Method Of Inducing The Production Of Protective Anti-hiv-1 Antibodies



This application claims priority from U.S. Provisional Application No. 61/282,526, filed February 25, 2010, U.S. Provisional Application

No. 61 /344,457, filed July 27, 2010, U.S. Provisional Application

No. 61/344,580, filed August 25, 2010 and U.S. Provisional Application

No. 61/344,622, filed September 1 , 2010, the entire contents of which are incorporated herein by reference.

This invention was made with government support under Grant Nos. AI067854, Al 24335 and Al 81579 awarded by the National Institutes of Health. The government has certain rights in the invention.


The present invention relates, in general, to an immunogen for HIV-1 vaccination and, in particular, to a method of inducing the production of protective anti-HIV- 1 antibodies by targeting B cell germline and clone intermediates using a combination of non-HIV-1 and HIV-1 immunogens. The invention also relates to compositions suitable for use in such a method.


The first antibody response to transmitted/founder HIV-1 envelope is non- neutralizing, targets Env gp41 and occurs at a mean of 13 days after appearance of plasma viremia (Tomaras et al, J. Virology 82: 12449-63 (2008)). While the initial T cell response to HIV- 1 that occurs at the same time as the initial antibody response drives mutations within T cell epitopes of HIV-1 , the initial gp41 antibody response to HIV- 1 does not. Rather, it is the autologous neutralizing antibody response, which is delayed until approximately three months after transmission, that is the first neutralizing antibody response associated with antibody escape mutants (McMichael et al, Nature Rev. Immunol. 10: 1 1 -23 (2010)).

The four epitopes on HIV-1 envelope to which rare broadly reactive neutralizing antibodies bind are the CD4 binding site (CD4BS) (mab (monoclonal antibody) IgGl bl 2) (Zwick et al, J. Virol. 77( 10):5863-5876 (2003)), the membrane proximal external region (MPER) epitopes defined by human mabs 2F5 and 4E10 (Armbruster et al, J. Antimicrob. Chemother. 54:915-920 (2004), Stiegler and Katinger, J. Antimicrob. Chemother. 51 :757-759 (2003), Zwick et al, Journal of Virology 79: 1252-1261 (2005), Purtscher et al, AIDS 10:587 (1996)), and the mannan glycan epitope defined by human mab 2G12 (Scanlan et al, Adv. Exper. Med. Biol. 535:205-218 (2003)). These four rare human mabs are all unusual: two are IgG3 (2F5 and 4E10), one has a unique Ig dimer structure (2G 12), one has a very hydrophobic CDR3 (2F5) (Ofek et al, J. Virol. 198: 10724 (2004)), and, in all four, the CDR3 is unusually long (Burton et al, Nature Immunol. 5(3):233-236 (2004), Kunert et al, AIDS Res. Hum. Retroviruses 20(7):755-762 (2004), Zwick et al, J. Virol. 78(6):3155-3161 (2004), Cardoso et al, Immunity 22: 163-172 (2005)). Of these, 2F5- and 4E10-like human mabs are quite rare. Acute HIV-1 patients do not make antibodies against the MPER or 2G 12 epitopes, MPER can be defined as amino acids 652 to 683 of HIV envelope (Cardoso et al, Immunity 22: 163-173 (2005) (e.g.,

QQEKNEQELLELDKWASLWNWFDITNWLWYIK). CD4 binding site (BS) antibodies are commonly made early in HIV-1 infection, but these antibodies generally do not have the broad spectrum of neutralization shown by mab IgGlbl 2 (Burton et al, Nat. Immunol. 5(3):233-236 (2004)).

To understand the pathogenesis of the ineffective initial antibody response to HIV-1 envelope (Env), PCR has been performed for amplification of immunoglobulin variable region of heavy- and light-chain (VH and VL) genes from single blood or bone marrow plasma cells from 5 acutely infected subjects from 17-30 days after HIV-1 transmission. The specificities of the plasma cell response induced by HIV-1 infection have been determined. Using PCR amplification of VH and VL genes of single human plasma cells induced by transmitted HIV-1 , the initial plasma cell/plasmablast response to HIV-1 has been studied. It has been found that the first antibody response to HIV-1 is induced to HIV-1 Env gp41 , and that gp41 induces an antibody response in pre-existing memory B cell clones, resulting in low-affinity, polyreactive anti-Env antibodies that cross-react with a number of host and bacterial molecules, particularly, of human gut bacterial flora.

The present invention results, at least in part, from studies designed to identify the source of both the initial anti-HIV-1 Env gp41 antibodies and the rare broadly neutralizing antibodies. The invention further results from the identification of a cellular protein expressed in most warm blooded vertebrates that is structurally similar to the 2F5, and possibly 4E10, epitopes of the HIV-1 gp41 MPER.

The invention provides an HIV- 1 vaccine designed to target a naive B cell pool that can be driven to give rise to broadly neutralizing antibodies to HIV-1 .


In general, the present invention relates to an immunogen for HIV vaccination. More specifically, the invention relates to a method of inducing the production of protective anti-HIV-1 antibodies by targeting B cell germline and clone intermediates using a combination of non-HIV-1 and HIV-1 immunogens. The invention also relates to compositions suitable for use in such a method.

Objects and advantages of the present invention will be clear from the description that follows. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1. A representative influenza antibody clone against HI Soloman Islands hemagglutinin.

Figure 2. Plasma cell antibody repertoire in patient 684-6, -20 days after HIV-1 transmission.

Figure 3. Production of inferred intermediate clone antibodies.

Figure 4. Inferred germline and clone member intermediates assayed for reactivity with clade B gp41 , autologous gpl40 and group M consensus gpl40 to determine where in the clone development reactivity with gp41 was acquired.

Figure 5. Reactivity of clone 684-6B acquired at the second intermediate precursor antibody (see also Figure 4).

Figure 6. Additional inferred intermediate antibody clones produced in mg quantities and analyzed for the dissociation constants (Kd) of antibody binding to gp41.

Figure 7. Acquisition of gp41 reactivity in patient 684-6 clone 684-6B germline and inferred intermediate antibodies.

Figure 8. Polyreactivity of 6846 clone 52 germline and inferred intermediate gp41 antibodies.

Figure 9. Reactivity of aerobic gut flora with clone 684-6B antibodies. Figures 10A and 10B. Blue Native-PAGE and western blot images of gut extract vs Mojo antibody. Figure 10A. Coomassie blue image. Figure 10B. Western blot image.

Figure 1 1 . Western blot image of gut extract vs Mojo antibody - non-reducing vs HV00276.

Figure 12. Western blot image of gut extract vs Mojo antibody - reducing vs HV00276.

Figures 13A and 13B. Ibl 2 germline antibody binds to lipids (PC:CL

liposomes). Figure 13A. Binding to HIV 89.6 gp 120. Figure 13B. Binding to lipids (PC:cardiolipin).

Figure 14. A large fraction of B cells expressing 4E10 VH are deleted in bone marrow at the pre-B to immature B cell stage in 4E10 VH knock-in mice.

Figure 15. Two roadblocks for induction of broad neutralizing antibodies. The first roadblock is that vaccines currently designed to stimulate B cells that produce rare broad neutralizing antibodies do not react with the germline B cell receptors of the naive B cells that are required to respond to the immunogen. While the initial B cell response to HIV-1 Env is made early on after infection, there is a cross reactivity of gp41 with host or pre-existing foreign molecules such that the B cell antibody clones that make the initial gp41 antibody response are derived from pre-exisitng polyreactive natural B cell clones whose germlines also do not react with gp41 and whose reactivity to gp41 is acquired later in clonal antibody development as cross-reactivity with gp41 is acquired through host or foreign antigen-driven clonal expansion. Once gp41 reactivity is acquired, gp41 then drives the clonal expansion. The second roadblock to vaccine development comes from work showing that both of these antibodies require the long hydrophobic CDR3s with lipid reactivity to neutralize (Alam et al, Proc. Natl. Acad. Sci. USA 106:20234-9 (2009)) and that the 2F5 and 4E10 VHs are sufficiently autoreactive to promote deletion in knock-in mice (Verkoczy et al, Proc. Natl. Acad. Sci. USA 107: 181 -186 (2010)).

Figures 16A-16F. Strategy for induction of broad neutralizing antibodies. Figure 16 A. Vaccines must be designed to stimulate B cell precursors by inclusion of either host (such as lipids) and/or foreign (such as gut flora) antigens to which the polyreactive naive B cell receptors (BCRs) bind (left-most arrow), and antigens (preferred Env constructs) to target intermediate clones of B cells that arise that cross-react with Env. The Env lead candidates for this component of the vaccine is the Malawi 1086 clade C gpl 40 oligomer that has induced in guinea pigs considerable breadth in neutralizing antibodies mixed with the clade B JRFL gpl 40 Env that selectively expresses the MPER neutralizing epitopes (middle arrow) and/or the transmitted founder Envs 6240, 040 and 63521 (see Figures 16B, 16C and 16D) that preferentially express epitopes bound by broadly neutralizing monoclonal antibodies. Finally, to overcome peripheral deletion and/or anergy of B cells that are driven to make polyreactive neutralizing antibodies, the vaccine contains potent TLR agonists or other adjuvants to drive activation of polyreactive B cells by germline and intermediate clone-targeted vaccines (right-most arrow). Figure. 16E. SDS-P AGE images of apoferritin. Figure 16F. Western blot images of apoferritin vs HV00274, HV00276. Acute HIV infection gp41 inferred intermediate antibodies 276 from clone 684-6B and 274 from clone 684-6A both bind to the 19Kd apoferritin subunit. Both mabs also bind to the 60Kd protein in the native marker. Figure 17. Design of HIV-1 Env gp 140 constructs.

Figure 18. Analysis of acute HIV-1 Envs and Group M consensus HIV-1 Env by Blue Native-PAGE and SDS-PAGE.

Figures 19A and 19B. Figure 19A. Immunogenicity of Group M Consensus HIV-1 Env, CON-S and Subtype C Acute HIV-1 Env, 1086C, Subtype B chronic HIV-1 Env, JRFL. Figure 19 B. Methods.

Figure 20. Deglycosylation of JRFL Env gp 140 CF with PNGase.

Figures 21 A and 21 B. Antigenicity of JRFL HIV Env gpHOCF in ELISA.

Figure 22. Antigenicity of JRFL gpl 40 Env in SPR.

Figure 23. Fusion-intermediate state of HIV-1 gp41 targeted by broadly neutralizing antibodies.

Figures 24 A and 24B. Figure 24A. Design of membrane anchored gp41 -inter. Figure 24B. 2F5 and 4E 10 mAbs bind to membrane conjugated gp-41 -inter with nM Kd and almost irreversible off-rates.

Figure 25. Lead candidate immunogens.

Figure 26. Gp41 -inter liposomes with TLR ligands and encapsulated

immunomodulatory ligands. Figure 27. Amino acid sequences for HIV-1 transmitted founder Envs 1086.C, 089.C, 040_C9, and 63521 , and codon optimized encoding sequences.

Figure 28. Clade B JRFL and 6240 gpl40 Env sequence and encoding sequence.

Figure 29. Early B cell response to HIV-1 : the role of innate B cells.

Figure 30. 2F5 and 4E10 broadly neutralizing antibodies react with self antigens that are phylogenetically conserved

Figure 31. 2F5 specifically binds to 43kDa, 50kDa, 70 kDa and 350 kDa 3T3 (mouse) cellular proteins on western blot

Figure 32. Conserved self-antigens that carry the 2F5 nominal epitope.

Figure 33. The H3 domain of kynuereninase (KYNU) is highly conserved.

Figure 34. Structure of human KYNU (PDB 2HZP) and location of ELDKWA motif.

Figure 35. Illustration of the DKW residues (ELDKWA) in human KYNU.

Figure 36. Binding of the 2F5 antibody to human KYNU may require distortion of the H3 domain.

Figure 37. KYNU dimers likely obscure the potential 2F5 binding site. Figure 38. 2F5 and possibly 4E 10 antibodies bind to recombinant human KYNU in western blots.

Figure 39. KYNU is recognized by 2F5-family antibodies.

Figure 40. 2F5 antibody avidly reacts with rhKYNU in a standard ELISA.

Figure 41. 2F5 antibody reacts with a peptide (DP178-Q16L) containing 2F5 epitope - anti-KYNU antibody does not.

Figure 42. 2F5 binding to rhKYNU and DP178-Q16L is comparable in a standard ELISA.

Figure 43. Antibody binding in ELISA plates is antigen specific.

Figure 44. 13H 1 1 does not bind rhKYNU.

Figure 45. 13H1 1 reacts with DPI 78-Q16L but not MPER-656.

Figure 46. Competitive inhibition of 2F5 binding to rhKYNU by JRFL, DPI 78- Q16L and R4A.

Figure 47. Comparable inhibition of 2F5 binding to rhKYNU and JRFL.

Figure 48. Soluble KYNU is bound by 2F5.

Figure 49. rhKYNU binding to surface-captured mAbs. Figures 50A-50C. Binding of 2F5 mAb and 2F5 RUA (reverted unmutated ancestor) antibodies to YNU, (Fig. 50A) 2F5, (Fig. 50B) 2F5-GL1 , (Fig. 50C) 2F5-GL3.

Figure 51. Inhibition of 2F5 binding to 3T3 cells by recombinant HIV-1 gpl 40 (JRFL), and the DP 178 and R4A peptides.

Figures 52A-52D. Enrichment and identification of protein band in intestinal bacterial lysate reactive with mAb HV00276. (Fig. 52A) Western blot analysis following Native PAGE gel run. (Fig. 52B) Protein fractions from bacterial lysate with molecular wt ~500kDa collected following size exclusion chromatography (SEC). (Fig. 52C) SEC fractions show enrichment of 520 kDa protein by Coomassie Blue (1 ) and silver staining (2) and western blotting (3, arrow). (Fig. 52D) Isoelectric zoom fractionation.

Figures 53A-53C. Liquid chromatography-mass spectrometry (LC-MS) identification of RNA polymerase. (Fig. 53A) LC-MS identification of RNA polymerase β subunit. (Fig. 53B) LC-MS identification of RNA polymerase β' subunit. (Fig. 53C) LC-MS identification of RNA polymerase a subunit.

Figure 54. Mab HV00276 binds to RNA polymerase core protein.

Figure 55. Mab HV00276 binds to the a subunit of RNA polymerase core protein.

Figure 56. Neutralization screening of primary memory B cell cultures. Memory B cells from peripheral blood of CHAVI08 chronically-HIV-1 infected volunteer 707-01 -021 -9 were EBV-transformed and stimulated for 14 days inpresence of CD40 ligand, oCpGs and CHK-2 inhibitor at a density of 8 cells/well. At the end of stimulation supernatants were tested for neutralizing activity against the reporter tier 2 clade C CAP45 virus. Solid dots represent the percentage of neutralization of each of the 3,600 cultures. Monoclonal antibodies CH01 -CH05 were isolated from the cultures represented with open dotted symbols. Positive controls (HIV Ig) are shown as open circles on the far right.

Figures 57A-57C. V-heavy and V-light chain alignments of monoclonal antibodies CH01 -CH05. Alignment of the sequences of the CH01 -CH05 V-heavy chains (Fig. 57A), CH01 -CH04 (Fig. 57B) and CH05 (Fig. 57C) V-light chains. The putative reverted unmutated ancestor sequence was used as template for both the V-heavy and the CH01 -CH04 V-light alignments. Since CH05 has an unrelated VK1~6 chain, it is shown separately.

Figure 58. Phylogenetic tree of the V-heavy chains of the CH01 -CH05 monoclonal antibodies.

Figure 59. Alignment of the inferred putative reverted unmutated ancestor antibodies. The alignment of all the putative reverted unmutated ancestor antibodies inferred by applying the V-heavy chains are separated from the V-light chains by "-— ~"

Figure 60. Binding of CHOI , CH02, CH03 quarternary broad neutralizing antibodies to A244 gp 120.

Figure 61. Binding of reverted unmutated ancestors of CHOI , CH02, CH03 quarternary broad neutralizing antibodies to A244 g l20.

Figure 62. PG9 and PG 16 bind to both A244 gpl 20 and 6420 T/F gpl 40. Figure 63. CHOI monoclonal antibodies decreased by binding affinity to A244gD- gp l 20 envelope compared to A244gD+ gpl 20 envelopes.

Figure 64. Forty-eight percent anti-gD IgA vaccine response (99 subjects).

Figure 65. Eight-one percent anti-gD IgG vaccine response (99 subjects).

Figure 66. Potential relevance of gD immunogenicity.

Figure 67. HEP-2 binding

Figure 68. Effect of kifunensine treatment on the ability of CHOI to mediate neutralization

Figure 69. Superimposition of the sequence of CHOI (here called 1 -27-G2) with the PG 16 Fab.


The present invention relates to a method of inducing the production in a subject (e.g., a human subject) of broadly neutralizing antibodies against HIV-1. The method comprises administering to the subject a non-HIV-l antigen that binds to a germline B cell receptor, the non-HIV-l antigen being administered in an amount and under conditions such that intermediate clones of B cells are produced that secrete antibodies that cross-react with HIV-1 Env. The method further comprises administering to the subject an HIV-1 antigen in an amount and under conditions such that naive B cells or their B cell intermediate clones are produced that secrete the broadly neutralizing anti-HIV-1 antibodies. It is likely that, for some epitopes on gpl 20, there will be rare na'ive B cells capable of binding to those epitopes while, for other epitopes, na'ive B cells that can give rise to broadly neutralizing antibodies will not bind Env and will need to be stimulated by additional non-Env epitopes. Roadblocks to inducing broadly neutralizing antibodies are described in Figure 15 and the present strategy for overcoming those roadblocks is described in Figure 16A.

Non-HIV-1 antigens suitable for use in the invention include host and/or foreign antigens. Non-HIV-1 antigens include, for example, lipids, such as cardiolipin, phosphatidyl serine, phosphatidylethanolamine, phosphatidylcholine, phosphotidylinositol, sphingomyelin, and derivatives thereof, e.g., l -palmitoyl-2- oleoyl-s«-glycero-3-[phospho-L-serine] (POPS), 1 -palmitoyl-2-oleoyl- phosphatidylethanolamine (POPE), and dioleoyl phosphatidylethanolamine (DOPE), or fragments thereof. Use of hexagonal II phases of phospholipids can be advantageous and phospholipids that readily form hexagonally packed cylinders of the hexagonal II tubular phase (e.g., under physiological conditions) are preferred, as are phospholipids that can be stabilized in the hexagonal II phase. (See Rauch et al, Proc. Natl. Acad. Sci. USA 87:41 12-41 14 ( 1990);

Aguilar et al et al, J. Biol. Chem. 274: 25193-25196 (1999)). Other suitable non- HIV-1 antigens include, for example, phycoerythrin (PE), C-phycocyanin (C-PC), or other phycobiliprotein, apoferritin, and anerobic or aerobic gut flora or component(s) thereof (for example, the 520Kd antigen (or the RNA polymerase holoenzyme or the RNA polymerase core protein, or subunit thereof, such as the a subunit of RNA polymerase core protein or portion thereof comprising the epitope to which mAb HV00276 binds), or the 60Kd or 50Kd antigen). The data presented in Example 2 indicates that mAb HV00276 binds to the a subunit of E. coli RNA polymerase core protein. The sequence homology is high between the a subunit of E. coli RNA polymerase core protein and a homologs from other bacteria (e.g., B. subtilis, S. dysenteriaea, S. enterica, M. tuberculosis, H. pylori and H. influenza) and eukaryotes (e.g., human and mouse proteins related to S. cerevisiae Rpb3 and Rpbl 1 ) (Zhang and Darst, Science 281 :262-266 (1998)). Accordingly, the invention includes the use of the 520Kd antigen (or subunit thereof, such as the a subunit of RNA polymerase core protein or portion thereof comprising the epitope to which mAb HV00276 binds) from eukaryotes and from bacteria in addition to E. coli. (See, for example, E. coli RNA polymerase a subunit : NP_289856 (gi/15803822); S. dysenteriaea: YP_404940(gi:82778591 ); H. influenzae: NP_438962 (gi: 16272744); Rpb3 : Swiss-Prot: P37382.2; Rpb3 (Homo sapiens): NPJ 16558.1 (gi: 14702171 ).)

ynureninase (KYNU) is a member of the family of pyridoxal 5'- phosphate (PLP)-dependent enzymes known as the aspartate aminotransferase superfamily. Eukaryotic constitutive kynureninases preferentially catalyze the hydrolytic cleavage of 3-hydroxy-l-kynurenine to produce 3-hydroxyanthranilate and 1-alanine. The cloning, expression, purification, characterization and crystallization of Homo sapiens KYNU has been reported (Lima et al,

Biochemistry 46(10):2735-2744 (2007). As described in Example 3 below, KYNU carries the core 2F5 epitope in its conserved H3 domain.

Based on the data provided in Example 3, it is anticipated that this endogenous ligand is responsible for tolerizing B and T lymphocytes and thereby inhibiting the production of effective immune responses against HIV- 1 in humans administered HIV-1 gp41 MPER epitope peptides. The invention provides, in one embodiment, methods of effecting immunization against HIV-1 comprising administering cross-reactive antigens that break this tolerance specifically, that is, without affecting tolerization against other, irrelevant self antigens. Suitable antigens include, for example, the recombinant KYNU molecule expressed in CHO or 293T cells with the ELDKWA sequence or a mutant gp41 or KYNU sequence with the ELEKWA sequence (ELEKWA is not present in human proteins and thus is not expected to be tolerizing). Other immunogens that can be used include transmitted/founder or wildtype chronic envelope gpl40s or gpl 60s or MPER peptides in liposomes with either the ELEKWA or the ELDKWA sequence. Immunogens with the ELDKWA sequence are, advantageously, administered with strong adjuvants, such as squalene based monophosphosphoryl lipid A, oligonucletides (oCpGs) and R848 (TRL-7/8 agonist). Liposomes with these TLR agonists and IFNa can also be used. (See also comments below.)

HIV-1 antigens suitable for use in the invention include membrane- proximal external region (MPER) antigens (Armbruster et al, J. Antimicrob. Chemother. 54:915-920 (2004), Stiegler and Katinger, J. Antimicrob. Chemother. 512:757-759 (2003), Zwick et al, Journal of Virology 79: 1252-1261 (2005), Purtscher et al, AIDS 10:587 (1996)) and variants thereof, for example, variants that confer higher neutralization sensitivity to MPER Mabs 2F5 and 4E10 or to other broadly neutralizing Envs, such as the MPER mutant Env peptide lipid complex containing a L669S mutation in the MPER (Shen et al, J. Virology 83:3617-25 (2009)). Preferred immunogens include those shown in Figures 25 and 26, as well as Figures 16B, 16C, Figure 17, Figure 18 and Figure 20. In another preferred embodiment, the variant is a MPER epitope peptide with an L669S mutation that confers higher neutralization sensitivity to MPER mAbs 2F5 and 4E10 (Shen et al, J. Virology 83: 3617-25 (2009)).

HIV-1 antigens suitable for use in the invention also include transmitted founder HIV-1 Envs, or fragments thereof. These fragments can be representative of portions of the CD4 binding site of gpl20 (Chen et al, Science

362(5956): 1 123-7 (2009)), MPER sequences, portions of gpl20 incorporating the V2, V3 regions of gpl20 (Walker et al, Science 326(5950):285-9 (2009) Epub 2009 Sept. 3), etc (e.g., see the sequences for 1086, 089, 6240, 040_C9 and 63521 set forth in Figs. 27 and 28). Preferred Env antigens include the Malawi 1086 clade C, 6321 and the US clade B 040_C9 gpl40 oligomers (Figures 17 and 18) (Keele et al, Proc. Natl. Acad. Sci. USA 105:7552-7 (2008)) produced as previously described (Liao et al, Virology 30:268-282 (2006)), which have induced in guinea pigs considerable breadth in neutralizing antibodies (Figure 19A), mixed with the clade B JRFL gpl40 Env, or fragment thereof, that selectively expresses the MPER neutralizing epitopes (see Figure 28 ). The JRFL gpl40 Env oligomer (Figures 19B, 20, 21 A and 2 I B) constitutively binds the 2F5 mAb. The JRFL oligomer deglycosylated using 500U of PNgase

endoglycosidase (New England BioLabs, Ipswich, MA) has enhanced binding of 2F5 and new binding of the 4E10 mAb (exposure of the 4E10 epitope on gp41) (Figures 21A and 21 B). The enhanced binding of 4E10 to deglycosylated JRFL is also shown in surface plasmon reasonance (SPR) analysis in Figure 22.

The method of the invention can be effected by administering to the subject a prime immunization comprising a non-HIV-1 immunogen followed by one or more boosts of an HIV-1 Env antigen. As pointed out above, suitable non- HIV-1 immunogens include lipids (e.g., cardiolipin, phosphotidylserine, or other anionic lipid), components of anaerobic or aerobic gut flora bacteria,

phycobiliproteins (e.g., PE) and KYNU or fragment thereof. As also pointed out above, suitable HIV-1 Env antigens include transmitted founder Env 1086.C from Malawi, 089.C from Malawi, 040_C9 from the U.S. and 63521 from a Clade B acute HIV-1 infected U.S. patient. Both the primes and the boosts suitable for use in the present method can comprise both non-HIV-1 and HIV-1 immunogens. Prime boost regimes can be readily optimized by one skilled in the art. DNA sequences encoding proteinaceous components of such regimens can be administered under conditions such that the proteinaceous component is produced in vivo.

As described in Example 5 below, 5 clonally related B cells have been isolated from a single patient that produce broadly neutralizing antibodies (CHOI through CH05). Possible reverted unmutated ancestors of the clonally-related antibodies have been inferred and expressed as real antibodies. The phylogenetic tree of these antibodies has been reconstructed. Both the natural and inferred ancestor antibodies have been characterized for their ability to bind a panel of HIV envelope proteins and to neutralize a panel of HIV isolates. It is important to note that the reverted unmutated ancestors (RUAs) bind to A244gD+ envelope. Therefore, such envelope, or other envelopes described to be neutralized by the RUAs, can be used as the "prime" in a preferred vaccine strategy of the invention. In accordance with this strategy, the "boost" can be effected, for example, using envelopes that are bound by the mature antibodies described herein. A further "boost" can be effected, for example, with 6420 or 63521 (or other protein, peptide or polypeptide that binds).

When a DNA prime or boost is used, suitable formulations include a DNA prime and a recombinant adenovirus boost and a DNA prime and a recombinant mycobacteria boost, where the DNA or the vectors encode, for example, either HIV- 1 envelope or a proteinaceous non-HIVl -1 antigen, such as a gut flora or KYNU component. Other combinations of these vectors can be used as primes or boosts, either with or without HIV-1 antigen and/or non-HIV-1 antigen.

In accordance with the invention, the non-HIV-1 antigen can be present in a liposome with the HIV- 1 Env antigen and one or more adjuvants. Alternatively, the non-HIV-1 antigen can be conjugated, for example, using a hetero- bifunctional agent such as DSSP, to the HIV-1 Env antigen and formulated with one or more adjuvants.

Liposomes expressing MPER antigens (Dennison, et al, J. Virology 83 : 1021 1 -23 (2009)) with or without Toll Like Receptor (TLR) agonists have been described (see, for example, WO 2008/127651). Gp41 intermediate state protein (Figure 23) has been described by (Frey et al, Proc. Natl. Acad. Sci. USA 105-3739-44 (2008)). The gp41 intermediates can be formulated with liposomes (Figures 24A and 24B) to form a stable immunogens that bind well to 2F5 and 4E10 (Figure 25). Gp41 MPER immunogens of the invention can be adjuvanted by incorporating, for example, monophosphorylipid A (MPL-A) (Avanti Polar Lipids, Alabaster, AL ) and a TLR 9 agonist, such as oCpGs 10103 (5' - TCGTCGTTTTTCGGTCGTTTT-3 ') and R848 TLR 7 agonist (Enzo Life Sciences, Farmingdale, NY ) (Figure 26). In addition, cytokine stimulators of B cell class switching, such as BAFF (BLYS) and/or APRIL (He et al, Immunity 26:812-26 (2007); Cerutti and Rescigno, Immunity 28: 740-50 (2008)) can be incorporated into the liposomes for optimal B cell stimulation.

Liposomes suitable for use in the invention include, but are not limited to, those comprising POPC, POPE, DMPA (or sphingomyelin (SM)),

lysophosphorylcholine, phosphatidylserine, and cholesterol (Ch). While optimum ratios can be determined by one skilled in the art, examples include POPC:POPE (or POPS):SM:Ch or POPC:POPE (or POPS):DMPA:Ch at ratios of 45:25:20: 10. Alternative formulations of liposomes that can be used include DMPC (1 ,2- dimyristoyl-sn-glycero-3-phosphocholine) (or lysophosphorylcholine), cholesterol (Ch) and DMPG (l ,2-dimyristoyl-sn-glycero-3-phoshpho-rac-(l -glycerol) formulated at a molar ratio of 9:7.5: 1 (Wassef et al, ImmunoMethods 4:217-222 (1994); Alving et al, G. Gregoriadis (ed.), Liposome technology 2nd ed., vol. Ill CRC Press, Inc., Boca Raton, FL (1993); Richards et al, Infect. Immun.

66(6):285902865 (1998)). The above-described lipid compositions can be complexed with lipid A and used as an immunogen to induce antibody responses against phospholipids (Schuster et al, J. Immunol. 122:900-905 (1979)). A preferred formulation comprises POPC:POPS:Ch at ratios of 60:30: 10 complexed with lipid A according to Schuster et al, J. Immunol. 122:900-905 (1979). The optimum ratio of peptide to total lipid can vary, for example, with the peptide and the liposome.

A variety of adjuvants can be used in the present invention (including those noted above). The peptide-liposome immunogens and the conjugates described above can be formulated with, and/or administered with, adjuvants such as squalene-based adjuvants (Kaldova, Biochem. Biophys. Res. Communication, Dec. 16, 2009 E-pub ahead of print) and/or TLR agonists (e.g., a TRL 3, TRL 5, TRL4, TRL9 or TRL7/8 agonst, or combination thereof) that facilitate robust antibody responses (Rao et al, Immunobiol. Cell Biol. 82(5):523 (2004)). Other adjuvants that can be used include alum and Q521 . Oligo CpGs in an oil emulsion such as Emulsigen (an oil in water emulsion) (Tran et al, Clin. Immunol. 109(3):278-287 (2003)) can also be used. Additional suitable adjuvants include those described in 1 1 /302,505, filed December 14, 2005, including the TRL agonists disclosed therein. (See also Tran et al, Clin. Immunol. 109:278-287 (2003), US Appln Nos. 20030181406, 20040006242, 20040006032,

20040092472, 20040067905, 20040053880, 20040152649, 20040171086,

20040198680, 200500059619). Immune response enhancing TLR ligands, such as Lipid A, oligo CpG and R-848 can be formulated individually or in

combination into liposomes that have HIV- 1 Env conjugated in them.

Liposomes loaded with strong adjuvants (e.g., potent TLR agonists) are examples of immunogens that can be used to overcome peripheral deletion and/or anergy of B cells that do get driven to make polyreactive neutralizing antibodies.

Transmembrane domain anchoring of HIV-1 gp41 peptides to liposomes can be used to achieve functional epitope display. The transmembrane domain of HIV- 1 gp41 can be used to anchor the peptide into liposomes comprising synthetic lipids. Induction of trimerization of the TMD can facilitate formation of trimeric forms of gp41 MPER. Alternatively, His-tagged (c-terminus end) versions of the Env gpl 40 can be anchored into liposomes as described for an intermediate form of HIV-1 gp41 (gp41 -inter).

The mode of administration of the non-HIV- 1 immunogen and/or HIV-1 protein/polypeptide/peptide, or encoding sequence, can vary with the immunogen, the patient and the effect sought, similarly, the dose administered. Typically, the administration route will be intramuscular, intravenous, intraperitoneal or subcutaneous injection. Additionally, the formulations can be administered via the intranasal route, or intrarectally or vaginally as a suppository-like vehicle.

Generally, the liposomes are suspended in an aqueous liquid such as normal saline or phosphate buffered saline pH 7.0. Optimum dosing regimens can be readily determined by one skilled in the art. The immunogens are preferred for use prophylactically, however, their administration to infected individuals may reduce viral load.

The human monoclonal antibodies (hu mAb) 2F5 and 4E 10 bind with high specificity and nanomolar (nM) affinities to polypeptides that correspond to the HIV-1 gp41 MPER. Both hu mAb also react with discrete human and mouse cellular antigens as determined by immunofluorescence microscopy and western blotting. These properties indicate that 2F5 and 4E10 are ideal for the isolation of cellular proteins, including denatured forms and polypeptides, biochemically extracted from mammalian cells and recovered by standard immunoprecipitation methods. The same properties of 2F5 and 4E 10 make them suitable for the identification of extracted cellular proteins/polypeptides by the standard methods of mass spectroscopy. Briefly, immunoprecipitated cellular proteins/polypeptides specifically bound to 2F5 or 4E10 can be subjected to enzymatic digestion and the mass and charge of the resulting fragments used to identify the parental


Certain aspects of the invention are described in greater detail in the non- limiting Examples that follow (see also Maksyutov et al, J. Clin. Virol. Dec; 31 Suppl l :S26-38 (2004), Haynes et al, Science 308: 1906 (2005), Gurgo et al, Virology 164:531 -536 (1988), USP 7,61 1 ,704, U.S. Appln. No. 1 1/812,992, filed June 22, 2007, U.S. Appln. No. 1 1/785,077, filed April 13, 2007,

PCT US2006/013684, filed April 12, 2006, PCT/US04/30397 (WO2005/028625), WO 2006/1 10831 , WO 2008/127651 , U.S. Published Application Nos. 2008/0031890 and 2008/0057075, U.S. Appln No. 1 1/918,219, filed December 22, 2008, U.S. Prov. Appln. No. 60/960,413, filed February 28, 2007, and U.S. Prov. Appln. Nos. 61/166,625, 61/166,648 and 61/202,778, all filed April 3, 2009, U.S. Prov. Appln. No. 61/282,526, filed February 25, 2010, U.S. Prov. Appln. No. 61/344,457, filed July 27, 2010, U.S. Prov. Appln. Client File No. 01579-1597, filed August 25, 2010, PCT/US2010/01018,

PCT US2010/03001 1 , and PCT/US2010/01017the entire contents of which are incorporated herein by reference).


Experimental Details

Acute HIV-1 Infected Patients. The patients selected for study were from 17 to 30 days following transmission with the dates of transmission estimated from patient history and Fiebig classification (Fiebig et al, AIDS 17: 1871 -1879 (2003) ). Patients 065-0 and FIKE were Fiebig Stage 1 , while patients 068-9, 684-6 and MCER were Fiebig stage 2.

Control Subjects. Single plasmablast/plasma cell sorts were performed on bone marrow, leukapheresis or peripheral blood mononuclear cells (PBMC) of uninfected subjects as well as those vaccinated with trivalent inactivated (TVI) influenza vaccine (FLUZONE® 2007 or 2008). Those immunized with TVI were studied 7 days after immunization (Liao et al, J. Virologic Methods 158: 171 -9, (2009); Wrammert et al, Nature 453 :667-71 (2008); Smith et al, Nature Protocols 4:372-84 (2009)).

Flow Sorting Strategy. PBMC, leukapheresis or bone marrow samples were reacted with anti-B cell antibodies as previously described (Liao et al, J. Virologic Methods 158: 1 71 -9 (2009)). Wrammert et al (Nature 453 :667-71 (2008)) have shown that the cells that are antibody secreting cells in human PBMC are those that are within the CD19+, CD38hi+, IgD", CD20lo+/- B cells. Thus, in both acute HIV infection (AHI) and in influenza vaccine vaccinated controls, to isolate single antibody secreting plasmablasts/plasma cells, CD19+, CD38hi+, IgD", CD20lo+/" cells were sorted by flow cytometry into single 96 well plates containing RNA extraction buffer as described (Liao et al, J. Virologic Methods 158: 171 -9 (2008); Wrammert et al, Nature 453 :667-71 (2008)). As positive controls for definition of successful isolation of the correct

plasmablast/plasma cell population, the same population was isolated from day 7 after trivalent influenza vaccine (FLUZONE® 2007 or 2008) vaccines. It was demonstrated that, as expected, 75% of those sorted cells were indeed influenza specific antibodies (Wrammert et al, Nature 453 :667-71 (2008)).

Identification and expression of the transmitted/founder envelope. The transmitted/founder Env of patients 684-6 and FI E were identified by single genome amplification and Env gene sequencing as previously described (Keele et al, Proc. Natl. Acad. Sci. USA 105:7552-7 (2008)). Env gpl 40C (gpl 20/41 cleavage site mutated), gpl 20 and gp41 proteins were expressed by transient transfections of 293T cells as described (Liao et al, J. Virologic Methods 158: 1 71 - 9 (2008)).

PCR amplification of plasmablast/plasma cell immunoglobulin VH and VL genes. The VH and VL Ig chains of sorted B plasmablast/plasma cells were isolated by single cell PCR and recombinant antibodies produced as described (Liao et al, J. Virologic Methods 158: 171 -9 (2009) ; Wrammert et al, Nature 453 :667-71 (2008) ; Smith et al, Nature Protocols 4:372-84 (2009)).

Sequencing, sequence annotation, quality control, and data management °f Ig VH and VL sequences. All PCR products of Ig VH and VL genes were purified using a Qiagen (Valencia, CA) PCR purification kit and sequenced in forward and reverse directions using an ABI 3700 instrument and BigDye® sequencing kit (Applied Biosystems, Foster City, CA). Base calling for each sample is done using Phred (Ewing et al, Genome Res. 8: 175-85 (1998); Ewing and Green, Genome Res. 8: 186-94 (1998)). The forward and reverse strands of the antibody genes are assembled to one final nucleotide sequence using a novel assembly algorithm based on the quality scores at each position (Kepler et al, submitted). The estimated PCR artifact rate was .28 or approximately 1 PCR artifact per 5 genes amplified. The isotype of the immunoglobulin is determined by a local alignment algorithm (Smith and Waterman, J. Mol. Biol. 147: 195-7 (1981 )). The germline rearrangement of the quality assured antibody sequence is determined using SoDA (Volpe et. al, Bioinformatics 22:438-44 (2006)).

Genomic information derived from SoDA, such as gene segment usage, somatic mutations and CDR3 regions, are stored in an ORACLE database for easy access.

To determine if antibodies from the same subject are clonally related, the following 3 criteria were utilized. First, the heavy chain of the antibodies in question must use the same VH and JH gene segments. Due to the length and high mutation in the D segment, these are more difficult to identify. Thus, similarity of D segments is not used as criteria for clonal relatedness. Similarly, both light chains must use the same Vic/νλ and ίκ/ίλ. Second, the heavy chains of the antibodies in question must have the same CDR3 length. This also applies to light chains. Third, the nucleotide sequence of the CDR3 of the heavy chains must be 70% identical. The same applies to the CDR3 of the light chain.

Antibodies 'that adhere to these three criteria are labeled as.being clonally related. Maximum Likelihood trees were constructed to determine the phylogenetic relationship between the clones using the PHYLIP 3.63 package (Felsenstein, Philos. Trans. R. Soc. Lond. B. Biol. Sci. 360: 1427-34 (2005)) using the inferred germline from SoDA as the root. The ancestral sequences were also inferred using the same package.

Design and generation of inferred germline and intermediate antibodies. For each member of an antibody clonal family, Maximum Likelihood analysis was used to infer the germline antibody precursor as well as multiple antibody intermediate forms (Felsenstein, J. Mol. Evol. 17: 368-76 (1981 ); Volpe et al, Bioinformatics 22:438-44 (2006)). These VH and VL genes were synthesized (GeneScript, Piscataway, NJ) and expressed as IgGl mAbs by recombinant techniques as above.

Expression of Vn and VL as recombinant mAbs. The isolated Ig VH and VL gene pairs were assembled by PCR into the linear full-length Ig heavy- and light- chain gene expression cassettes for production of recombinant mAbs by transfection in human embryonic kidney cell line, 293T (ATCC , Manassas, VA) using the methods as described (Liao et al, J. Virol. Methods 158: 171 -9 (2009)). The purified PCR products of the paired Ig heavy- and light-chain gene expression cassettes were co-transfected into 80-90% confluent 293T cells grown in 6-well ^g of each per well) tissue culture plates (Becton Dickson, Franklin Lakes, NJ) using PolyFect (Qiagen, Valencia, CA) and the protocol recommended by the manufacturer. Six to eight hours after transfection, the 293T cells were fed with fresh culture medium supplemented with 2% fetal calf serum (FCS) and were incubated at 37°C in a 5% C02 incubator. Culture supematants were harvested three days after transfection and quantified for IgG levels expressed and screened for antibody specificity. For future characterization of select antibodies identified through screening assays, the linear Ig heavy and light chain gene constructs were cloned into pcDNA 3.3 for production of purified recombinant mAbs using standard molecular protocols.

For production of purified recombinant mAbs derived from the isolated VH and VL genes and the inferred germline and intermediate precursor antibody sequences, 293T cells cultured in T175 flasks were co-transfected with the heavy and light chain Ig gene-containing plasmids using using PolyFect (Qiagen, Valencia, CA), cultured in DMEM supplemented with 2% FCS. Recombinant mAbs were purified from culture supematants of the transfected-293T cells using anti-human Ig heavy chain specific antibody-agarose beads (Sigma, St. Louis, MO).

Screening for antibody specificity by ELISA and Luminex assays.

Concentration of recombinant mAbs in the supernatants were determined using the method as described (Liao et al, J. Virol. Meth. 158: 171 -179 (2009)).

Specificity of the expressed recombinant mAb were assayed for antibody reactivity to HIV-1 antigens and to a panel of non-HIV-1 antigens. HIV antigens included Env peptides gp41 immunodominant region


(Immunodiagnostics, Woburn, MA), HIV-1 group M consensus gpl 20 (Liao et al, Virology 353 :268-82 (2006)), HIV-1 group M consensus gpl 40 CFI (Liao et al, Virology 353 :268-82 (2006)), p66 (Worthington Biochemical, Lakewood, NJ), p55 (Protein Sciences, Meriden, CT), p31 (Genway, San Diego, CA), nef (Genway, San Diego, CA), tat (Advanced Bioscience, Kensington, MD) and AT- 2 inactivated HIV-1 ADA virions (Rutebemberwa et al, AIDS Res. Human Retrovirol. 23:532-42 (2007)); gift of Jeffrey Lifson, NIH, NCI, Frederick Cancer Research Facility). In addition, 684-6 mAbs were assayed against autologous gpl 40, gpl 20 and gp41 , and FIKE mAbs were assayed against autologous gpl40 and gp l 20. Non-HIV- 1 antigens included trivalent influenza vaccine 2007 (FLUZONE 2007), recombinant influenza HA protein from H I A/Solomon Islands/03/2006 (Protein Sciences Corp. Meriden, CT), tetanus toxiod

(Calbiochem, San Diego, CA), HEP-2 cells (Inverness Medical Professional Diagnostics, Princeton, NJ), cardiolipin (Avanti Polar Lipids, location (Alabaster, AL) (Haynes et al, Science 308: 1906-8 (2005)) and lipid A (Avanti Polar Lipids, Alabaster, AL). Whole cell lysates of anaerobic and aerobic bacterial extracts termed as gut flora were prepared as described below. Briefly, bacteria were inoculated from 4 stool specimens from patients and grown on agar plates under anaerobic or aerobic conditions at 30°C. Confluent bacteria were harvested, washed twice with phosphate-buffered saline (PBS) and treated with a

commercially available bacterial protein extraction reagent (Pierce, Rockford, IL). The resulting extracts were filtered with a 0.22μπι filter and stored at -80°C until use (Kawatsu et al, J. Clin. Microbiol. 46: 1226-31 (2008)). Assays against FLUZONE®, influenza HA, gp41 immunodominant and MPER regions, as well as gut flora whole cell lysates, were performed by both ELISA (Tomaras et al, J. Virology 82: 12449-63 (2008)) and Luminex bead assays (Tomaras et al, J.

Virology 82: 12449-63 (2008)). Assays against tetanus toxoid, cardiolipin (Sigma, St Louis, MO), killed Cryptococcus and Candida albicans were ELISA Assays for reactivity with Hep-2 epithelial cells were indirect immunofluoresence assays (Mietzner et al, Proc. Natl. Acad.. Sci. USA 105:9727-32 (2009)).

Surface Plasmon reasonance (SPR) analysis of antibody reactivity. SPR binding assays were performed on a BIAcore 3000 (BIAcore Inc, Piscattaway, J) maintained at 20°C. HIV- 1 gp41 or oligomeric gpl 40 proteins (Con S gpl 40, autologous Env gpl40) were immobilized on a CM5 sensor chip by standard amine coupling as previously described (Alam et al, J. Immunol.

178:4424-35 (2007)). Human mAbs were captured on anti-human Fc antibody coupled surfaces and then each human mAbs were captured to about 200-500 RU. Specific binding responses of mAb binding were obtained following subtraction of non-specific binding on control surfaces (HIV-1 gpl 20 for Env immobilized surfaces and human IgG, IS6, for mAb captured surfaces). Rate constants were measured using the bivalent analyte model (to account for the avidity of bivalent Ig molecules) and global curve fitting to binding curves obtained from mAb titrations. MAbs were injected at 30 μΕ/πήη for 2-6 min and glycine-HCl pH 2.0 and surfactant P20 (0.01 %) were used as the regeneration buffer. Results

Influenza vaccination. Clones of antibodies from influenza vaccinated subjects derived from single cell sorted plasma cells/plasmablasts were studied and the response was found to be highly clonal. The clones members almost all reacted with the influenza antigen tested. Figure 1 shows a representative influenza antibody clone against HI Soloman Islands hemagglutinin. A total of 450 antibodies were isolated from plasma cells/plasmblasts of three influenza vaccinated subjects and, of these, 57.7% were influenza-specific. Of all the 265 antibodies isolated from influenza infected subjects, twenty independent clones of clonally related antibodies were identified, among which, 1 15 antibodies (92%) reacted with influenza antigens.

Clonal antibody response in acute HIV infection. In contrast to influenza vaccination, where -75% of plasma cells/plasmablasts were influenza specific, out of a total of 1074 recombinant antibodies that have been isolated from plasmablasts/plasma cells of 5 AHI patients, 89 or 8.3% expressed antibodies (range 3.3% to 13.4%) were HIV-1 specific, while the majority of the remainder of the mAbs either were against non-HIV antigens (~6%) or had unknown specificity (882 or 82.1 %). With the panel of non-HIV-1 related antigen assays, it was possible to demonstrate high affinity antibodies to Hep-2 epithelial cells (27 or 2.5%), gut flora (5 or 0.5%), cardiolipin (4 or 0.4%), influenza (9 or 0.8%), Cryptococcus (4 or 0.4%), Candida albicans (2 or 0.2%), and tetanus toxoid (8 or 0.7%). An additional 38 or 3.5% reacted with at least 2 of these antigens. Three of the patients had lipid A and one patient had gut flora antibodies suggesting the very early onset of gut damage, microbial translocation and induction of anti-lipid A and gut flora antibodies. Remarkably, none of these early AHI patients had any mAbs detected with HIV-1 specificities other than gp41 within days 17-30 after HIV transmission.

It was previously reported that consensus Envs were equal to autologous Envs in detecting the AHI response to gp41 (Tomaras et al, J. Virology 82:

12449-63 (2008)). However, to rule out the possibility that responses were being missed in AHI B cell analysis, the mAbs from 684-6 and FIKE were screened with their autologous recombinant gpl40 Envs. In general, the response to the autologous gpl 40 envs was much less than to the clade B gp41 .

Thus, the initial plasmablast/plasma cell repertoire response to the transmitted/founder virus, like the plasma antibody response (Tomaras et al, J. Virol. 82: 12440-63 (2008)), was. focused on Env gp41 epitopes. In addition, HIV-1 activates and drives to terminal differentiation preexisting memory B cells from previous vaccination or infectious agent antigens, such as Cryptococcus, Candida albicans, and tetanus toxoid. Moreover, in the course of AHI, polyreactive clones of Hep-2 cell autoreactive B cells are triggered to join the initial plasmablast/plasma cell response.

Analysis of antibody clones within the AHI plasmablast/plasma cell repertoire. In general, there few clones isolated from the AHI

plasmablast/plasma cell repertoire compared to the reported plasmablast/plasma cell repertoires induced by influenza vaccination (Wrammert et al, Nature 453 :667-72 (2009)) or the memory B cell repertoire of gpl 40+ B cells in subjects with broad neutralizing antibody activity in plasma (Scheid et al, Nature 458:636- 40 (2009)). In chronic HIV- 1 infection in six patients with broad neutralizing antibodies, Scheid et al (Nature 458:636-40 (2009)) found the number of B-cell clones varied among patients from 22 to 50 in 502 antibodies isolated from those six patients.

In the study of AHI, only 8 clones of antibodies were found in 1074 mAbs isolated from 5 AHI patients. These included three clones of antibodies that reacted with gp41 among 6 independent clones of antibodies identified in one of AHI patients. Of interest, of all 52 clonal members of the 3 AH1 gp41 clones, only 1 7 (37%) reacted with gp41 . This is in contrast to 94% of influenza-reactive influenza clone members.

Figure 2 shows AHI clone 684-6B— a remarkable VH3-7, DH1 -26, JH5, VK1 -39, J 4, IgG3 mutated clone with 52 members, with no unmutated members. Out of the 57 antibodies, only 4 (8%) reacted with gp41.

Analysis of the gp41 reactivity with clone inferred germline and intermediate antibodies. It was reasoned that either HIV-1 gp41 was reacting with the germline B cell receptor of naive B cells and was stimulating low affinity clones with poor antigen drive, or that gp41 may cross-react with pre-existing clones of memory B cells and enjoin clonal members to undergo simultaneous gp41 and self antigen drive. To distinguish between these two possibilities, Maximum Likehood analysis was used to infer the germline unmutated antibody and partially mutated clone intermediates were used to determine their reactivity with gp41 (Figure 3). To determine where in the clone development reactivity with gp41 was acquired (i.e., germline VH+VL or later intermediates), inferred ' germline and clone member intermediates were assayed for reactivity with clade B gp41 , autologous gp 140 and group M consensus gp 140 (Figure 4). It was found that reactivity of clone 684-6B was acquired at the second intermediate precursor antibody (Figures 4 and 5).

The next question asked was whether the reactivity with gp41 represented antigen drive by gp41. Figure 6 shows more inferred intermediate antibody clones were produced in mg quantities and analyzed for the dissociation constants ( d) of antibody binding to gp41 . Figure 7 shows a heat map plot with the dissociation constants plotted as log 10 of the Kds, and demonstrates that, indeed, as the intermediates progress to actual isolated antibodies, there is progression of affinity maturation for binding to gp41. Given the induction of polyreactive non-HIV-1 gp41 clones during AHI, the next question asked was whether clone 684-6B members were polyreactive by reactivity with cardiolipin and Hep-2 epithelial cells. In the Hep-2 indirect immunofluoresence assay, reactivity of clone 684-6B was acquired at the same inferred intermediate precursor stage as gp41 reactivity (Figure 8). All clone members of 684-6B reacted with cardiolipin, including the germline unmutated antibody, and while Hep-2 reactivity waxed and waned during clone

development, reactivity with cardiolipin was relatively stable throughout the intermediates until the end clones 307 and 350. The polyreactivity of the germline and other clone members with cardiolipin stongly suggests that the initial antibody response to HIV is derived from HIV gp41 stimulating a preexisting, polyreactive clone of natural antibodies and gp41 recruits clones of B cells to become polyreactive gp41 clones as soon as the original clone acquires cross reactivity to gp41 by somatic hypermutations. This finding has considerable ramifications to HIV vaccine design.

The nature of the germline reactivity to non-HIV-1 antigens. Given the surprising result of the acquisition of reactivity of the 684-6B clone not in the germline antibody of each clone but in inferred clone intermediates, an effort was made to identify host antigens against which the germline might react to identify likely origins of the antibody clones activated in HIV.

It was hypothesized that because there is early gut microbial translocation in the gut due to AHI and because much of the initial antigenic stimulation in AHI comes at mucosal surfaces, the initial antibody response may in some manner be tied to or related to the gut microbial antibody response. To study this, a determination was made as to whether there were measurable reactivity of the clonal antibodies and inferred germline and inferred intermediates from 684-6B clone to the whole cell lysates of anerobic and aerobic gut flora. In addition, EBV transformation was used to isolate a panel of pentameric IgM mAbs from intestine, bone marrow or blood of AH1 or uninfected subjects.

First, a series of IgM antibodies was isolated from AHI and two from uninfected subjects that were either gp41 reactive or gp41 non-reactive. The question asked was whether the IgMs that were reactive with gp41 also were reactive with gut flora. Table 1 shows that, indeed, all the mAbs that were gp41 reactive were also reactive with gut flora antigens while those mAbs that were not reactive with gp41 were not gut flora reactive.

Table 1 : All HIV- Env gp41 IgM abs Isolated from Infected or Uninfected also Bind to

Either Anerobic or Aerobic Gut Bacterial Whole Cell Lysates

MAb HIV-1 Env gp41 Anerobic Gut Aerobic Gut Source of Mab

Bacteria WCL Bacteria WCL

Reactivity in Luminex Units

21B10 173 272 1012 AHI intestine

2C3 148 210 591 AHI intestine

F3 177 671 2237 AHI intestine

F8 1023 372 5433 AHI intestine

1E7 17153 259 133 AHI bone marrow

2B9 24886 742 347 AHI bone marrow

ALL8 13031 1816 1584 AHI intestine

C1 -2 2500 172 >80 uninfected intestine

C08 3673 241 >80 uninfected blood

XM-1 <80 <80 <80 uninfected blood

XM-2 <80 <80 <80 AHI intestine

XM-3 <80 «80 <80 AHI intestine

AHI = acute/early HIV-1 infection.

ab = monoclonal antibody.

WCL = whole cell lysate.

<80 = no reactivity over background in Luminex assay with gp41 or gut flora whole cell lysates.

Remarkably, when the germline and intermediate precursors from all clones tested were assayed with whole cell lysate of aerobic and anerobic gut flora, all of the antibodies in all of the clones reacted with gut flora whole cell lysate. Figure 9 shows a heat map of the 684-6B clone reacting at each mAb with aerobic whole cell lysate (WCL). Similar results were obtained with anerobic WCL. When analyses were performed to determine antigen drive mediated by gut flora, it was found that, indeed, there were increases in antibody affinity coincident with progressive somatic hypermutation in the AHI clones, though less so than for gp41.

Western blot of AHI gp41 mAbs with anerobic and aerobic gut flora whole cell lysates. Next, the reactivity of the inferred intermediate #2 in Figure 6 (HV00276) was determined with both anerobic and aerobic WCL in blue native PAGE (Figures 10A and B) and in SDS-PAGE (Figures 1 1 and 12). In blue native gel analysis, the 684-6B clone mAb reacted with a 520,000 Da molecule in both aerobic and anerobic gut samples (Figures 10 A and 10B). Moreover, mAb 276 also reacted with the 480KDa MW marker that is phycoerthryn (Figures 10A and 10B). Figures 1 1 and 12 show that under SDS-PAGE non-reducing (Figure 1 1 ) and reducing (Figure 12) conditions, strong bands are seen again at -520,000 Da. Also smaller band is seen at approx 60 and 50 Kd as well as in the native marker under reducing conditions (Figure 12). The native marker is again phycoerythrin (PE) showing polyreactivity against PE by the 684-6B clone mabs.

Importantly, the somatically mutated original 2F5 and 4E10 broad neutralizing antibodies also reacted with protein bands in gut flora WCL with 2F5 reacting with ~300,000Da molecule and ~80,000 Da molecules in aerobic WCL and 4E10 reacting with -80,000 and 100,000 Da molecules in aerobic WCL. In Figure 12 (SDS-PAGE under reducing conditions), it is seen that HV00276 . (intermediate 684-6 ab #2) binds to an ~520,000Da band in aerobic and anerobic WCL while 2F5 reacts with an -80,000 Da band and 4E10 with an approximately 60,000da band in aerobic WCL.

It has been shown previously that the broad neutralizing antibodies 2F5, 4E 10 andl b l 2 are polyreactive antibodies that bind to multiple host antigens. Thus, the question is, if the initial response to HIV is by a polyreactive antibody response, why are not polyreactive antibodies made that broadly neutralize? Two possibilities have been considered.

First, it has been shown that the germline of lbl2, 2F5 and 2G12 do not bind to HIV gpl20 or gp41 while the somatically hypermutated antibodies do bind (Xiao et al, Biochem. Biophys. Res. Commun. 390:404-9 (2009)). Thus, the notion is for many of the epitopes of broad neutralizing antibodies, the

immunogens the field has been using do not target the B cell receptors of the naive B cells they are targeting. The germline of the lbl2 has now been studied for lipid reactivity and for gut flora whole cell lysate activity and it has been found that, indeed, the germline lbl2 reactivity is negative to HIV gpl20

envelope while the reactivity of the somatically mutated lb 12 is very high to HIV gpl20 (Figure 13). In contrast, the reactivity of the germline of 1 b 12 is very high to cardiolipin while the somatically mutated polyreactive original lbl 2 mAb reactivity to cardiolipin is very low though not negative (Figure 13). Moreover, the germline of lb 12 is reactive as well with gut flora whole cell lysate, while the mature original somatically mutated 1 B12 mAb is only weakly reactive (Table 2).

Table 2. Reactivity of Broadly Neutralizing Monoclonal Antibodies 2F5, 4EI 0, 1612, and 2G 12 with Gut

Flora and Their Germline Antibodies With Gut Flora

MAb gp41 gp120 Anerobic Gut Aerobic Gut

Flora WCL Flora WCL

Reactivity in Luminex Units

1b12 original NA 5106 148 384 1 b12 germline NA <80 524 1 127

2F5 original 32717 9237 103 100 2F5 germline NA NA NA NA

4E10 original

4E10 germline NA NA NA NA

2612 original

2612 germline <80 <80 <80 <80

17b original 1433 <80 <80 <80 CCR5 binding site antibody Second, it has been hypothesized that the polyreactivity of 2F5, 4E10 and 1 bl 2 target the B cells making these types of antibodies for deletion or anergy (Haynes et al, Science 308: 1906-8 (2005); Haynes et al, Human Antibodies 14:59-67 (2005); Alam et al, J. Immunol. 178:4424-35 (2007)). This hypothesis has recently been proven for the 2F5 VH in 2F5 FH homozygous knock-in mice (Verkoczy et al, Proc. Natl. Acad. Sci. USA 107: 181-6 (2010)) and now in 4E10 VH homozygous mice (Figure 14). In both animal models of knock-in of the broadly reactive somatically mutated VHs, the mutated VHs are sufficiently autoreactive to cause deletion in the bone marrow and to invoke multiple tolerance mechanisms in the periphery.

In summary, the results described above demonstrate:

The initial antibody response to HIV is focused on non- neutralizing Env gp41 epitopes.

The initial gp41 antibody response arises from preexisting somatically mutated, polyreactive "natural" antibody clones whose germline Ab do not react with gp41 but whose inferred intermediate Abs do react with gp41.

While the antibody members of gp41 antibody-reactive clones are polyreactive and cross-react with lipids and other self cellular antigens, the affinity of anti-gp41 antibodies increases as somatic hypermutation occurs, indicating gp41 antigen drive.

Initial HIV-induce clonal development however is not efficient nor high affinity— perhaps due to self mimicry, leading to a mixture of HIV Env- reactive and non-reactive antibody clone members. • The germline of broad neutralizing antibodies lbl 2, 2F5 and 2G 12 do not appear to react with their inferred germline antibodies.

IgM antibodies isolated from AHI or uninfected subjects that bind to gp41 also bind to gut flora whereas gp41 negative IgMs do not bind gut flora antigens

The germline of lb 12 reacted with lipids and gut flora, implying origin from pre-existing polyreactive natural antibody producing na'ive B cells that likely originated from B cell clones originally targeted against gut flora.

The somatically mutated 2F5, 4E10 and lbl 2 broadly neutralizing antibodies all react with antigens in gut flora whole cell lysates, indicating that these antibodies likely derived from clones of na'ive B cells originally targeted to gut flora.


The enrichment and identification of a protein band in intestinal bacterial lysate reactive with mAb HV00276 is shown in Fig. 52. Western blot analysis following a Native PAGE gel run shows that mAb HV00276 binds to a -520 kDa protein band in an anaerobe and aerobe intestinal bacterial lysate. Protein fractions from the bacterial lysate having a molecular weight of ~500kDa were collected following size exclusion chromatography (SEC). SEC fractions show enrichment of the 520 kDa protein by Coomassie Blue (1), silver staining (2) and western blotting (3, arrow) with mAb HV00276. Isoelectric zoom fractionation shows migration of the mAb reactive protein to gel compartment A4 with pH6.2- 7. The 520 kDa band from the enriched fractions was subjected to LC-MS analysis for protein identification. RNA polymerase β, β' and a subunits were identified (see Fig. 53).

E.coli RNA polymerase core protein and holoenzyme (core protein+ σ subunit) (Epicentre Biotechnologies, Madison, WI) were run on a NativePAGE gel, and the reactivity of mAb HV00276 was detected using western blotting. Reactivity to both core and holoenzyme was detected indicating that mAb HV00276 binds to RNA polymerase core protein.

E.coli RNA polymerase core protein (Epicentre Biotechnologies,

Madison, WI) was run on a denaturing SDS-PAGE gel under both reducing (Red) and non-reducing (NR) conditions (left panel). On denaturing SDS-PAGE, the individual subunits (β, β', a and ω) of the core protein can be resolved and visualized following Coomassie Blue staining (right panel). Western blot analysis of the transferred gel shows that the 276 mAb binds only to the 37 kDa oc-subunit of the RNA polymerase core protein. No reactivity of HV00503 mAb, which was negative for intestinal bacterial lysate proteins, was observed with any of the core protein subunits.


To understand how self tolerance may influence protective humoral responses to HIV-1 , it is crucial to determine which self antigens are mimicked by HIV-1 epitopes and where/when these self antigens are exposed to T- and B lymphocytes. It is shown in Fig. 30 that monoclonal human antibodies specific for epitopes of the HIV-1 gp41 MPER also react with self-antigens present in acetone fixed mouse 3T3 cells. As shown in Fig. 31 , at least four discrete molecules can be immunoprecipitated from mouse 3T3 cells by biotinylated 2F5 antibody. The dominant species precipitated has an apparent molecular mass of approximately 50 - 54 kDa.

A conserved mammalian protein, KY U, carries the core 2F5 epitope and has a molecular mass of 51 kDa (Fig: 32). The 2F5 core epitope is present in the KYNU of many vertebrate species (Fig. 33) and is present in the conserved H3 domain of KYNU (Fig. 34). As shown in Fig. 35, the ELDKWA region is in a well-ordered alpha helix. The DKW motif is not surface-exposed.

Binding of the 2F5 antibody to human KYNU may require a distortion of the H3 domain, potentially resulting in a slowed K^n. As shown in Fig. 36, in H3, the D and W residues likely have exposed side chains but K is buried. The 2F5 antibody may necessarily "distort" the H3 helix to bind the ELDKWA epitope. Under physiological conditions, KYNU is thought to be a homodimer. The ELDKWA motif may be available to KYNU monomers but is unlikely to be accessible when KYNU forms dimers (Figs. 37 and 38).

Putative germline 2F5 antibodies also react with rhKYNU (Figs. 39 and 40). This is an important point in that it demonstrates that KYNU could be the original ligand of B cells that eventually produced the mutated, high affinity 2F5 antibody. As shown in Fig. 41 , /the 2F5 antibody avidly reacts with a peptide (DP178-Q16L) containing the 2F5 epitope whereas anti-KYNU antibody does not (see also Figs. 42 and 43).

13H 1 1 , a non-neutralizing mouse HIV-1 MPER monoclonal antibody that recognizes an epitope proximal to the 2F5 determinant, does not bind rhKYNU (Fig. 44). Fig. 45 provides a mapping of residues that distinguish the binding sites of 2F5 and 13H 1 1 monoclonal antibodies to the HIV-1 gp41 MPER. The data shown in Fig. 46 demonstrate competitive inhibition of 2F5 binding to rhKYNU by recombinant HIV-1 gpl40 env (JRFL), DP178-Q16L, and an irrelevant peptide antigen, R4A. JRFL recombinant HIV-1 gpl 40 comparably inhibits the binding of 2F5 to JRFL (homologous inhibition) and to rhKYNU (heterologous inhibition) (Fig. 47). The similarity of the inhibition curves indicates that a single, common epitope is responsible for 2F5 binding to both JRFL and rhKYNU.

As shown in Fig. 48, 2F5 monoclonal antibody binds both plate-bound and soluble rhKYNU comparably. Surface plasmon resonance studies demonstrate that both 2F5 and its unmutated precursors are capable of binding avidly to rhKYNU (Fig. 49). The slower Ko„ is consistent with the 2F5 antibodies distorting the native KYNU structure in order to achieve maximal interaction. Kon rates are very slow indicating that the bound KYNU interacts stably with all 2F5 types.

SPR binding analysis shows that the 2F5 mAb and its RUA (2F5-GL1 and 2F5-GL3) bind to KYNU (Fig. 50). Each of the antibodies was captured on a human anti-Fc immobilized sensor surface and soluble KYNY was injected at concentrations 50, 30, 20, and 0μg mL. Overlay of the binding curves show specific binding of KYNU to each antibody. Non-specific binding was measured using a control mAb ( Synagis, anti-RSV) which showed no binding to KYNU.


To determine whether 2F5 reactivity to fixed 3T3 cells could be inhibited by proteins/polypeptides containing the 2F5 MPER core epitope (ELDKWA), 2F5 monoclonal (10μg/ml) antibody was reacted with increasing molar concentrations of homologous (JRFL and DP I 78) or heterologous (R4A) inhibitors (1 hr, 25°C). These mixtures were subsequently added to

hydrated/blocked slides covered with methanol/acetone fixed 3T3 cells for (2 hr, 25 °C). Slides were rinsed and then washed overnight in 250 ml (PBS with 0.1 % Tween-20 and 0.5% BSA). Washed slides were overlayed with goat anti-human IgG-FITC (1 :400 in PBS with 0.1 % Tween-20 and 0.5% BSA). After 1 hr., slides were washed, coversliped in Fluoromount-G. Twenty-four hr. later, fluorescence images were acquired using a Zeiss Axiovert 200 confpcal microscope at 200x magnification and a fixed 300 msec exposure time.

Homologous inhibitors, the JRFL protein and, to a lesser extent, DP 178 polypeptide, inhibited 2F5 binding to 3T3 cells. An irrelevant polypeptide, R4A, showed no inhibition. (See Fig. 51.) These data demonstrate that a substantial amount of 2F5 reactivity to fixed 3T3 cells is determined by protein-protein interaction rather than un-specific lipid binding. Thus, proteins, like KYNU, may be primary autoligands for 2F5.


As described above, the present invention relates to a vaccine strategy that comprises administering HIV envelope proteins (peptides or polypeptides) to, first, target B cells that express unmutated ancestor antibodies that are able to give rise to broadly neutralizing matured antibodies and, then, drive maturation of the B cell clones toward the desired breadth of neutralization by boosting the B cells that are undergoing somatic maturation with selected HIV envelope proteins (peptides or polypeptides). The development of the strategy involved

reconstruction of this maturation pathway. Desired final (mature) antibodies were isolated from a patient who produces broadly neutralizing antibodies and the antibodies were characterized. The respective putative ancestral antibodies were inferred and expressed as real antibodies and a determination was made as to what they bind. The notion is that the B cells expressing unmutated "ancestral" and intermediate antibodies will affinity mature when triggered with the appropriate proteins (peptides or polypeptdes) to yield the broadly neutralizing antibody- secreting B cells observed in the patient. Selection and isolation of cross-clade neutralizing monoclonal antibodies CHOI, CHOI, CHO , CH04 and CH05

Approximately 30,000 memory B cells obtained from frozen PBMCs of subject 707-01 -021 -9 were screened and 28 cultures were found that neutralized >50% of CAP45 infectivity (Fig. 56). Monoclonal antibodies CHOI , CH02, CH03, CH04 and CH05 (CH01 -CH05) were isolated from four of these culture wells (1 -27-G2, 1 -27-G1 1 , 1 -19-Fl O and 1 - 19-B7) (Fig. 56).

Amplification and sequencing were carried out of the V-heavy and V-light chains obtained from the RNA-later-treated memory B cells frozen at the time of screening. Cultures 1 -27-G2 and 1 -19-F10 contained only one pair (3~20/κ3~20; CHOI and CH02 monoclonal antibodies, respectively), which indicates that the cultures were monoclonal and that the CHO I and CH02 are natural antibodies. Conversely, 1 -27-G1 1 and 1 -19-B7 contained multiple V-heavy and V-light chains, indicating that the cultures were oligoclonal.

To identify the natural pairs from these latter cultures, single-cell sorted memory B cells, collected at the time of initial screening, were amplified and sequenced. CH03 and CH04 (both 3~20/κ3~20) were natural pairs isolated from cultures 1 -27-G1 1 and 1 -19-B7, respectively.

Human B-cell hybridomas were generated from culture 1 -19-B7 by further expanding and cloning by sequential limiting dilutions the memory B cells for approximately 4 weeks. By this means, the CH04 natural antibody was obtained and CH05 was identified, which was produced by a lesser population of expanded memory B cells and expressed the same 3-20 V-heavy of CH04 but paired with a different 1~6 V-light chain.

The CHOI - CH03 monoclonal antibodies were obtained by transfecting the V-heavy and V-light pairs into 293T cells and expressed in an IgGl backbone as previously described (Liao et al, J Virol Methods. 158(1 -2): 171 -9 (2009)). Monoclonal antibodies CH04 and CH05 were instead purified from the hybridoma B cell lines.

These data demonstrate that the strategy allows quick identification of neutralizing monoclonal antibodies in approximately 2 weeks and production of natural monoclonal antibodies as early as one month. Furthermore, this method resolves the uncertainties of the classic phage display libraries related to the precise characterization of a monoclonal antibody being true to the natural antibodies that are represented in the in vivo repertoire. Finally, reported for the first time is the production of two natural human B-cell hybridomas that broadly neutralize HIV-1.

Genomic characterization of the CH01-CH05 antibodies

It was determined that the CHOI - CH05 antibodies are all member of the same clonal family based on the following factors: (1 ) V(D)J families; (2) length of the HCDR3; (3) nucleotide sequences of the HCDR3 region and of the «- insertions.

The analysis of the heavy chains showed that CHOI - CH05 are IgG l antibodies, sharing the same V 3-20* 1/J 2*01 rearrangement (Table 3). They also share the same D region which resulted from the D-D fusion of the 3~10* 1 and the 20F15*2/inv regions (Table 3). The HCDR3 is 26 amino acids long (Table 3). N-insertions were also of the same length and shared a nucleotide makeup compatible with the notion that CH01 -CH05 monoclonal antibodies are clonally related (Fig. 57A). The V-heavy sequences of CH04 and CH05 are identical (Fig. 57A), which suggests that the moment in which the V-light chain peripheral editing occurred was intercepted. Table 3 - Main characteristics of the CH01-CH05 VH and VL sequences

' V-heavy chain V-light chain

j HCDR3 Mutation LCDR3 Mutation

V D Isotype V J k/1

length rate* length rate

CHOI 3-20 3~3, 20F15/inv 2 26 0.120 IgGl 3-20 1 k 9 0.091

CH02 3-20 3~3, 20F15/inv 2 26 0.1 18 IgGl 3-20 1 k 9 0.1 16

CH03 3-20 3~3, 20F15/inv 2 26 . 0.152 IgGl 3-20 1 k 9 0.138

CH04 3-20 3~3, 20F15/inv 2 26 0.153 IgGl 3-20 1 k 9 0.1 10

CH05 3-20 3~3, 20F15/inv 2 26 0.153 IgGl 1-6 2 k 9 -

* Mutation rates are calculated from putative reverted unmutated ancestor variable heavy and variable light chains inferred from the sequences of each individual monoclonal antibody independently.

Seemingly to the V-heavy chains, CHOI - CH04 shared the same

VLK3~20/JLK1 rearrangement (Fig. 57B), an LCDR3 of the same length (9 aminoacids) and similar ^-insertions (Fig. 57B). The V-light chain of monoclonal antibody CH05 was instead unrelated (Fig. 57C), with a different VLK1/JLK2 rearrangment, LCDR3 length and M-insertions. It is contemplated that the biology underlying the pairing of the V-light chains to the VH3-20 chain is that the VH3~20/VLK3~20 chain pairs (CHOI - CH04) preceded the VH3~20/VLK1~6 pairing (CH05) because higher VL numbers are closer to the JK locus and, therefore, ancestor antibodies would have had to rearrange VLK3 first and then VLKI . Furthermore, the low-numbered JK loci have to come before the high- numbered. Therefore, the transition from VLK3/JLK1 to VLK1/JLK2 is consistent with simple editing. Finally, the phylogenetic tree shown in Fig. 58, and discussed below, provides further very strong evidence that the VLK3/JLK2 rearrangement happened first.

Next, a determination was made of the genetic relationship of the CHOI - CH05 monoclonal antibodies by constructing the phylogenetic tree of the V- heavy chains (Fig. 58). To do so, the putative reverted unmutated ancestors of the CHOI - CH05 antibodies were inferred by applying the maximum likelihood analysis on the observed antibodies as a whole. Using this method, two possible RUAs (0219-RUA1 and 0219-RUA2) were predicted that differed only for a single silent nucleotide substitution (G or T) in position 329 (Fig. 59). The putative RUAs were also predicted by analyzing each observed monoclonal antibody independently. With this method, 9 RUA antibody candidates were identified: one for CHOI (CH01 -RUA1 ), two for CH02 (CH02-RUA1 and CH02- RUA2), four for CH03 (CH03-RUA1 , CH03-RUA2, CH03-RUA3 and CH03- RUA4) and two for CH04 (CH04-RUA1 and CH04-RUA2). The alignment of all the computed putative RUAs is shown in Fig. 59. The phylogenetic tree of the V-heavy chains (Fig. 58) shows that CH02 and CH03 are genetically close to each other and that CH03 is the most somatically mutated monoclonal antibody of the family.

Taken together, these data demonstrate that CHOI - CH05 are clonally- related heavily somatically mutated monoclonal antibodies that share a long HCDR3 and harbor a D-D fusion rearrangement. Moreover, this is the first description of peripheral light chain editing in humans.

CHOI - CH05 monoclonal antibodies broadly neutralize tier 2 HIV-1 isolates and bind to a limited set of monomeric gpl20/gpl40 HIV-1 envelope proteins.

The neutralization breadth of the CH01 -CH05 antibodies was tested against a panel of 96 HIV-1 primary isolates. The panel comprised 4 tier 1 A isolates, 3 tier IB isolates (2 clade B and 1 clade AE) and 89 tier 2 isolates which included 10 clade A, 21 clade B, 27 clade C, 4 clade D, 7 clade G, 1 clade AE, 1 clade AD, 9 CRFOl _AE and 9 CRF02_AG viruses.

As predicted by the genetic analysis, CHOI - CH05 shared a very similar pattern of neutralization (Table 4). All the antibodies neutralized viruses from multiple clades and the breadth of neutralization ranged from 44.9% (43/96 isolates) of CHOI to 34.7% (33/95 isolates) of CH02. CH03, CH04 and CH05 neutralized 43.2% (41/95), 43.2% (41/95) and 44.2% (42/95) isolates, respectively. None of the antibodies neutralized tier 1 A isolates. Tier I B isolates were neutralized only by CHOI (2 out of 3), CH02 and CH03 ( 1 out of 3) but not by CH04 or CH05. Conversely, CH01 -CH05 showed larger breadth of neutralization against tier 2 viruses. CHOI preferentially neutralized CRF02_AG isolates (7/9; 77.8%), followed by clade A (7/10; 70%), CRF01_AE (5/9; 55.6%), clade B (9/21 ; 42.9%), clade C (1 1/27; 40.7%), and clade G (1 /7; 14.3%) isolates. Clade D viruses were not neutralized. Conversely, it is important to note that the CH01 -CH05 monoclonal antibodies strongly neutralized AE.CM244.ecl (Table 4). The preferential neutralization of tier 2 viruses over tier 1 viruses is important in that previous work demonstrated that broad neutralization of easy-to- neutralize tier 1 isolates does not translate into breadth against more difficult-to- neutralize tier 2 isolates and, therefore, those kinds of antibodies could be of limited help in preventing or controlling HIV-1 infection.

(Continued) Virus Clade Tier Fiebig

CAP45.2.00.G3 C 2 6

Dul72.17 c 2 6

ZM214M.PL15 c 2 5

ZM233 .PB6 c 2 5

Z 249M.PL1 c 2 2

ZM53M.PB12 c 2 5

ZM109F.PB4 c 2 5

ZM135 .PL10a c 2 5

CAP210.2.OO.E8 c 2 4

Cel086_B2 C (T/F) 2 1

Ce0393_C3 C(T/F) 2 4

Cell76_A3 C (T/F) 2 1

Cell72_Hl C(T/F) 2 1

Ce2010_F5 C(T/F) 2 4

Ce0682_E4 C(T/F) 2 1

Ce2060_G9 C (T/F) 2 1

Ce703010054_2A2 C(T/F) 2 5

BF1266.431a C(T/F) 2 1

246F C1G C (T/F) 2 2

249M BIO C(T/F) 2 4

ZM247vl(Rev-) C (T/F) 2 2

7030102001E5(Rev-) C (T/F) 2 1

1394C9Gl(Rev-) C(T/F) 2 1

Ce704809221_llB3 C(T/F) 2 1

Q23.17 A 2 6

Q259.d2.17 A 2 5

Q769.d22 A 2 5

Q842.dl2 A 2 5

Q461.e2 AD 2 5

191955_A11 A (T/F) 2 4

191084 B7-19 A (T/F) 2 4

(Continued) In comparison, the recently described PG9 and PG16 quaternary antibodies, shown in the table, neutralized 73/83 (88%) and 69/83 (83.1 %) tier 2 isolates, respectively. Interestingly, with only one exception (T251 -18), PG16 neutralizes a subset of the isolates neutralized by PG 9 and the CH01 -CH05 broadly neutralizing antibodies neutralize a subset of viruses neutralized by PG16. This finding is compatible with the hypothesis that the CH01 -CH05 epitope is related to that of PG9/PG16.

Next, the potency of the CH01 -CH05 antibodies against the neutralization- sensitive isolates was evaluated. Overall, the median IC50 was approximately 1 μg/ml with an average 1C50 ranging from 2.4 to 5.6 μg/ml. CH03 showed the strongest potency among the CH01 -CH05 antibodies with a mean IC50 of 2.4 μg/ml and a median IC50 of 0.46 μg/ml, comparable to those of PG9 (mean IC50 = 2.1 μg/ml; median IC50 = 0.1 1 μg/ml) but weaker than those of PG16 (mean = 0.67 μg/ml; median <0.02 μg/ml). CHOI , the broadest neutralizer, showed a mean and median IC50s of 3.7 and 1 .1 g ml, respectively. CH02, CH04 and CH05 mean lC50s were 4.9, 4.7 and 4.3 μ^ηιΐ, and median IC50s were 0.97, 0.8 and 0.79 μ§ ηι1, respectively.

The ability to neutralize transmitted founder viruses is another critical parameter to evaluate. As shown in Table 4, CH01 -CH05 bNabs were able to neutralize 3/3 (100%) clade A, 2/9 (22.2%) clade B and 2/3 (66.7%) clade C transmitted founder viruses.

Taken together, these data indicate that the clonal family of CH01 -CH05 antibodies broadly neutralize tier 2 isolate from multiple clades, including transmitted founder viruses. This is the first report of a clonal family of broadly neutralizing antibodies. Since there was no significant differences in the pattern of neutralization of CH05 compared to that of the other broadly neutralizing antibodies of the clonal family, these results also indicate that the edited VLK1 ~6 chain permitted the neutralization of the tested isolates at comparable levels to the VLK3~20 chain.

In contrast to the mature antibodies, the inferred putative RUAs did not show such breadth of neutralization. Yet, few isolates were potently neutralized. The neutralization profile of 6 inferred RUAs tested on a panel of 24 isolates is shown in Table 5. It is important to note that CH03-RUA1 , CH03-RUA4 and CH03-RUA3 neutralized AE. CM244.ecl isolate with IC50 of 4.45, 5.26 and 18.8 μ§/Γη1, respectively. Also, B.WIT04160.33 was potently neutralized by all the RUAs tested (IC50s from 0.06 to 0.47 μ^ιηΐ). A.Q23.17 isolate was also neutralized very potently by CH01 -RUA1 , CH03-RUA1 , CH03-RUA3 and CH03-RUA4 with IC50s <0.02 μ^πιΐ. Conversely, CH02-RUA1 and CH03- RUA2 neutralized A.Q23.17 at IC50s three orders of magnitude higher, showing the same pattern of neutralization of C.ZM233M.PB6.


LEGEND (IC50 ug/ml) Binding of CH0J-CH05 antibodies to monomeric gpl 20/gpl40 HIV-1 envelopes.

To determine which monomeric envelope could be used in a vaccine formulation to bind to B cells and trigger the production of CHOI -CH05-like antibodies, CH01 -CH05 monoclonal antibodies and RUAs were tested for binding to a panel of 32 monomeric envelopes. Table 6 shows the EC50s expressed in μΜ.

Table 6 - CH01-CH05 ELISA binding to monomel ic gpl20/gpl40 envelope proteins

Source Clade Env En v Name CH01 CH02 CH03 CH04 CH05 synagis

Chronic A gp140 OOM SA 4076 NB NB NB NB NB NB NB NB NB NB


Chronic Anc gp140 US-1* NB NB NB NB NB NB NB NB NB NB



Chronic C gp140 97CNGX2F 140 CF NB NB NB NB NB NB NB NB NB NB

Chronic C gp140 DU 123 NB NB NB NB NB NB NB NB NB NB

Chronic C gp140 CN54 NB NB NB NB NB NB NB NB NB NB

Chronic G gp140 HV 14000 (DRCBL) NB NB NB NB NB NB NB NB NB NB

Chronic B gp120 W61D NB NB NB NB NB NB NB NB NB NB

Chronic B gp120 MN NB NB NB NB NB NB NB NB NB NB

Chronic B gp120 VBD2" NB NB NB NB NB NB NB NB NB NB

Chronic E gp120 A244gr 7.8 150 34.5 23.1 28.7 >666.7 >666.7 NB NB NB

Chronic C gp120 Z 651 NB NB NB NB NB NB NB NB NB NB

Chronic AE gp120 CM 243 12.7 >666.7 97.3 >666.7 >666.7 NB NB NB NB NB

Consensu: A1.CON gp140 A1.con.env03 140 CF NB NB NB NB NB NB NB NB NB NB

Consensu! AECON gp140 HV 13700 (AEcon.env03 140 CF) NB NB NB NB NB NB NB NB NB NB

Consensu! B.CON gp140 B.con.env03 140 CF NB NB NB NB NB NB NB NB NB NB

Consensu! C.CON gp140 C.con.env03 140 CF NB NB NB NB NB NB NB NB NB NB

Consensu! M gp140 Con 6 140 CF NB NB NB NB NB NB NB NB NB NB

Consensu! M gp140 Con S 140 CFI NB NB ' NB NB . NB NB NB NB NB NB

T/F A gp140 HV13341 (0219) NB NB NB NB NB NB NB NB NB


T/F B gp140 HV00043 (63521 TC21 140C) NB NB NB NB NB NB NB NB NB NB

T/F B gp140 HV00044 (6240 TZ5 140C) NB NB NB NB NB NB NB NB NB NB

T/F B gp140 HV00045 (6235714 D3 140C) NB NB NB NB NB NB NB NB NB NB

T/F B gp140 HV00046 (902114 B2 140C) 63.2 240 NB NB NB NB NB NB NB NB

T/F B gp140 HV00049 (700010040 C9 140C) NB NB NB NB NB NB NB NB NB NB


T/F C gp140 HV00047 (089C 140C.) NB NB NB NB NB NB NB NB NB NB

T/F C gp140 HV00048 (1086C 140C) NB NB NB NB NB NB NB NB NB NB


' SIV gp140

" gp120 no MFER

RUA2 Binding to monomeric envelope was weak with the exception of gpl 20 A244gD+, which was bound by the CH01 -CH05 antibodies with EC50s ranging from 7.8μΜ (CHOI ) to 150μΜ (CH02). In addition, and of extreme relevance for the selective targeting of precursors of B cells capable of secreting broadly neutralizing antibodies, also two putative RUAs showed some binding (Table 6). The other HIV- 1 envelope that was bound by all the five mature antibodies was gpl 20 CM243, even though the mean EC50 was higher. The sequence of the A244 (CM244) Envelope is from McCutchan et al (AIDS Res. Hum. Retrovir. 8(1 1 ): 1887-1895 (1992)) with the exception of aa substitutions of L124P, N196S, 198E, A212P and D284 N. In addition, there is a 30AA sequence from the gD protein of herpes simplex virus KYALVDASLKMADPNRFRGKDLPVLDQ at the N-terminus of gp 120 A 144 (CM244). This sequences comprises the receptor binding sites of the gD protein required for HSV entry and infection (Yoon et al, J. Virol. 77:9221 (2003), Connolly et al, J. Virol. 79: 1282-1295 (2005),

Campadelli-Fiume et al, Rev. Med. Virol. 17: 313-326 (2007)). In the RV144 Thai vaccine trial where this A244 gpl20 was used as an immunogen, the subjects responded to the gD protein in both the MN gpl 20 and the A244 gpl 20 with both IgA (Fig. 60) and IgG (Fig. 61 ) gD antibodies. Fig. 62 shows that there are two potential sites of interest in the gD peptide that may mimic the alpha 4 beta 7 binding site of gpl 20 LPV and LDQ. Thus this raises three possibilities:

1 . Motif for gpl 20 binding to a4b7 is LDV and LDI


This raises the question whether antibodies to gD can block binding of HIV gp 120 to a4b7.

2. LDQ of HSV-gD is a receptor binding site for host cellular receptor heparan sulfate (Yoon et al, J.Virol. 77:9221 (2003)).

This raises the question whether antibodies to gD can block binding HIV Env to heparan sulfate. 3. The LDQ is also the receptor binding site for the second HSV receptor HVEM. The anti-HSV antibody response to LDQ could be protective against HSV (Yoon et al, J.Virol. 77:9221 (2003)).

Therefore, an anti-gD response could be protective for HIV by reducing active infection.

Lack of binding to most monomeric gpl 20/gpl40 envelopes indicates that CH01 -CH05 bind to a conformation-sensitive, quaternary antibody, preferentially expressed on trimeric envelopes. Similar findings have been reported for PG9 and PG 16 antibodies (Walker et al, Science 326(5950):285-9 (2009)).

Conversely, the strong binding to the A244gD+ gpl 20 envelopes strongly suggested that the co-expression of the HSV-1 glycoprotein D restored the functional epitope.

To investigate the role of HSV-1 glycoprotein D in enhancing the binding of the CH01 -CH05 antibodies and to detect binding of the RUAs to envelopes at levels that can be below the threshold of detection of standard ELlSAs but still physiologically relevant, the constant of dissociation (kd) of the CH01 -CH05 antibodies and RUAs to A244gD+ and A244gD" gpl 20 envelopes was measured using surface plasmon resonance (Table 7). A244gD+ consistently showed a kd at least an order of magnitude lower than A244gD" but, even more importantly, all the RUAs bound to A244gD+ gpl20 with kd's ranging from 790nM to 26.7 nM. The surface plasmon reasonance patterns for these data are shown in Figs. 63-66. Also seen in Fig. 66 is that the transmitted founder virus 6240 bound with sub- nanomolar Kd to PG9. Taken together these data demonstrate that the A244 gD+ envelope as well as the 6240 transmitted founder envelope were in a similar conformation as the gpl 20 found in the native Env trimer, and thus should be in the correct confirmations for use as immunogens. Table 7 - Constant of dissociation of the CH01-CH05 monoclonal antibodies and RUAs detected by surface plasnion resonance

Constant of dissociation Kd (nM)

Env Clade CHOI CH02 CH03 CH04 CH05 PG9 PG16 CH01-RUA1 CH02-RUA1 CH02-RUA2 CH03-RUA1

A244 gD+ gpl20 E 8.8 450 15.6 25.5 26.1 4.8 450 26.7 790 410 83.5

A244 gD' gpl20 E 340 N/A 728 410 340 N/A N/A 2460 N/A N/A 270

63521 TC21 gpl40 B (T/F) N/A N/A N/A N/A N/A 0.75 160 N/A N/A N/A N/A

624008 TA5 gpl40 B (T/F) 300 N/A 1060 360 1450 210 110 N/A N/A N/A N/A

Autoreactivity and polyreactivity profile of CHOI - CH05 broadly neutralizing antibodies.

Table 8 shows that CH03 is autoreactive with RNP, histone and centromere B autoantigens. Presence of antibodies binding to centromere in CH03 was also found using indirect fluorescent antibody staining on HEp-2 cells (Fig. 67). Table 12 reports the binding (measured by Luminex assay) of CHOI - CH05 to 4 non-HIV antigens. The data show that CHOl -CH-3 are strongly polyreactive. CH04 and CH05 polyreactivity is still detectable even through at a much lower level. Conversely, PG9 and PG16 showed no polyreactive abilities. These data point out a potentially relevant difference on the biology of the respective developments between the CH01 -CH05 and PG9 and PG16 antibodies.

Table 8 - Autoreactivity (Athena)

Criteria for positive: >50


ug/ml SSA SSB Sm RNP Scl 70 Jo l dsDNA Cent B Histone

Neg Control - - - - - - - - - -

Pos Control 1 - 397

Pos Control 2 - 631 699 1073 441

Pos Control 3 - 544 458 402 575

4E10 50 306 254 9 20 3 156 19 31 333

25 247 206 7 15 4 138 8 19 .274

12.5 169 124 5 9 3 87 6 13 160

6.25 115 93 4 6 2 65 3 9 113

CHOI 50 8 5.5 4 8 4 3.5 32 22 26

25 6.5 5.5 4 5.5 3 2.5 17 14 17.5

12.5 5 5 3 4.5 2 2 9.5 10 13.5

6.25 5.5 5.5 2.5 4 1.5 2 6.5 7 10

CH02 50 5.5 4.5 3.5 12.5 2.5 2 16 12.5 15.5

25 6 4.5 3 9 2 1.5 10 9.5 11.5

12.5 5 3.5 2.5 6 1.5 1 5 7 8.5

6.25 5 5 2.5 5 2 1.5 3 6 7

CH03 50 9 7 24 132 12 5 0 98 844

25 6 4 10 70 7 4 38 34 386

12.5 6 6 9 74 8 4 0 30 359

6.25 5 5 7 51 4 3 0 19 231

CH05 50 27 13 2.5 9 2 9 5.5 14.5 21.5

25 12.5 5.5 3 4.5 1.5 3.5 5.5 9 14

12.5 8 5 3.5 3.5 1.5 2.5 1.5 6.5 11

6.25 7 6 2.5 3 1.5 2.5 2 6 10

CH04 50 19 11 3 9 3 5 7 18 26

25 11 8 2 6 2 3 6 13 16

12.5 7 4.5 2.5 4 2 2.5 2.5 8 11

6.25 6 6 3 3 1.5 2 5 6.5 9.5

PG9 50 5 5 2 3 2 1 2 5 7

25 5 6 3 2 1 2 1 5 9

12.5 5 4 3 2 2 2 3 4 6

6.25 4 4 2 3 2 2 3 4 5

PG16 - 50 4 2 0 2 1 0 1 3 2


12.5 4 4 2 3 1 1 5 4 8

6.25 5 5 2 2 2 2 6 5 8 Evaluation of the polyreactivity of the CH01 -CH05 antibodies measured by Luminex

Anaerobic Aerobic Gut Influenza HA

HCV E2 Gut Flora Flora (Wisconsin)

CHOI 17.5 196.7 19.2 0

CH02 333.3 258.8 28.5 0.3 .

CH03 2909.5 286.3 97.5 1

CH04 0.5 17 3.5 2.5

CH05 1.3 23.5 9 5.5

PG9 0 0 0 0

PG 16 0 0 0 0

4E 10 89.2 83 96.2 0

Synegis 0.5 2.5 1 .5 1

Results are expressed as background substracted RFUs using an antibody concentration of 50 Hg/ml

Characterization of the epitope targeted by the CHOI - CH05 broadly neutralizing antibodies. PG16-like phehotype

It was determined that the CH01 -CH05 antibodies share unique characteristics with the quaternary broadly neutralizing PG9 and PG16 antibodies recently described by Walker et al (PLoS Pathog. 6(8).pii:el 001028 (August 5, 2010)). In particular, the CH01 -CH05 bNabs were characterized as "PG-like" antibodies based on the following four criteria: (1 ) the point mutation of the asparagine into a lysine at position 160 (N160K) of the gpl 20 protein abrogates the neutralization of an otherwise neutralization-sensitive isolate, (2)

neutralization of otherwise neutralization-sensitive isolates is abrogated when the virus is partially deglycosilated due to its production in cells treated with the mannosidase I-inhibitor kifunensine, (3) the epitopes are preferentially displayed in the context of envelope trimers but are not found on monomelic gpl 20 or gpl 40 envelopes, and (4) threading shows a high similarity with PG9 or PG16 bNAbs. As a representative of the CH01-CH05 clonal family of bNabs, CHOI was tested to determine if it met all the four criteria.

Table 9 shows the effect of the 160 point mutation on the CHOI neutralizing activity (IC50 and IC80) compared to that of PG9 and PG16 against a panel of wild-type and mutated isolates: clade A Q23.17 and clade B JR-CSF JRFL and 7165.18 isolates. CHOI , PG9 and PG16 all strongly neutralize the wild-type Q23.17 (IC50s = 0.014, 0.002 and 0.001 μ§/ιη1, respectively) and JR- CSF (IC50s = 0.07, 0.003 and 0.003 μg ml, respectively) isolates. The introduction of the N160K mutation in the gpl20 protein of Q23.17 and JR-CSF equally leads to complete abrogation of neutralizing activity by the three antibodies (IC50 > 50 μg/ml). CHOI , PG9 and PG16 also share the same neutralization pattern against JRFL and its mutants. Neither of them neutralizes wild-type JRFL. A single mutation at position 168 (E168K) reconstitutes a properly conformed epitope and results in potent neutralization (IC50s = 0.044, 0.008 and 0.003 μg/ml for CHOI , PG9 and PG16, respectively) but the subsequent introduction of the N160 mutation reverts the effect of the E168 mutation, making the JRFL/E168K/N160K isolate neutralization resistant (IC50 > 50 μg/ml) to all the three bNabs. Finally, 7165.18 is neutralized by CHOI (IC50 = 5.82 μg/ml) and PG16 (IC50 = 1 1.8 μg/ml) but not PG9 (IC50 > 50 μg/ml) and, again, the N160K mutation abrogates neutralization by both CHOI and PG16. Taken together, these data indicate that the neutralization activity of CHOI is similarly affected by the signature N160K mutation in gp!20 as PG9 and PG16.

Table 9 - Effect of point mutations on sensitive glycosilation sites for PG9/PG16-like


* curve reached plateau at 78%. curve reached plateau at 50-55%. " curve reached plateau at 50-55%.

Another characteristic of PG9 and PG16 is that otherwise neutralization- sensitive viruses become resistant when 293T cells used to produce the virus are treated with kifunensine. Fig. 68 shows that CHOI neutralization of YU2

produced in293T cells is seemingly negated by treatment with 50μΜ of


Broadly neutralizing antibodies with a limited breadth of binding to

monomelic gpl 20 and gpl 40 envelopes described above is typical of quaternary antibodies, whose epitope is correctly exposed in the context of the trimeric


Superimposition of CHOI onto threads of 7 distinct monoclonal antibodies showed that the structure of PG16 was the best fit to predict the 3D conformation of HCOl (Table 10). Figure 69 shows the superimposition of CHOI onto the

PG 16 thread. PG9 and PG16 are characterized by a unique shape of the HCDR3 region that protrudes from the tip of the antibody structure in a "hammer-like"

shape (Pancera et al, J. Virol. 84(16):8098- 1 10 (2010)). No other antibody had

been previously described with such characteristics. Notably, CHOI structure is very similar and the "head" of the "hammer" superimposes well with that of

PG 16 (Fig. 69). Being the HCDR3 shorter than PG9 and PG16, the sequence differs in some parts and this might be the structural explanation of the different breadth of reactivities between the CH01 -CH05 antibodies and PG9/PG 16.

An interesting feature of quarternary antibodies is that they may be tyrosine sulfated in the same way as the CD4i antibodies (Huang et al, PNAS 10 (9):2706- 271 1 (2004) Epub 2004 February 23 and Pejchal et al, PNAS 1 07(25): 1 1483-8 (201 0)). Sequence analysis of CHOI performed with "sulfinator", a tyrosine sulfation prediction program, predicted one tyrosine that is likely to be sulfated (ARGTDYTIDDAGIHYQGSGTFWYFDL) (Table 1 1 ). (Note that CHOI is called 1 -27-G2.)

Taken together these data strongly support the notion that CH01 -CH05 bNabs are PG-like antibodies that recognize a quaternary epitope involving the V2 region of gpl 20.

In summary, the data presented above demonstrate: ( 1 ) a strategy has been developed that allows the rapid identification and isolation of natural antibodies without the need of generating phage display libraries; (2) a family of five clonally related broadly neutralizing antibodies has been described and their development tracked; (3) preliminary evidence of peripheral receptor editing in humans has been provided; (4) novel members of broadly neutralizing antibodies of the PG-like family have been described that are not genetically related to the previously described PG9 and PG 16 broadly neutralizing antibodies; and (5) a method has been developed to increase accuracy of predicting putative reverted unmutated ancestors when more than a single monoclonal antibody is available. For immunogen design for induction of quarternary V2, V3 antibodies , it is demonstrated in Example 5 that the gpl20 Env A244 with a herpes simplex gD sequence can both bind well to the V2,V3 conformational determinant broad neutralizing Abs PG9, PG16, CH01-CH05, and also bind to reverted unmutated ancestors of CHOI , 02 and 03 antibodies. Moreover, the 6240 transmitted founder Env can bind well to PG9, and PG16 mabs. Thus, a potent immunization regimen for induction of V2, V3 broad neutralizing antibodies is to prime several times (for example, from 1 -3) with the A244 gpl 20 envelope with the gD sequence at the N-terminus and then boost, for example, with the 6240 transmitted founder gpl40 (for example, from 1 -3 times) either systemically (e.g., IM or subcutaneously) or mucosally (e.g., intranasally, sublingualy, intravaginally or rectally) . Given the immunogenicity of the HSV receptor binding region in the A244 g l20, this construct containing the gD peptide can also be used for a HSV vaccine construct. Similarly, the gD peptide inserted at the N-terminus of any HlV-1 envelope in a similar manner can be used for inducing protective antibodies to herpes simplex virus types 1 and 2.

All documents and other information sources cited above are hereby incorporated in their entirety by reference.


1. A method of inducing the production in a subject of broadly neutralizing antibodies against HIV-1 comprising:

i) administering to said subject a non-HIV-1 antigen that binds to a germline B cell receptor, said non-HIV-1 antigen being administered in an amount and under conditions such that intermediate clones of B cells are produced that secrete antibodies that cross-react with HIV- 1 Env, and

ii) administering to said subject an HIV-1 antigen in an amount and under conditions such that na'ive B cells or said intermediate clones of B cells are produced that secrete said broadly neutralizing anti-HIV-1 antibodies.

2. The method according to claim 1 wherein said subject is a human.

3. The method according to claim 1 wherein said non-HIV-1 antigen is a lipid.

4. The method according to claim 3 wherein said lipid is cardiolipin,

phosphatidylserine, phosphatidylethanolamine, phosphatidylcholine, phosphotidylinositol, sphingomyelin, or derivative thereof.

5. The method according to claim 4 wherein said lipid is l-palmitoyl-2-oleoyl-SH- glycero-3-[phospho-L-serine] (POPS), l -palmitoyl-2-oleoyl-phosphatidylethanolamine (POPE), or dioleoyl phosphatidylethanolamine (DOPE).

6. The method according to claim 3 wherein said lipid is a hexagonal II phase of a phospholipid.

7. The method according to claim 1 wherein said non-HIV-1 antigen is

phycoerythrin (PE), C-phycocyanin (C-PC), apoferritin, or anerobic or aerobic gut flora or component thereof.

8. The method according to claim 1 wherein said non-HIV antigen comprises the a subunit of RNA polymerase core protein of a bacteria or eukaryote.

9. The method according to claim 1 wherein said non-HIV antigen is kynureninase ( YNU) or antigenic fragment thereof.

10. The method according to claim 9 wherein said KYNU is recombinant KYNU expressed in CHO or 293T cells, or antigenic fragment thereof.

1 1. The method according to claim 1 further comprising administering an adjuvant.

12. The method according to claim 1 1 wherein said adjuvant is squalene based adjuvant, a TRL agonist, an oligonucletide (oCpGs) or R848.

13. The method according to claim 1 wherein said HIV-1 antigen is a membrane- proximal external region (MPER) antigen, or variant thereof.

14. The method according to claim 13 wherein said HIV-1 antigen is an immunogen shown in Figure 16B, 16C, 17, 18, 20, 25 or 26.

15. The method according to claim 13 wherein said variant is a MPER epitope peptide with an L669S mutation.

16. The method according to claim 1 wherein said HIV-1 antigen is a transmitted founder HIV-1 Env, or antigenic fragment thereof.

17. The method according to claim 16 wherein said fragment comprises a portion of the CD4 binding site of gpl 20, an MPER sequence, or a portion of gpl20 comprising the V2 or V3 region of gpl 20.

18. The method according to claim 1 wherein the method is effected by administering to said subject a prime immunization comprising said non-HIV-1 antigen followed by one or more boosts comprising said HIV-1 antigen.

19. The method according to claim 1 wherein said non-HIV-1 antigen comprises a lipid, a component of anaerobic or aerobic gut flora bacteria, phycobiliprotein, or KYNU or fragment thereof, and said HIV-1 antigen comprises an HIV-1 Env antigen selected from the group consisting of transmitted founder Env 1086.C from Malawi, 089. C from Malawi, 040_C9 from the U.S. and 63521 from a Clade B acute HIV-1 infected U.S. patient.

20. The method according to claim 1 wherein said non-HIV antigen or said HIV antigen comprises a protein and a DNA sequence encoding said protein is administered to said subject under conditions such that said DNA sequence is expressed and said protein is thereby produced.

21. The method according to claim 1 wherein A244gD+ envelope is administered as a prime and an envelope bound by CHOI, CH02, CH03, CH04 or CH05 is administered as a boost.

22. The method according to claim 1 wherein said non-HIV-1 antigen is present in a liposome with said HIV-1 antigen and at least one adjuvant.

23. The method according to claim 1 wherein said non-HIV-1 antigen is conjugated to said HIV-1 antigen and formulated with one or more adjuvants.

24. An antibody selected from the group consisting of CHOI , CH02, CH03, CH04, and CH05, or antigen binding fragment thereof.

Sign in to the Lens