody concentrations (0.1, 1, 3, 10 micrograms/ml).

Summary of the Invention It has now been discovered that antibodies raised pursuant to immunization with a complex of (a) the HIV antigen gpl20 and (b) CD4-Immunoadhesin have the ability to inhibit propagation of HIV infection to healthy human T-lymphocytes (by inhibiting lymphocyte-lymphocyte fusion and in~asion of lymphocytes by HIV).

:
... . . ~ :

, :
.:

W092l05799 PCT/US91/06079 ~2~ ~2~ 6 Detailed Desçription of the Invention As used in this disclosure, each of the following terms shall have the mea~ing ascribed to it below.
"gpl20" shall mean not only gpl20 itself b~lt also any other molecule that binds with CD4 in a si;milar manner and that when so bound has the same conformation. For example, the te~n will include fragments of gpl20 that bind to CD4 as well as analogs and derivatives of gpl20 that possess the ability to bind CD4 and to generate anti~odies with the cell-fusion blocking ability of the antibodies of the present invention7 In addition, this definition of gpl20 shall include gpl20 from any HIV isolate since methods for sequencing this protein are known and are independent of the particular isolate of ~IV from which the native protein is derived.
"CD4" shall mean CD4 and/or fragments, derivatives or analogs containing the gpl20-binding site of CD4, such as CD4 isolated from the lymphocytic surface and CD4 or CD4 deriva-tives (such as CD4-Immunoadhesin and analogs (e.g. soluble CD4) produced by synthetic (including but not limited to recombinant DNA) techniques.
"gpl20/CD4 complex" shall mean a bimolecular (i.e., noncovalent) conjugate or complex between gpl20 and CD4 (or between gpl20 and antiidiotypic antibody bearing a CD4 internal image). A simple mixture of gpl20 and CD4 contains this complex because of the high affinity between gpl20 and its cellular receptor~ It is not necessary that the complex be made of isolated gpl20 and CD4. For example, whole lymphocytes having gpl20 bound to the CD4 on thPir surface are envisioned as a possible form of an immunogen encompassing the gpl20/CD4 complex of the present invention. The native lymphocytes of an HIV-infected human would not act as such an immunogen in that human because of the ability of autologous CD4+ lymphocytes to act as antigen-presenting cells (APC), as reported by Lanzavec-chia, A. et al/ Nature, 334:530-532, 1988. (Consequently, the gpl20/CD4 complex even if it exists on the surface of these human lymphocytes could not act as an immunogen.) ~owever, antibodies of the type of the present invention could be ; , .
: .

' . ~' ' ~ ~: .

W092t~799 PCT/US9~/06079 ~ 7 2 ~
induced in a human by immunization with the foregoing complex in one or more of its forms contemplated herein.
"Antibodies of the present inven~ion" or ~the present antibodies" shall mean (a) polyclonal, group~specific an-tibodies raised by immunization with the qpl20/CD4 complex andcapable of inhibiting lymphocyte fUSiOII; and/or (b) monoclonal anti~odies raised against this immunogen and possessing the same fusion inhibiting property.
"Neutralization of HIV" shall meaLn inhibition of the lQ ability o~ ~IV to bind to and invade a susceptible lymphocyte.
"Cell fusion inhibition" shall mean inhibition of the ability of HIV-infected lymphocytes to for~ syncytia with healthy CD4+ lymphocytes (i.e. lymphocytes possessing surface CD4) in the absence of free HIV.
As stated above, although HIV-infected humans do produce antibodies to gpl20 that inhibit the binding of HIV to CD4 and hence prevent invasion of CD4~ T lymphocytes (i.e., lymphocytes possessing surface CD4) by the virus, the titers of such antibodies are not very high either in absolute terms or by comparison to these individuals' titers of nonneutralizing antibodies.
The majority of sera from seropositive indi~iduals contain some HIV-neutralizing antibodies directed against an epitope located within a more conserved area of gpl20, and distinct from the immunodominant epitope. Such antibodies are group-specific and constitute useful investigative tools in the pathogenesis of HIV and in research efforts to produce abate-ment or prevention of ~IV infection. However, even though these human group-specific antibodies neutralized HIV in vitro and inhibited lymphocyte fusion in vitro, they failed to inhibit lymphocyte invasion by free ~IV introduced in the serum of primates passively immunized with the human antibodies.
This f~ilure has been attributed to the fact that these antibodies are not directed towards an immunodominant epitope and/or to the insufficient affinity of these antibodies for their antigenic determinant compared with the extremely high affinity of gpl20 for its lymphocytic receptor CD4 (the :, ~. . ~ . ' ':

WOg2/~5799 PCT/US9~/~079 29~2~3 8 ~-equilibrium constant of the latter is of the order of 10-9 M
which is significantly higher than the affinity of most antigens for their antibodies). Also, nonprimate mammalian immune systems do not recognize this epitope of gpl20 and consequently animals do not produce such antibodies. There-fore, the need still exists for different antibodies that (a) are group-specific, (b) neutralize the virus and inhibit cell fusion and (c) do not compete with the high-affinity binding of gpl20 to CD4.
In accordance with the present invention, comparative experiments were conducted in which animals were immunized with (a) CD4, (b) gpl20, and (c) a mixture of CD4 and gpl20 which is believed to result in a bimolecular (noncovalent) conjugate or complex between CD4 and gpl20 because of the high affinity (10-9 ~) of gpl20 for its receptor.
Animals immunized with CD4 alone exhibited a very high titer of anti-CD4 antibodies. ~owever, only a relatively very small portion of these antibodies ~nhibited the ability of HIV
infection to spread among T4 lymphocytes as measured by a cell fusion assay which tests the ability of HIV-infected lym-phocytes to fuse with healthy lymphocytes bearing CD4.
In parallel, another group of animals were immunized with gpl20 alone. The anti-gpl20 immune response was sig-nificant (although not nearly as high as the anti-CD4 response) but the anti-gpl20 had no HIV-neutralizing ability (as measured by the same assay).
A third group of animals were immunized with a mixture of CD4 and gpl20. The immune response showed the presence of anti-CD4 antibodies ~although the titer was significantly lower compared to the anti-CD4 elicited by immunization with CD4 alone) and anti-gpl20 antibodies (in amounts comparable to thoss elicited by immunization with gpl20 alone). However, the sera showed a very high titer of cell-fusion inhibiting antibodies.
The antibodies of the present invention are not simply anti (CD4), since the cell-fusion inhibiting titer does not correlate with the titer obtained by the immunization with CD4 , . . .
, " - , ' : ,, ~ 2~`7 ~3 alone and since anti CD4 do not possess the cell-fusion inhibiting ability of the antibodies of the present invention.
For the same reasons, the antibodies of the present invention do not appear to be simply anti gpl20. Furthermore, the present antibodies are not the same as previously observed antibodies which inhibit the event of binding between CD4 and gpl20 because the titers of the present antibodies do not correlate with the titers of the previously observed binding-inhibiting antibodies. Finally, the antibodies of the present invention are not elicited except in the presence of the gpl20~CD4 complex, as will be illustrated below, and therefore constitute novel and distinct entities.
The HIV-neutralizing ability of the antibodies of the present invention has been measured by a cell-fusion assay developed by Skinner, M.A. et al, J. Virol., 62:4195, 1988.
This assay exploits the ability of HIV-infected lymphocytes to form syncytia (fused cells) with healthy but HIV-susceptible (CD4~) lymphocytes (in the absence of free HIV), a process that starts by the expression (on the surface o infected lym-phocytes) of gpl20, which then binds to the CD4 of heal-thy cells. The assay thus compares the ability of HIv-infected lymphocytes to form such syncytia under experimental conditions with the ability of the infected cells to form syncytia under control conditions, i.e. in the absence of a potential fusion-inhibitor.
This assay is a stringent indicator of HIV-infection inhibition ability by a given inhibitor and in particular by an antibody. Another fusion assay is available that measures the a~ility of free ~IV to invade lymphocytes ~i.e., the fusion takes place between the virion and the lymphocytes). However, although many antibodies are available that can inhibit lymphocyte infection by free virus, very few antibodies also inhibit lymphocyte-lymphocyte fusion in the absence of free virus. On the other hand, most, if not all, antibodies that inhibit cell fusion also inhibit infection by Eree virus.
For these reasons, the performance of the present antibodies in the lymphocyte-lymphocyte cell fusion assay - - -.: . - , .. ,, '' ' ' , : . ' ' ,:

~ .... .: , ' 2a~2~'~ 10 ~

constitutes good evidence of the ~IV-inhibiting ability of such lymphocytes.
The antibodies of the present invention may be thus used to inhibit both invasion of lymphocytes by ~IV and spread of HIV infection via lymphocyte fusion and hence constitute good candidates for passive immunization (both prophylactic and therapeutic). Such passive immunization may be combined with cytotoxic agents or coadministered with other HIV inhibitors, such as Iicin toxin A chain which when linked to recombinant CD4 has been shown to be selectively toxic to 'nfected T-lymphocytes. Till, M.A., et al, Science, 242:1166-1168, 1988.
In addit;on, immunization of susceptible mammals with a combination of gpl20/CD4 is expected to improve the effective-ness of the immune response of these mammals against HIV infec-tion both preventively and therapeutically.
Other uses for the antibodies of the present invention include use in screening tests for the presence of the gpl20/CD4 complex; as research tools to identify new epitopes of gpl20 and specifically epitopes that are available only by changes in the conformation of gpl20 by CD4 binding and/or vice versa. Monoclonal antibodies in accordance with the invention and especially human monoclonal antibodies represent a preferred form of the present invention and can be used for passive immuniæation in humans.
In addition, many other uses are contemplated as will be apparent to those of ordinary skill in the art.
Amounts used for immunization in mammals can generally vary from a~out 10 to about 100 micrograms CD4/kg body weight and from about 13 to about 130 micrograms of gpl20/kg body weight. The foregoing amounts are based on the assumption that equivalent amounts of CD4 and gpl20 will be used, which is preferred but not necessary. It will be appreciated of course by those of ordinary skill in the field that an excess of one or the other constituent of the complex (i.e., an amount in addition to that sufficient to form a complex with the avail-able amount of the other constituent) is not fatal to the operability of the present invention but an equimolar mixture ,~

~ 2 c~ 3 of CD4 and gpl20 is preferred.
Well-known immunization protocols may be used with or without adjuvant. One preferred protocol involved use of the complex in complete Freund's adjuvant as s,et forth in Example 1 (of course any other well-known immunization adjuvant can be used or adjuvant can be omitted altogether).
A sin~le immunization is sufficient, but immunization may be repeated 4 weeks after the first injection with an additional optional booster 4 weeks after the second injection in incomplete adjuvant (or without adjuvant). In addition, the immunogenic ability o~ the complex of the present invention can be boosted by use of carriers such as tetanus toxoid, keyhole limpet hemocyanin, vaccinia virus, diphtheria toxoid, etc., as i5 well known in the art.
Concentrations of the antibodies of the present invention effective in inhibiting lymphocyte-lymphocyte fusion should be at least sufficient to prevent successful carrying out of the sequence of events that lead to either invasion of the lymphocytes by free HIV or lymphocyte-lymphocyte fusion between any available gpl20 (whether on the viral envelope or on the surface of an infected lymphocyte) and CD4~ lymphocytes.
The upper limit of the effective concentration is irrelevant ln vitro. In vivo, the upper limit of the effective antibody concentration may be limited by factors outside the binding mechanism, such as on immune response of the host against the antibodies.
The antibodies of the present invention (whether monoclonal or polyclonal) may be purified by well-known techni-ques for purification of immunoglobulins, including but not limited to use of precipitation techniques (such as ammonium sulfate precipitation) and/or immunoaffinity chromatography methods (with an antigen as the adsorbent) wherein the desired antibody is preferentially bound to the column or excluded in the eluant; protein A sepharose chromatography; Affigel-blue chromatography; high performance liquid chromatography and combinations of these techniques.
Monoclonal antibodies to the complex of the present .. , . . . . . - .
.,:' . .: ' .- :

- :

WOg2/05799 PCT/US91/~079 20'~12~'~ 12 ~
invention may be raised according to well-known techniques, such as those described in Kohler and Milstein, Nature, 256:495, 1975 and in Goding, infra.
In addition, it is believed that human monoclonal antibodies may be raised from immortalized human l~mphocytes sensitized against the complex o~ the present invention ln vitro (as described in Reading, C.L., 1982, J. Immunol. Meth., 53:261; Hoffman, M.K. et al in Engleman, E.G. et al (Eds) Human Hybridomas and Monoclonal Antibodies, Plenum Press, New York, 1985, p. 466; Borrebaeck, C.A.K., Trends in BiotechnoloqY, 4:147, 1986) or in viv_ ~by immunizing humans with the complex as described herein and selecting ~-cells with the appropriate specificity). ~uman C3D+ peripheral blood mononuclear lym-phocytes can be exposed to Epstein-Barr virus for immortaliza-tion. Epstein-Barr virus (EBV) can be obtained from the filtered supernatant of the marmoset cell line B95-8 (Miller, G., and Lipman, M., 1973, PNAS (US~) 70:190) available from the ATCC under Accession No. CRL-1612. Infected lymphocytes are then washed and cultured in RPMI-1640 medium in the presence of fetal bovine serum, glutamine, penicillin and streptomycin in 96-well plates at 104 cells per well. After screening for antibody production, positive cultures can be expanded and cultured further. Cultures with supernatants showing specific reactivity to the complex of the present invention can be subcultured on feeder layers of GK5 human lymphoblastoid cells (derived from GM1500 as described in Kearnyl J., N. Enql~ J-Med./ 309:217, 1983) irradiated with 3000 rads of gamma-radiation. Stable clones can then be subcultured at low densities (10-100 cells) on feeder cells and the subcultures can be expanded. The specificity of the antibodies can be tested by the Skinner et al assay referenced and described herein.
Antiidiotype antibodies can be obtained by immunizing syngenic mice with anti-CD4 monoclonal antibody. A most efficient protocol of immunization is to inject the monoclonal antibodies four times a day at weekly intervals. The total amount injected is 13 doses of 20 micrograms per mouse at each W092/057g9 ~ 9 ~ 1 2 ~ 3 PcT/us91/o6o79 "il 13 time. In particular, on day 1, the antibody is injected coupled to RLH (keyhole limpet hemocyanln) in complete Freund's adjuvant; on day 6 the antibody is injected co~pled to KLH in incomplete Freund's adjuvant ~IFA); on day 13 the antibody is injected coupled to KLH in saline; on day 27 the antibody is injected alone in incomplete Freund's adjuvant. To obtain the monoclonal antibodies, the animals are boosted after 4 to 12 weeks after the last injection by injecting again the monoclona~ antibody in incomplete Freund's adjuvant in saline.
Spleen fusion with an appropriate myeloma partner is performed 3 days after the boost. Fusion and hybrid growth and ~election are then performed in accordance with well-known techniques.
The antiidiotypic antibodies bearing CD4 internal image can he identified by testing for binding to gpl20 (e.g., by radioim-muno assay, enzyme-linked immunosorbent assay, etc.) and purified using well-known methods including immunoaffinity chromatography using gpl20 as the adsorbent.
Again, the complex of the present invention when antiidiotype antihodies ("AA") are used can be formed by mixing gpl20 and AA preferably in equimolar amounts and waiting for the two constituents to complex with each other.
The materials used in the present invention may be purified from natural sources or synthesized by well-known Irecombinank and other~ techniques as described above. In addition, recombinant CD4 and CD4-Immunoadhesin are available from Genentech, Inc. and so is recombinant gpl20. If not already obtained in purified form, these materials should preferably be purified prior to use in immunization. Techni~
ques for purification are well known; see, e.g., U.S. Patent No. 4,725,669 of Es~ex et al. issued February 16, 1988.
To be effective in inhibiting cell invasion by HIV and cell fusion, the antibodies of the present invention must be present in the vicinity of infected T4 lymphocytes (or ~IV-susceptible uninfect~d T4 lymphocytes) before the gpl20 ex-pressed on the surface of infected T4 lymphocytes ~or the gpl20 on the surface of the virion) binds to its receptor (CD4) on the surface of an uninfected lymphocyte (or on the surface of a ,"

:
. . . . ~ ,:

2~ 3 14 ~
susceptible lymphocyte in the case of infection by free ~irus).
Preferably, the antibodies of the present invention will be present in such vicinity prior to encounter between the gpl20-bearing infected lymphocyte (or virus) and the target CD4+
lymphocyte.
The inYention is further described in the Examples that follow. The purpose of these examples is to illustrate the present invention and not to limit its scope.

Example 1: Immunization of Mice and Antibody Titers Three groups of 4 mice (each weighing 25 g) were im-munized as follows:
The first group was injected once intraperitoneally li.p.) with 10 micrograms of CD4 in 0.2 ml of complete Freund's adjuvant.
The second group was injected once i.p. with 12.5 micrograms of gpl20 in 0.2 ml of complete Freund's adjuvant.
The third group was injected once i.p. with a mixture of 10 micrograms of CD4 and 12.5 micrograms of gpl20 (previously incubated together for 30 minutes) in 0.4 ml of complete Freund's adjuvant.
The mice were bled weekly over a three-month period and the sera were monitored for CD4- and gpl20-binding titers. In addition, the sera were monitored for their ability to block lymphocyte-lymphocyte fusion, according to an assay described in Example 2 belvw, and for their ability to inhibit binding between CD4 and gpl20, according to an assay described in Example 3.
CD4 and gpl20 titers were determined as follows:
Binding antibodies were assayed by ELISA. Microtiter plates were coated with the appropriate antigen (CD4 or gpl20, respectively~ at a concentration of 3 micrograms/ml in car-bonate buffer 0.1 M, pH 9.6 overnight at 4C, washed and blocked by incubating them with 1% bovine serum albumin (BSA) in phosphate buffer saline (PBS) for 45 min. at room tempera-ture. A 0.1 ml sample of test serum at various dilutions in PBS were then added and incubated for two hours at room , . . . :

W092/0~799 2 ~ 2 ~ 3 PCT/US9~/~6079 ~; 15 temperature. The plates were then treatecl with goat-antimouse antibodies labelled with alkaline phosphatase and finally incubated with PNPP (disodium p-nitrophenyl phosphate) at a concentration of 1 mg/ml in diethanolamine buffer, pH 9.8.
Each step was followed by a wash in PBS-Tween 20 buffer.
Absorption at OD405 was determined 1 hour after incubation with substrate in a Titertek Multiskan MCC 340. The OD readings were converted to micrograms/ml of undiluted serum by inter-polation to OD-0.5 and assuming a l:l ratio between bound mouse antibody and ~oat and mouse tracer.
Figure lA is a plot of the anti-gpl20 titer (in microgramstml serum) of sera elicited by immunization of mice with gpl20 (solid line), CD4 (broken line) and the gpl20/CD4 complex (- - - -) for each weekly bleeding.
As evident from Figure lA, immunization with gpl20 alone elicited anti-gpl20 antibodies; immunization with CD4 alone elicited no anti-gpl20 antibodies, immunization with the complex also elicited anti-gpl20 antibodies (i.e. the complex generated formation of antibodies against gpl20 alone).
Figure lB is a plot of the anti-CD4 titer (in micrograms/ml serum) of sera elicited by immunization with CD4 alone (broken line), gpl20 alone (solid line) and complex (- -_ ) .
In Figure lB, immunization with CD4 alone elicited a very high ant~-CD4 immune response; immuni~ation with the complex elicited a lower but significant anti-CD4 response; and immunization with gpl20 alone elicited no anti-CD4 response except after week 7 when a small antiidiotype (anti-(anti gpl20)) response was observed.
Example 2: Iymphocyte Fusion Assay The assay to determine the immune serum capacity to block lymphocyte-lymphocyte fusion was described in Skinner, supra. Ten microliters of different dilutions (at least 1:10) of the test serum were distributed in Costar 96 A/2 (half~well~
plates.
5x103 or 10x103 infected cell partners (from CBJIIIB or . .
. , ,. ~ :
, ' ' -W092/0~799 PCT/US91/06079 2~9:~2~ 16 ~
CEM/RF lymphocyte cell line publicly available from National Institute of Allergy and Infectious Disease, Reagent Program respectively infected with ~TLV IIIB or HTLV RF isolate available from Dr. Gallo at the National Institute of ~ealth) were added to each well contained in 40 microliters of culture m~dium. Then, 7x104 uninfected CD4-~ human lymphocytes (Molt4 available from American Type Culture Collection, Rockville, MD, Accession No. CRL-1582) in 40 microliters of culture medium containing fetal bovine serum were ~dded to the wells and the plates were incubated at 37OC in a 5% CO2 atmosphe~e for 20-24 hours. The plates were then read for the occurrence and the number of lymphocyte syncytia in an inverted microscope at a 40-fold enlargement. Giant cells having a size of at least five times the area of normal cells were scored as syncytia produced by cell fusion. In the absence of any antiserum, the syncytia were 50-80/well (control). Fusion-blocking units (FBV) were calculated hy converting the percent decrease in syncytia scored relative to the control value and taking into account the serum dilution (1 FBU is defined as the amount of antibody that reduces the number of syncytia to 50% of the control value). The results, in FBU, are plotted in Figure lC.
Again, the solid line represents FBU achieved by immune sera of mice immunized with gpl20 alone. In Figure lC, this value is essentially the same as the control. The broken line represents the fusion-blocking ability of immune sera elicited by immunization with CD4 alone. In Figure lC, this value is positive but not very high. This is attributable to the fact that anti-CD4 will bind some of the CD4 on the uninfected lym-phocyte surfaces and thus prevent the gpl20 of the infected lymphocytes from binding to the CD4.
In Figure lC, the line ----- - represents the fusion-blocking ability of immune sera elicited by immunization with gpl20/CD4 complex. The FB~ of this sera starts at about that of the anti-CD4 sera in week 1 and extends to about 50 times that of the anti-CD4 sera.

f -` 17 2~912~3 ExamPle 3- Monoclonal_Antibodies Monoclonal antibodias will be raised by immunizing mice with complex in accordance with the method of Ex~mple 1 ~except that two injections can be used spaced 4 weeks apart) optional-ly with a booster using incomplete Freundi's adjuvant 4 weeks after the second injection. Three days after the last im-muni~ation, spleen cells will be obtained, purified and fused with myeloma cells. The spleen of the mouse with the highest fusion-blocking ability will be excised using well-known dissertion techniques. A single-cell suspension ~ill be made up by teasing the spleen as described in Goding, J.W., Monoclonal Antibodies: Principles and Practice, Academic Press, Inc., New York 1983, pp. 50-97 and specifically on p. 64. The spleen cells will be harvested by centrifugation (e.g. 400xg for 5 min.) and wash~d. Erythrocytes will be removed by ammonium chloride lysis followed by centrifugation. The spleen cells will be counted and approximately 108 cells will be used for fusion with commercially available mouse myeloma cells te.g. SP 20 from American Type Culture Collection under Accession No. CRL-1581). (2 - 3)x107 myeloma cells will be mixed with the spleen cells in serumi-free media and centrifuged at 400xg for 5 min. Any remaining medium will be removed by suction. The cell pellet will be suspended in 0.5-1 ml of warm fusion medium containing 10 g of 50% w/v PEG (m~WO 1500) and 10 ml Dulbecco's modified Eagles' Minimum Essential Medium, pH
7.6. The mixture will be stirred, centrifuged and resuspended in fetal bovine serumi-containing medium using normal spleen cells as feeders. The cells will then be exposed to XAT
selective media and grown in such media in an atmosphere containing 5% CO2. The cultures will be pulled and fed.
Hybrids will be growing and screenable at 10-15 days after incubation begins. Positive clones (i.e. clones secreting fusion-blocking antibody) will be identified and recloned until their secretion of the desirable immunoglobulin is steady and reliable. Monoclonals will thus be obtained. The Skinner et al assay can be used to determine specificity of the desired antibody.

' .

W092/05799 PCT/US91/~079 2C~3 18 Difficulty in obtaining such monoclo~als is not expected because the titers of fusion-blocking antibody are relatively high as demonstrated in previous Examples. If desired, spleen cells se~reting anti-CD4 and/or anti-gpl20 will be separated first before fusion to maximize the probability of obtaining fusion-blocking monoclonals~

Example 4:
It is known that the HIV infection of CD4+ cells can be prevented by antibodies specific for the gpl20 binding site on CD4, defined as the V1 domains of CD4 involved in the initial CD4-gpl20 binding event, i.e., the region homologous to CDR-2, amino and residues 41-52 (Peterson, A., et al., Cell. 54:65, 1988; Landau, N.R. et al., Nature. 334:159, 1988; Clayton, L.X., Na~ . 335:363, 1988; Jameson, B.A. et al., Science.
240:1355, 1988), and in part CDR-3, amino and residues 83-92 (Nara, P~L., et al., Proc. Natl. Acad. Sci. USA. 86:7139, 1989;
Sattentau, Q.J., et al., J. Exp. Med. 170:131~, 1989). These antibodies prevent infection by sterically interfering with the 20 binding site.
Poly- and monoclonal antibodies from mice immunized with CD4 complexed to gpl20, their binding characteristics and capacity to prevent the formation of ~IV-dependent syncytia, are described below. The data presented below demonstrate that there are epitopes on CD4, unrelated to the binding site of gpl20, which antibodies can recognize, thus affecting post-virus-binding events that usually lead to infection.
The following materials and methods were used as described below.
Recombinant molecules. Solllble recombinant CD4 (rCD4), CD4 immunoadhesin (containing V1 and V2 domains of CD4 spliced to the CH2 and CH3 domains of human IgG as described in Byrn, R.A., et al., Nature 344: 667, 1990, and recombinant gpl20 were obtained from Genentech Inc., South San Francisco, CA.
35 Animals. Balb-c female mice 10-15 weeks old (Jackson Labs, Bar ~arber, ME) were used both for immuni~ation and for production of monoclonal antibodies.

W092/05799 PCTtUS91/06079 ~ 19 ~ 2~
Immunizations. Mice were injected intraperitoneally (i.p.) with antigen emulsified in Complete Freund's Adjuvant (CFA, Difco, Detroit, MI). The antigens and doses were a) CD4, 16 micrograms/mouse; b) gpl20, 12.5 micrograms/mouse, c) CD4-gpl20, 16 micrograms CD4 and 12.5 micrograms gpl20, thoroughly mixed and incubated for 20 min, and then emulsified in CFA.
The mice were bled prior to immuni~ation and every week after, for 13 weeks. Serum samples were stored at -20Co Enzymes and substrates. Alkaline phosphatase and glutaral-dehyde, used to label antibodies for Enzyme Linked Immunosor-bent Assay (ELISA) tests, and substrate paranitrophanyl-phosphate (PNPP) were acquired from Sigma Chemicals (St. Louis, MO).
Monoclonal antibodies. The monoclonal antibodies (mAbs) mentioned below are F-91-36, F-91-55 and F-91-94, hereinafter called 36, 55 and 94 respectively. They were derived from a fusion of mouse 91, immunized with CD4-gpl20 complex as described above. All these mAbs axe IgGl subclass.
Antibodv bindinq of qpl20 and CD4. ELISA tests were performed by coating plates with 10 micrograms/ml gpl20 or, respectively, with 3 micrograms/ml rCD4 and incubating them with serial dilutions of the mouse sera starting at 1:100 dilution.
Phosphatase-labeled host anti-mouse IgG (obtained from Sigma Chemicals) was used to reveal bound antibodies. Readings were performed in a Titertek automated photometer.
Inhibition of qpL20 bindinq to solid phase CD4. After coating the plates with 3-10 micrograms/ml rCD4, 50 micrograms of gpl20 (5 micrograms/ml) were added and incubated for 1 hour, +/-serial two-fold dilution of test antiserum or antibody starting at 100 micrograms/ml concentration. After washing, phosphatase-labeled monoclonal anti-gpl20 was applied and incubated 1 hr. The phosphatase activity was then measured by the rate of PNPP hydrolysisO The inhibition caused by the test antibody was expressed in percent decrease from the contxol gpl20 binding.
Effect of mAbs on CD4 bindinq_to solid phase qpl20. Plates were coated with 3 micrograms~ml gpl20. CD4-phosphatase at 10 - ~ . . .
, ~ ..

.
' . ' ', ' ' ': .. ~: ~ ~ '' ,, :
:. . --- ~
:

W~92/0579g PCT/US9l/OfiO79 2 ~ 12~ 20 ~
micrograms/ml and 3 micrograms~ml were incubated separately for one hour with PTH or with the test mAbs at concentrations 0~3, 1, 3 and lO micrograms/ml. The mixtures were then adcled to the coated wells and incubated l hr. After washing, the amount of bound CD4 was revealed by PNPP.
Effect of qpl20 on CD4 bindinq_to solid-phase captured mAbs.
In separate 96-well plates rgpl20 and phosphatase-labeled rCD4 were serially titrated by 10-fold dilution from 80 and 40 micrograms/ml respectively, in different directions on each plate. Equal volumes of each were then combined and incubated 2 hr at room te~peratures ~RT). Simultaneously with the above incubation, the test monoclonals were added to a goat anti-mouse IgG coated plate for 2 hrs at RT, then washed of excess sample. Fifty microliters of the titrated complex was trans-ferred to the captured mAb-GAM plate and incubated it RT for 2 hrs, then washed. PNPP substrate was added and color developed.
Effect of gpl20_on mAbs bindinq to solid-phase CD4. P 1 a t e s were coated with 3/micrograms/ml rCD4; 25 microliters of 0.1-1.0 micrograms/ml rgpl20 and 25 microliters of scalar con-centrations (0.3-10.0 micrograms/ml) of test mAb were added together to the wells and incubated for 2 hr. After washing, the bound mAb was revealed using goat anti-mouse Ig phosphatase-labeled antibody.
Cross-inhibltion of CD4 bindinq. Competition mapping of mAb specificities was performed by coating plates with rCD4 and layering alkaline phosphatase-labeled mAbs in the absence and presence of graded concentrations of unlabeled test mAbs whose specific binding sites are known from the literature. OKT4A
(binding in Vl~ and OXT4 (binding in V4) were obtained from Ortho (Diagnostics Systems, Raritan, NJ), anti-Leu3a (binding in V1), L83 (V1-V2), L88 tVl-V2), L120 (V4) were a gift from Dr. David Buck, Becton-Dickinson Laboratories, Mountain View, CA.
Inhibition_of sYnC~ ia formation. The cell fusion among CD4+
cell lines acutely infected by the virus requires gpl20-CD4 specific binding. The test w~s performed according to Matthews W092/05799 21 ` 3 et al. (Proc. Natl. Acad. Sci USA 84:5424, 1987). Briefly, GEM
cells chronically infected with either HThV3-:~IIB or HTLV3-RF
were used for each determination. Sera diluted 1:10 were distributed in 96-well A/2 plates (Costar). Five to 10 x 103 HIV-infected cells were added. 70 x 103 uninfected Molt 4 cells were admixed and the plates were inc:ubated overnight at 37C, after which the number of giant cells (>5 times the size of the parental cells) were counted at 40 X magnification. The mouse serum control varied from 50 to 85 giant cells per well.
Results 1. Polyclonal responses to in-iection of complexed sCD4-rq~120 in mice.
Three groups of four mice were injected once, either with rCD4 alone (Fig. 2, panels A,D,G), with sCD4-rgpl20 complex (Fig. 2, panels B,E,H), or with rgpl20 alone (Fig.2, panels C,F,I). The weekly bleedings were titrated over a 3-month period for capacity to bind CD4 (panels D-F~ and gpl20 (panels A-C), and for capacity to inhibit the formation of syncytia (panels G-I). The overall results were strikingly different among the three groups. Figure 2 displays the individual titers for each parameter and for each group. There were often multiple peaks during the response (panel D) and mice of a single group showed different timings for their peaks (panels D and H). When the timing and magnitude of the binding and neutralizing responses were examined, (a) there was no clear correlation between the titers of gpl20 binding a~d syncytia blocking (R versus H); (b) there was inverse relation-ship between rCD4 and rCD4-rgpl20 immunized groups when rCD4 ~inding titers and syncytia blocking were compared (D vs~ G and E vs. H), ~c) there was a small CD4 binding response in the group immunized with rgpl20, which could be attributed to anti-idiotypes (panel F).
2. Polvclonal syncytia-blockinq responses by mice receivinq CD4-qpl20 comDlexed are not type-specific.
The antisera from mice injected with CD4-gpl20 effi-ciently block syncytia formation caused by such widely dif-ferent HIV isolates as HTLV3-IIIB and ~TLV3-RF. Figure 3 shows . , ~ . .: . .

.
' .' W092/0~799 PCTtUS91/06079 3 22 ~
individual titrations of the sera from these mice, revealing a similar rank order o~ the individual responses against the two isolates (93>94>91>92)~
3. The sync~tia-blockin~ capacity of CD4-q~120 responders_ is absorbed by CD4 and not by qpl20.
To determine whether the high capacity of sera from complex-immunized mice to inhibit syncytia formation was due to anti-CD4 or anti-gpl20 antibodies (or to a cooperative action of the two) a series of absorption experiments was performed using c~oluble and solid phase bound antigens. The samples were subsequently monitored for changes in binding or syncytia-blocking titers. The results are se-t forth in Table 1 below.

~ ~3 ~ r9 1 2 ~ 3 able 1 Absorption of pooled sera from mice immunized with CD4-gpl20, demonstrating that syncytia blocking is associ.ated with anti-CD4.
.
CD4 binding gpl20 binding # syncytia**
Absorbent O.D.*_ (~) O.D.*~) IIIB RF
none 1.216 (100) .940 (100) 0 3-gpl20 1.180 (97) .188 (20) 0 5 CD4 .420 (34) .936 (99) 11 24 no antiserum - - - - 52 60 *O.D. 405/60 min (color developed by phosphatase hydrolysis of PNPP).
**Syncytia blocking assay as described above.
The results unequivocally attributed the syncytia-blocking capacity to antibodies that recognize the CD4 moiety of the complex rather than anti-gpl20, since absorbed anti-gpl20 serum showed unaltered syncytia blocking, while the decrease of CD4 binding was accompanied by a significant decrease of syncytia blocking with both isolates tested (Table 1) .

4. The_ca~acit~ to inhibit CD4-~pl20 bindinq in vitro corre-lates with the CD4-bindinq titer and not with the syncytia-blockinq titer.
Since the syncytia-blocking capacity observed was mediated by anti-CD4, the inverse relationship of binding and syncytia-blocking titers in mice receiving CD4 versus CD4-gpl20 had to be attributed to a difference in the fine speciicity of the anti-CD4 antibodies in the two groups. To further charac-terize the anti-CD4 antibodies, one test was to compare their relative capacity to inhibit gpl20 binding to CD4. The results are set forth in Table 2 below.

. :

.

:,, W092/05799 P~T/US9~06079 2~ 2~ ~
~3~ Difference in fine specificity distribution among polyclonal anti-CD4 anti~odies.
CD4 binding (microgram ~locking gpl20 binding Fusion blocking Ab/ml~ ~Uso/ml)* _ us0/ml~
Pool #9 (immuni~ed CD4~gpl202.4 80.0 45.0 Pool #18 (immunized CD4) 31.0 890.0 7.0 *One blocking U50 is the amount of antibody that reduces to 50~ the amount of gpl20 bound to CD4-coated wells or, respectively, the number of syncytia in the fusion test. U5p/ml was calculated by multiplying 1 U50 unit by the dllution factor for each test.
The results (Table 2) showed high inhibition titers in all mice immunized with CD4 alone and low titers in those immunized with the CD4-gpl20 complex, indicating that the blocking of syncytia by the latter appeared to be mediated by a mechanism other than prevention of binding of HIV to its receptor on the cPll surface.

5. Study of anti sCD4-rpl20 res~onses usina monoclonal anti-bodies.
In order to dissect the polyclonal response of mice im-munized with the CD4-gpl20 complex, hybridomas were produced from one of these animals, and the resulting antihodies were characterized for capacity to bind CD4, to bind gpl20 and to block syncytia formation. The results are set forth in Table 3 below.

W092/05799 P~T/US911~6~79 ~ 2~
Table 3 List of hybridomas obtained from fusion 91, hierarchically ordered according to their supernatant's capacity to bind CD4, and tested for binding gpl20 and blocking syncytia formation.
Binding Bindiny # SyncytiaSyncytia Clone CD4 spl20 at 1/2 Blocking Desiqnation IO-D.)* (O.D.)* dilution** (%) 1 135 1.363 0 144 1.208 0 148 1.149 .007 1.059 0 215 1.049 0 145 .982 0 84 .932 0 142 .930 .012 140 .937 o ~ .
210 .724 .015 185 .672 0 143 .5g9 .OOg 201 .482 0 J .437 .006 203 .398 .006 o385 .010 156 .372 ~028 58 .313 .011 224 .308 0 146 .255 .182 0 9 80 165 .157 0 36 .153 .041 0 100 172 .145 .006 9~ .136 0 0 100 223 .125 0 59 .121 0 V . 109 0 R .106 .012 48 .106 0 0 .361 0 .358 M (o023) .176 68 0 .116 116 (.012 0 16 65 [control 46]

17Q 0 ~ _ .
*O.D. 405/120 min when the supernatant was tested in ELISA on plates ooated with CD4 (respectively, gpl20~. See methods above 2~1253 ** Lower than the control are shown.
Table 3 shows an early test of 170 wells with hybridoma clones, ordered according to their CD4 binding capacity~
Thirty (i.e~, the majority) of the positive clones produced anti-CD; only four produced anti-gpl20, and the remainder were negative for both. Three of the anti-CD4 clones (and none of the anti-gpl20) exhibited capacity to block syncytia. All borderline positives (i.e. those showing less than 0.100 in the gpl20 binding) and those with partial blocking of syncytia (clones 116, 95 and 32) became negative af-ter subcloning.
Table 4 below shows a further characterization of the hybridomas when the inhibition of gpl20-CD4 binding test was performed.
Table 4 Classification of mA~s from mice immunized with CD4-gpl20 complex.
Inhibiting Binding ~inding CD4-gpl20 Group-Specific qpl20 CD4Bindinq* Syncitia Blockinq mAbs:
48, 35, 40 +

135, 144, 148 - +
75, 215, 145 210, 185, 143 94, 36 _ + +
- + _ +

*Cut-off for a positive response was 25% blocking at greater than or equal to a 1:2 dilution.
The anti-CD4 mAbs can be divided into three categories:
a) those that do not inhibit gpl20 binding and do not block syncytia; b) those that do not inhibit gpl20 binding and block syncytia; and c) those that inhibit gpl20 binding and block syncytiaO MAbs 55 and 94 were further studied, as representa-tives of the latter two categories, respectiv~ly. Both were of IgG1 isotype.

W092/0~799 ~ 3 P~T/US~1/06079 ~; 27 6. Preliminary_ma~pinq experiments. The binding site of both mAbs 55 and 94 was localized within the first two domains (V1-V2) of CD4 by binding experiments using the CD4 IgG immunoad-hesin (Genentech) which contains the two external domains of CD4 spliced to an immunoglobulin constant region (data not shown). A cross-inhibition experiment using labeled mAbs 94 and 55, was performed. The binding of these m~bs -to solid phase CD4 was tested in the presence of a s~ries of anti-CD4 monoclonal antibodies whose epitopes and binding characteris-tics are shown or partially known from the literature. Theresults are set forth in Table 5 below.
Table 5 Cross-inhibition of CD4 bindina bv mAbs 55 and 94.
_ Inhibitor mAbs mAb 55 - - +/- - - - - + - +
mAb 94 - - - - _ _ _ _ ~ _ *Cut-off point for positivity was 25% or larger decrease in hinding when the inhibitor was 6 times more concentrated than the test mAb.
The two mAbs were not inhibited by any of the tested antibodies with the exception of L83, which produced a partial but consistent competition (Table 5). These results do not provide a precise mapping of the binding sites of the two mAbs.
OKT4A and anti-Leu3a bind in the first and second domains of CD4 and the test showed that neither binding site overlaps with mAbs 5~ and 94. L83 recognizes a conformational determinant which is affected by mutations both in region 8-40 ~Vl) and region 119-188 (V2). These data taken together indicated that the fine specificity of mAbs 55 and 94 are different from most studied antibodies and different from each other.

7. Interference of qP120 with CD4 bindinq by mAbs. Preliminary experiments had shown that mAb 94 blocked gpl20 binding to CD4, while mAb 55 did not. In order to be able to detect both li .

.

, ~ ' . .

W092/0~799 PCTt~S9~/06079 2~ 28 ~ ~
inhibition and possible cooperativity between antibody and gpl20, a series of three ELISA tests were performed by which the ternary intera~tion of CD4, antibody and gpl20 was examined by keeping in turn one of the reactants in solid phase, and varying the concentrations of the other two. The results are shown in Figures 4, 5 and 6, which also include curves obtained with reference antibodies OKT4A twhich competitively interferes with CD4-gpl20 binding) and OKT4A (which does not). In Figure 4 thP binding of labeled CD4 to gpl20 was s:Lightly enhanced or non-significantly changed in the presence of increasing concentrations of 55, while it was progressively inhibited by 94. In Figure 5 the binding labeled CD4 to solid phase-captured antibody was increased 40% by gpl20 in the case of 55 at the highest concentration of CD4, but was unaffected or slightly decreased at lower concentrations of CDq. In the case of 94 there was a progressive decrease of CD4 binding in the presence of increased gpl20 concentrations, more evident when CD4 was limiting. In Figure 6, the binding of labeled 55 to solid phase CD4 was moderately enhanced in the presence of 0.1 or 1.0 micrograms/ml gpl20 while the binding of 94 was depressed in these condi.tions.
The conclusions of this series of experiments are that a) mAb 94 behaves always as an inhibitor/competitor of the CD4-gpl20 binding, and b) mAb 55 in certain conditions does not interfere with the binding in a way similar to OKT4, while in other conditions it showed a degree of cooperativity with CD4-gpl20 binding.
Example 5 The ef f ect of immunizing mice with CD4 im~unoadhesin (12.5 micrograms), rgpl20 (16.5 micrograms) and the combination of CD4 and rgpl20 (12.5 micrograms and 16.5 micrograms, respectively) was studied. Three groups of f our mice were immunized as described in Example 4 above, and their antibody response to the immunogens was examine~ 30 days post-immuniza-tion using the techniques described in Example 4 above. The results are set forth in Table 6 below.

W092/05799 PCT/US91/n6079 2~ ~ 2~j3 I M M U N I Z A T I O N
16.5 microgram 12.5 microgram 16.5 microgram gpl20 gpl20 Immunoadhesin 12.5 microgram Immunoadhesin 51 1 3.652 1 053 1 1.7 2 3.0 2 0 2 3.3 3 3.6 3 0 3 3.7 4 3.~ 4 0 4 3.4 gpl20 Binding 55 1 ~.357 1 056 1 4.8 2 2.8 2 0 2 3.6 3 3.8 3 0 3 ~.0 4 4.3 4 0 4 3.3 .
51 1 0 52 1 2.053 1 2.0 2 0 2 3.9 2 2.0 3 0 3 1.7 3 2.4 4 0 4 2.0 4 2.3 , CD4 Binding 1 0 57 1 3.35~ 1 3.8 2 0 2 4.0 2 3.7 3 0 3 4.0 3 3.9 4 0 4 3.9 4 3.6 51 1 0 52 1 053 1 1.0 ~ 0 2 0 2 1.9 -3 0 3 0 3 2.2 4 0 4 0 4 1.6 Syncitia Blocking 55 1 0 57 1 056 1 1.6 2 0 2 0 2 1.3 4 ~ 4 0 4 1.0 As can be seen in the date set forth in Table 6 above, mice receiving the complex Immunoadhesin-rgpl20 responded to gpl20 with antibody titers similar to those receiving only rgpl20 while these mice responded to CD4 with antibody titers sLmilar to those receiving Immunoadhesin alone. ~owever, mice which were immunized with Immunoadhesin-rgpl20 were the only group of mice which showed significant-titers of syncitia-blocking antibodies.
All cited literature and patents or patent applications are incorporated by reference in their entirety.

.. . . . . . . . .
-:, , ' : . ~ .
. . . . .
, .

2~ %~3 30 App-endix of Materials Sources CD4, CD4-Immunoadhesin Genentech Inc., South San Francisco, CA
Freund's Adjuvant Difco Laboratories, Surrey, England goat antimouse immunoqlobulin Sigma Chemical Co., St. Louis, Missouri gpl20 Genentech Inc., South San Francisco, CA
microtiter plates Linbro available from Fisher Scientific, lS Springfield, N.J.
96A half-well plates Cos-tar available from Fisher Scientific Springfield, N.J.
PNPP Sigma Chem~cal Co., St. Louis~ Missouri Tween Sigma Chemical Co., St. Louis, Missouri Titertrek Multiskan Apparatus Flow Laboratories, McLean, Virginia RPMI 1640 GIBCO, Grand Island, N.Y.

' . . ' ' : ' :' : ,: ~ :
~ . .

: , :, . . , ': ,' ,:

Claims (10)

WHAT IS CLAIMED IS:

1. A method for inhibiting cell fusion between human T-lymphocytes infected by human immunodeficiency virus and uninfected healthy human T-lymphocytes, the method comprising:
exposing said healthy lymphocytes prior to their binding to said infected lymphocytes to the presence of an inhibitory effective amount of antibodies that have been raised against a complex, said complex comprising (a) HIV protein gp120 and (b) a member selected from the group consisting of (i) CD4-Immunoadhesin, (ii) cell-free CD4 and (iii) an-tiidiotypic antibody to CD4 bearing a CD4 internal image.

2. A method for inhibiting invasion of human T-lymphocytes by human immunodeficiency virus (HIV) the method comprising:
prior to or after binding of said virus and said lymphocyte, exposing said virus to the presence of an in-hibitory effective amount of antibodies that have been raised against a complex, said complex comprising (a) HIV protein gp120 and (b) a member selected from the group consisting of (i) CD4-Immunoadhesin, (ii) cell-free CD4 and (iii) an-tiidiotypic antibody to CD4 bearing a CD4 internal image.

3. A method for raising antibodies that inhibit cell fusion between uninfected human lymphocytes and at least one of (a) human lymphocytes infected with human immunodeficiency virus (HIV) and (b) HIV virus particles, the method comprising:
immunizing a mammal with both (HIV protein gp120 and a member selected from the group consisting of (i) CD4-Immunoadhesin, (ii) cell-free CD4 and (iii) antiidiotypic antibody to CD4 bearing a CD4 internal image.
waiting for said mammals to mount an immune response; and collecting said antibodies.

4. The method of claim 3 comprising immunizing said mammal with a mixture of said gp120 and said member.

5. The method of claim 3 comprising simultaneously immunizing said mammal with said gp120 and said member.

6. The method of claim 3 comprising successively immunizing said mammal with said gp120 and said member.

7. An immunogenic composition comprising as an active ingredient an immunogenically effective amount of a complex of gp120 and a member selected from the group consisting of (i) CD4-Immunoadhesin, (ii) cell-free CD4 and (iii) antiidiotypic antibody to CD4 bearing a CD4 internal image.

8. The composition of claim 7 further comprising a physiologically acceptable medium.

9. The composition of claim 7 further comprising an immunization adjuvant.

10. The composition of claim 7 further comprising an immunogenicity-enhancing carrier.