WO1993019786A1 - High affinity, strongly neutralizing monoclonal antibodies against the cd-4 binding site of gp120 of human immunodeficiency virus - Google Patents

Abstract

mAbs are described which have high affinity for, and strongly neutralize, HIV-1. The mAbs are specific for an epitope that is partially or completely within the CD-4 binding site of gp120.

Description

3/19 -V

HI_GH AFFINITY, STRONGLY NEUTRALIZING MONOCLONAL ANTIBODIES AGAINST THE CD-4 BINDING SITE OF GP12^{0 O}F HUMAN IMMUNODEFICIENCY VIRUS

This application is a continuation-in-part of U.S Serial No. 07/715,336 filed June 14, 1991, which is a continuation-in-part of U.S Serial No. 07/604,146, filed October 26, 1990, both of which parent applications are hereby incorporated by reference.

This invention was made with government support under grant nos. AI26081 and AI23884 awarded by the National Institute of Allergy and Infectious Diseases. The government has certain rights in the invention.

Field of the Invention

This invention relates to monoclonal antibodies ("mAbs") that strongly recognize the CD-4 binding site of human immunodeficiency virus, type 1 ("HIV-l") .

Background of the Invention

The HIV-l envelope is composed of two glycoproteins, gpl20 and gp41. These glycoproteins are initially synthesized in virus-infected cells as a precursor called gpl60. This molecule is cleaved into gpl20 and gp41 prior to assembly of virions. The latter two glycoproteins are non-covalently associated with each other and are anchored to the viral membrane via gp41, a transmembrane protein (reviewed in (Olshevsky et al. 1990)).

One region which has been shown to elicit neutralizing antibodies is the V3 region hypervariable loop (hvl-V3) of the gpl20 (amino acids 307-330) . This* is an immunodominant epitope cluster eliciting potent neutralizing Abs in man and experimental animals (summarized in (Javaherian et al. 1990) ) .

Another epitope cluster of HIV-l envelope that has been shown to elicit neutralizing antibodies is the CD-4 binding site of gpl20. The CD-4 binding site is believed to be __ _ _

-2-

formed by non-contiguous protein loops from multiple regions of gpl20 (Olshevsky et al. 1990) . However, the precise structure of the CD-4 binding site and its contact residues have yet to be defined.

Earlier in the AIDS epidemic, there was skepticism about the protective function of neutralizing Abs against HIV-l, since such Abs could be found in seropositive individuals who went on to develop AIDS. Now it is understood that the titers of neutralizing Abs developed in humans during the course of HIV-l infection are generally not very high (Robert-Guroff et al. 1985, Weiss et al. 1985) , that higher titers of certain anti-HIV-1 Abs do correlate with a better prognosis (Robert-Guroff et al. 1985, Rook et al. 1987, Ljunggren et al. 1987, Ho et al. 1987, Devash et al. 1990) , and that deleterious Abs against HIV-l that actually enhance viral infection may be present in seropositive individuals (Robinson et al. 1990, Homsy et al. 1988, Takeda et al. 1988, Jouault et al. 1989). Furthermore, recent studies demonstrate the protective effects of certain anti-HIV-1 Abs in vivo. In one such study, passive administration of hyperim une plasma from healthy HIV-l-infected humans to ARC and AIDS patients resulted in sustained, clearance of p24 antigen and a maintenance or increase in the recipients' anti-viral Ab titer, and clinical improvement was noted in 5 of 9 recipients (Karpas et al. 1988) . In another study, chimpanzees were challenged with a stock of the IIIB strain of HIV-l that had previously been incubated with neutralizing serum Ab from an HIV-1-seropositive chimpanzee. The challenged animals were protected against viral infection, as assessed by lack of serum Ab response to virus and attempts at viral isolation (Emini et al. 1990) . Successful long term protection of two chimpanzees against HIV-l infection has been demonstrated by immunization with recombinant gpl60 followed by a V3 loop peptide (Girard et al. 1991) . In a different study, chimpanzees immunized with -3-

recombinant gpl20 and challenged with HIV-l were also protected from infection (Berman et al. 1990) . In both of these vaccine trials, significant titers of strain-specific neutralizing Ab were induced prior to challenge with virus. The protection obtained is believed to be due primarily to this neutralizing Ab, since subunit vaccines are thought to be poor inducers of cytotoxic T cells (see (Berman et al. 1990)) .

Most recently Emini et al. (Emini et al. 1992) have demonstrated in vivo protection against HIV-l using a mouse- human IgGl chimeric monoclonal antibody ("mAb") , CBl, which contains the intact variable region of a previously described antiV3_IIIB loop mouse mAb, 0.5β (Matsushita et al. 1988). Emini's 'experiments demonstrated that a "humanized" anti-V3 loop mAb with the neutralizing activity of 0.5β can safely be administered in vivo at a concentration sufficient to achieve protection from HIV-l infection.

Summary of the Invention

The mAbs of the invention are specific for HIV- envelope glycoprotein gpl20, and achieve at least about 50% neutralization in vitro of about 2 x 10* infectious units of the RF HIV-l strain at a concentration of about 1 μg/ l. This reflects a higher binding and neutralization ability with respect to this strain than has been seen with previous anti-CD-4 binding site mAbs. The mAb is specific for an epitope that is partially or completely within the CD-4 binding site of gpl20.

In another aspect of the invention, certain anti-gpl20 Abs are combined to obtain a synergistic result. In particular, the invention includes a combination of mAbs, the first being one of the novel anti-CD-4 binding site mAbs of the invention, and the second being an Ab that is specific for the V3 region of HIV-l envelope glycoprotein gpl20, the combination being effective at synergistically binding or neutralizing HIV-l. -4-

Included as well in the invention are cell lines producing mAbs of the invention.

Brief Description of the Drawings

Figure 1 shows the results of a radioimmunoprecipitation/SDS gel analysis discussed in the examples.

Figure 2 shows the results of an im unoblot experiment described in the examples.

Figure 3 is a graph of the results of a competitive inhibition experiment described in the examples.

Figures 4-6 are graphs of neutralization experiments described in the examples.

Figure 7 is a graph of combination index values calculated from the results shown in figures 5 and 6, and discussed in the examples.

Detailed Description of the Invention

The mAbs of the invention (whether singly or in the synergistic combination) can be used in many ways, including immunoassays, affinity chromatography and any other procedures that use anti-HIV mAbs. The strong recognition of HIV makes the mAbs especially useful in detection, either indirectly through a competitive assay for HIV antibody, or directly in an assay for HIV antigen. The mAbs described below are especially useful in affinity chromatography procedures for purifying a wide variety of HIV strains.

Thus, the mAbs of this invention are useful in a conventional format competitive immunoassay, such as described below, to determine antibodies, such as human antibodies, to the HIV-l epitopes recognized by the mAbs. The present invention also includes test kits to measure the presence of HIV antigens or of human Abs against the epitopes recognized by the novel mAbs. For example, a kit for a competitive ELISA contains mAbs of the invention, a solid phase on which is coated an antigen which the mAbs are -5-

specific for, and means for detecting the formation of a complex between the mAbs and the antigen. An ELISA using biotin labeled Abs can be performed similarly to the assay described below. Such an assay determines whether the sample has any antibody competing with the antibody of the invention.

In a typical direct assay the mAb is reacted with a sample and the formation of a complex between the mAb and antigen from the sample is determined using well known conventional methods. An assay for determining the presence of an antigen which the mAb of the invention binds to can be performed using, for example, a sandwich format wherein a solid phase is coated with antibody to HIV envelope, the sample is added, and then labeled mAb of the invention (such as with biotin) is added. Following a wash, enzyme labeled avidin can be added as well as enzyme substrate. Such general types of assays are well known in the art and many variations are possible. Kits for determining an antigen for which mAbs of the invention are specific may comprise a mAb (or mAbs) of the invention, a solid phase on which is coated an antibody specific for HIV-l env, and means for detecting the formation of a complex among the mAb of the invention, the antibody specific for HIV-l env, and an HIV antigen for which the mAbs are specific.

As noted, the mAb of the invention has high neutralization ability with respect to the RF strain, achieving at least about 50% neutralization in vitro of about 2 x 10⁴ infectious units of the RF HIV-l strain at a mAb concentration of about 1 μg/ml. Preferably, the mAb achieves that degree of neutralization at a concentration of about 0.5 μg/ml. Stated in a comparative manner, the mAb of the invention achieves at least about 50% neutralization in vitro of the RF strain in a concentration of about 0.5 μg/ml under conditions where another highly neutralizing anti-CD- 4 binding site mAb, mAb 1125H, achieves 50% neutralization in a mAb concentration of about 5 μg/ml. As used throughout , ___,__, - 3/19786

-6-

herein, the amounts of neutralization is as determined using a standard 24 hr. neutralization assay described in the examples. Other assays may arrive at different results. It is not necessary for a mAb of the invention to have in vivo the neutralization activity measured in the in vitro test described.

While the mAb of the invention is against the CD-4 binding site region of gpl20 of HIV-l, it does not necessarily bind only within the CD-4 binding site. The CD- 4 binding site is a conformational region that encompasses many disparate areas of gpl20. It has been found that addition of soluble CD4 at about a 1000 fold excess over mAb inhibits binding to HIV-l_ιπB of the antibody of the invention to its epitope, by about 90%.

In a preferred embodiment, the mAb of the invention is specific for a highly conserved gpl20 epitope. In particular, the epitope is conserved among strains MN, IIIB, SF-2 and RF. Human mAbs are also preferred.

Most preferably, the mAb of the invention substantially has the epitope specificity of human mAb 5145A. mAb 5145A is produced by an EBV-transformed human B-cell line deposited on March 10, 1992 at the American Type Culture Collection, (ATCC) , located at 12301 Parklawn Dr., Rockville, MD, 20852, United States of America, and assigned accession number CRL 10982. Methods for obtaining 5145A and its further characterization are described below. mAb 5145A is specific for an epitope overlapping the CD-4 binding site that is destroyed by reduction of disulfide bonds. It does not react with LAV-2. 5145A recognizes a wide variety of strains, including almost all African strains tested. It is also notable for its ability to strongly neutralize in vitro all four of the MN, IIIB, SF-2 and RF strains at about an equivalent level. A preferred mAb neutralizes in vitro about 2 x 10⁴ infectious units of virus of all four strains to a level of 50% at concentrations of antibody that do not differ by more than about 3X. Using the standard 24 hr. -7-

neutralization assay described below, 50% neutralization of approximately 2 X 10⁴ infectious units of virus was achieved at about 0.2-0.5 μg/ml of mAb for the four strains and 99- 100% neutralization of this amount of virus (for the four strains) at 10 μg/ml.

The strong in vitro neutralization results indicate high affinity for all of those strains. Binding of a preferred mAb of the invention to gpl20 is markedly decreased by the following amino acid changes from HXB2 wild type: 368 D/R, 368 D/T and 370 E/R, while the following amino acid changes did not markedly decrease binding: 257 T/R, 257 T/G, 381 E/P and 427 W/S.

Using an immortalized cell line of the invention such as the deposited cell line CRL 10982, one can isolate all or a portion of the genes coding for the mAb expressed by the cells. These genes may be altered and resulting mAbs screened to obtain mAbs with even greater affinity for antigen. The genes may also be altered to change the isotype, idiotype, or effector functions of the mAbs. Expression systems have been developed to allow expression and secretion of genetically engineered non-murine mAbs, such as human mAbs, in mouse cells. In a most preferred embodiment the mAb substantially has the complementarity determining regions of mAb 5145A.

Aside from other uses described herein, mAb 5145A is also useful in screening to determine whether cultures obtained of anti gpl20 mAbs (human or otherwise) are producing mAbs to an epitope closely related to one recognized by antibody 5145A, such as by using a competitive ELISA assay with labeled 5145A. Such an assay determines if mAbs from the culture being screened compete with 5145A in binding to an epitope presented on, for example, recombinant gpl60 coated ELISA plates. MAbs which compete highly would be specific for the same or adjacent epitopes. A suitable competitive assay is described below, as is methodology for obtaining mAb-producing cells for screening. 3 1

-8-

As noted above, the invention also includes a synergistically neutralizing combination of mAbs.

New or already known antibodies against the V3 region can be screened for use in the synergistic combination of the invention. To obtain new antibodies against the V-3 region for use in the synergistic combination, a peptide consisting of amino acids 305-328 of the MN strain (or the corresponding sequence in other strains) can be substituted for recombinant gpl60 in screening using the procedures of the Examples. That (MN) sequence is NYNKRKRIHIGPGRAFYTTKNIIGC, described in Gurgo et al. 1988.

The anti-V3 antibody of the combination can be specific for any part of the V3 region, so long as it synergistically neutralizes with the anti-CD4 binding site antibody (as determined in the test for synergistic neutralization described herein) . It preferably competitively inhibits the binding to gpl60»»ι, or V_m of antibodies produced by the cell line deposited under accession no. CRL 10770, the inhibition preferable being more than about 80%, most preferably more than about 95%.

Examination of the reactivity of antibodies against specific peptides can be used to determine the strain specificity of the screened antibodies. mAb 4117C is characterized by its reactivity with the V3 peptide of the following strains: MN, SF-2, NY-5, CD-451, WMJ-1, WMJ-3, Z- 3, Z-321, and SC; and by its lack of reactivity with the following strains: WMJ-2, LAV-MA, BR, LAV-IIIB, PV-22, ELI, Z-6, NX3-3, JY-1, HXB-2 and MAL. mAb 4117 is produced by EBV transformed human B-cells deposited on June 14, 1991 at the American Type Culture Collection, 12301 Parklawn Dr., Rockville, MD 20852 USA and assigned accession No. CRL 10770.

In another embodiment, the mAb specific for the V3 region has the epitope specificity of mAb 4117c. Such a mAb can be produced, for example, by cells derived from the deposited cell line. -9-

The anti-CD4 mAb of the combination is the novel mAb that is described herein, in all its different aspects.

For example, the human mAb specific for the CD-4 binding site used in the combination can be one that achieves at least about 50% neutralization of about 2 x 10⁴ infectious units of RF HIV-l strain at a concentration of about 1 μg/ml.

Simarlarly, the preferred mAb specific for the CD-4 binding site of the combination has the epitope specificity of the 5145A, or results from the manipulation of the genes obtained from the cell line producing 5145A (such as described herein) .

A given combination of a mAb against the V-3 region and for example, CRL 10982 can be screened in a standard neutralization assay for synergistic neutralizing activity. The individual neutralizing activity of each antibody individually is compared with the neutralization activity of the antibodies combined. A suitable neutralization assay is described below. The ability of the antibody combination to synergize will be evidenced by a significant increase in neutralization activity over that obtained in the presence of equivalent concentrations of the individual antibodies. The extent of synergy can be guantitated by calculating the Combination Index using known statistical methods (Chou et al. 1984; Chou et al. 1989; Chou 1991). The combination index value at greater than 20% neutralization preferably is less than 0.8 to indicate significant synergy.

Polyclonal compositions can also be obtained which are specific for V3 and which synergize with the anti-CD-4 binding site mAb of the invention. Immunization and screening of polyclonal sera from, for example chimpanzees, rabbits or sheep can be accomplished using the immunogens described herein for generating mAbs, and by using the screening techniques described herein. An important advantage of mAbs instead of polyclonal antibodies, however, -10-

is that mAb technology allows production of unlimited amounts of homogeneous reagents.

Antibodies from different sources may be employed to obtain antibodies for use in the invention. Methods have been described in the literature for inducing neutralizing antibodies against different epitopes of HIV-l gpl20 in both rodents and chimpanzees. Antibodies against the V3 loop have been induced in both rodents (Javaherian et al, 1990) and chimps (Girard et al. , 1991) by immunizing animals with synthetic V3 peptides either in free form, or conjugated to KLH. Anti-V3 antibodies have also been induced by immunizing chimps with purified gpl20 and gpl60 (Berman et al., 1990). Monoclonal antibodies against these gpl20 epitopes can be prepared from immunized mice by standard techniques, and monoclonal antibodies can be prepared from chimps by .following the EBV-transfor ation procedure described herein.

Antibodies can be purified by immunoaffinity chromatography. For example, to purify anti-V3 antibodies AH-Sepharose beads are activated by treatment with glutaraldehyde, and conjugated either to purified V3 peptide or gpl20. Antibodies against V3 can be obtained by passing 10-fold diluted hyperimmune serum through the columns to allow the antibodies to bind, and washing off unbound antibodies with saline and 0.5M NaCI solutions. V3-specific antibodies can be eluted from these columns by washing with tris-glycine buffer, pH2.7 or by passing through excess V3 peptide.

Antibodies of the invention can be used, if desired, in methods of neutralizing infection of T-cells by HIV. The antibodies are used in the manner described to reduce the possibility of infection. For example, passive administration of the human mAbs of the invention can be performed to decrease the chance of HIV-l infection in cases of acute exposure to HIV. For example, the HIV-l neutralizing mAbs of this invention could be passively -11-

administered to pregnant seropositive women to prevent their fetuses from becoming HIV-l infected. In addition, mAbs can be used to prevent HIV-l infection by administering them to individuals near the time of their exposure to HIV-l, such as following a needle stick.

Human, or other primate, mAbs have distinct advantages over, for example, rodent mAbs, in administration to humans for prevention, or therapy, of viral infection. They have increased stability and very low immunogenicity in humans. Thus, human mAbs are much less likely to create deleterious anti-immunoglobulin responses than are non-primate mAbs and it should be possible to obtain stable levels of therapeutic doses of the human mAbs in humans.

If a neutralizing mAb is desired to be used for treating HIV-infected individuals or preventing infection by HIV, the mAb should be extremely potent, so that neutralizing concentrations can be attained in vivo following administration of milligram amounts of the mAb. It has been estimated that between 0.03 to 3 mg/ml of a neutralizing Ab with similar affinity to that of CD-4 for gpl20 would be required to eliminate HIV infection in vivo (Layne et al. 1989) . This would necessitate administration of approximately 0.15 to 15 g of Ab per patient, the higher ranges of which are not desirable because of the side-effects associated with administering such high protein doses and the difficulties and cost of producing such large amounts of purified antibodies. The high affinity and neutralization ability of the synergizing mAbs of the invention, however, make possible a reduction in the concentration of mAb required when used for this purpose. The anti-CD-4 binding site mAb of the invention can also be used in combination with other anti-HIV-1 mAbs, in order to attain an additive neutralization effect, if desired.

In order to administer the mAb of the invention the mAb can^"be adjusted to 5% solution in sterile saline, yielding a concentration of 50 mg/ml. Where synergizing antibodies of the invention are used, the best ratio of the synergizing antibodies is determined experimentally, using the 24 hour fluorescent focus assay described below. For example, a 1:1 ratio of 5145A and an anti-V3 antibody with the neutralization ability of 4117C can be used. The concentration of each should be determined to give very roughly comparable levels of neutralization to each other (i.e. an equipotent ratio is preferably used) . The amount of this solution required for protection can be determined in animal experiments, performed first in Hu-SCID mice (Mosier et al. 1988, McCune et al. 1988) and subsequently in chimpanzees. The reagent can consist of as few as one type of antibody, although it is believed that the most efficient composition will contain a large number of different antibodies directed against major antigenic sites, preferably including synergizing antibodies. This is in order to increase the affinities and cross-reactivity of the antibodies to different HIV variants which may exist and to decrease the likelihood of a deleterious anti-idiotype response.

It is believed that it may also be beneficial to mix engineered antibodies of different isotypes, including IgGs, IgMs, and IgAs, in order to increase the affinities and effector activities of the antibodies. It is also believed to be beneficial to include antibodies conjugated to toxins, mentioned below, to increase the killing of infected cells, and engineered bispecific antibodies, to increase targeting of infected cells to immune cell-mediated cytotoxic mechanisms.

Based on available data and theoretical considerations, it is believed that to prevent virus spread in vivo requires achieving plasma concentrations of neutralizing antibody combinations of 1 to 30 ug/ml. 1-5 mis of the 5% solution (50-250 mg total Ig) can be given by intravenous injection to patients. Assuming a total blood volume of 5L, and assuming that all of the delivered Ig remains in the plasma -13-

with a half life. of 2 weeks, this likely results in an initial plasma concentration of the mAbs ranging from 10-50 ug/ml.

Excellent properties of the preferred mAb of this invention are: 1) its demonstrated HIV-l neutralizing activity in vitro at low mAb concentrations, 2) its high affinity for antigen (HIV-l gp 120), 3) its broad HIV-l strain specificity, 4) the fact that it is of human origin and will, therefore, elicit few, if any, deleterious immune reactions when administered to humans, and 5) the heavy chain isotype of the preferred mAb is IgG, which is significant because human IgG Abs are the only class of Ab able to cross the placenta, and Abs of the IgGl subclass can potentially kill HIV-1-infected cells in vivo via Ab- and complement-dependent cytotoxicity (ACC) and/or Ab-dependent cellular cytotoxicity (ADCC) .

If desired, the mAbs of this invention may be modified by covalent attachment of a toxin such as ricin A, pokeweed antiviral protein, poisonous lectins, abrin, diphtheria toxin, or other toxins to the mAbs. It has been demonstrated that such anti-HIV-1 mAbs-toxins are dapable of specifically killing HIV-l infected cells in vitro.

In considering the use of these mAbs to prevent HIV-l infection, the killing of HIV-l infected cells via ACC, ADCC, or following mAb conjugation with a toxin, could complement the neutralizing activity of our mAbs by eliminating a very small percentage of HIV-l infected cells which might result if 100% neutralization of HIV-l by the mAbs is not obtained.

Gram quantities of mAbs are preferably obtained for the various in vitro or in vivo uses possible. These amounts can be obtained by growth of cell lines producing the mAbs of the invention in a mini-bioreactor. Additionally, cost- effective methods to increase mAb production are: 1) fusion of EBV-transformed lines with a human/mouse heteromyeloma (Teng et al. 1983; Kazbor et al. 1982) and 2) PCR 3/1

-14-

amplification of expressed immunoglobulin V_H and V_L genes from cell lines using published primer sequences (Larrick et al. 1989) , followed by cloning of these genes into available eukaryotic expression vectors containing constant region genes (Orlandi et al. 1989) . Alternatively the genes could be directly cloned from cDNAs or genomic DNA of antibody producing cells. The latter constructs can then be expressed as mAbs at high levels in mouse myeloma cell lines. Transformed cell lines producing recombinant monoclonal antibodies are included within the scope of the invention, as are the antibodies produced thereby. Antibodies are preferred that result from cloning of variable region genes from 5145A, and thus substantially have the complementarity determining regions of that mAb. The invention also includes moieties having the same function as mAbs, such as Fab fragments, F(ab')₂, Fd or other fragments, modified proteins such as chimeras with altered Fc regions, or having mutagenized idiotypic regions, or the heavy or light chains alone, so long as they bind to the same epitopes as the mAbs of the invention. Techniques for producing such fragments or modified antibodies are known to one skilled in the art (e.g. , Parham 1986) .

The invention also includes cell lines producing the mAbs of the invention. Those cell lines can be, for example, conventionally immortalized cell lines such as the EBV transformed cell line described herein, or cell lines producing mAbs of the invention that result from the cloning of immunoglobulin genes or portions thereof using immunoglobulin expression vectors into an appropriate antibody producing cell line.

Previous studies have shown that the affinities of mAbs to their antigens, and the abilities of those antibodies to bind infectious agents can be significantly enhanced by changing their isotypes. For example, a human IgG mAb against group B streptococci was converted to an IgM by standard recombinant DNA methods (Shuford et al. 1991) . The IgM form of the antibody showed an approximately 100-fold greater level of binding in an ELISA assay than the IgG, and was able to protect mice from lethal effects of the bacteria at a greater than 16 fold reduction in concentration.

Significant enhancement of the activity of antibodies described herein can be expected to be achieved by changing their isotypes. It is believed that the preferred reagent of this invention consists of mixtures of engineered antibodies of different isotypes, including IgGs, IgMs, and IgAs, in order to increase the affinities and effector activities of the antibodies.

The invention is illustrated by the following examples, which are intended only to be illustrative of the invention, and not limiting of the invention's scope.

Examples

Preparation and Screening of Cell Lines

Producing mAbs According to the Invention

Human monoclonal antibodies ("HumAbs") were obtained by EBV-transformation of peripheral blood mononuclear cells from seropositive hemophiliacs.

The peripheral blood mononuclear cells ("PBMC") were isolated by centrifugation of fresh, heparinized blood, diluted 1:3 with RPMI1640 medium (Flow®), on Histopaque (Sigma®) at 400 x g for 30 min. at room temperature. Cells at the medium/ Histopaque interface were recovered and diluted 7-8 fold with RPMI 1640 medium. The cells were spun down (400 x g, 20 min.) , and then resuspended in 50 ml of RPMI 1640 medium, counted, and spun down again as before. Cells were then resuspended at a density of 2 x 10⁶ cells/ml in RPMI 1640 medium supplemented with 15% (vol/vol) fetal calf serum (HyClone®) , 2 mM L-glutamine, penicillin (50 units/ml) , and streptomycin (50μg/ml) (complete medium) . Epstein-Barr virus (EBV) 100X stock, (Raubitschek et al. 1985) was then added so that it constituted 1/10 of the final volume of the cell suspension, and the cells were incubated overnight at 37°C in 5% C0₂ in a 25 cm² flask. The following day, the cells were gently resuspended, diluted approximately 10-fold with RPMI 1640 medium, and spun down. The pellet was resuspended at a final density of 10⁴ cells/ml in complete medium. The cells were then plated in U bottom 96-well plates at lOOμl (1000 cells) per well onto lOOμl of irradiated (3500 rads) rat embryo fibroblasts in complete medium. The cultures were fed weekly for 4 weeks at which time approximately 45% of the wells exhibited growth. Then their supernatants were assayed for anti-^'env Ab production (see below) . Those cultures testing positive were picked onto fresh irradiated rat embryo fibroblasts in 96-well plates and re-assayed the following week. Cultures remaining positive were then sublined onto irradiated rat embryo fibroblasts at densities ranging from 1 to 100 cells/well. Those cultures growing in plates in which the number of wells with growth indicated > 95% probability of monoclonality as determined by the Poisson distribution (Coller et al. 1987) were re-tested for anti-envelope antibody production, and those testing positive were expanded into bulk culture. In this way we obtained cell lines 5145A and 4117C.

Southern Blot Analysis to Determine Clonality of Cell Line

DNA was isolated fro the 4117C cell line followed by restriction enzyme digestion, agarose gel electrophoresis, blotting to nitrocellulose, and hybridization to ³²P-labeled nick-translated probe. (Eckhardt et al. 1982) . The DNA was cut with Hind III, which allows visualization of rearrangements due to V-D-J joining upon hybridization with an immunoglobulin J_H region probe (Ravetch et al. 1981) . The J_H probe used was a EcoRI-Hindlll fragment approximately 3.3 kilobases in length from the germ line J_H locus; the Hindlll site at its 3' end is present in the germ line DNA [Ravetch, J. V., U. Siebenlist, S. Korsmeyer, T. Waldmann, and P. Leder. (1981) Cell 27:583-591], whereas the EcoRI site at its 5¹ end was created upon cloning. The monoclonality of the cell line was confirmed.

Methods Used for Detection of Human Anti-gpl20 mAbs Produced by Cell Line

ELISA assays were used to detect HIV-l env-specific Abs. The initial screening of EBV-transformed human cultures for production of anti-env Ab was done using recombinant gpl60_BRU (Kieny et al. 1988) or V3_MN to coat PVC ELISA plates (Flow/ICN) . This version of gpl60, supplied by Pasteur Merieux, lacks the site which is normally cleaved to form gpl20 and gp41. In later assays on supernatants from cultures identified as positive in the initial screening, a variety of other HIV-l proteins or peptides can be used to determine the specificity of the mAbs. These include recombinant gpl20 of the IIIB strain produced by Celltech, Inc. and available through the AIDS Research and Reference Reagent Program (NIH) or described by (Leonard et al. 1990) synthetic V3 peptides from a variety of strains (strain specificity is described above) ; pl21, a gp41 peptide (amino acids 565-646) sold commercially by Dupont or described in Chang, et al. , European Patent Application 0199438 published October 29, 1986, or other peptides described below. 4117C is negative for gpl20 of the IIIB strain and for pl21. 5145A is negative for V3 peptides and gp41.

Unless noted otherwise, 50 ng/well of protein diluted in Na₂C0₃/NaHC0₃ buffer, pH 9.8 was incubated in the plates overnight at 4°C. The following day, the plate was washed 3 times with PBS/tween/azide (Sigma® PBS with 0.05% Tween 20, 1 mM NaN₃) . Next, the wells of the plate were blocked (to prevent nonspecific binding) by incubation with 50μl of 2% BSA in PBS for 1.5 hr. , 37" C. After washing as before, 50μl of supernatant from human cell lines was added to the wells and incubated for 1.5 hr. , 37°C. Unbound Ab was washed from the wells, and 50μl of a 1/500 dilution of goat anti-human IgG conjugated to alkaline phosphatase (Zymed®) in 2% BSA was added to each well. After an incubation and wash identical to those discussed above, 50μl of alkaline phosphatase substrate (disodium p-nitrophenyl phosphate) , 1 mg/ml in diethanola ine buffer (1M diethanolamine, 0.5mM MgCl_2f 3mM NaN₃, pH 9.8) was added. The absorbance at 405nm was read in a Titertek Multiskan Plus® ELISA reader (Flow®) at times ranging from 5 min. to 2 hr. following substrate addition. The background obtained when culture media was used rather than supernatant from human cell lines was automatically subtracted from the results by the ELISA reader. Radioimmunoprecipitation and Western Blot Assays

For radioimmunoprecipitation assays, glycoproteins in HIV-1-infected cells at 5-7 x 10⁵ cells/ml were labeled with ³H-glucosamine (lOOμ Ci/ml) (Pinter et al. 1989) . The cells were then lysed and immunoprecipitated as previously described (Pinter et al. 1988) . Briefly, the cell pellet was brought up in lysis buffer at a concentration of 5 x 10⁶ cells per ml. The lysate was then precleared with fixed, killed staphylococcus aureus cells (Staph A) , and 70μl of pre-cleared lysate was added to 70μl of supernatant from human Ab-producing cell lines or 1/400 dilution of human sera. Following an incubation and precipitation by Staph A, the pellet was brought up in Laemmli sample buffer containing 1% DTT and run on an 11% polyacrylamide gel as described (Laemmli 1970) . Fluorography (Bonner et al. 1974) then allowed detection of radiolabeled, immunoprecipitated glycoproteins in the gel.

Western blot analysis was performed using strips prepared with HIV-l lysate essentially as described by Pinter et al. (Pinter et al. 1989) . The lysate was diluted in buffer composed of 0.01M Tris hydrochloride (pH 7.4) ,10% glycerol, 0.01% bromophenol blue, either 0 or 1% DTT, and 1% SDS. The Western blot strips were incubated with a 1/2 dilution of supernatant from human Ab-producing cell lines or a 1/100 dilution of human serum, and bound Ab was detected (Pinter et al. 1989) .

Purification and Quantitation of Human mAbs

Human mAbs were purified from cell supernatants on MASS protein A filters (Nygene Corp., Yonkers, N.Y.) and concentrated using CentriCell ultrafilters (Polysciences, Inc.) (Tilley et al 1991). The mAbs were guantitated by ELISA using affinity-purified human IgG (Cappel) of known concentration as standard.

HIV Strains

HIV-l strains IIIB (Popovic et al. 1984; Ratner et al. 1985) and SF2 (Levy et al. 1984; Sanchez-Pescador et al. 1985) were obtained from Dr. Jeffrey Laurence, Cornell University School of Medicine; strains MN (Gallo et al. 1984; Shaw et al. 1984) and RF (Popovic et al. 1984; Starcich et al. 1986) were obtained from the NIH AIDS Research and Reagent Repository. The identities of strains IIIB, MN, and RF were confirmed by us using strain-specific antisera against the hypervariable V3 loop (hvl-v3) of each strain in an immunofluorescence assay. The IIIB-specific chimpanzee antiserum was obtained through a collaboration with Dr. Marc Girard, Pasteur Institute, whereas the MN- and RF-specific rabbit antisera were provided by Dr. Robert Neurath, New York Blood Center. An HIV-2 strain, LAV-2 (Clavel et al. 1986) was obtained from Luc Montagnier, Pasteur Institute via Dr. Alvin Friedman-Kien, New York University School of Medicine. African strains were isolated by Dr. Ellen Murphy at PHRI from blood samples obtained from Bangui (Central Africa) . The Haitian isolate, AL, was a gift from Dr. David Ho, Aaron Diamond AIDS Research Center, New York.

Immunofluorescence Assays for HIV Strain Specificity of mAbs

Prior to attachment of cells to Multi-spot microscope slides (Shandon) for immunofluorescence analysis, the slides were treated with poly-L-lysine (lOOμg/ml in PBS, 50ml per well) for 30 min at room temperature. The slides were then washed with distilled water and dried. Cells that were 100% HIV-1-infected or uninfected were then washed in sterile PBS, resuspended in PBS at a density of 1-2 x 10⁶ cells/ml, and incubated on the poly-L lysine-coated slides (50μl cell suspension/well) at 37°C for 30 min.. The slides were then washed 2X in 100-200 ml PBS, using a slide-holder and trays. Following the 2 washes in PBS discussed above, the slides were washed IX in distilled water and then incubated in 100-200 ml of acetone or methanol for 8 mins. The slides were then removed from the fixative and allowed to air dry.

Prior to addition of human Abs to the fixed cells on slides, non-specific binding was blocked by incubation of the slides with 1 mg/ml bovine gamma globulin in PBS for 30 min. at 37°C. After washing 2 times in PBS and once in distilled H₂0, the slides were allowed to air dry. Undiluted supernatant from human Ab- nproducing cell lines or serum diluted 1/100 to 1/200 in 1 mg/ml bovine gamma globulin in PBS was then incubated at 25-50μl well for 1 hr. at 37°C with the fixed cells on the slides. After washing and drying the slides as discussed above, a 1/50 dilution of goat anti-human IgG conjugated to FITC (Zymed) in 1 mg/ml bovine gamma globulin in PBS was incubated with the slides as in the previous step. After washing and air drying the slides as discussed above, the cells on the slides were counterstained with 0.05% Evans Blue for 10 min. at room temperature. The slides were then washed extensively with distilled H₂0 and air dried. Finally 2μl per well of 0.033M DTT in 50% glycerol in PBS was added as preservative, a coverslip was placed over the wells, and the slides were viewed under a Nikon Diaphot immunofluorescence microscope.

Determination of mAb Isotypes

The heaving chain class was selected to be IgG by the use of goat anti-human IgG in the primary screening. Heavy chain subclass was determined for 4117C using a variation of the -immunofluorescence assay. Human mAb-producing cells were attached to slides and fixed with acetone. The slides -22-

were blocked with bovine gamma globulin and washed as discussed above. Next, a 1/5000 dilution of human IgG subclass-specific mouse monoclonal Ab (Zymed) (specifically, anti-IgGl and anti-lgG2 were used in these experiments) was added and incubated for 1 hr. at 37° C. Following washing and drying of the slides, biotinylated goat antimouse IgG (Zymed), 1/200 dilution, was added and incubated for 1 hr., 37° C. After washing and drying the slides, a 1/50 dilution of FITC/streptavidin (Zymed) , was added and incubated for 1 hr. , 37°C. After washing and drying the slides, the cells were counterstained, and viewed as discussed above.

Light chain isotype was determined by a variation of the ELISA assay discussed above. Following incubation of supernatant from mAb-producing human cells with gpl60 in duplicate ELISA wells, the mAb isotype was determined by development of one well with goat anti-human kappa Ab conjugated to alkaline phosphatase and the other well with goat anti-human lambda Ab conjugated to alkaline phosphatase. Both of the latter reagents (Tago) were used at 1/3250 dilution.

Competitive Inhibition Assays

These assays were done using a variation of the ELISA procedure discussed above. ELISA plates were coated with gpl60, blocked with BSA, and washed. In the CD-4 inhibition experiments, a constant volume of biotinylated Ab from human Ab-producing cells was added (70 ng/ml) to varying amounts of soluble CD-4 (in 1% BSA in PBS) in eppendorf tubes, and RPMI was then added to yield a constant total volume. After mixing, the Ab/CD-4 mixtures were pipetted at 50μl/well into the ELISA plates. The remainder of the ELISA procedure was carried out as discussed above. Neutralization Assay

Prior to conducting neutralization assays, the mAbs were purified on recombinant protein A Sepharose columns essentially as described (Harlow et al. 1988) . The column fractions containing mAb (as determined by ELISA assay of fraction aliquots) were concentrated in an AMICON centriprep 30 column and dialyzed against PBS.

The neutralization assay was carried out as follows. Purified Abs, or combinations of Abs, were diluted in complete media containing 10% FCS to obtain concentrations ranging from 0.1 to 20 μg/ml in a total volume of lOOμl. Included in this volume was approximately 2 x 10⁴ tissue culture infectious units of HIV-l. After a 30 min. preincubation of virus and mAb at room temperature, the mixtures were each added to 1 x 10⁵ H9 cells in a final volume of 200 μl. Following a 24 hr. incubation at 37° C, the cells in each well were plated onto separate wells of poly L-lysine-coated slides and stained sequentially with a rat anti-nef serum (1/200) or serum from a seropositive individual (1/200) followed by a species specific anti-IgG Ab conjugated to FITC (1/50) (Zymed) . The latter two antibodies were diluted in 1 mg/ml bovine gamma globulin in PBS. The cells were counterstained with Evan's Blue, and the percentage of infected cells from each culture relative to the control (no mAb added) was assessed by counting immunofluorescent cells versus total counterstained cells under the fluorescence microscope. Under the conditions discussed above, approximately 3% of the cells (3000 cells) are infected at the end of the assay in the absence of neutralizing antibody.

Affinities of mAbs

The affinity of mAb for gpl60 was determined by diluting mAbs of known concentration and assaying the -24-

various dilutions on gpl60 coated plates by ELISA as discussed above. It has been demonstrated that the concentration at which half-maximal Ab binding is observed is a rough value of 1/K (van Heyningen et al. 1987) .

Epitope Mapping

Wells of Immulon II plates (Dynatech) were coated overnight with 100 μl of sheep polyclonal antibody, D7324, against the carboxy terminus of gpl20 (Aalto, Dublin) . This antibody was used at 5 μg/ml in sodium bicarbonate buffer, pH 8.5. This and all subsequent steps of the procedure were carried out at room temperature. The wells were then washed 2 times with TBS (144 mM NaCI, 25 mM Tris HC1, pH 7.5) and blocked with 2% milk powder (Cadbury's Marvel) in TBS. After removal of the 2% milk powder from the wells, 100 μl of supernatants from gpl20 wild type- or mutant-transfected COS-1 cells were added. Then panel of gp 120 mutants used in these experiments was derived from HXB2 (a III B clone) and was a subset of those reported by others (Thali et al 1992, Olshevsky et al. 1992, Thali et al. 1991). Specifically, all of the mutants reported in Thali et al. 1992 were used except 76 P/Y, 256 S/Y, 257 T/A, 297-329, 368 D/P, 368 D/N, 368 D/K, 368 D/E, 370 E/D, 380-381 GE/YV, 386 N/Q, 457 D/N, 457 D/E, and 470 P/G. In addition, mutants described by Olshevsky et al. 1992 were used — 36 V/L 45 W/S, 429 K/L, 470 P/L, AND 500-501KA/KGIPKA — and mutants described by Thali et al. — 356 N/I, 386 N/R, AND 450 T/N. Following a l to 2 hr. incubation with the supernatants from gp 120 wild type- or mutant-transfected cells, the wells were washed 2 times with TBS. At this time, 100 μl of each human mAb at 0.1 to 0.5 μg/ml in TMSS (2% milk powder, 20% sheep serum in TBS) was added to each well. One hundred μl of a 1/4000 dilution in TMSS with 0.5% Tween-20 of a pool of three sera from HIV-l seropositive individuals was also included as a positive control for the presence and -25-

reactivity of wild type gpl20 and each gpl20 mutant at this step. Following a 1 hr. incubation, the wells were washed 2 times as before. Next, 100 μl of a 1/4000 dilution of alkaline phosphatase-conjugated goat anti-human IgG (Accurate Chemicals) in TMSS was added to each well. Following a 1 hr. incubation, the wells were washed 4 times with AMPAK was buffer (Dako) , and bound alkaline phosphatase was detected with the AMPAK system (Moore et al. 1988) . The optical density (OD) at 492 nm was then determined spectrophotometrically. The recognition index for each human mAb against a given gpl20 mutant was calculated as: OD (mAb on mutant gpl20)/OD (mAb on wild type gpl20) x OD (seropositive serum on wild type gpl20/OD (seropositive serum on mutant gpl20) . This is similar to the calculation of recognition index used by Thali et al.

Specificity of anti-V3 antibody 4117C

4117C was found to recognize a variety of divergent HIV strains, including MN, SF-2, FV (New York), 11699 (Central Africa) , and the JR-CSF primary isolate (Los Angeles) (Koyanagi et al. 1987) . 4117C human mAb is less strain specific than other anti-V3 human mAbs that have been described (Scott et al. 1990, Zolla-Pazner et al. 1990). Comparison of the V3 sequences of the isolates recognized by 4117C reveals that the sequence GPGR at the tip of the loop is shared by all of them. In addition, the sequence IXI just to the left of the GPGR is highly conserved among these isolates. These observations indicate that 4117C may be directed against a relatively conserved sequence near the tip of the loop. The GPGRAF sequence at the tip of the loop has recently been shown to induce broadly reactive anti-V3 Abs in experimental animals (Javaherian et al. 1990) . Figs. 6 and 7 show that human mAb 4117C exhibits potent neutralizing activity against the MN and SF-2 strains of HIV, respectively. -26-

Further Characterization of 5145A

" The apparent affinity constant, K, of 5145A was measured against recombinant gp 160 from either BRU or MN strains and found to be approximately 2 x 10⁹ L/mole. This is comparable to the apparent affinity constants of two of other potent neutralizing HuMAbs, 1125H (anti-CD4 binding site), 1.3 x 10⁹ L/mole, and 4117C (anti-V3 loop), 0.9 x 10⁹ L/mole. mAb 1125H is described herein for comparative purposes. It is produced by EBV- transformed human B-cells and was deposited with the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Md. , 20852, USA, on October 23, 1990, and has been assigned accession number CRL 10582.

Fig. 1 demonstrates the specificity of the 5145A (and 1125H) for gpl20 by radioimmunoprecipitation/SDS gel analysis. Fig. 2 shows, by immunoblot analysis, that the epitope of 5145A (and the epitope of 1125H) is destroyed by reduction of gpl20 disulfide bonds. The latter feature is common to all HuMAbs against the CD4-binding site that we are aware of. Fig. 3 is a graph of the results of the study of competitive inhibition of 5145A binding to gpl60 by soluble CD-4. The figure shows that 5145A binding to its epitope is inhibited in a concentration-dependent manner by soluble CD-4. The efficiency of this inhibition is not as great as that obtained in CD4 inhibition of HuMAb 1125H binding to gpl20, also shown. Approximately 6-fold more sCD4 is required for 50% inhibition of binding of 5145A to its epitope than is required for this level of inhibition of 1125H binding to its epitope in this assay. Taken together, these results show that 5145A is specific for an epitope of gpl20 that is conformation dependent and substantially overlapping the CD-4-binding site. However, additional data presented below demonstrate that the epitope of 5145A is not identical to that of any other anti-CD4 binding site mAb described.

Immunofluorescence analysis of 5145A reactivity'with fixed, virus-infected cells shows that 5145A strongly recognizes all HIV-l strains tested thus far by us. Strains tested include five North American strains (MN, SF-2, IIIB, RF and AL) and 10 Central African strains listed in Table 1 (no's 6-15) .

Table 1

REACTIVITY OF HUMAN mAbS WITH HIV ISOLATES BY IMMUNOFLUORESCENCE ASSAY AGAINST VIRUS-INFECTED CELLS

-28-

The V3 loop of each of these Central African strains has been sequenced. Among the Central African Republic HIV- 1 strains are representatives of African subtype A, African subtype D, and Northern Thailand sub-type E based on their V3 region sequences. Table l shows that both 5145A and 1125H reacted with all of the North American (MN, SF-2, IIIB) and Haitian (RF, AL) strains tested. 5145A also recognized 9 of 10 Central African Republic strains tested, while 1125H reacted with only 4 of these 10 strains. The group of African strains recognized by 1125H includes African subtype A and Northern Thailand subtype E strains, while the group of strains not recognized by 1125H includes subtypes A and E as well as a single African subtype D strain. Interestingly, the one HIV-l strain that was not recognized by either mAb, i.e., 4067, could not be classified among the known subtypes of HIV-l based on its V3 sequence. Neither 1125H nor 5145A recognized LAV-2. These results indicate that the 5145A epitope is, indeed, surprisingly conserved across HIV-l strains and is significantly more conserved than the 1125H epitope. Neither 1125H nor 51454A recognize LAV-2, an HIV-2 strain (Table 1).

Fig. 4 shows that 5145A has potent neutralizing activity against the four North American HIV-l isolates that it reacts with by immunofluorescence assay. Fifty percent of approximately 2 x 10⁴ infectious units of virus in this assay is neutralized by approximately 0.2-0.5 μg/ml of 5145A. Furthermore, 90-100% of each of these viral strains is neutralized by 10 μg/ml of 5145A. These neutralization results contrast with those seen using other anti-CD4 binding site HuMAbs known to us. 1125H HuMAb has similar neutralizing activity against MN and IIIB strains to 5145A, but significantly less neutralizing activity against SF-2 and RF strains than 5145A (Tilley et al. 1991) . Specifically, a 5-10 fold greater concentration (5 μg/ml) of 1125H is required to neutralize 50% of the SF-2 and RF virus than was required for this level of neutralization of the -29-

IIIB and MN strains. 1125H neutralizes a maximum of only about 75% of the SF-2 and RF virus at a concentration of 20 μg/ml. Three other anti-CD4 binding site HuMAbs (Ho et al. 1991; Posner et al. 1991; Lake et al.) exhibit more narrow strain specificity and/or lower neutralizing activity than either 1125H or 5145A. For example, the 1.5e HuMAb does not recognize nor neutralize the RF strain (Ho et al. 1991) and the Sl-1 HuMAb neutralizes RF only in the presence of complement (Lake et al.), whereas both 1125H and 5145A neutralize RF significantly in the absence of complement (Tilley et al. 1991). The F105 HuMAb has orders of magnitude lower neutralizing activity against the MN strain than does either 1125H or 5145A (Posner et al.). These differences among other anti-CD4 binding site HuMAbs in strain specificity and neutralizing activity indicate that each of these other HuMAbs is directed against different epitopes than that of 5145A that are within or overlapping the CD4-binding site.

We utilized a panel of gpl20 mutants for partial epitope mapping of 5145A. This panel was created primarily to study the effects of mutations within the constant regions of gpl20 upon binding of sCD4 and anti CD4-binding site mAbs to gpl20. Mutations that altered the global conformation of gpl20 were excluded from the panel. Previous studies were done using radioimmunoprecipitation/SDS gel analysis to detect binding of sCD4 or mAbs to mutant and wild-type gpl20/gpl60 molecules. We utilized an ELISA to assess this type of binding. Mutant or wild-type gpl20 molecules present in the supernatant of COS-1 cells in which they were expressed were captured by antibody against their C-terminus onto ELISA plates. Anti-gpl20 mAbs were then reacted with the bound gpl20 molecules, and binding of the mAbs was detected with enzyme-conjugated anti-Ig antibodies via standard techniques. This assay yields very similar results to those obtained by the radioimmunoprecipitation/SDS gel technique. -30-

as illustrated in Table 2 where results of 1125H binding to various gpl20 mutants in these two assays are compared.

In this determination of 5145A binding to this panel of mutants in ELISA, we sought to identify primarily mutations in gpl20 that would result in markedly decreased binding of 5145A, i.e. less than or equal to 20% of the binding to the mutant as seen to wild-type gpl20. Mutations in only two residues, 368 D and 370 E, had such an effect on 5145 binding. Interestingly, mutations in residues 256 T and 427 W that affected CD4 binding and binding of other anti-CD4- binding site mAbs did not result in a substantial decrease in 5145A binding to gpl20.

-31-

Table 2

Relative

Recognition Index CD-4-binding

Mutant 1125H ability Wild type 1.00 1.00 257 T/R 0.008 0.16 257 T/G 0.08 1.04 368 D/R <0.03 <0.004 368 D/T 0.03 0.33 370 B/R 0.02 <0.003 381 E/P ≥l.OO 1.09 427 W/S

0.03 <0.006

Determined by ELISA as described. Results for 5145A represent the mean of three independent experiments.

Standard deviation from the mean.

5145A did not lose substantial reactivity with gpl20 and gpl60 upon removal of N-linked carbohydrates from these glycoproteins under non-reducing conditions. This experiment was carried out by isolating virus from cells grown in the presence of 1-MNN, a substance that inhibits the processing of high mannose sugars to complex carbohydrates. The high mannose sugars on viral glycoproteins obtained from these 1-MNN-treated cells were then digested with endoglycosidase I (EH) under non-reducing conditions, resulting in the removal of all but one N- linked sugar residue (N-acetylglucosamine) from each glycan attachment site on the proteins. The fact that 5145A retains most of its reactivity with gpl60- and gpl20-derived viral glycoproteins deglycosylated by this procedure indicates that N-linked glycans are not essential components of the 5145A epitope nor are they required for maintaining its conformation. These studies do not rule out the possibility that the presence of N-linked glycans is required for the initial folding of the 5145A epitope into its correct conformation.

Synergistic Combination with 4117C

Figs. 5 and 6 show neutralization curves demonstrating that 5145A and 4117C (anti-V3 loop HuMAb) synergistically neutralize the MN and SF-2 strains, respectively. The plots of combination index vs. F. (% neutralization / 100) for the two synergistic neutralization experiments shown in Figs. 5 and 6 are given in Figs. 7A and 7B, respectively. Other parameters calculated from these two experiments using a computer program developed by Chou and Chou (Chou and Chou 1989), are shown in Table 3. The linear correlation coefficient, r, for the data shown in Figs. 5 and 6 is shown in Table 3 below. /19786

Table 3

DRI at % neutralization =

Human mAbs r m^c 50 70 90 95 97 99

MN neutralization

5145A 0.976 1.73 ± 0.19 6.5 6.0 5.2 4.8 4.6 4.1

4117C 0.997 1.11 ± 0.04 5.1 6.1 8.3 9.7 10.914.0

5145A + 4117C 0.993 1.47 ± 0.09 S

SF-2 neutralization

5145A 0.983 2.01 ± 0.19 5.8 5.2 4.5 4.2 3.9 3.5

4117C 0.982 1.15 ± 0.11 7.8 9.7 13.9 17.0 19.626.4

5145A + 4117C 0.994 1.65 ± 0.09 S

Dose reduction index = (D_x)n/(D)_n (ref. 20) ' Linear correlation coefficient ° Slope of the neutralization curve

Not applicable

3 1

-34-

R is high for each of the neutralization curves (where 1 is a perfect fit) , indicating that the data fits the median effect equation very well. Two conclusions can be reached from this data that further distinguish 5145A and its neutralizing activity from HuMAb 1125H. First, the slope of the neutralization curves, m, for 5145A alone vs. MN or SF-2 is closer to 2 than 1, indicating that 5145A exhibits cooperativity with itself in neutralizing HIV-l. In contrast, the slopes of comparable neutralization curves using 1125H against these same strains approximate l, indicating that 1125H does not exhibit this type of cooperativity. Secondly, Fig. 7 shows that the Cl values remain approximately constant with different F_Λ (% neutralization / 100) values, indicating similar degrees of synergism between 5145A and 4117C at different levels of viral neutralization (Chou et al. 1984) . In contrast, in synergistic neutralization between 1125H and 4117C, synergism increases dramatically (i.e., Cl values decrease dramatically) , as the % neutralization increases. Thus, while 1125H + 4117C exhibit greater synergism than 5145A + 4117C at higher levels of neutralization, the latter combination exhibits greater synergism than the former at low levels of neutralization (<25% neutralization) . These observations indicate that 51 5A and 1125H both exhibit strong synergistic neutralization of HIV-l when combined with 4117C. On a scale of 0 to +4 synergism, the 1125H and 4117C combination exhibits +3 to +4 synergism, whereas the 5145A + 4117C combination exhibits +3 synergism. However, the mechanism(s) by which 5145A and 1125H each neutralize virus alone and in combination with 4117C appear to be very different based on the m values and Cl v. F„ plots discussed above. LITERATURE CITED

All periodicals, patents and patent applications, and other references cited herein are hereby incorporated by reference.

Adachi, A., H.E. Gendelman, S. Keonig, T. Folks, R. Willey, A. Rabson, and M.A. Martin. (1986) J. Virol. 59:284- 291.

Anderson, L.J., Bingham, P., and Hierholzer, J.D. (1988) J. Virol. 62:4232-4238.

Berman, P.W. , T.J. Gregory, L. Riddle, G.R. Nakamura, M.A. Champe, J.P. Porter,, F.M. Wurm, R.D. Hershberg, E.K. Cobb, and J.W. Eichberg. (1990) Nature 345:622-625.

Berman, P.W., Gregory, T.J., Riddle, L. , Nakamura, G.R., Champe, M.A. , Porter, J.P., Wurm, F.M. , Hershberg, R.D., Cobb, E.K., and Eichberg, J.W. (1990) Nature. 345:622- 625.

Bonner, W. M. , and R. A. Laskey. (1974) Eur. J. Bioche . 46:83-88.

Chou, T.C. (1991) The median-effect principle and the combination index for quantitation of synergism and antagonism, p. 61-102 In T.C. Chou, and D.C. Rideout (eds.) Synergism and antagonism in chemotherapy. Academic Press, San Diego, CA.

Chou, J. and T.C. Chou. (1989) Dose-Effect Analysis with Microcomputers: Dose, Effect, Bindings and Kinetics, computer software for the IBM PC series (2.1 DOS, 5.25 floppy disk) , Biosoft, Cambridge, England .

Chou, T.C. and P. Talalay. (1984) Adv. in Enz. Regulation 22:27-55.

Clavel, F., D. Guetard, F. Brun-Vezinet, S. Chamaret, M. A. Rey, M. 0. Santos-Ferreira, A. G. Laurent, C. Dauguet, C. Katlama, C. Rouzioux, D. Klatz ann, J. L. Champalimaud, and L. Montagnier. (1986) Science 233:343-346. 3/19786

-36-

Cleff, J.C.S., Chanas, A.C. , and Gould, E.A. (1983) J. Gen. Virol. 64:11221-1126.

Co and Queen (1991) Nature 351:501-502.

Coller, H.A. , and Coller, B.S. (1987) Methods of Hybridoma Formation, A.H. Bartal, and Y. Hirshaut, eds. (Clifton, New Jersey: Humana Press) , pp. 231-236.

Devash, Y. , Calvelli, T.A. , Wood, D.G., Reagan, K.L., and Rubinstein, A. (1990) Proc. Natl. Acad. Sci. USA. 87, 3445-3449.

Dubuisson, J. , Guillaume, J. , Boulanger, D. , Thiry, E. , Bublot, M. , and Pastoret, P. (1990) J. Gen. Virol. 71:647- 653.

Eckhardt, L. A., S. A. Tilley, R. B. Lang, K. B. Marcu, and B. K. Birshtein. (1982) Proc. Natl. Acad. Sci. USA 79:3006-3010.

Ehrlich, P.H., Z.A. Moustafa, K.E. Harfeldt, C. Isaacson, and L. Ostberg, (1990) Hum. Antibod. Hybridomas 1:23-26.

Ehrlich, P.H., Z.A. Moustafa, J.C. Justice, K.E. Harfeldt, I.K. Gadi, L.J. Sciorra, F.P. Uhl, C. Isaacson, and L. Ostberg. (1988) Clin. Chem. 34:1681-1688.

Emini, E.A. , W.A. Schlief, J.H. Nunberg, A.J. Conley, Y. Eda, S. Tokiyoshi, S.D. Putney, S. Matsuchita, K.E. Cobb, CM. Jett, J.W. Eichbert, and K.K. Murthy. (1992) Nature 355:728-730

Emini, E.A. , Nara, P.L., Schleif, W.A. , Lewis, J.A. , Davide, J.P., Lee, D.R. Kessler, J. , Conley, S., Matsushita, S., Putney, S.D., Gerety, R.J. , and Eichberg, J.W. (1990) J. Virology. 64:3674-3678.

Gallo, R. , S.Salahuddin, M. Popovic, Q. Shearer, M. Kaplan, B. Haynes, T. Palker, R. Redfield, J. Oleske, B. Safai, G. White, P. Foster, and P. Markham. (1984) Science 224:500-503. -37-

Gerna, G. , Revello, M.G., Dovis, M. , Petruzzelli, E. , Achilli, G., Percivalle, E. , and Torsellini, M. (1987) J. Gen. Virol. 68:2007-2012.

Girard, M. , Kieny, M.P., Pinter, A., Barre-Sinoussi, F. , Nara, P., Kolbe, H. , Kusu i, K. , Chaput, A., Reinhart, T. , Muchmore, E. , Ronco, J. , Kaczorek, M. , Gomard, E. , Gluckman, J. , and Fultz, P. (1991) Proc. Natl. Acad. Sci, USA. 88:542-546.

Gorny, M.K. , V.Gianakakos, S. Sharpe, and S. Zolla- Pasner. (1989) Proc. Natl. Acad. Sci. USA 86:1624-1628.

Harada, S., Y. Koyanagi, and N. Yamamoto (1985) Science 229:563-566.

Harlow, E. , and D. Lane. (1988) in Antibodies: A laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, p. 310.

Ho, D. D., J.A. McKeating, X.L. Li, T. Moudgil, E.S. Daar, N. Sun, and J.E. Robinson (1991) J. Virol. 65:489- 493.

Ho, D.D. , Sarngadharan, M.G. Hirsch, M.S., Schooley, R.T., Rota, T.R., Kennedy, R.C., Chanh, T.C, and Sato, V.L. (1987) J. Virol. 61:2024-2038.

Homsy, J., Tateno, M. , and Levy, J.A. (1988) Lancet, i: 1285-1286.

Javaherian, K. , Langlois, A.J. , LaRosa, G.J. , Profy, A.T. , Bolognesi, D.P., Herlihy, W.C, Putney, S.D., and Matthews, T.J. (1990) Science. 250:1590-1593.

Jouault, T., Chapuis, F., Olivier, R. , Parravicini, C, Bahraoui, E. , and Gluckman, J.C. (1989) AIDS. 3:125-133.

Karpas, A., Hill, F., Youle, M. , Cullen, V., Gray, J. , Byron, N. Hayhoe, F. , Tenant-Flowers, M. , Howard, L. , Gilgen, D., J.K. Oates, Hawkins, D. , and Gazzard, B. (1988) Proc. Natl. Acad. Sci. USA. 85:9234-9237.

Gurgo C. , Guo H.G., Franchini G., Aldovini A., Collatti E., Ferrell K. , Wong-Staal F. , Gallo R.G., Reitz M.S. Jr. (1988) Virology 4164:531-536. Kazbor, D., A. E. Lagarde, and J. C. Roder. (1982) Proc. Natl. Acad. Sci. USA 79:6651-6655.

Kieny, M.P., Lathe, R. , Riviere, Y. , Dott, K. , Schmitt, D., Girard, M. , Montagnier, L. , and Lecocq, J. (1988) Protein Engineering. 2:219-225.

Kimura-Kuroda, J. , and Yasui, K. (1983) J. Virol. 45:124-132.

Kingsford, L. (1984) Virology. 136:265-273. Koyanagi, Y., Miles, S., Mitsuyasu, R.T., Merrill, J.E., Vinters, H.V. , and Chen, I.S.Y. (1987) Science 236:819-822.

Laemmli, U. K. (1970) Nature 227:680-683. LaRosa, G.J., J.P. Davide, K. Weinhold, J.A. Waterbury, A.T. Profy, J.A. Lewis, A.J. Langlois, G.R. Dressman, R.N. Bosell, P. Shadduck, L.H. Holley, M. Karplus, D.P. Bolognesi,. T.J. Matthews, E.A. Emini, and S.D. Putney. (1990) Science 249:932-935.

Larrick, J. W. , L. Danielsson, C. A. Brenner, E. F. Wallace, M. Abrahamson, K. E. Fry, and C. A. K. Borrebaeck. (1989) Biotechnology 7:934-938.

Leonard, C K. , M. W. Spellman, L. Riddle, J. N. Thomas, and T. J. Gregory. (1990) J. Biol. Chem. 265:10373-10382.

Ljunggren, K. , Bottiger, B., Biberfeld, G. , Karlson, A., Fenyo, E. , and Jondal, M. (1987) J. Immunol. 139:2263- 2267.

Matsushita, S., M. Robert-Guroff, J. Rusche, A. Koito, T. Hattori, H. Hoshino, K. Javaherian, K. Takatsuki, and S. Putney. (1988) J. Virol. 62:2107-2114.

McCune, J.M. , R. Namikawa, H.Kaneshima, L.D. Schultz, M. Lieber an, and I.L. Weisaman. (1988) Science 241:1632- 1639.

Moore, J.P., and R.F. Jarrett. (1988) Human Retroviruses 4:369-379.

Mosier, D.E., R. J. Gulizia, S.M. Baird and D.B. Wilson. (1988) Nature 335:256-259. Olshevsky, U. , Helseth, E. , Gurman, C. , Li, J. , Haseltine, W. , and Sodroski, J. (1990) J. Virol. 64:5701- 5707.

Orlandi, R. , D. H. Gussow, P. T. Jones, and G. Winter. (1989) Proc. Natl. Acad. Sci. USA 86:3833-3837.

Parham, P. (1986) In Cellular Immunology 4th Ed. (ed. D.M Weir), vol. 1, chap. 14. Blackwell Scientific Publications, California.

Peiris, J.S.M., Porterfield, J.S., and Roehrig, J.T., (1982) J. Gen. Virol. 58:283-289.

Pinter, A., W. J. Honnen, S. A. Tilley, C Bona, H. Zaghouani, M. K. Gorny, and S. Zolla-Pazner (1989) J. Virol. 63:2674-2679.

Pinter, A., and W. J. Honnen. (1988) J. Virology 62:1016-1021.

Popovic, M. , M. G. Sarngadharan, E. Read, and R. c Gallo. (1984) Science 224:497-500.

Putney, S.D., T.J. Matthews, W.G. Robey, D.L. Lynn, M. Robert-Guroff, W.T. Mueller, A.J. Langlois, J. Ghrayeb, S.R. Petteway, K.J. Weinhold, P.J. Fischinger, F. Wong-Staal, R.C Gallo, and D.P. Bolongesi. (1986) Science 234:1392- 1395.

Ratner, L. , A. Fisher, L.L. Jagodzinski, H. Mitsuya, R.S. Liou, R.C. Gallo, and F. Wong-Staal. (1987) AIDS Res. Human Retroviruses 3:57-69.

Ratner, L. , W. Haseltine, R. Patarca, K. J. Livak, B. Starcich, S. F. Josephs, E. R. Doran, J. A. Rafalski, E. A. Whitwhoen, K. Baumeister, L. Ivanoff, S. R. Petteway, M. L. Pearson, J. A. Lautenberger, T. S. Papas, J. Ghrayeb, N. T. Chang, R. C Gallo, and F. Wong-Staal. (1985) Nature 313:277-284. '

Raubitschek, A. A. (1985) in Human Hybridomas and Monoclonal Antibodies, E.G. Engleman, S.K.H. Foung, J. Larrick, and A. Raubitschek (ed.) Plenum Press, New York, P. 454-455. -40-

Ravetch, J. V., U. Siebenlist, S. Korsmeyer, T. Waldmann, and P. Leder. (1981) Cell 27:583-591.

Robert-Guroff, M. , Brown, M. , and Gallo, R.C (1985) Nature 316:72-74.

Robinson, W.E., Kawamura, T. , Lake, D. , Masuho, Y., Mitchell, W.M., and Hershy, E.M. (1990) J. Virol. 64:5301- 5305.

Robinson, J.E., D. Holton, S. Pacheco-Morell, J. Liu, and H. McMurdo. (1990) AIDS 4:11-19

Robinson, J.E., Holton, D. , Pacheco-Morell, S., Liu, J.and McMurdo, H. (1990) AIDS Res. Human Retroviruses 6:567-579.

Rook, A.H. , Lane, H.C., Folks, T., McCoy, S., Alter, H. , and Fauci, A.S. (1987) J. Immunol. 138:1064-1067. Russell, P.H. (1986) Arch. Virol. 90:135-144. Sakoyama, Y. , K.J. Hong, S.M. Byun, S. Ueda, Y. Yaoita, H. Hayashida, T. Miyata, and T. Honjo. (1987) Proc. Natl. Acad. Sci. USA 84:1080-1084.

Sanchez-Pescador , R. , M. D. Power, P. J. Barr, K. S. Steimer, M. M. Stempten, S. L. Brown-Shimer, W. W. Gee, A. Renard, A. Randolph, J. A. Levy, D. Dina, and P. A. Luciw. (1985) Science 227:484-492.

Scott, C.F., Silver, S., Profy, A.T., Putney, S.D., Langlois, A., Weinhold, K. , and Robinson, J.E. (1990) Proc. Natl. Acad. Sci. USA. 87:8597-8601.

Shaw, G. M. , B. H. Hahn, S. K. Arya, J. E. Groopman, R. C Gallo, and F. Wong-Staal. (1984) Science 226:1165-1171.

Shuford, W. , H.V. Raff, J.W. Finley, J. Esseltyn and L. Harris. (1991) Science 252:724-727.

Starcich, B. R. , B. H. Hahn, G. M. Shaw, P. D. McNeely, S. Modrow, H. Wolf, E. S. Parks, W. P. Parks, S. F. Josephs, R. C Gallo, and F. Wong-Staal. (1986) Cell 45:637-648.

Takeda, A., Tuazon, C.U., and Ennis, F.A. (1988) Science 242:580-583.

Teng, N. N. H. , K. S. Lam, F. c Riera, and H. S. Kaplan. (1983) Proc. Natl. Acad. Sci. USA 80:7306-7312. 3/19786

-41-

Thali, M. , U. Olshevsky, C. Fur an, D. Gabuzda, M. Posner, and J. Sodroski. (1991) J. Virol. 65:6188-6193.

Thali, M. , C Furman, D.D. Ho, J. Robinson, S.A. Tilley, A. Pinter, and J. Sodroski. (1992) J. Virol. 66:5635-5641.

Tilley, S.A. , W.J. Honnen, M. Racho, M. Hilgartner, and A. Pinter. (1991) Res. Virol. 142:247-259.

Ueda, S., F Matsuda, and T. Honjo. (1988) J. Mol. Evol. 27:77-83.

Volk, W.A. , Snyder, R.M. , Benjamin, D.C, and Wagner, R.R. (1982) J.Virol. 42:220-227. van Heyningen, V., and S. van Heyningen. (1987) in Methods of Hybridoma Formation, A.H. Bartal, and Y. Hirshaut (ed.). Humana Press, Clifton, New Jersey, p. 399-411.

Wain-Hobson, S. , P. Sonigo, 0. Danos, S. Cole, and M. Alizon. (1985) Cell 40:9-17.

Weiss, R.A. , Clapham, P.R., Cheingsong-Popov, R. , Dalgleish, A.G., Came, C.A., Weller, I.V.D., and Tedder, R.S. (1985) Nature 316:69-71.

Zolla-Pazner, S., Gianakakos, V., Williams, C, Sharpe, S., and Gorny, M.K. (1990) Production of human monoclonal antibodies against the V3 loop of gpl20. Sixth International Conference on AIDS, San Francisco. Abstract, Th.A. 75.

Claims

n r__- /19786-42-What is claimed is:

1. An anti-CD-4 binding site mAb which is specific for HIV-envelope glycoprotein gpl20, which mAb achieves at least about 50% in vitro neutralization of about 2 x 10⁴ infectious units of the RF HIV-l strain at a mAb concentration of about 1 μg/ml.

2. The mAb of claim 1 that achieves at least about 50% neutralization in vitro of about 2 x 10⁴ infectious units of the RF HIV-l strain at a mAb concentration of about 0.5 μg/ml.

3. The mAb of claim 2 that achieves at least about 50% neutralization in vitro of about 2 x 10⁴ infectious units of the SF-1 HIV-l strain at a mAb concentration of about 1.5 μg/ml

4. The mAb of claim 1 that is also capable of neutralizing in vitro the MN, IIIB, and SF-2 strains.

5. The mAb of claim 1 capable of neutralizing the RF, MN, IIIB and SF-2 strains substantially as shown in Figure 4.

6. The mAb of claim 1 which is about 90% competitively inhibited from binding to gpl60_IIIB by CD-4, wherein said CD- 4 is at about 1000 fold excess over the mAb.

7. The mAb of claim 1 which substantially has the epitope specificity of the mAb produced by the cells deposited at the A.T.C.C under accession no. CRL 10982.

8. The mAb of claim 1 which is derived from the cell line deposited at the A.T.C.C. under accession no. CRL 10982. -43 -

9. The mAb of claim 1 which is capable of neutralizing in vitro at least 50% of RF strain HIV-l at a concentration of about 0.5 μg/ml under conditions where the mAb produced by cells deposited at the A.T.C.C. under accession no. 10582 neutralizes at least about 50% of RF strain HIV-l at a concentration of about 5.0 μg/ml.

10. The mAb of claim 1 which is capable of neutralizing at least in vitro 50% of SF-2 strain HIV-l at a concentration of 0.5 μg/ml under conditions where the mAb produced by cells deposited at the A.T.C.C. under accession no. 10582 neutralizes 50% of SF-2 strain HIV-l at a concentration of 5.0 μg/ml.

11. The mAb of any of claims 1 through 10 which is a human mAb.

12. A human mAb of claim 1 specific for a gpl20 epitope which is conserved among the IIIB, MN, SF-2, and RF strains, which human mAb is specific for an epitope that is partially or completely within the CD-4 binding site, and which human mAb is capable of neutralizing in vitro about 2 x 10⁴ infectious units of virus of those four strains to a level of 50% at concentrations of mAb that do not differ by more than about three times.

13. The mAb of claim 12 which is capable of 50% neutralization of the four strains at concentrations of between about 0.2 - 0.5 μg/ml and about 99-100% neutralization at concentrations of about 10 μg/ml.

14. The mAb of claim 12 having an affinity as measured against recombinant gpl60_IIIB of at least about l x 10⁹ L/mole. -44-

15. The mAb of claim 14 having an affinity of 2 x 10⁹ L/mole.

16. A cell line producing the mAb of any of claims 1 through 15.

17. A combination of mAbs comprising

(a) an anti-CD4 binding site of any of claims 1 through 15; and

(b) a mAb that is specific for the V3 region of HIV-l envelope glycoprotein gpl20; which combination of mAbs is capable of synergistically neutralizing HIV-l infectivity in vitro.

18. The combination of claim 17 having a combination index value of less than 0.8 at greater than 20% neutralization of 10⁴ infectious units of HIV-l at a concentration no greater than about 3 μg/ml of the combination at an equipotent ratio.

19. The combination of claim 17 wherein the antibody specific for the V3 region competitively inhibits the binding to gpl60_HN or V3_HN of antibodies produced by the cell line deposited at the A.T.C.C. under accession no. CRL 10770.

20. The combination of claim 17 wherein the mAb specific for the V3 region has the epitope specificity of mAbs produced by the cell deposited at the A.T.C.C. under accession no. CRL 10770.

21. The combination of claim 17 comprising human mAbs produced by cells derived from cells deposited at the A.T.C.C under accession no. CRL 10982 and mAbs produced by cells-^{^}derived from cells deposited under accession no. CRL 10770.

22. The combination of claim 17 wherein the antibodies are in an equipotent ratio.

23. A kit for the detection of HIV antibodies comprising the mAb or mAbs of any of claims 1 through 15 and 17 through 22, a solid phase on which is coated an antigen which the mAb is specific for, and means for detecting the formation of a complex between the mAb and the antigen.

24. A reagent for neutralizing HIV-l infection comprising a mAb or mAbs of any of claims 1 through 15 and 17 through 22 in a physiologically compatible solution.

25. A method of preventing HIV infection comprising administering the mAb or mAbs of any of claims 1 through 15 and 17 through 22 to an individual.