Background of the Invention This invention relates to antibodies specific for Human Immunodeficiency Virus (HIV) .

HIV is the proposed causative agent of Acquired Immune Deficiency Syndrome (AIDS). (Popovic et al . ,

1984, Science 224:497). Different strains of HIV differ in the amino acid sequences of proteins encoded by the viral genome, particularly in the amino acid sequence of the external envelope glycoprotein gpl20 (Starcich, 1986, Cell 45:637; Hahn et al . , 1986, Science 232:1548). gpl20 binds the cellular receptor of the virus, CD . Cells expressing the envelope protein f_.se with CD4-bearing cells in culture (Lipson et al. 1986, Nature 323:725; Sodroski et al., 1986, Nature 322:470), resulting in the formation of multinucleate syncytia. Both native gpl20 and recombinant gpl20 elicit antibodies that are capable of neutralizing HIV in cell culture (Robey et al., 1986, Proc. Nat. Aca. Sci.

83:7023; athewε et al . , 1986, Proc. Nat. Aca. Sci. 83:9709; Laskey et al., 1986, Science 233:209; and Putney et al. , 1986, Science 234: 1392), These antibodies generally neutralize only the viral variant from which gpl20 was derived.

Over 100 HIV variants have been identified; among them are RF (Popovic et al . , supra) , WMJ-1 (Hahn et al., supra) , LAV (Wain-Hobson et al., 1985, Cell 40:9), ARV-2 (Sanchez-Pescador et al . , 1985, Science 227:484), and III-B (Ratner et al , , 1985, Nature

313:277). The majority of monoclonal antibodies that neutralize the HIV-III _B variant bind a specific region

of the III _Q gpl20 molecule referred to as the principle neutralizing domain (PNDT. which has been mapped to a 24 amino acid highly variable region of gpl20 (Matsushita et al., 1988, J. Virol. 62:2107; and Skinner et al. , 1988, AIDS Research and Human Retroviruses 4:187).

The principle neutralizing domain of the HIV gpl20 molecule is a 36 amino acid region of the gpl20 molecule between amino acids 303 and 338, inclusive, according to the gpl20 numbering convention of Ratner et al., supra. Over its entire length, the gpl20 polypeptide sequence varies from one HIV variant to the next by approximately 20-25%, whereas, the amino acid sequence variation among principle neutralizing determinant regions is approximately 40-50%. This highly variable region is flanked by conserved cysteine residues which may form a disulfide bond and define a "loop" region containing the largely conserved sequence Gly-Pro-Gly in its center. Synthetic loop region peptides, 8 amino acids or more in length, have been found to elicit the production of antibodies that neutralize.virus only from the isolates or variants of it from which the amino acid sequence of the peptide was derived. Human monoclonal antibodies directed against

HIV-l proteins have been produced by hybridoma formation or EBV transformation (Banapour et al., 1987, J. Immunol. 139:4027; Sugano et al., 1988, Biochem. and Biophys. Res. Co m. 155:1105; Morrow et al, 1988, J. Immunol. 140:941; Gorny et al., 1989, Proc. Nat. Aca. Sci. 86:1624; and Amadori et al. , 1989, AIDS Res. and Human Retroviruses 5:73).

Approximately 30% of humans infected with HIV are infected with a particular HIV variant,- MN, in which the gpl20 loop region contains the sequence I-G-P-G-R.

Summary of the Invention The invention features a human monoclonal antibody which neutralizes MN variants of Human Immunodeficiency Virus Type I, the antibody being produced by a cell line having A.T.C.C. Accession No■ Q _> ioι<jD , and the immortalized human cell line that produces the antibody.

The antibody of the invention can be used to inhibit HIV infection in a human patient infected with or suspected of having been infected with HIV. Administration of the antibody to a patient shortly after exposure or suspected exposure to the infectious agent may prevent the establishment of infection by the virus. For example, a patient may have accidently come into contact with HIV-contaminated blood, blood products, or bodily secretions. The antibodies may also prevent the transfer of HIV from a seropositive gravid female to her offspring by administering the antibody prior to or during pregnancy, and/or by administration to the offspring at birth and thereafter. The antibodies may also be used for passive immunization therapy; e.g., members of high risk groups who are still HIV-seronegative can be treated at regular intervals with an antibody preparation in order to prevent the establishment of a chronic HIV infection.

The antibody of the invention is, because of the widespread distribution of MN variants in infected persons, useful for detecting HIV in biological samples, for screening blood supplies, and, potentially, for treating a large percentage of HIV-infected patients.

Other features and advantages of the invention will be apparent from the following- description of the prefered embodiments thereof, and from the claims.

Description of Preferred Embodiments The drawings will be described briefly.

Fig. 1 is a Western blot analysis of human monoclonal antibody (HMab) reactivity with two strains of HIV-l.

Fig. 2 is a dot blot showing reactivity of four HMabs with gpl20 from different HIV strains.

Fig. 3 is a graph showing ELISA reactivity of

K24-3b and N70-2.3a HMabs with Con-A immobilized gpl20 from nine strains of HIV-l.

Fig. 4 is a graph showing ELISA reactivity of N70-2.3a, N70-1.5e, N70-1.9b (of the invention) HMabs with Con-A immobilized gpl20 from eight strains of HIV-l

Fig. 5 is a Western blot showing reactivity of

K24-3b and N70-2.3a HMabs with eight independent HIV-l strains. Isolation of MN-πeutralizing antibody producing cell line N70l9b

The procedure for isolating cell line N70-l.9b involved the steps of isolating lymphoid cells from a human patient who was asymptomatic for HIV infection but was HIV-1-seropositive, transforming those cells with Epstein Barr Virus (EBV) to immortalize them, and screening resultant lymphoblastoid cell lines for anti-HIV._, antibody production. Transformation of Human Lymphoid Cells

It has been observed (e.g., Gorny et al., 1989, supra; Yarchoan et al., J. Clin. Invest., 1986;

78:439-447) that peripheral blood B cells from HIV-l infected subjects vary greatly in their susceptibility

to EBV transformation. In general, B cells from patients with severely impaired immune function and relatively low CD4 cell counts are the most resistant to transformation, whereas B cells from asymptomatic patients with relatively high CD4 cell counts tend to transform more readily. However, transformation rates even within the population of apparently healthy asymptomatic patients are variable, and not all attempts to produce HMab's from this group have been successful. Human Monoclonal Antibodies

Cell line N70-1.9b (A.T.C.C. Accession No. ) was obtained by EBV transformation of peripheral blood B cells obtained from an asymptomatic but HIV-1- seropositive patient. (Herein, N70-l.9b is used to designate both the cell line and the antibody it produces.) The antigenic specificity of the N70-l.9b antibody was screened by ELISA using the gpl20 envelope protein containing the principle neutralizing domain from HIV™, and the antibody was investigated for HIV neutralization activity by inhibition of syncitium formation. The epitope recognized by N70-1.9b would also be expected to be expressed in the virus strain infecting the N70 donor.. Immortalization of Human Lymphoid Cells

Peripheral blood mononuclear cells (PBMC) were isolated on Ficoll-Hypaque gradients and were depleted of CD3 positive T cells using an indirect panning technique (Wysocki et al., Proc. Natl. Acad. Sci. USA, 1980, 75:2844-2848) in which cells reacting with the OKT3 monoclonal antibody were absorbed to petri dishes coated with F(ab) ₂ antibodies to mouse IgG. Non-adherent cells, enriched in B cells, were inoculated with the B95-8 strain of EBV (Miller et al . , Proc. Natl.

Acad. Sci. USA, 1973, 70:190-194) and plated at 10 ³ or 10 cells per well in 96 we'll tissue culture plates with irradiated human umbilical cord blood lymphocytes (HUCL) (10 cells per well) as feeder cells. Cultures were maintained in RPMI 1640 containing 5% fetal calf serum (FCS) and 1% Nutridoma-Hu ((Boehringer-Mannheim) , a serum substitute of low protein content. Isolation of Antibody Producing B cell Lines

A critical factor in HMab production is the availability of an efficient and sensitive im unoassay for screening hundreds of microwell cultures for antibody production. In the conventional ELISA, which is the basis of most commercial ELISA kits for serologic testing, purified viral antigens are passively coated in wells of ELISA plates. The preparation of antigens for this assay requires the production of very large amounts of virus, which then must be purified and inactivated. The process of virus purification may result in significant losses of gpl20. Hence, this assay may be inefficient in detecting antibodies to gpl20 and favor detection of antibodies to other ^' HIV antigens. This may explain in part the predominance of HMabs reacting with gag proteins or gp41 (Banapour et al . , 1987, supra; Sugano et al., 1988, supra; Morrow et al., 1988, supra; Gorny et al . , 1989, supra; Amadori et al. , 1989 supra) . Using the conventional antibody screening assay, 1 HMab was isolated, as follows.

In the first transformation experiment, EBV exposed, T cell-depleted PBMC from an HIV-l infected donor were plated at 10 cells per well in 96 well culture plates with irradiated HUCL feeder cells. Approximately 50% of the cultures were transformed after 4-5 weeks of culture. Culture fluids were then screened

by ELISA for IgG antibodies reacting with fixed, immobilized HIV-infected H9 cells, as f-oliows.

HIV-l infected H9 cells were immobilized in Concanavalin-A (Con-A) coated assay wells and then fixed with 1:1 acetone-methanol . The wells were blocked with RPMI-10% FCS for 1 hour. Fluids from 96 well cultures were transferred to wells in the assay plates. After 1 hour, wells were washed with phosphate buffered saline (PBS) containing 0.1% Triton-X 100 (PBS-TX) and then reacted with peroxidase-conjugated antibody to human IgG (Protos Labs, San Francisco, CA) . Color was developed with 100 ul tetramethylbenzidine (TMB)-H ₂ 0 ₂ as substrate. The reaction was stopped by the addition of H ₂ SO. and color was read as Optical Density at 450 n in a Titertek Multiskan ELISA reader.

One transformed culture, designated K24-3b, was found to be a stable producer of .antibody, which on further testing reacted by indirect immunofluorescence with both fixed and unfixed HIV-l infected cells but not with uninfected cells. Multiple subcultures of K24-3b cells were established a ^' t low cell density and all continued to produce antibody, although they ceased to grow after about 8 months. Because the original cells were plated at a relatively low cell density and the incidence of transformation was less than 50%, it is likely that the K24-3b cell line was established as a clone.

Because the initial conventional ELISA screening gave only one HIV specific HMab, a novel immunoassay was used to screen EBV transformed B cells from another HIV-1-seropositive patient. This immunoassay is based on the observation that HIV envelope glycoproteins bind via their carbohydrate

moieties to Con-A (Montagnier et al . , 1985, Virology 144:283). In this immunoassay, HIV-l glycoproteins released by infected cells grown in serum free medium are affinity-immobilized in Con-A coated assay wells. This procedure greatly simplifies the preparation of solid-phase glycoprotein antigens for large scale antibody screening. The assay is highly sensitive and selective in detecting antibodies to gpl20. Virus need not be purified; only small volumes of cells grown in serum free medium are needed to yield ample quantities of antigen for Con-A immobilization. Indeed, many serum free virus stocks can be diluted 1:2 or 1:4 without diminished antigen activity and thus, as little as 100 ul of supernatant fluid can be used to prepare 20-40 96-well ELISA plates.

In the second experiment, EBV exposed T cell-depleted PBMC from another HIV positive patient were seeded at 10 cells/well with irradiated HUCL in two 96 well plates. Transformation occurred in 100% of the wells. Culture fluids were screened by the novel

ELISA for IgG antibodies ^' reacting with Con-A immobilized viral glycoproteins derived from the J62 strain of HIV-l grown in MT4 cells in serum free medium, as follows.

Wells of Immulon-2 assay plates (Dynatech) were coated with 200 ug/ml Con-A in PBS and then incubated with 100 ul of detergent disrupted supernatant fluids from HIV-l producer cell lines grown for 2-3 days in serum free RPMI supplemented with 1% Nutridoma-Hu. In absence of serum components, disrupted viral glycoproteins present in such culture fluids bind to Con-A in amounts sufficient to function as solid phase antigens in a highly sensitive ELISA. Unreacted Con-A binding sites were blocked with RPMI-10% FCS for 1

hour, Control antigens were sin.il-a.rly_ prepared from culture fluids of uninfected MT4 cells. Transformed B cell culture fluid were transferred to both antigen coated and control wells of assay plates which were incubated at room temperature for 1 hour. Binding of antibodies was measured as described above. This ELISA was also used in later experiments to test the reactivity of HMabs with glycoproteins from different virus strains.

Ten transformed cultures produced IgG antibodies reacting with J62 glycoproteins but not with control antigen. Seven cultures produced antibodies for less than two months. Three cell lines, designated N70-2.3a, N70-1.5e, and 70-1.9b, respectively, were stable antibody producers and were cloned at 10 cells per well. Clones of each line were stable with respect to growth and antibody production for over 10 months.

IgG subclass and light chain type of each antibody was determined by reactivity with murine monoclonal antibodies to the four heavy chain subclasses (Behring Diagnostics) or polycloήal goat antibodies to lambda and kappa light chains in a sandwich ELISA, according to conventional isotyping techniques. All four HMabs are of the IgGl subclass; K24-3b, N70-1.5e, and N70-1.9b contain kappa light chains and N70-2.3a contains lambda light chains.

Characterization of HMab Specificity by Western Blot and Dot Blot Assays

The antigenic specificity of each HMab was determined using dot blot and Western blot assays. Twelve HIV-l strains were used as target antigens: strains C39, J62, SA90, SA96, and L86 were isolated from mitogen activated T cells of five asymptomatic HIV-l

infected subjects by co-cultivation with activated normal T cells in medium supplemented with interleukin-2; strain SA3 was similarly isolated from a patient with AIDS; strain HiTi is described in Rasheed et al., Virology, 1986, 154:395-400; strain K3 was obtained from the Tulane Delta Primate Center, New

Orleans, LA; HTLV-IIIB (Popovic et al . , Science, 1984,

224:497-500), the prototype HIV-l strain, was obtained from American Type Culture Collection; HTLV-III^.

(Gallo et al., Science, 1984, 224:500-502; Shaw et al. ,

Science, 1984, 226:1165-1170); baculovirus produced recombinant LAV gpl20 (American Biotechnologies, Inc.,

Cambridge, MA); as well as glycosylated recombinant gpl20 from HIV-l _gF2 (Levy et al., Science, 1984,

225:840-842), were obtained from the AIDS Research and

Reference Reagent Program. Strains C39, J62, SA96, and

SA90 were grown in MT4 cells (Harada et al., Science,

1985, 229:563-566); HTLV-IIIB, HTLV-III.^, SA3, HiTi, and K3 were grown in H9 cells.

Strain L86, isolated from the B cell donor of one monoclonal antibody (K24-3b), ^' did not replicate in continuous T cell lines and was propagated in mitogen activated cord blood T cells in medium containing 100 units per ml recombinant IL-2. To prepare antigens for

Con-A immobilization, cells infected with each virus strain were grown for 2-3 days in serum free medium RPMI supplemented with 1% Nutridoma-Hu. Clarified fluids were treated with 1% Triton-X and stored in aliquots at

-20°C until use.

Western blots were performed as follows. ₇

Extracts of 1-2 x.10 HIV-l infected cells prepared by solubilizing cells for 30 min in 1% Triton-X followed by removal of insoluble material by centrifugation in a

microcentrifuge. Samples were mixed 1:1 with SDS sample buffer without reducing agents and ^" heated for 5 min at 95°C, Cell lysates of uninfected H9 and MT4 cells were similarly prepared. Samples were fractionated by electrophoresis in 7.5% sodium dodecyl sulfate-polyacrylamide gels, in a BioRad mini-gel apparatus. Proteins were then electrophoretically transferred to nitrocellulose membranes. Western blot strips were incubated with blocking buffer (1% bovine serum albumin, 0.5% Tween 20, in 0.5 M NaCl, 10 mM Tris, pH 8), reacted first with each antibody preparation and then with alkaline phosphatase-conjugated antibodies to human or sheep IgG, as appropriate. Colored bands were developed using nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl-phosphate (NBT-BCIP, Sigma,

St. Louis, MO) as substrate. A sheep antiseru to gpl20 of HTLV-IIIB, obtained from the AIDS Research and Reference Reagent Program, was used as positive control in detecting gpl20/160. Figure 1 shows the reactivity of four HMabs on

Western blots of antigens of two HIV-l strains, HTLV-IIIB and J62. Lanes 1-5 of each of panels A and B are as follows: Lane 1, K24-3b; Lane 2, N70-2.3a; Lane 3, N70-ι.5e; Lane 4, N70-l.9b; Lane 5, sheep anti-HTLV-IIIB gpl20. On blots of HTLV-IIIB as the target antigen (Panel A), three HMabs (K24-3B, N70-2.3a, and N70-1.5e) reacted strongly with a prominent band of approximately 120 Kd and with a less intense band of 160 Kd, Although N70-1.9b appeared to react weakly with gpl20 on this blot (Panel A, lane 4), in ^: other assays it did not react with HTLV-IIIB at all. On blots prepared from strain J62 as the target antigen (Panel B), all four HMabs showed identical binding to a prominent band

at 160 Kd below which was a diffuse band extending to approximately 120 Kd; this pattern is characteristic for this strain. The staining patterns obtained with a polyclonal sheep antibody to gpl20 on blots of both strains were identical to that observed with the monoclonals (lane 5 in panels A and B) . The HMabs did not react with blots of uninfected MT4 or H9 cells (not shown) .

These results indicated that these four HMabs react with gpl20 and its uncleaved cellular precursor, gpl60. However, in view of the possibility that bands identified as gpl60/120 in some commercially available HIV-l Western blot strips are actually multimers of gp41 (Zolla-Pazner et al., New Engl . J. Med., 1989, 320:1280), the specificity of the four Hmabs for gpl20 was tested using dot blots of recombinant LAV gpl20 and lentil lectin purified J62 glycoproteins.

For dot blot assays, strips of nitrocellulose were dotted with 1 ul of baculovirus-produced recombinant LAV gpl20 at 100 ug/ml and J62 envelope glycoproteins, which were partially purified from detergent treated serum-free culture medium by lentil lectin affinity chromatography (Montagnier et al . , Virology, 1985, 144:283-289) and concentrated to 10 ug/ml . Recombinant gpl20 was also dotted after being heated for 5 min at 95°C in the presence or absence 2-mercaptoethanol. Antibody assays on dot blot strips were performed as for Western blots, except a goat antiseru to gpl60 of HTLV-IIIB (Rusche et al. , Proc. Natl. Acad. Sci. (USA), 1987, 84:6924-6928) was used as a positive control.

As shown in Figure 2, three of the four HMabs (K24-3b, N70-2.3a and N70-1.5e) reacted strongly with

recombinant gpl20. N70-1.9b did not bind to LAV gpl20 but did bind to J62 antigen. Fig. 2 i ^" s Keyed as follows: "gpl20 J62" is purified gpl20 of strain J62, "rgpl20 LAV is non-reduced recombinant LAV gpl20, "rgpl20 reduced" is reduced recombinant LAV gpl20, and "rgpl20 heated" is nonreduced, heated LAV gpl20. The amount of J62 antigen dotted was about 10 fold less than the recombinant antigen, explaining the weaker staining observed with this antigen. These results, therefore, indicate that the bands of 120 and 160 Kd observed on our Western blots indeed represent gpl20/160,

In preliminary Western blot studies, neither K24-3b nor N70-2.3a reacted with blots prepared from cell lysates heated in sample buffer containing 2-mercaptoethanol (not shown), suggesting that the epitopes identified were sensitive to reduction. To further test the effect of reduction on these epitopes, the antibodies were tested on dot blots of recombinant gpl20 LAV that was heated at 95°C in the presence or absence of 2-mercaptoethanol. The results shown in Figure 2 demonstrate that K24-3b ^' and N70-1.5e did not bind to reduce antigen and binding of N70-2.3a to reduced antigen was significantly diminished, while heating alone only slightly diminished antigenic activity. As N70-1.9b did not bind to LAV gpl20, the effect of reduction on its epitope was not determined in this experiment. N70-1.9b was subsequently tested on dot blots of reduced and non-reduced J62 glycoproteins and no reactivity was observed with reduced antigen. Thus, all four HMabs identify reduction sensitive epitopes.

Analysis of Strain Specificity of HMabs by ELISA

Because multiple virus strains-isolated from each asymptomatic B cell donor generally do not replicate in continuous cell lines (Cheng-Mayer et al., Science, 1988, 240:80-82) and usually can be propagated only in IL-2 dependent activated primary T cells or in monocytes, the preparation of solid phase antigens in sufficient amounts to develop screening immunoassays based on the passive coating method presented a problem. The Con-A immobilization technique offered a potential solution to this problem, because only small volumes of virus are needed rather than large quantities of purified viral antigen.

The four HMabs were tested by ELISA for reactivity with Con-A immobilized viral glycoproteins from different HIV-l strains. Theoretically, the binding of gpl20 to Con-A could block access of antibodies to some epitopes. However, we found that murine monoclonals known to react either with the CD4 binding region or the V3 hypervariable domain react strongly with Con-A immobilized gpl20 (unpublished) , indicating that epitopes within these two regions are represented in the assay.

As illustrated in Figure 3, one strain (L86) was grown in a serum free culture of IL-2 dependent, activated primary T cells and gpl20 released into the medium functioned well in the Con-A immobilization assay. Similar results may be achieved with other strains isolated from asymptomatic B cell donors; thus, it may become feasible to screen for antibodies reacting with antigens of homologous isolates.

In one experiment (Figure 3), culture fluids of K24-3b and N70-2.3a, and a HIV-l positive control serum

(H72) were tested on a panel- of nine different strains, which included L86, the strain isolated from the B cell donor of K24-3b. N70-2.3a reacted with all nine strains. Results shown as mean O.D. of triplicate determinations; standard deviation bars are shown. Although some differences in binding of this antibody on the panel were observed, generally parallel differences were observed with the positive control serum. Thus it is likely that the binding levels of both N70-2.3a and the H72 serum provide a relative measure of the amounts of gpl20 immobilized from each strain. We have no explanation for the much weaker reactivity of N70-2.3a with the L86 strain compared to the positive serum. However, L86 was the only strain grown in IL-2 dependent primary T cells, which release less virus than continous cell lines. It is possible that more gp41 than gpl20 was immobilized in the L86 virus preparation and antibodies to gp41 account for the greater serum reactivity.

By comparison to N70-2.3a, the K24-3b monoclonal showed remarkable variability in reactivity with these viruses. This antibody reacted with six of the nine strains but did not bind to strains SA3 or K3, and showed minimal binding to strain SA96. Whereas the reactivity of both N70-2.3a and K24-3b with strains L86 and J62 were very nearly the same, the binding of K24-3b to strains SA90 and C39 was much less than N70-2.3a, the difference being greatest with SA90. These observations have been reproducible in assays performed with different batches of antigens. Smaller differences in binding of these two antibodies was also apparent with strains HiTi and HTLV-III. These differences were not related to antibody concentrations, since preparations

of both antibodies used in 'these a ^~ ssays appeared to saturate available antigenic sites on immobilized antigens; optical densities obtained with serial dilutions of both antibodies up to 1:32 were very nearly the same (data not shown) when tested against the J62 isolate. These data indicate that N70-2.3a identifies a conserved epitope, while K24-3b identifies a variant epitope which is heterogeneously expressed in this panel of virus strains.

Figure 4 illustrates the results of a similar experiment comparing the reactivities of N70-1.5e and N70-1.9b with Con-A immobilized glycoproteins derived from eight strains. (Results are a single determination.) In this experiment, N70-2.3a served as a positive control. Both N70-1.5e and N70-2.3a reacted strongly with all eight strains, whereas N70-1.9b reacted only with J62, the strain that was used in the screening the original B cell cultures for antibody production. The results indicate that N70-1.5e, like N70-2.3a, reacts with an epitope shared by all strains tested thus far, while N70-1.9b reacts with a strain-restricted epitope. The reactivity of the four HMabs was investigated further using two additional target gpl60 antigens, HIV-I-^ and recombinant HIV-l _gF2 . The results, presented in Table 1, ^' show that N70-1.9b, as well as the other three HMabs, reacted strongly by ELISA with Con-A immobilized glycoproteins from HIV-I^ and HIv-ι _gF2 . Strain Specificity of HMabs by Western Blot Analysis

The strain specificity of two of the four HMabs, K24-3b and N70-2.3a, were also tested on Western blots prepared from the above panel of HIV-l strains. The results, shown in Figure 5, are in agreement with

results obtained by ELISA. Fig. 5 _^ panel A, shows reactivity of K24-3b; panel B, reactivity of N70-2.3a. The different strains are indicated at the top of each blot lane; "IIIB" refers to HTLV-IIIB. N70-2.3a reacted with gpl20/160 of all eight strains. K24-3b reacted with gpl20/160 of the same strains it identified by ELISA. Similarly, K24-3b failed to react with SA3 and K3; its minimal reactivity with strain SA96 was below the sensitivity of photography. Although N70-1.5e and N70-1.9b have not been similarly tested by Western blots on all of viruses, the strain restricted reactivity of N70-1.9b observed by ELISA is corroborated by its failure to react with recombinant LAV gpl20 in dot blot assays. Two additional transformation experiments, one involving B cells from the N70 donor, yielded over 60 transformed B cell cultures, including the anti-HIV-antibody producing clone N70-II.3a discussed below, that produce IgG antibodies reacting specifically with Con-A immobilized HIV glycoproteins. Approximately 30% of these cultures may yield stable antibody •producing clones, thus indicating the feasibility of generating sizeable numbers of HMabs which together represent a broad representation of human antibody responses to variant and conserved epitopes of gpl20 during asymptomatic infection. Screening for HIV^ Specificity

ELISA and Western Blot (WB) assays were performed on four human monoclonal antibodies, N70-1.9b, N70-1.5e, K24-3b, and N70-II.3a, to determine their antigenic specificity. In the ELISA, five different recombinant proteins or protein fragments were used as test antigens: gpl60-IIIB, gpl60-RF, PB-1-IIIB, PB-l-RF,

and PB-l-MN. In the Western Blot, three-different test antigens were used: gpl60-IIIB, PB-1-IIIB, and PB-l-MN. Intact gpl60 polypeptide was produced in insect cells using a baculoviruε expression system. A 180 amino acid protein fragment with an amino acid sequence spanning the principle neutralizing domain, denoted PB1 (Putney et al., 1986, Science 234:1392), was expressed from an approximately 540bp DNA fragment generated by PvuII + Bσlll cleavage of the full-length env gene.

ELISA was performed as follows. Each well of a 96-well Costar flat-bottom microtiter plate was coated with the antigen by placing a fifty microliter aliquot of a PBS solution containing the antigen at a final concentration of 2-10 ug/ml in each well. The Con-A method described above was not used here because the antigens (proteins or peptides) are purified and, therefore, immobilized in sufficient amounts for antibody binding. The antigen solution was aspirated and replaced with PBS + 0.5% BSA and incubated for 1 hour. Following incubation, the wells were then aspirated, washed, and 50 ul of the antibody was added. Following incubation, the wells were washed 3 times with PBS, and then incubated for 30 min. with 50 ul of an appropriate dilution of goat anti-human immunoglobulin conjugated with horseradish peroxidase (HRP, Boehringer Mannheim, West Germany). _. The wells were washed again 3 times with PBS and 50 ul of 1 mM ABTS (2,2 azino-bis (3-ethyl benzth, azoline 6-sulfonic acid) in 0.1M Na-Citrate, pH 4.2, to which a 1:1000 dilution of 30%

H ₂ 0 ₂ had been added), the substrate for HRP, was added to detect bound antibody. The ABTS samples were then read at OD. _1Q on a Dynatech spectrophotometric autoreader (Virginia) .

Tables 2-5 give ELISA results-. -Table 2 also gives results of a Western blot. Numbers in Table 2 indicate the number of times positive/number of times tested. Positive and negative controls for the ELISA < were HIV + serum and HIV - serum, respectively; the positive control in the Western Blot was goat anti-gpl60 IIIB antisera.

In the ELISAs, one of the five antibodies, N70-1.9b, was positive for the recombinant protein PB-1 MN. This result was clearly confirmed in the Western Blot (Table 2) and Tables 3 and 4 show results of ELISAs in which the PB-1 fragments from the envelope protein from different HIV variant strains (IIIB, RF, and MN) were test antigens for binding. The results in Tables 2-4 demonstrate that N70-1.9b binds specifically to the MN prototype but not to he IIIB or RF prototypes.

The ELISA results shown .in Table 5 demonstrate that the N70-1.9b monoclonal binds the principle neutralizing domain, or loop region, of the HIV,-, gpiβo molecule. N70-l.9b, N70-l.5e, and N70-II.3a were tested for their ability to bind ^" a fragment of the envelope protein from either the HIV-MN or the HIV-IIIB strain. "RP70" is the "full-loop closed" and "RP142" is the open 24mer from the principle neutralizing domain (PND) of the MN envelope protein; and "RP135" is a 24mer from the PND of the IIIB strain. These fragments contain amino acid sequences in the neutralizing domain sub-sequence of the gpl20 loop region as follows:

RP142: Y N K R K R I H I G P G R A F Y T T K N I I G (C)

RP70: I N C T R ^' P N Y N K R K R I H I G P G R A F Y T T K N I I G T I R Q A H C N I S

RP135 (IHg isolate) :

N N T R K S I R I Q R G P G R A F V T I G K I G (C)

Peptide RP70 was formed into a closed loop by creation of a disulfide bond between the two cysteine residues near the ends of the amino acid sequence. A method for creating such a bond is described in (Zhang et al . , 1988, Biochemistry 22:3785-3794). The results in Table 5 show that N70-1.9b binds to the principle neutralizing domain of the MN variant (RP70, RP142) but not to the PND of the IIIB variant (RP135).

In the experiments presented in Tables 2-5, the reagents were in reducing buffer. This may explain the apparently contradictory results of antibody binding in Table 1 versus Tables 2-5; i.e., the ELISA of Table 1 did not contain reducing buffer and, consequently, all of the antibodies bound gpl20/160. We conclude that the presence of reducing buffer results in more selective antibody binding. Neutralization of HIV„„ Two.of the antibodies, one of which was

N70-1.9b, were then assayed for inhibition of syncitium formation by HIV-^, infected cells. In this assay, recombinant Vaccinia Virus expressing the envelope gene of HIV^ was used to infect cells of CD4+ human T-lymphoma line CEM (A.T.C.C. Accession No. CCL119), and the antibody was then added to the cells to screen for blockage of HIV envelope mediated cell fusion. A positive result, indicating the ability of the antibody to neutralize the virus, was defined to be at least a 90% inhibition of syncytia formation.

CEM cells were infected with r-ecombinant Vaccinia Virus expressing HIV^ gpl60 derived from plasmid pSCR2502, which contains the PB-l fragment of MN; the remainder of gpl60 was of IIIB origin. In this assay, syncytia are induced which are inhibitable by antisera or monoclonal antibodies directed against the PND. The results, shown in Table 6, show that N70-1.9b inhibits syncytia induced by vaccinia gpl60-MN over a range .of concentrations of the antibody, whereas N70-l.5e does not inhibit the formation of syncytia. Use

When employed to treat individuals infected by HIV or suffering from AIDS, the human monoclonal antibody of the present invention may be administered as a passive immunization agent in effective amounts broadly ranging between about 200 mg and about 15 grams and preferably between 50 mg and .1 gram.

The antibody of the invention is administered parenterally, either via the intravenous or intramuscular route. A typical treatment regimen would comprise administration of an effective amount of antibody administered over between about one week and about 6 months, The number of treatments required to control a patient's disease may vary from individual to individual, depending upon the severity and stage of the illness and the individual characteristics of each patient being treated. The total dose required for each treatment may be administered by multiple doses or in a single dose. The human monoclonal antibodies may be administered alone or in conjunction with other HIV treatments, such as AZT, in order to control a patient's disease. The anti-HIV treatment may be administered one or two times a week or more as determined by the

patient's condition and the stage -of the_patient' s disease.

The human monoclonal antibody of the present invention can be incorporated into conventional pharmaceutical formulations for use in treating individuals that are afflicted with HIV or for prophylaxis in individuals at risk for such infections. The pharmaceutical formulations of the invention comprising an anti-HIV effective amount, range between about 200 mg and about 15 grams, of the human monoclonal antibody of the present invention. The quantity of effective dose applied by each injection is relatively unimportant since the total dosage can be reached by administration of one or a plurality of injections. In addition, such formulations may comprise pharmaceutically-acceptable carriers, diluents, salts and other materials well-known in the art. Isotonic saline, sterile water, 10% maltose, human serum albumin, glycine or other pharmaceutically-acceptable materials may be used as diluents, carriers or solvents in preparing the pharmaceutical formulations comprising the human monoclonal antibody of the present invention.

Deposit Cell line N70-1.9b was deposited in the American Type Culture Collection on November 1, 1989, and assigned Accession Number tf£ ^ø _.

Applicants' assignees, Louisiana State University and Repligen Corporation, represent that the A.T.C.C. is a depository affording permanence of the deposit and ready accessibility thereto by the public if a patent is granted. All restrictions on the availability to the public of the material so deposited will be irrevocably removed upon the

granting of a patent. The material- ill _be available during the pendency of the patent application to one determined by the Comissioner to be entitled thereto under 37 C.F.R. 1.15 and 35 U.S.C. 122. The deposited material will be maintained with all the care necessary to keep it viable and uncontaminated for a period of at least five years after the most recent request for the furnishing of a sample of the deposited microorganism, and in any case, for a period of at least thirty (30) years after the date of deposit or for the enforceable life of the patent, whichever period is longer. Applicants' assignees acknowledge its duty to replace the deposit should the depository be unable to furnish a sample when requested due to the condition of the deposit. A copy of the A.T.C.C. Budapest Treaty deposit receipt will be furnished upon request. -

Table 1

Reactivity of Four HMabs with Con-A Immobilized gpl20 of Three HIV-l Strains

Optical Density with Indicated Strain

HMab HTLV-IIIMN** rgpl20/HIV-l _SF *** J62 **

Background reactivity of HMabs with blocked Con-A coated wells without antigen was < 0.100

** HTLV-III,,-. and J62 viruses grown in serum free medium as in Figs. 4 and 5

*** Recombinant glycos ^' ylated gpl20 from HIV-l _gF produced in Chinese hamster ovary cells; gpl20 incubated at 1 ug/ml in Con-A coated wells

Table 2

ELISA W.B. gpl60 gplδO PB-1 PB-1 PB- gpl60 PB-1 PB-1 IIIB RF IIIB RF MN IIIB IIIB MN

HIV+ sera 4/4 3/3 1/3 3/3 3/3 N.D.

HIV- sera 0/4 0/3 0/3 0/3 0/3 N.D. goat αgplβO(IIIB) sera N.D. 2/2 1/1 N.D

N70-1.9b 0/4 0/3 0/3 0/3 2_/3 0/2 0/1* 1_/1

N70-II.3a 0/4 0/3 0/3 0/3 0/3 2/2 0/1 0/1

K24-3B 0/4 0/3 0/3 0/3 0/3 0/2 0/1 0/1

N70-1.5e 0/4 0/3 0/3 0/3 0/3 .1/2 0/1 0/1

*very weak positive

Table 3

Exp. 1 Capture Antigen

Table 4

Exp. 2 Capture Antigen

104 95