IL102092A - Use of recombinant hiv envelope protein in medicament for treating hiv and therapeutic composition containing the agglomerated protein - Google Patents

Description

102092/2 HlV-n iD> noi-ira >oiJ >:ιαϊρ>-ι nioy HIV in n t?w \yia>\y -irmon iin nn on mnpn -ι>νοηι USE OF RECOMBINANT HIV ENVELOPE PROTEIN IN MEDICAMENT FOR TREATING HIV AND THERAPEUTIC COMPOSITION CONTAINING THE AGGLOMERATED PROTEIN MICROGENESYS , INC.

C: 14630 I-IO O 1 BACKGROUND OF THE INVENTION The Human Inununodeficiency Virus Type-1 (HIV-1) is a retrovirus which causes a systemic infection with a major pathology in the immune system and is the etiological agent responsible for Acquired Immune Deficiency Syndrome (AIDS) . Barre-Sinoussi , et al., Science, 220 : 860-871 (1983); Popovic et al.f Science. 224 : 497-500 (1984). Clinical isolates of HIV-1 have also been refnrred to as Lymphadenopathy-Associated Virus (Feorino, et al., jcience. 225; 69-72 (1984) and AIDS-related Virus (Levy et al.,- Science 225: 840-842 (1984)).

AIDS has become pandemic and the development of a vaccine has become a major priority for world public health. A high percentage of persons infected with HIV-1 show a progressive loss of immune function due to the depletion of T4 lymphocytes. These T4 cells, as well as certain nerve cells, have a molecule on their surface called CD4. IIIV-1 recognizes the CD4 molecule through a receptor located on the envelope of the virus particles, enters these cells, and eventually replicates and kills the cell. An effective AIDS vaccine might be expected to elicit antibodies which would bind to the envelope of HIV-1 and prevent it from infecting T4 lymphocytes or other susceptible cells.

Vaccines are generally given to healthy individuals before they are exposed to a disease organism as an immune prophylactic. However, it is also reasonable to consider using an effective AIDS vaccine in post-exposure immunization as immunotherapy against the disease. Salk, J., Nature . 327 ; 473-476 (1987) .

It is widely believed that the HIV-1 envelope f "env") is the most promising candidate in the development of an AIDS vaccine. Francis and Petricciani. New Eng. J. Med.. 1586-1559 (1985) ; Vogt and Hirsh, Reviews of Infectious Disease, 8: 991-1000 (1986); Fauci, Proc. Natl. Acad. Sci. USA. 83: 9278-9283. The HIV-1 envelope protein is initially synthesized as a 160,000 molecular weight glycoprotein (gpl60) . The gpl60 precursor is then cleaved into a 120,000 molecular weight external glycoprotein (gpl20) and a 41,000 molecular weight transmembrane glycoprotein (gp41) . These envelope proteins are the major target antigens for antibodies in AIDS patients. Barin, et al., Science. 228: 1094-1096 (1985). The native HIV-1 gpl20 has been shown to be immunogenic and capable of inducing neutralizing antibodies in rodents, goats, rhesus monkeys and chimpanzees. Robey, et al., Proc. Natl. Acad. Sci. USA 83:7023-7027 (1986).

Due to the very low levels of native HIV-1 envelope protein in infected cells and the risks associated with preparing an AIDS vaccine from HIV-1 infected cells, recombinant DNA methods have been employed to produce HIV-1 envelope antigens for use as AIDS vaccines. Recombinant DNA technology appears to present the best option for the production of an AIDS subunit vaccine because of the ability to produce large quantities of safe and economical immunogens. The HIV-1 envelope protein has been expressed in genetically altered vaccinia virus recombinants. Chakrabarti, et al., Nature, 320: 535-537 (1986); Hu, et al., Nature , 320: 537-540 (1986); Kieny, et al., Biotechnology . 4:790-795 (1986) . The envelope protein has also been expressed in bacterial cells (Putney, et al., Science, 234 : 3 1392-1395 (1986)), in mammalian cells (Lasky, et al., Science, 2_3:209-12 (1986)), and in insect cells. Synthetic peptides derived from amino acid secpaences in an HIV-1 gp41 have also been considered as candidate AIDS vaccines. Kennedy, et al. (1986). However, e. successful AIDS vaccine has not been produced using these materials and methods.

The use of a baculovirus-insect cell vector system to produce recombinant HIV-1 envelope proteins is one aspect of the invention disclosed in coassigned E* 327,180 and IL Q l^.

• - .' ■ (See also. BP 265,785 and IL 98118.) The baculovirus system has been demonstrated to be of general utility in producing HIV-1 proteins and other proteins. As examples, the baculovirus Autographa californica nuclear polyhedrosis virus (AcNPV) has been used as a vector for the expression of the full length gpl60 and various portions of the HIV-1 envelope gene in infected Spodoptera frugiperda (fall armyworm) cells (Sf9 cells) . Also disclosed in the prior copending patent applications is the truncated gpl,60 gene (recombinant number Ac3046) , the protein produced from recombinant Ac3046, and a purification technique for the Ac3046 gene product that includes lentil lectin affinity chromatography and gel filtration chromatography. The gpl60 protein purified in this manner and aggregated to form particles was found to be highly immunogenic in rodent and primate species.

The ideal AIDS vaccine, in addition to the requirements of being substantially biologically pure and non-pyrogenic, should provide life-long protection against infection with HIV-1 after a single or a few injections. This is usually the case with live attenuated vaccines. When killed bacteria or viruses, or materials isolated from them, such as toxoids or proteins, are used to make a vaccine, there often results a poor antibody response and only short term immunity. To overcome or minimize these deficiencies in a vaccine, an additional component, called an adjuvant, may be added. Adjuvants are materials which help 1 1Θ2092/2 4 stimulate the immune response. Adjuvants in common use in human vaccines are gels of aluminum salts (aluminum phosphate or aluminum hydroxide) , usually referred to as alum adjuvants. Bomford, et al., "Adjuvants," Animal Cell Biotech. 2: 235~250. Academic Press Inc. (London: 1985) .

ACKNOWLEDGMENT OF CITED PRIOR ART EP Ο265785 describes use of the baculovirus-insect cell vector system for cloning and expression of recombinant proteins such as HIV envelope protein, said to be a candidate in the development of vaccines and diagnostics for AIDS.

EP O327I8O describes undenatured or native HIV-1 envelope protein and an AIDS vaccine which contains it.

SUMMARY OF THE INVENTION In one aspect', the present invention provides use of a recombinant HIV envelope protein in the manufacture of a medicament for treating an individual infected with human immunode iciency virus (HIV) , substantially as described herein.

In another aspect, the present invention provides a therapeutic composition for use in treating an individual infected with human immunodeficiency virus (HIV) , which comprises a recombinant HIV envelope protein and an alum adjuvant, wherein the recombinant protein is agglomerated into particles having a molecular weight of at least about 2,000,000.

NOTICE UNDER THE REGISTRAR'S CIRCULAR NO. 23 (P) OF APRIL 5, 1 92 Since the invention is defined in the appended claims, it will be apparent that the passages of the present specification which fall outside the scope of the claims do not relate directly to the invention. 1020921 BRIEF DESCRIPTION OF THE DRAWINGS Details of this invention are set forth below with reference to the accompanying drawings: Fig. 1 illustrates the cloning strategy used to isolate the HIV-1 envelope gene (env) from the E. coli plasmid pNA2. The hatched regions are HIV-1 DNA sequences and the open regions are from the cloning vectors. The black region in the plasmid pl774 is constructed from synthetic oligonucleotides and was introduced as an Smal-Kpnl fragment into the Smal-Kpnl sites of plasmid pl614. The sequence of this synthetic oligonucleotide is shown.

Fig. 2 illustrates the strategy used to construct the recombinant plasmid vector (p3046) , which in turn is used to construct the baculovirus expression vector Ac3046. The plasmid pMGS3 contains sequences (cross-hatched areas) from the baculovirus AcNPV on either side of a cloning site at position 4.00. This site has the unique restriction endonuclease sites for Smal, Kpnl, and Bglll. The AcNPV polyhedrin promoter is in the 5· direction from the 4.00 position. The sequence 5 · -TAATTAATTAA-31 is in the 3' direction, and has a translational termination codon in all three reading frames. The plasmid pl774 and the sequence of the synthetic oligonucleotide region is as described in Fig. l. The plasmid p3046 contains all of pMGS3 except for the 5 102092/2 sequences between the Smal and Bglll sites, where the HIV-1 envelope gene of pl774 is inserted. Fig. 3 shows the nucleotide sequences of the DNA flanking the Ac3046 gpl60 coding sequences. The 3046 env DNA sequence between +1 and [+2267 is shown in Fig. 4.

Figs. f4a-4t? show the actual DNA sequence of the HIV-1 env gene segment along with the synthetic oglionucleotide sequences at the 5' end of the env gene in Ac3046 (between +1 and +2267')·. The locations of restriction endonuclease sites are listed above the DNA sequence and the predicted amino acid sequence is listed below the DNA sequence. The bases are numbered on the right and on the left.

Figs. /5a-5h compare the DNA sequences of the env gene from Ac3046 with a published env gene sequence from LAV-1. The LAV-1 sequence is on the top and Ac3046 is on the bottom. A line (1) below the LAV-1 sequence indicates that the sequence in Ac3046 is the same in this position. The DNA sequence numbering used is that described by Wain-Hobson, et al., Cell . 40.: 9-17 (1985) for LAV-1.

Fig. 6 shows the ELISA end point dilution titers of human HIV-1 antibody positive sera (top graph) and rhesus monkey sera (bottom graph) from animals immunized with gpl60 (IJ55, KL55) or gpl20 (AB55, CD55, GH55) . The ELISA titers were measured against highly purified gpl20 and gpl60 proteins. The specifically bound antibody was measured with a goat anti-human IgG HRP conjugate. The highest dilution of serum that gives a positive response in the test is the titer.

Fig. 7 is a Table summarizing the gpl60 Vaccine-induced immune responses of vaccinated seropositive patients.

Fig. 8 (A and B) shows vaccine-induced antibody responses directed against specific HIV envelope epitopes.

Fig. 9 shows the vaccine-induced T-cell proliferative responses to gpl60 in vaccinated seropositive individuals.

Fig. 10 (A.C) shows the lymphocyte proliferation responses associated with vaccination. 6 102092/ Fig. 11 is a graph showing the percent change in CD4 cells in responders and non-responders over time.

In accordance with the present invention it has been discovered that recombinant IIIV-I gpl60 envelope protein ( "rgpl60" ) , especially when adsorbed onto an adjuvant such as, alum (e.g., aluminum phosphate) is particularly useful as an AIDS vaccine. One aspect of this invention is an AcNPV expression vector having the coding sequence for a portion of the HIV-1 envelope gene which encompasses the amino acids equivalent to 1-757 of the sequence in Wain-llobson (ibid) , which are amino acids 1-752 found in the recombinant clone No. 3046. Another aspect of the invention is the production of that recombinant HIV-1 envelope protein (and the protein itself) in insect cells — especially the rgpl60 protein , coded for by the amino acid sequences 1-757 of Wain-Hobson (i.e., 03046 having 752 amino acid residues) .

Other aspects of this invention comprise purification and formation of recombinant envelope protein particles from the gene product of the recombinant baculovirus that produces the 3046 protein and adsorption of the 3046 particles to aggregates of aluminum phosphate. .

The invention also comprises prophylactic and/or therapeutic vaccines for AIDS or HIV infection and methods of preventing or treating AIDS or HIV infection.

DETAILED DESCRIPTION OF THE INVENTION The following examples illustrate the invention without limiting its scope.

The recombinant baculovirus Autographa californica nuclear polyhedrosis virus (AcNPV) which contains a truncated HIV-1 gpl60 gene coding for amino acids 1-757 of the HIV envelope protein (recombinant Ac3046) is described in coassigned E* 327,100 and IL Qblfr.

The cloning steps employed to construct the recombinant baculovirus-containing genes or portions of genes from HIV-1 are also disclosed there and are incorporated by. 7 reference.

The following is a detailed description of the genetic engineering steps used to construct the Ac3046 expression vector. The materials employed, including enzymes and immunological reagents, were obtained from commercial sources. Examples showing how to make and use the invention are also provided.

Other recombinant envelope proteins, referred to collectively as rgpl60, are also contemplated, and include recombinant gpl20 and gp41 proteins. Ac3046 is just one example of an expression vector and recombinant envelope protein according to the invention.

EXAMPLE 1 Construction of the baculovirus recombinant Ac30 6 bearing the HIV-1 coding sequences for amino acids 1-757 Cloning and expression of foreign protein coding sequences in a baculovirus vector requires that the coding sequence be aligned with the polyhedrin promoter and upstream sequences on one side and with baculovirus coding sequences on the other side. The alignment is such that homologous recombination with the baculovirus genome results in transfer of the foreign coding sequence aligned with the polyhedrin promoter and an inactive polyhedrin gene.

Accordingly, a variety of insertion vectors were designed for use in HIV envelope gene constructions. The insertion vector MGS3, described below, was designed to supply the ATG translational initiating codon. Insertion of foreign sequences into this vector must be engineered such that the translational frame established by the initiating codon is maintained correctly through the foreign sequences.

The insertion vector MGS3 was constructed from an EcoRI-I restriction fragment clone of DNA isolated from a plaque purified AcMNPV isolate (WT-1) . MGS3 was designed to consist of the following structural features: (a) 4000 bp of sequence upstream from the ATG initiating codon of the polyhedrin gene; (b) a polylinker introduced by site-directed mutagenesis, which 8 consists of an ATG initiating codon at a position of the corresponding polyhedrin codon, and restriction sites Smal, Kpnl, Bglll and a universal stop codon segment; (c) 1700 bp of sequence extending from the Kpnl restriction site (which is internal to the polyhedrin gene) through to the terminal EcoRI restriction site of the EcoRI-I clone. See, e.g.. Fig. 2.

EXAMPLE 2 Construction of baculovirus recombinants bearing LAV env coding sequences A recombinant plasmid designated NA2 (Fig. 1) consists of a 21.8 kb segment of an entire HIV-1 provirus inserted into pUC18. This clone was reportedly infectious since it could produce virus following transfection of certain human cells. Adachi, et al., J. Virol. 59:284-291 (1986). The complete envelope gene sequences contained in NA2 were derived from the LAV strain of HIV.. Barre-Sinoussi (1983) .

The HIV-1 envelope gene was isolated and engineered as described below, and as shown in Fig. 1. The envelope gene was initially isolated from NA2 as a 3846 bp EcoRI/SacI restriction fragment and cloned into the EcoRI/SacI restriction site pUC19. The resultant plasmid was designated p708.

The envelope gene was subsequently reisolated as a 2800 bp Kpnl restriction fragment and cloned into the Kpnl restriction site of pUC18. The resulting clone was designated pl614.

The Kpnl restriction fragment in pl614 contained a slightly truncated piece of the HIV envelope gene such that 121 bp of the N-terminal corresponding sequence was missing. This missing part in the gene, which included the signal peptide sequences, was replaced by insertion of a double-stranded synthetic oligomer. The inserted oligomer was designed from the LAV amino acid sequence using preferred polyhedrin gene codon usage. To facilitate further manipulation, a new Smal restriction sequence was concomitantly introduced in place of the ATG initiating codon. The ATG initiation codon will be supplied by the baculovirus insertion vector. The resultant plasmid was 9 designated pl774.

Referring to Fig. 2, restriction fragments from pl774 containing coding sequences of various domains of the HIV-1 envelope were cloned into the MGS insertion vectors. (e.g.. MGS3) such that the ATG initiating codon of the insertion vector was in-frame with the codons of the envelope gene. Construct p30 6 consisted of the Smal/BamHI restriction fragment isolated from pl774 inserted into the Smal/Bglll site of the plasmid vector pMGS3. This clone contains sequences coding for amino acids l through 757 of gpl60 and uses a termination codon supplied by the MGS3 vector.

EXAMPLE 3 Preparation and Selection of Recombinant Baculovirus The HIV env gene recombination plasmid p3046 was calcium phosphate precipitated with AcMNPV DNA (WT-1) and added to uninfected Spodoptera frugiperda cells. The chimeric gene was then inserted into the AcMNPV genome by homologous recombination. Recombinant viruses were identified by an occlusion negative plaque morphology. Such plaques exhibit an identifiable cytopat-hic effect but no nuclear occlusions. Two additional successive plaque purifications were carried out to obtain pure recombinant virus. Recombinant viral DNA was analyzed for site-specific insertion of the HIV env sequences by comparing their restric-tions and hybridization characteristics to wild-type viral DNA.

EXAMPLE 4 Expression of HIV env from recombinant baculoviruses in infected insect cells Expression of HIV env sequences from the recombinant viruses in insect cells should result in the synthesis of primary translational product. This primary product will consist of amino acids translated from the codons supplied by the recombination vector. The result is a protein containing all the -amino acids coded for from the ATG initiating codon of the expression vector downstream from the polyhedrin promoter to the 10 102092/2 translational termination signal on the expression vector (e.g., rgpl60) . The primary translation product of Ac3046 should read Met-Pro-Gly-Arg-Val at the terminus where Arg (position 4) is the Arg at position 2 in the original LAV clone. The Met-Pro-Gly codons are supplied as a result of the cloning strategy.

EXAMPLE 5 Nucleotide sequence of the σρΐ θ insert and flanking DNA.

The nucleotide sequence of the gpl60 insert and flanking DNA was determined from restriction fragments isolated from viral expression vector Ac3046 DNA. The sequencing strategy involved the following steps. The 3.9 kb EcoRV-BamHI fragment was purified by restriction digestion of Ac3046 viral DNA. The Ac3046 viral DNA had been prepared from extracellular virus present in the media of cells being used for a productio lot of vaccine.

As shown in Fig. 2, the 3.9 kb EcoRV-BamHI fragment consists of the entire gpl60 gene and 100 bp of upstream and about 1000 bp of downstream flanking DNA. Of this, the nucleotide sequence of the entire gpl60 gene was determined, including 100 bp of upstream and 100 bp of downstream flanking DNA.

Briefly, the results of the sequencing revealed a chimeric construct as predicted from the cloning strategy. The sequence of the gpl60 was essentially as reported by Wain-Hobson, et al. (1985). The sequence of ij[256~/bases between the presumed translation initiation and termination codons predicts ^752"? amino acid codons and 28 potential N-linked glycosylation sites. The estimated molecular weight of this rgpl60, including the sugar residues, is approximately 145,000.

Sequence analysis of 200 bases of flanking DNA indicated correct insertion as shown in Figs. 3, 4 and 5.

EXAMPLE 6 Amino Acid Sequence of qpl60 Using standard automated Edman degradation and HPLC procedures, the N-terminal sequence of the first 15 residues of gpl60 was determined to be identical to that predicted from the DNA sequence. The N-terminal methionine is not present on the gpl60 protein. This is consistent with the observation that AcNPV polyhedrin protein is also produced without an N-terminal methionine. A summary of the actual gpl60 DNA and N-terminal protein sequences, as has been determined by analysis of the AcNPV 3046 DNA and purified gpl60, is as follows (Table 1) .

TABLE 1 LAV env gene in the AcNPV 3046 expression vector Residue 2 3 4 5 6 7 8 9 10 11 12 13 14 Pro Gly Arg Val Lys Glu Lys Tyr Gin His Leu Trp Arg Trp Gly ATG CCC GGG CGT GTG' AAG GAG AAG TAC CAA CAC CTG TGG CGT TGG GGC These results compare to the original LAV-1 clone as follows (Table 2) .

TABLE 2 LAV env gene in the original LAV-1 clone Residue 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Met Arg Val Lys Glu Lys Tyr Gin His Leu Trp Arg Trp Gly ATG AGA GTG AAG GAG AAG TAT CAG CAC TTG TGG AGA TGG GGG EXAMPLE 7 Purification of Recombinant qpl60 One aspect of the present invention is the procedure used to extract and purify the recombinant HIV-1 envelope protein coded for in the Ac30 6 expression vector. The recombinant HIV-1 envelope protein gpl60 -is produced in S. fru iperda cells during 12 4-5 days after infection with Ac3046. Purification of this rgpl60 protein involves the steps: 1. Washing the Cells 2. Cell Lysis 3. Gel Filtration Chromatography 4. Lentil Lectin Affinity Chromatography 5. Dialysis This example describes the purification of the recombinant gpl60 from about 2 x 10 Ac3046 infected cells. 1. Washing the cells. Infected cells are washed in a buffer containing 50 mM Tris buffer (pH 7.5), 1 mM EDTA and 1% Triton X-100. The cells are resuspended in this buffer, homogenized using standard methods, and centrifuged at 5000 rpm for 20 minutes. This process is repeated 3 times. 2. Cell Lysis. The washed cells are lysed by sonication in 50 mM Tris buffer (pH 8.0-8.5), 4% deoxycholate and 1% beta mercaptoethanol . Sonication is done using standard methods. After sonication, only remnants of the nuclear membrane are intact and these are removed by centrifugation at 5000 rpm for 30 minutes. The supernatant containing the extracted gpl60 has no intact cells, as determined by light microscopy observations. 3. Gel filtration. Gel filtration is done in a Pharmacia 5.0 x 50 cm glass column packed with a Sephacryl resin (Pharmacia). The total bed volume is about 1750 ml. To depyrogenate and sanitize the column and tubing connections, at least 6 liters of 0.1 N NaOH is run through the column over a period cr 24 hours. The effluent from the column is connected to a UV flow cell and monitor and a chart recorder (Pharmacia) and then is equilibrated with 4 liters of Gel Filtration Buffer. The crude gpl60 is loaded onto the column and is developed with Gel Filtration Buffer.

The column separates the crude mixture into three major UV absorbing fractions. The first peak comes off between about 500 and 700 ml, the second between 700 and 1400 ml and the third between 1400 and 1900 ml buffer. This same profile is observed on small analytical columns from which it has been determined that the first peak is material that has a molecular weight of 2,000,000.

This peak is translucent due to a concentration of high molecular weight lipids and lipid complexes. This peak also contains from 10% to 20% of the gpl60 extracted from the infected cells. Apparently this fraction of gpl60 is complexed to itself or other cell components to form high molecular weight aggregates.

The second broad peak contains the majority of the gpl60 and proteins with molecular weights of between about 18,000 and 200,000.

The third peak contains little protein and the majority of the UV absorption is due to the beta mercaptoethanol in the sample.

When the second peak is first detected from the tracing 14 of the UV absorbance, the effluent from the column is applied directly onto the lentil lectin column. Once the second peak has come off the column, the effluent is disconnected from the lentil lectin column and directed to waste. 4. Lentil Lectin. The lentil lectin affinity gel media (Lentil Lectin-Sepharose 4B) was purchased in bulk from Pharmacia. The lentil lectin was isolated by affinity chromatography on Sephadex to greater than 98% purity and then was immobilized by coupling to Sepharose 4B using cyanogen bromide. The matrix contains about 2 mg ligand per ml of gel. The lentil lectin column is a 5.0 x 30 cm glass column (Pharmacia) containing 125 ml lentil lectin-Sepharose 4B gel. The affinity matrix is reused after being thoroughly washed and regenerated by a procedure recommended by the supplier. When not in use, the gel is stored in the column in a solution of 0.9% NaCl, 1 mM MnCl2, 1 mM CaCl2, and 0.01% thimerosal. The column is washed and equilibrated with 250 ml lentil lectin buffer described above before each use.

The crude gpl60 is applied to the column directly as it is eluting from the gel filtration column as described above. Once the crude gpl60 is bound to the column, it is washed with 800 ml lentil lectin buffer containing 0.1% deoxycholate. Under these conditions all of the gpl60 binds to the column. Lentil lectin buffer plus 0.3M alpha-methyl mannoside is used to elute the bound glycoproteins which is monitored through a UV monitor at a wavelength of 280 nm. 15 5. Dialysis . Sugars and deoxycholates are removed by conventional dialysis.

The purification of gpl60 from 1 liter of infected cells can be summarized in the following table (Table 3) .

In another embodiment, conventional ion exchange chromatography (anionic or cationic) may be used in place of gel filtration. Similarly, the order of steps is not critical: For example, gel filtration or ion exchange chromatography may follow the lentil lectin purification step. Other reagents may also be used according to the invention. For example, other detergents may be used to purify the recombinant protein in place of deoxycholate. These include nonionic detergents such as Tween 20 (polysorbate 20), Tween 80, Lubrol, and Triton X-100.

TABLE 3 - Purification Summary Total protein was estimated from absorbance at 280nm.

EXAMPLE 8 A. Assembly of σρΐβθ Particles.

As one aspect of the present invention, it has been discovered that the gpl60 antigen can be assembled intp particles of > 2,000,000 molecular weight during purification. The gpl60 protein is extracted from the cell as a mixture of 80-90% monomeric (160,000 molecular weight) and 10-20% polymeric (particle form) . The gel filtration step removes the aggregated forms of gpl60. Attempts to purify the gpl60 from this fraction (first peak off the gel filtration column) suggest that it is complexed with other cell proteins, possibly even with membrane fragments. However, the gpl60 antigen in the second peak off the gel filtration column has a molecular weight of about 160,000-300,000 and is, therefore, in predominantly monomeric or dimeric form.

The formation of aggregates or polymers of gpl60 occurs during the development of the lentil lectin column. It has been determined that the antigen forms aggregates whether it is eluted from the lectin column in 0.5% deoxycholate, which is about the 0.2% critical micelle concentration (CMC) for deoxycholate, or whether the gpl60 is eluted from the column in 0.1% deoxycholate.

The size of the aggregates are measured on a high resolution FPLC Superose 12 column (Pharmacia). Samples from representative lots of purified gpl60 have a size that is predominantly equal to or greater than the 2,000,000 molecular weight of a blue dextran size standard.

A cross-linking study by Schwaller, et al. (1989) , demonstrated that gpl60 produced in insect cells is a tetramer of identical submits. The study also shows that gpl60 in HIV-infected cells and virus particles is tetrameric. Thus, the recombinant gpl60 particles may have tertiary and quaternary structures that are similar to those found in the native HIV gpl60.

Proper 3-dimensional structure could be important for the formation of epitopes that require correct folding of gpl60. 17 It is likely that, as non-glycosylated proteins are removed from association with the gpl60 antigen during the binding and washing to the lentil lectin column, the hydrophobic portions of gpl60 begin to form intermolecular associations. The deoxycholate is probably not bound to the gpl60 as the concentration can be kept above the CMC and the antigen will still form complexes. The assembly of this antigen into aggregates appears to be an intrinsic property of this protein once it is purified according to the invention. It is possible that the very hydrophobic N-terminal sequence that is present on the gpl60 protein contributes to the natural ability of this protein to form particles. After purification, the gpi60 complexes can be sterile filtered through a 0.2 micron cellulose acetate filter without significant loss of protein.

B. Analysis of Particle Formation.

An analysis of purified gpl60 particles by electron microscopy demonstrates that they are protein-like, spherical particles of 30-100 nM.

As an additional test for the presence of particles, purified gpl60 was analyzed by gel filtration. About 100 micrograms of gpl60 was applied to a Superose 12, FPLC gel filtration HR 10/30 column (Pharmacia, Inc.). This column was first calibrated with protein molecular weight standards. The protein profile from this column is highly reproducible; the elution volume is inversely proportional to the molecular weight of the protein standards. The column separates the monomeric gpl60 from the polymeric forms and excludes globular proteins of > 2 x 106 molecular weight. When developed on this column, essentially all of the purified gpl60 elutes in the void volume and is, therefore, > 2 x 106 (2,000,000) molecular weight in size. \ 18 EXAMPLE 9 .

A. Adsorption of qpl60 to Alum.

The effectiveness of insoluble aluminum compounds as immunologic adjuvants depends on the completeness of adsorption of the antigens on the solid phase. As part of the present invention it was discovered that alum compositions could be made that would efficiently adsorb the gpl60 but at a pH that would not reduce the potency of the gpl60-alum complex as an im unogen. The factors controlled during the formation of this alum (aluminum phosphate gel) composition are: 1. The optimal pH for adsorption of antigens to alum is about 5.0. However, it was discovered that the gpl60 lost immunogenicity at a pH of 6.5 in comparison to a pH of 7.5 so the alum is made at a pH of 7.1 + 0.1. It was discovered that essentially 100% of the gpl60 will still adsorb to the alum at this pH. 2. The ionic strength from the NaCl present is relatively low and is less than 0.15 M. 3. There is a molar excess of aluminum chloride relative to sodium phosphate to assure that there is an absence of free phosphate ions in the supernatant. 4. The gpl60 antigen is added to freshly formed alum to stop crystal growth and minimize the size of the particles.

The procedure to make 200 ml alum and adsorb purified gpl60 to the alum is such that the final concentration of antigen is 40 μg/ l , as outlined below.

B. Preparation of Reagents (200 ml total formulated lot) .

Prepare the following solutions in 100 ml sterile, 19 pyrogen-free bottles or beakers. Mix the salts for Solution 1 and Solution 2 and the sodium hydroxide and filter through 0.2 micron cellulose acetate filters into 100 ml sterile, pyrogen-free bottles.

Autoclave the solutions for 30 min; slow exhaust.

Cool to room temperature.

C. Formation of Alum 1. Add Solution 1 (aluminum chloride-sodium acetate to the formulation vessel using 25 ml sterile, disposable pipets. Note the volume of Solution 1 and begin stirring the solution. 2. Add Solution 2 (sodium phosphate) to the vessel using 25 ml sterile, disposable pipets and continue stirring as the precipitate forms and note the volume of Solution 2. 3. Add 3 ml Solution 3 (sodium hydroxide) and continue stirring for 5 min. Take a 0.5 ml sample and measure the pH. If the pH is less than 7.0, add an additional 20 0.5 ml sodium hydroxide, stir for another 5 minutes and measure the pH again. Continue until the pH is between 7.0 and 7.2. 4. Determine the total volume added to the formulation vessel (Solution 1 + Solution 2 + Solution 3) , then add sterile FI to bring the volume to 100 ml. 5. Immediately add 8,000 micrograms of purified gpl60 in 100 ml of 1 mM Tris pH 7.5 directly into the formulation vessel. 6. Continue stirring for a minimum of 20 minutes, then dispense the formulated vaccine into sterile vials.

EXAMPLE 10 Immunoaenicity of Alum Absorbed QP160 (Specific Ab Response) An accepted method to determine the immunogenicity of an antigen preparation (vaccine) is to measure the specific antibody response in groups of mice which have been given a single dose of antigen. At the end of 4 weeks the mice are bled and the serum antibody levels to a specified antigen (usually the antigen used to immunize the animal) are measured by a standard antibody test, e.g. an ELISA (enzyme linked immunosorbent assay) .

The immunogenicity in mice of purified gpl60 with no adjuvant at pH 6.0 and pH 7.5 adsorbed with alum (as described in Example 9) or mixed with Freund's Complete Adjuvant are summarized below (Table 4) .

TABLE 4 Group gpl60 Seroconversion Mean ELISA _ qpieo Adjuvant Lot$ bl % fP/N) 2 The mice were bled 28 days post immunization and the sera tested at 1:10 dilution in an ELISA assay against gel-purified gpl60. Similar results were obtained using a commercial ELISA (Genetic Systems Inc.; EIA m ELISA) assay against the native HIV-1 proteins at a serum dilution of 1:400. 21 None, pH 7,5 8702 0.140 57% 4/6 None, pH 6.0 8702 0.110 26% 2/7 Alum 8702 1.000 90% 9/10 Alum 8705 2.285 100% 6/6 Freund ' 8604 1.108 83% 5/6 Freund ' s 8702 1.396 100% 7/7 Freund ' s 8604 0.434 67% 4/6 Alum 8705 1.003 67% 4/6 Mice immunized with a single 1.0 microgram dose of gpl60 antigen without any added adjuvant will elicit an antibody response against gpl60 (see table above) . However, a much stronger antibody response is seen in groups of mice immunized with 1.0 microgram of gpl60 adsorbed to the alum adjuvant. A single dose of less than 0.1 microgram of gpl60 mixed with complete Freund 's or formulated with alum will seroconvert > 50% of the immunized mice. Although less so, the gpl60 antigen was immunogenic in mice as an unformulated antigen at pH 7.5 and at pH 6.0, but there was a loss of immunogenicity at the lower pH.

EXAMPLE 11 Immunogenicity of Alum Absorbed gpl60 (ELISA Serum Study) The ability of a candidate vaccine to elicit an immune response is a very important biological property. To confirm that the alum formulated gpl60 vaccine was immunogenic in animals and to confirm that the alum adjuvant increased this immunogenicity, the following experiment was performed. 3 The number of seroconverted mice (P) to the total number tested (N) .

On day 0, mice (groups of 10) were injected with a single dose (0.5 micrograms, 1.0 micrograms, or 5.0 micrograms) of gpl60 alone, gpl60 adsorbed to alum or gpl60 in complete Freund's adjuvant (CFA) . On day 28 the mice were bled and the sera examined by ELISA (1:10 dilution) for the presence of antibodies to gpl60.

Results from the sera drawn on day 28 are summarized in the table below (Table 5) . In all groups, greater than 50% of the mice showed seroconversion. At all doses the number of seroconversions and the average serum absorbance (OD50 nm at a 1:10 dilution in the ELISA assay) were higher with gpl60 adsorbed to alum than those obtained in mice immunized with gpl60 alone.

These results demonstrate that the alum adjuvant significantly increased the immunogenicity of the gpl60 antigen.

TABLE 5 - 28 Days Post-Iniection 0.5 μg Dose 1.0 g Dose 5.0 g Dose Mean Mean Mean 5 P/N 4 OD OD P/N OD gpl60 9/10 .407 7/10 .699 7/10 .430 gpl60 (alum) 9/10 .547 8/10 .797 10/10 1.347 gpl60 (CFA) 10/10 1.130 10/10 1.967 10/10 1.317 The number of mice that seroconverted (P) compared to total number tested (N) at 28 days after being immunized with 0.5 micrograms, 1 micrograms or 5 micrograms of VaxSyn m HIV-1.

The mean absorbance (OD450) of the mice that seroconverted as measured by the sponsor's ELISA assay against gpl60 at a 1:10 dilution of serum. 23 EXAMPLE 12 Neutralization Data HIV-1 neutralization assays are an accepted method to determine whether an antibody preparation will inhibit the HIV-1 virus from infecting susceptible human cultured lymphocyte cells. Antisera from animals immunized with gpl60 were tested in an HIV- 1 neutralization assay and the results are summarized in the table below (Table 6) .

TABLE 6 Guinea pigs, rabbits and rhesus monkeys have also been immunized with gpl60 (using alum or Freund's as an adjuvant) . In general, the immunization of these animals has produced a good antibody response against the HIV-1 envelope proteins.

Micrograms of gpl60 or gpl20 administered during the first/second/third immunization.

The highest dilution of antisera that will inhibit the infection by 50% relative to HIV-1 infected cells that were exposed to serum from non-immunized animals.

EXAMPLE 13 IiTununoqeriicitv in Chimpanzees Genetically, the chimpanzee is man's closest relative and is currently the only animal model for infection of HIV-1. In a safety/immunogenicity trial in three chimpanzees, two chimpanzees were immunized with 40 micrograms or 80 micrograms of gpl60 in an alum formulated vaccine. Each received a booster immunization at 4 weeks with 40 micrograms and 80 micrograms of gpl60, respectively. A control animal was vaccinated at the same time with a 1 ml saline solution. Weekly serum samples were analyzed from each of the three chimpanzees for antibodies to gpl60 and to HIV-1 viral antigens using three immunological assays, an ELISA assay against purified gpl60 developed by MicroGeneSys, Inc., Western Blot analysis, and a commercial HIV-1 ELISA assay. The results of these analyses are described below.

A. ELISA (MGSearch HIV 160) The ELISA assay, MGSearch HIV 160, MGSearch being a trademark of MicroGeneSys, Inc. of Meriden, Connecticut, U.S.A., is an immunosorbent assay against gpl60 and is described in copending coassigned U.S. patent Application Serial No. 920,197 (now No. 585,266) .

Serum samples taken before immunization and for the 11 weeks following the primary immunization were diluted from 1:10 to 1:100,000 and then incubated with nitrocellulose strips containing a 100 μg purified gpl60 in a spot. The end point dilution titer is the highest dilution in which the test was positive for anti-gpl60 antibody as detected with a goat anti-human IgG-alkaline phosphatase conjugate.

The serum samples from the control animal and from the pre-immune sera of the immunized animal were negative. The chimp which received the 80 microgram dose was positive at a 1:100 dilution by week 2 and the chimp which received a 40 microgram dose was positive at a 1:10 dilution by week 4. The antibody titers to gpl60 continued to increase until week 5, at which time 25 the end point dilution titers were approximately 1:100,000 and 1:2,000,000 respectively. The antibody titer in both animals dropped just slightly during weeks 6-11.

This type of response is similar both quantitatively and qualitatively to antibody responses commonly observed in chimps that have been vaccinated with a human Hepatitis B Virus vaccine.

B. Commercial ELISA Test It was clear from the MGSearch HIV 160 ELISA and Western blot analyses of sera from the VaxSyn immunized chimpanzees, that they had seroconverted and have antibodies against the recombinant gpl60. To determine if they were also making anti- HIV antibody which recognized the native viral envelope proteins, the pre-immune sera and sera from weeks 1 through 11 were tested in a licensed, commercial ELISA test kit, the LAV EIA™ test kit of Genetic System Corporation, Seattle, Washington. The animal immunized with 80 micrograms of gpl60 was positive at a 1:100 dilution by week 2 and continued to show an increase in antibody level through week 6. The animal immunized with 40 micrograms was positive at a 1:100 dilution by week 6.

EXAMPLE 14 Distribution of Antibodies Between gpl20 and qp41 It is important to determine whether the antibody responses against gpl60 in a vaccinated animal is directed against gp41, gpl20 or both. A variety of immunological methods, including radioimmunoprecipitation (RIP) , immunofluorescence (IF) , Western blot analysis (WB) , and quantitative ELISA against three different recombinant envelope antigens were employed to detect and measure for the distribution of antibodies against various regions of the HIV-1 envelope proteins.

Fig. 6 summarizes the immunoreactivity of three VaxSyn is a trademark of MicroGeneSys, Inc. for the AIDS vaccine described herein. different recombinant antigens: [ART] [TAB] (1) gpl20-delta (truncated recombinant HIV-1 gpl20 with about 40 amino acids missing from the C-terminus of the molecule) ; [ART] [TAB] (2) gpl20 (full length recombinant HIV-1 gpl20 and [ART] [TAB] (3) gpl60.

Human sera from 50 HIV-1 antibody positive individuals and 3 pooled human sera were highly reactive with gpl60, moderately reactive with gpl20 and little or no antibody reacted with truncated gpl20. It is likely that the truncated gpl20, which represents more than 90% of the HIV-1 external glycoprotein, contains protective determinants. The observation that human AIDS positive sera have few antibodies to this region of the envelope is consistent with the fact that the immune response to viral infection is not fully protective and that human positive sera usually exhibit a low-level of neutralizing activity in vitro.

In contrast, rhesus monkeys immunized with either the gpl60 immunogen or with the truncated gpl20 have antibodies that react strongly with the truncated gpl20 portion of the HIV-1 envelope. This difference in distribution of antibody recognition sites along the viral envelope and the higher titers observed in the monkeys may account for the fact that the monkey sera had high neutralizing titers.

A quantitative assessment of the immunoreactivity of these three recombinant envelope antigens with human and immune rhesus sera is presented in Fig. 7. All the monkey sera tested had high titer antibody against the truncated gpl20 antigen (gpl20-delta) , including those from animals immunized with gpl60.

These results demonstrate that the recombinant gpl60 elicits an antibody response in rhesus monkeys that is different than what often occurs during natural infection. There are epitopes in the gpl20-delta region of gp-160 that are efficiently recognized in the immunized monkeys that are not seen by the human immune system during infection. These new epitopes may be important for protection against HIV-l, and could be an important property of the recombinant gpl60 for prevention and treatment of HIV-infection .

EXAMPLE 15 Therapeutic Vaccina Administration A clinical trial with 30 HIV-seropositive human patients was conducted to determine the effects of vaccination with cloned HIV gpl60 (produced in the baculovirus system as described above) on HIV infected individuals.

Vaccination with the recombinant gpl60 led to an augmentation in the gpl60 HIV-specific humoral and cellular immune responses of 19 out of 30 (63%) HIV seropositive volunteers. Fourteen out of 15 (93%) volunteers receiving 6 doses of the vaccine demonstrated an increase in their total gpl60 antibody. Therefore, recombinant HIV proteins (i.e., rgp41, rgpl20, rgpl60 and admixtures thereof) can be advantageously administered in a method to treat a human patient infected by HIV.

The effective amounts of HIV protein used in this embodiment of the invention can be determined according to techniques well known in the art, such as those presented below. In general such effective amounts may range between about 1 microgram and about 100 micrograms per kilogram body weight of the patient. The frequency of administration can also be determined by known means. In a preferred embodiment, administration is via the parenteral route, i.e., intravenously, intraperitoneally , intramuscularly, intradermally, etc., as is well known by those of ordinary skill in the art.

A. Volunteer Selection Thirty volunteers with HIV infection were recruited. Only seropositive volunteers with early stage HIV infection, defined as Walter Reed Stage 1 or 2 (CD4 cell count not less than 400 for greater than 3 months, with or without lymphadenopathy) were eligible for enrollment. (Redfield, et al., New Engl. J\ 28 Med. 31 : 131-132 (1986) . Additional entry criteria limited volunteers to adults between the ages of 18 and 50, with a normal complete blood count, no evidence of end organ disease, no alcohol or drug abuse over the preceding 12 months, and who were not receiving anti-retroviral or immunomodulatory drugs. All patients underwent a 2 month baseline evaluation prior to randomization into treatment groups. No volunteers received any antiretroviral or immunomodulatory drugs during the trial.

Twenty-six of the 30 volunteers were men; 4 were women. Fourteen were Caucasian, 13 Black, and 3 Hispanic. The mean age was 29 (range 18-49) . At enrollment 8 volunteers were Walter Reed Stage 1 and 22 volunteers were Walter Reed Stage 2. The baseline mean CD4 count was 668 (range 388-1639) . The mean time between initial diagnosis and study entry was 24 months (range 3 months to 49 months) .

B. Vaccine Product and Immunization Schedule As described herein, the test vaccine comprises a noninfectious subunit glycoprotein derived from gpl60 as a baculovirus expressed recombinant protein. The immunogenic protein was produced in Lepidopteran insect cells, was biochemically purified, and was adsorbed to aluminum phosphate for final vaccine formulation.

Three dose formulations of gpl60 were used: 40 micrograms per milliliter, 160 micrograms per milliliter and 320 micrograms per milliliter. The injection volume for both the 40 μg and 160 μg dosages was 1 ml; 2 ml of 320 μg per milliliter was used to deliver the 640 g dose injections.

The thirty volunteers were distributed into six groups of five volunteers each. Two immunization schedules were investigated: Schedule A, with vaccination on days 0, 30, and 120; and Schedule B, with vaccination on days 0, 30, 60, 120, 150 and 180. Within each immunization Schedule (A or B) there were three groups which received different dosages of vaccine (Table 7 below) . All vaccinations were administered by intramuscular injection into the deltoid muscle. The duration of the trial was 10 months: a 2 month baseline evaluation, and an 8 month follow-up evaluation after the initial vaccination.

TABLE 7 - Immunization Schedule Amount of gpl60 Administered (pg) Dav 0 30 60 120 150 180 Schedule A Group 1 40 40 40 Group 3 160 160 160 Group 5 640 640 640 Schedule B Group 2 40 40 40 160 160 160 Group 4 160 160 160 640 640 640 Group 6 640 640 640 640 640 640 C. Assessment of Safety and Toxicity Each volunteer was interviewed and examined on days 0, 1, 2, 3, 15 and 30 after each injection. Volunteers were queried concerning fever, chills, nausea, vomiting, arthralgia (painful joints) , myalgia (muscular pain) , malaise, urticaria (hives) , wheezing, dizziness, or headache. Examinations to assess local reactions at the site of injection included erythema, swelling, itching, pain and tenderness, skin discoloration, skin breakdown, change in regional lymphadenopathy , change in function of the injected extremity, and subcutaneous nodule formation at the site of injection. Monthly complete blood counts, serum chemistries, coagulation profile and urine analysis were also assessed.

In vitro cellular immune function was assessed by T-cell phenotyping (total lymphocyte, CD4 and CD8 cell phenotypes) as described in Rickman, et al., Clinical Immuno . 52: 85-95, 1989; Birx, et al., J^. Acquir. Immune Defic. Syndr. 4.: 188-196, 1991) . T-cell proliferative response to mitogens (pokeweed and Con A) and control antigens (Candida albicans and tetanus) was also evaluated. Birx et al, supra . In vivo cellular immune function was assessed by delayed hypersensitivity skin testing to control antigens (i.e. , mumps, tetanus toxoid, Candida albicans \ 30 and trichophyton) .

Quantitative viral cultures of peripheral blood mononuclear cells (PBMC) and plasma were assessed as described in Burke, et al., J^. Acquir. Immune Defic. Svndr. 3: 1159-1157, 1991. DNA polymerase chain reaction (Wages, et all, J^. Med. Virol. 33.: 58-63, 1991) and serum p24 antigen levels were assessed to monitor in vivo HIV viral load.

No evidence of systemic toxicity was observed, but local reactogenicity was noted in 87 percent of the subjects (13 in each vaccination group) . Local reactions included induration, tenderness, and transient subcutaneous nodule formation at the injection site; an increase in regional adenopathy was rarely noted. No subject refused a booster injection. No difference in the frequency of local reactions was observed for primary immunization, booster injection, or dosage.

No evidence of adverse effects on the immune system was demonstrated as measured in vitro by mitogen and antigen specific proliferative responses, in vivo by delayed hypersensitivity skin testing responses, or by acceleration of quantitative CD4 cell depletion. Baseline mean CD4 cell counts were 716 and 605 for vaccine responders and non-responders, respectively. Mean CD4 cell counts from study days 180-240 were 714 and 561, for vaccine responders and non-responders, respectively. During the course of the 240-day trial, the net change in mean CD4 cell counts for vaccine responders was a minus 0.2 percent, while among vaccine non-responders the mean CD4 cell count declined by 7.3 percent (Figure 11) . Vaccine induced HIV immunogenicity was not associated with evidence of accelerated CD4 decline in any individual subject throughout the entire course of the trial.

To assess the possibility of increased HIV replication and viral load in subjects as a consequence of vaccination, in vivo viral activity was measured by quantitative plasma and PBMC viral cultures, PBMC DNA polymerase chain reaction, and serum levels of p24 antigen. Quantitative cultures and DNA polymerase chain reaction assays demonstrated no alteration during this 31 trial. Serum p24 antigen was undetectable in the subjects.

D. Assessment of Immunooenicitv Antibodies directed against whole HIV proteins were measured using both recombinant produced viral gene products gpl60, p66, p24 and whole viral lysate of prototype HIV strain MN. Dot blot and Western Blot techniques were used, as described in Toubin, et al., Proc. Natl . Acad. Sci. USA 76; 4350-4354 (1979) . Antibody responses to specific envelope epitopes were also measured (see Fig. 7) .

In Fig. 7 epitopes 88 (amino acids 88-98 in gpl20) and 448C (amino acids 448-514 in gpl20) were selected because antibody directed against these regions of gpl20 are reported to correlate with early stage HIV infection.

Epitopes 106 (amino acids 106-121 in gpl20) , 241 (amino acids 241-272), 254 (amino acids 254-272), 300 (amino acids 300-340) , 308 (amino acids 308-322) , 422 (amino acids 422-454) and 735 (amino acids 735-752) were selected because of their putative functional importance. Epitopes 106 and 422 have been implicated in CD4 binding; epitopes 241, 254 and 735 have been implicated in group specific neutralization; and epitopes 300 and 308 have been implicated in type-specific neutralization) .

Epitope 582 (amino acids 582-602) was selected as a control because it represents the immunodominant envelope domain in natural HIV infection. Additional epitopes investigated included 49 (amino acids 49-128) ; and 342 (amino acids 342-405) .

In Fig. 7, a shaded box signifies a documented change in the HIV envelope-directed immune response. Shaded boxes with (=) signify a primary humoral response; shaded boxes with (+) signify a secondary humoral response; (-) signifies antibody negative to specific epitope pre and post immunization; and a (+) signifies antibody positive to specific epitope pre and post immunization, but without a quantitative change. Shaded boxes with (.) signify new T-cell proliferative response to gpl60 following immunization. A (.) alone signifies no cellular 32 response to gpl60; while hb signifies "high background" (not interpretable) ; and nd signifies "not done." Neutralization activity was measured against three prototype isolates (HIV-IIIB, RF and MN) in a syncytium inhibition assay as described in Nara, Nature, 333 : 469-470 (1988) . HIV specific cellular responses were measured by known lymphocyte proliferation assay techniques using gpl60, p24 and baculoviral expression system control protein (Birx, supra) .

E. Vaccine Responders and Non-Responders Subjects were classified as vaccine responders only if a reproducible selective increase of both a cellular and humoral immune response against HIV envelope specific epitopes were associated with the vaccination series (Fig. 7) . Vaccine induced humoral immunity was defined as seroconversion to HIV envelope specific epitopes and/or a secondary booster immune response to envelope specific epitopes. Vaccine induced cellular immunity was defined as the development of a new, reproducible, vaccine associated, proliferative response to gpl60.9 Subjects who developed neither a humoral nor a cellular proliferative response or who developed only a humoral or only a cellular proliferative response to gpl60 epitopes or HIV envelope were classified as non-responders .

F. Vaccine Induced Humoral Responses Referring to Fig. 7, 19 of the 30 subjects (63 percent) demonstrated a vaccine induced augmentation of both gpl60 HIV specific humoral and a cellular immune responses. These 19 were classified as "vaccine responders". Four of the 11 "non- responders" developed only a humoral or a cellular immune response. All 7 subjects who failed to demonstrate any detectable vaccine induced response received only 3 doses 9 . . . . . .

This definition of a vaccine responder is highly restrictive in light of the scientific objectives of this trial: e.g. , to assess the feasibility of post-infection immunization.

(Schedule A) . No changes in antibody binding to HIV polymerase (p66) , or structural (p24) gene products or the non-HIV control antigen tetanus were detected. No anti-baculoviral Lepidopteran cell control protein antibody developed in any subject.

Increases in envelope antibody (gpl60) were detected in 13 subjects by Western Blot using the whole virus lysate HIV-MN. The changes were related to the immunization schedule. Three of 15 subjects (20 percent) on Schedule A, and 10 of 15 subjects (67 percent) . Schedule B developed an antibody increase to envelope proteins (P=0.025 by Fisher's exact test, two-tailed). All 13 subjects also seroconverted to specific envelope epitopes.

Conversely, of the 10 subjects who failed to seroconvert to any envelope specific epitope, none exhibited an increase in envelope antibody by Western Blot. The remaining 7 subjects who seroconverted to specific envelope epitopes demonstrated no change in whole virus envelope antibody by Western Blot. No changes in antibody directed against non envelope HIV proteins were observed in any subject.

Fourteen of 15 subjects (93 percent) on Schedule B (6 doses) demonstrated an increase in total gpl60 antibody, as opposed to only 7 of 15 subjects (47 percent) on Schedule A (3 doses) (P=0.01 Fisher's, two-tailed). (Fig. 7).

As shown in Fig. 8, the pre-immunization to post-vaccination prevalence of each gpl60 specific epitope respectively was as follows: Epitope 49 (27 to 70 percent) , Epitope 88 (28 to 52 percent) , Epitope 106 (50 to 87 percent) , Epitope 214 (0 to 14 percent) , Epitope 254 (0 to 13 percent) , Epitope 300 (47 to 77 percent) , Epitope 308 (42 to 69 percent) , Epitope 342 (0 to 27 percent) , Epitope 422 (3 to 10 percent), Epitope 448C (73 to 87 percent) , and Epitope 735 (17 to 33 percent) . Vaccine induced seroconversion was noted against all of the specific epitopes except 582 (Fig. 7) . Antibodies (seroconversion) directed against Epitopes 241, 254 or 342 were only detected following vaccination.

Secondary immune responses were detected to the following epitopes: 88, 106, 300, 448C, and 582. The prevalence of antibody directed against epitope 582 was 100 percent pre- vaccination and only one subject (3 percent) demonstrated a secondary immune response .

The pattern of vaccine induced HIV antibody to envelope epitopes was variable (Fig. 7) . Primary antibody responses (seroconversion) to at least one epitope occurred in 20 subjects; 14 of 15 receiving Schedule B, and 6 of 15 randomized to Schedule A (P=0.005 Fisher's, two-tailed). Schedule A subjects seroconverted to only 15 of 110 (14 percent) of the potential epitopes to which they had no preimmunization antibodies. Schedule B subjects seroconverted to 60 of 129 (47 percent) (P<0.0001 Fisher's, two-tailed) . Seroconversion to three or more envelope epitopes occurred in 9 subjects (60 percent) randomized to Schedule B but only 2 subjects (13 percent) randomized to Schedule A (P=0.02 Fisher's, two-tailed).

Serum neutralization activity against three distinct strains (HIV-IIIB, MN, and RF) was determined on days 0, 90 and 195 in 7 subjects. Four of 5 vaccine responders demonstrated increasing neutralizing activity to one or more isolate. The - vaccine responders also demonstrated an increased ability to inhibit syncytium formation compared to non-responders .

G. Vaccine Induced Cellular Responses Changes in cellular immune response were based on a comparison of mean pre-vaccination (baseline) and post- vaccination lymphocyte stimulation indices (LSI) using a Wilcoxon rank sum test.

Twenty-one of 30 subjects (70 percent) developed a new T cell proliferative response to gpl60 post-immunization (Fig. 7).

Figure 9 illustrates proliferative responses to gpl60, p24 and a baculovirus control protein in four typical vaccine responders over time. For all · subjects the gpl60 induced proliferation increased from a baseline mean LSI of 3 to an LSI of 10 (calculated utilizing the mean of 4 values following the last immunization) . In contrast, no change was noted for proliferative responses directed against HIV p24 protein or the control baculovirus protein.

Vaccine induced changes in mean LSI values for all subjects, for subjects subgrouped by vaccine responsiveness, and for subjects grouped by immunization schedule are illustrated in Figure 10.

The change in proliferative response to gpl60 was significantly different between vaccine responders and non-responders (<0.001, ilcoxon, one tailed) . The gpl60 proliferation responses induced by Schedule B (6 doses) were greater than those induced by Schedule A (3 doses) (P<0.10, ilcoxon, one tailed) .

Nineteen of the 21 subjects who developed proliferative responses to gpl60 also developed a humoral response (vaccine responders) . The maximum mean lymphocyte stimulation index (LSI) to gpl60 observed for all vaccine responders was 50.1. However, each vaccine responder's response was variable (peak values ranging from a LSI of 3 to 171) (Fig. 7) , as was the temporal relationship to vaccination of the magnitude and duration of the cellular responses to gpl60 (Figure 9) .

H. Discussion of Results Despite the limited sample size of this trial, several factors were demonstrated to be associated with vaccine immunogenicity. Six of 15 (40 percent) of the subjects on Schedule A versus 13 of 15 (87 percent) of the subjects on Schedule B were vaccine responders (P=0.02 Fisher's, two-tailed) (Fig. 7) . Of the 16 subjects with a mean baseline CD4 count greater than 600 per milliliter, 13 (81 percent) were vaccine responders, as opposed to 6 of 14 (43 percent) subjects whose mean entry CD4 count was less than 600 cells per milliliter (P=0.07 Fisher's, two-tailed). As summarized in Table 8, multiple immunizations improved immunogenicity, even among \ patients with baseline CD4 counts less than 600 cells per milliliter. For example, 5 of 6 subjects on Schedule B (6 injections) were vaccine responders as compared to only 1 of 8 who received the 3 injection regimen (Schedule A) P=0.0J Fisher's, two-tailed) (Table 8).

TABLE 8 GP 160 Vaccine Immune Responsiveness by Baseline CD4 Count and Immunization Schedule CD4 Count N # Responders (%) # Non-Responders f¾^ SCHEDULE A >600 7 5 (71%) 2 (29%) 500-600 5 1 (20%) 4 (80%) <500 3 0 (0%) 3 (100%) Subtotal 15 6 (40%) 9 (60%) SCHEDULE B >600 9 8 (89%) 1 (11%) 500-600 2 2 (100%) 0 (0%) <500 4 3 (75%) 1 (25%) Subtotal 15 13 (87%) 2 (13%) TOTAL 30 19 (63%) 11 (37%) The therapeutic use of vaccines was introduced by Pasteur in the 19th century for the treatment of acute rabies infection. But the utility of this approach in the treatment of other infections has not been extensively explored. Although there are other examples of post infection modification of viral-specific immunity (such as after hepatitis A or B exposure) , there are no well documented studies in man which demonstrate the feasibility of this approach for an established or chronic viral infection.

Here, the invention provides virus-specific immune modification by active immunization after infection. Specifically, an HIV envelope gene derived gpl60 vaccine 37 augmented the human host directed viral-specific humoral and cellular responses in 19 of 30 early HIV infected persons.

This study qualitatively and quantitatively measured distinct antibody responses to specific HIV epitopes in natural infection versus post infection immunization. In this way, an accurate determination of vaccine induced humoral immunogenicity in already infected persons was documented in 70 percent of the subjects. For example, twenty subjects (19 vaccine responders and 1 vaccine non-responder) seroconverted to specific envelope epitopes. Seroconversion associated only with vaccination (epitopes 241, 254, and 342) occurred in 10 subjects.

Additionally, variations in humoral responses to this vaccine, as characterized by epitope mapping, will permit prospective cause and effect analysis of specific antibody responses, and allow unique opportunities to characterize potential immunoregulatory mechanisms not elicited during a natural infection.

Although the in vivo relevance of serum neutralizing activity is presently unknown, the observation of increased neutralizing activity against disparate HIV strains (IIIB, RF , M ) in 4 of 5 vaccine responders suggests that post-infection immunization induced changes in functional antibodies. The test vaccine induced increases in serum neutralization capacity against distinct HIV strains and will potentially aid in the definition of group specific neutralization epitopes.

A proliferative response to HIV envelope proteins rarely occurs in natural HIV infection. However, after immunization with gpl60, specific T-cell proliferative responses were documented in 21 (70 percent) of the subjects. The reason for this difference is unclear. One possibility is that the new proliferative response may be directed against an envelope epitope (s) unique to the vaccine (as a result of vaccine production methodology or alternative in vivo antigen processing) . Alternatively, the protein used in the proliferation assay may not stimulate primary T-cell 38 proliferative responses against homologous "wild type" envelopes of natural virus. However, additional evidence that vaccination boosts the host cellular immune response has been obtained: selected vaccine responders demonstrated HIV-IIIB type-specific cytotoxic T-cell responses following booster immunization.

The factors responsible for vaccine immunoresponsive-ness in HIV infected persons remain to be clarified. Even in early HIV infection, individuals respond suboptimally to a variety of vaccines as compared to matched controls. This hyporesponsiveness has been related to early B cell dysregulation and T-cell dysfunction. Here, vaccine immunoresponsiveness was associated with baseline CD4 cell count, which is consistent with the hypothesis that the immunological status of the host is an important determinant of vaccine responsiveness. However, the immunization schedule within specific T-cell count intervals also influenced vaccine responsiveness: Schedule B (6 injections) was superior. Indeed, the decreased vaccine response observed in the subjects with lower CD4 cell counts could be improved by an increased number of vaccinations which suggests that further modifications in dosage, regimen, adjuvants or formulation, could be anticipated to further improve host immunoresponsiveness.

Although concerns have been raised about the safety of active immunization of HIV infected persons with HIV specific vaccine products, there was no evidence of immune-specific toxicity. Quantitative cultures, DNA polymerase chain reaction assays and serum antigen assays show an increased in vivo HIV load An excellent in vivo marker of HIV replication, the rate of CD4 cell decline, was favorably influenced among the subjects, especially those classified as vaccine responders. The change in mean CD4 counts for responders was -0.2 percent and was -7.3 percent for non-responders . The data demonstrates that postinfection immune responsiveness was not associated with an increase in CD4 destruction and suggests an association with decreased HIV replication in vivo.

The vaccination results in this study were also 39 compared with a database of ten infected and untreated individuals matched for age, ethnic group, and baseline CD4 cell count. The mean CD4 count decreased by 8.7 percent in this reference group, decreased by 7.2 percent in the subjects assigned to Schedule A, and increased by 0.6 percent in subjects assigned to Schedule B. These results indicate that postinfection vaccination with recombinant HIV envelope protein is feasible, and furthermore the result are encouraging with respect to the prophylactic uses of such vaccines. 1G2092/2

Claims (23)

1. Use of a recombinant HIV envelope protein in the manufacture of a medicament for treating an individual infected with human immunodeficiency virus (HIV) , substantially as described in the specification.

2. Use according to claim 1, where the medicament is in unit dosage form containing about 1 to 1G0 micrograms per kilogram of body weight recombinant protein per dosage unit.

3. Use according to- claim 1, where the medicament is in unit dosage. form containing about 10 to about 4000 pg recombinant protein per dosage unit.

4. Use according to claim 1, where the medicament is in unit dosage form containing about 40 to about 1200 pg recombinant protein per dosage unit.

5. Use according to claim 1, where the medicament is in unit dosage form containing about 13· 3 to about 400 pg recombinant protein per dosage unit.

6. Use according to claim 1, where the medicament is in nit dosage form containing about 6.65 to about 200 pg recombinant protein per dosage unit.

7. Use according to claim 1 , wherein the medicament is in unit dosage form and each dosage unit contains an amount of recombinant protein sufficient to elicit an increase in HIV-specific cellular or humoral immune responses. 40 102092/2

8. Use according to any of the preceding claims, wherein the recombinant protein is produced by a baculovirus insect cell expression system.

9. · Use according to claim 8, wherein said expression system is the baculovirus insect cell vector Ac3046.

10. Use according to any of the preceding claims, wherein the recombinant protein has a molecular weight of approximately 145,000.

11. Use according to any of the preceding claims, wherein the recombinant protein is agglomerated into particles having a molecular weight Of at least about 2,000,000.

12. Use according to any of the preceding claims, wherein the recombinant protein is at least one of gpl60, gpl20 and gp4l.

13. · Use according to any of the preceding claims, wherein the recombinant protein is combined with an adjuvant.

14. Use according to claim 13, wherein the adjuvant is an alum adjuvant.

15. · Use according to any of the preceding claims, wherein the recombinant protein comprises about 757 successive amino acids of gpl60 and substantially excludes about 40 terminal amino acids of gpl60. 41 102092/2

16. A therapeutic composition for use in treating an individual infected with human immunodeficiency virus (HIV) , which comprises a recombinant HIV envelope protein and an alum adjuvant, wherein the recombinant protein is agglomerated into particles having a molecular weight of at least about 2,000,000.

17. · A composition according to claim 16, which is in unit dosage form.

18. A composition according to claim 17, wherein each dosage unit contains about 10pg to about 4000ug of recombinant HIV envelope protein.

19. A composition according to any of claims 16 to 18, wherein the recombinant protein is produced by a baculovirus insect cell expression system.

20. A composition according to claim 19, wherein said expression system is the baculovirus insect cell vector Ac3046.

21. A composition according to any of claims 16 to 20, wherein the recombinant protein has a molecular weight of approximately 1^5 000.

22. A composition according to any of claims 16 to 21, wherein the recombinant protein is at least one of gpl60, gpl20 and gp4l. 42 i - 102092/2

23. A composition according to any of claims 16 to 22, wherein the recombinant protein comprises about 757 successive amino acids of gpl60, the gpl60 peptide being truncated to about position 757. an substantially excluding at least about 40 terminal amino acids of gpl60. For the Applicants, Sanford T. Colb & Co. 43