EP0973933A1 - Immunological tolerance to hiv epitopes - Google Patents

Abstract

Fusion immunoglobulin (fIg) proteins comprising one or more heterologous epitopes associated with a disease in which immune responsiveness is deleterious are useful to induce tolerance to these epitopes. HIV-1 gp120 epitopes linked in frame with an Ig heavy (H) chain are useful constructs for the induction of epitope-specific tolerance to HIV. Treatment of a subject with such a construct, or with lymphoid or hematopoietic cells expressing or secreting such fIg molecules induces specific immunological tolerance to those epitopes. Such tolerance, by preventing production of antibodies to selected gp120 epitopes, can prevent or inhibit 'bystander' apoptosis of uninfected host T cells which have bound the HIV gp120 protein to their surface CD4 molecules and are subsequently cross-linked by undesired anti-gp120 antibodies, thereby priming them for apoptisis in the presence of antigens which activate those T cells. gp120 epitopes corresponding to non-neutralizing B cell epitopes or certain T helper cell epitopes are preferred for producing the fIg molecules. In addition to fIg H chains and complete Ig molecules, DNA encoding such H chain and cells transformed with such DNA are provided.

Description

IMMUNOLOGICAL TOLERANCE TO HIV EPITOPES

Statement of Rights to Inventions Made Under Federally Sponsored Research

This invention was funded in part by a grant from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, which provides to the United States government certain rights in this invention.

BACKGROUND OF THE INVENTION

Field of the Invention

The invention in the fields of immunology, molecular biology and medicine relates to compositions, primarily fusion immunoglobulins, and methods useful for inducing a state of immunological tolerance to selected epitopes of human immunodeficiency virus (HIV) gpl20 or target epitopes assocaited with other diseases. Administration of these composition will induce and maintain tolerance to the epitopes in a subject infected with (or at high risk for) HIV, or in whom an immune response to a different target epitope is deleterious. Prevention of antibody responses to the selected HIV epitopes promotes survival of the host immune system and contributes to treatment of HIV disease. The compositions are also useful as adjuncts to HIV or other virus vaccines in modulating the immune response to maximize induction of protective anti-viral T cell immunity.

Description of the Background Art Immunological tolerance (hereinafter "tolerance"), the basis of the lack of reactivity of the immune system to self components, can also be induced artificially by a wide variety of manipulations. Hence, an animals can be rendered tolerant to antigens which are foreign. Autoimmunity is thought to result in part from the breakdown of tolerance to previously tolerated antigens. A variety of experimental procedures are known for inducing antigen-specific tolerance in neonates and adults (Billingham, R.E. et al. (1953) Nature 172:603-606; Chiller, J.M. et al. (1970) Proc. Natl. Acad. Sci. USA. 55:551-556; Borel, Y. et al. (1973) Science 752:76-78). In the immunocompetent adult, tolerance induction has been generally more difficult. Tolerance to foreign transplantation antigens or viral CTL epitopes, for example, was most effective in models where hematopoietic or lymphoid ablation was followed by reconstitution with antigen-expressing bone marrow (BM)-derived antigen-presenting cells (APC) (Ildstad, S.T. et al. (1984) Nature 307:168-170; Cobbold, S.P. et al. (1984) Nature 572:548-551; Roberts, J.L. et al. (1990) J. Exp. Med. 777:935-940; Oehen, S.V., et al. (1994) Cell. Immunol. 755:342-352; Nemazee, D. et al. (1989) Proc. Natl. Acad. Sci. USA 5(5:8039-8043).

For autoimmune diseases, studies have focused on the acquired induction of tolerance to autoantigens to prevent and/or ameliorate disease. For example, in murine models of multiple sclerosis or diabetes, prevention of disease has been accomplished with intrathymic, oral, or intravenous administration of , high doses of target autoantigens (Tisch, R. et al. (1993) Nature 366:12-15; Higgins, P.J. et al. (1988) J.Immunol.

140-A40-445; Critchfield, J.M., et al. (1994) Science 263:1139-1143).

One well-known way to induce tolerance is by attaching the antigenic determinant or epitope to be tolerated to isologous or heterologous immunoglobulin (Ig) molecules, primarily of the IgG isotype. Such molecules are termed "tolerogenic carriers" or "tolerogens" (Scott, D.W. (1979) Immunol. Rev. 43:241). Igs of different origin may vary in their persistence in an animal after administration and/or in the mechanism by which they induce tolerance. However, IgG carriers have been by far the most efficacious inducers in adult animals of tolerance to haptens, nucleosides and peptides (Borel, Y. (1980) Immunol. Rev. 50:11; Scott, D.W. (1976) Cell Immunol. 22:311). These carriers owe their superior tolerogenicity to their persistence in vivo and the ability of an epitope chemically attached to the IgG molecule to crosslink membrane IgM (mlgM) on the surface of B lymphocytes with surface Fc receptors. However, chemical coupling of epitopes to IgG carriers can be limi^Jed by the availability of free reactive amino groups, structural change of the epitope as a result of the coupling reaction, and the uncontrolled targeting of the added determinant to different portions of the IgG.

Protein engineering strategies have been used to create molecules containing heterologous epitopes for the amplification of specific immune responses. For example, heterologous oligopeptide epitopes of immunological interest have been inserted in-frame into bacterial flagellin (Newton, S. et al, (1989) Science 244:10-12; Jennings et al, (1989) Protein Eng. 2:365), influenza virus nucleoprotein (Chimini, G. et al. (1989) J. Exp. Med. 169:91-302), hepatitis B surface antigen (Rutgers et al, (1988) Bio/Technology 5:1065) and in the complementarity determining regions (CDR) of immunoglobulins (Billetta, R. et al, (1991) Proc. Natl. Acad. Sci. USA 55:4713-4717; Zanetti et al. (1992) Nature, 355:416; Zanetti et α/.WO90/090804);

Zaghouani, H.et al. (1993) Science 259:224-221; Zaghouani, H. et al, (1993) Int. Rev. Immunol. 70:265-278; Zaghouani, H. et al. (1995) Proc. Natl. Acad. Sci. USA 92: 631-635).

Attempts have been made to test the ability of such a recombinant protein to induce an enhanced immune response to the heterologous oligopeptide. A peptide immunoglobulin fusion Ig protein or referred to herein as a "fusion Ig" or "fig" has been used to induce immunity. For example, a fig was made which expressed in the CDR3 of its V_H region the repetitive tetrapeptide Asn-Ala-Asn-Pro (SEQ ID NO:l), designated (NANP)_n (in single letter amino acid code), of the circumsporozoite protein of Plasmodium falciparum, an etiologic agent of malaria (Billetta et al., supra). A monoclonal antibody (mAb) specific for (NANP)_n which was made against P. falciparum bound to the above fig and was blocked by a synthetic (NANP)₃ peptide. Immunization of rabbits and mice with the engineered fig in adjuvant elicited antibodies to the (NANP)₃ synthetic peptide and to P. falciparum parasite. Such antibodies efficiently inhibited the invasion of cultured liver cells by P. falciparum. Thus, immunity to malaria was induced in the absence of the parasite using antibody V regions engineered to mimic the parasite's molecular structure. The authors suggested that antibody (idiotype) mimicry of an exogenous antigen is possible and may only require a discrete stretch of identity for successful mimicry. An alternate and simpler explanation of these results by the present inventors is that this material, when administered in adjuvant, simply acted as an immunogenic hapten- carrier conjugate. C. Bona et al, (1994) Cell Mol. Biol. 40 Suppl 7:21-30 expressed viral epitopes on Ig molecules by replacing the D segment of a VJJ gene with a B cell epitope from the V3-loop of HIV- 1 envelope glycoprotein gpl20, a cytotoxic T lymphocyte (CTL)-epitope from influenza virus nucleoprotein or a T helper epitope from influenza hemagglutinin. The T cell-targeted peptides in the form of fig molecules produced by cells transfected with chimeric V genes, activated specific T cells. The authors speculated about possible practical applications for Ig molecules bearing foreign epitopes for the development of prophylactic and immunotherapeutic reagents.

It is noteworthy that Zanetti et al. (supra) and Bona et al. (supra) produced chimeric Ig molecules (which are figs as the term is used herein) for the purpose of immunization (vaccination), not tolerization. Although WO90/09084 casually preferred a speculative notion, lacking any particularity or evidence, that this type of construct could be used for tolerization, the authors provided no scientific basis for such a utility. In fact, the way in which their exogenous epitope was inserted into the Ig framework region resulted only in immunogenic, not tolerogenic, constructs. The Zanetti et al. reference therefore lacks any proof that its inventors were in possession of a tolerogenic fig and provides no enabling support for any tolerogenic molecule or preparation. Hence, the induction and maintenance of tolerance to oligopeptides presented to the immune system via engineered Ig proteins has not been demonstrated prior to the invention as described herein. In particular, the art has not seen the use of cells (expressing such fig molecules) as agents of epitope-specific tolerance induction or maintenance. The present invention is the first discovery of tolerogenic cellular engineering to achieve a meaningful effect with therapeutic utility.

In summary, the art recognizes that recombinant fusion proteins, including fig proteins, may be useful as immunogens to induce immune responses to the heterologous oligopeptide. However, there remains a recognized need to develop general and specific methods of inducing stable, long-lasting tolerance to any of a number of epitopes of clinical significance in a subject. Also needed are vectors that can introduce the target epitope to which tolerance is desired into a host cell or whole animal, such that the epitope (a) induces tolerance and (b) persists in vivo so that it maintains the tolerant state. It is essential that any tolerization protocol include a means to maintain the specific state of tolerance. Maintenance of tolerance is understood to require the persistence of the tolerogenic epitope in vivo (Smith, R.T. (1961) Adv. Immunol. 1:61; Golub, E.S. et al. (1967) J. Immunol. 99:6; Ramsdell, F. et al. (1992) Science 257:1130-1134).

The present inventors were the first to discover an approach that not only could induce tolerance to an oligopeptide presented to the immune system in the form of a recombinant fig protein but also could maintain a tolerant state in the subject.

See also Scott and Zambidis, co-pending, commonly assigned application U.S.S.N. 08/195,874 (allowed) and WO95/21926, which applications are hereby expressly incorporated by reference in their entirety.

One of the present inventors' central hypotheses for explaining the signalling process in tolerance is that crosslinking with anti-μ chain antibodies provides "signal

1" to B cells, which, in the absence of T cell help (signal 2), leads to anergy. At high concentrations of anti-μ, extensive crosslinking of IgM leads to a significant level of B-cell apoptosis because a greater proportion of the B cells are forced to exit the G₀ phase and enter the cell cycle. This effect can be mimicked by multivalent antigen in specific B cells (Carsetti, R. et al, (1993) Eur. J. Immunol. 25:168). A unifying explanation for various experimental results is that that multiple crosslinking events are necessary for the induction of apoptosis. (See, also, Warner, G. et al. (1991) Cell. Immunol, 138:404; Scott, D.W. et al. (1987) Immunol. Today, 5:105; Ales-Martinez, J.-E. et al. (1992) Sem. in Immunol. 4:195; Scott et al. (1996) Intern. Immunol. 9:1375-1385).

The Immune Response to HIV gp!20 and its Role in AIDS The immune response to HIV has been studied extensively. Early studies suggested a role for neutralizing antibodies in protection or containment of HIV infection. This is particularly true in the case of simian immunodeficiency virus (SIV), a relative of HIV, where a cloned virus could be employed (Burns, D. et al.

(1993) J. Virol. 67:4104). Neutralizing antibodies to the epitopes of the envelope glycoprotein g l20, especially the V3 loop, have been described in infected individuals. However, more recent evidence suggests that the antibody response to HIV may not be protective and may, in fact, contribute to the progression of disease (Fust, G. et al. (1995) Immunol. Today, 16:161; Wang, S. et al. (1994) Virology 199:241; Banda, N. et al. (1992) J. Exp. Med. 775:1099). Thus, while antibody responses against various epitopes of HIV clearly occur, the effectiveness and therapeutic significance of these responses is unclear.

The CD4 molecule on T lymphocyte serves as receptor for major histocompatibility complex (MHC) class II antigens and is referred to as "coreceptor" because its engagement synergizes with engagement of the T cell receptor for antigen (TCR) in activating the cells. When CD4 molecules were engaged by antibody independently of the TCR (in murine studies), the T cells were induced to undergo apoptosis (Wang, Z.Q. et al. (1994) Eur. J. Immunol. 24:1549-1552). Thus, besides functioning as a coreceptor with the TCR, CD4 has a function of its own in facilitating the induction of apoptosis. CD4 also serves as a cellular binding site or receptor for the HIV gpl20. In transgenic mice expressing a human CD4 transgene, appropriate crosslinking of gpl20 caused massive deletion of HIV-reactive T cells in vivo (Wang, Z.Q. et al. (1994) Europ. J. Immunol. 24:1553-1551). If T cells in which CD4 is engaged by anti-CD4 antibody administration are capable of expressing functional Fas protein on their surface, they degrade their DNA and disintegrate rapidly.

Antibodies to gpl20 can lead to enhancement of HIV entry into non-T cells via Fc receptors (Homsy, J. et al. (1989) Science 244:1357, supra). Uptake of complexes between HIV and anti gpl20 antibody by cells of the immune system, particularly monocytes, can result in establishment of a latent, subclinical infection and a virus reservoir susceptible to later activation(Kliks, S.C., (1993) Proc. Natl. Acad. Sci. USA 90:11518)). HIV-infected patient sera frequently contain antibodies against a peptide of the gpl20 C5 region which cross-react with HLA-C monomoφhic determinants (DeSantis, C. et al. (1993) J. Infec. Dis. 755:1396; Palker,

TJ. et al. (1987) Proc. Nat'lAcad. Sci. USA 84:2419). Though apparently not causing autoimmune damage, the antibodies are an example of non-protective antibodies produced during HIV disease. In an equine retroviral disease model, an antibody response to a variant virus (EIAV) may end in more extensive disease (Cook, R. et al. (1995) J. Virology 6 :1493). Moreover, production of non- neutralizing anti-HIV antibodies (specific for "irrelevant" or "ineffective" epitopes) may pre-empt the formation of antibodies to important, neutralizing epitopes.

In AIDS, shifts in cytokines from those produced by T_H2 cells to cytokines made by T_H1 cells were observed (Clerici et al. (1994) Proc. Natl. Acad. Sci. USA 97:11811; Gougeon, M-L. et al. (1993) Science. 268:1269; Ameisen, J-C. (1992)

Immunol. Today. 73:388). While controversial (Fauci, A. (1993) Science 262:1011), such shifts may contribute both to apoptosis and to hypergammaglobulinemia. The findings discussed above have led those developing the next generation of vaccines to re-direct the immune response in patients and to devise T-cell peptides that serve as stimulatory ("vaccine") and target epitopes for cytotoxic T lymphocytes (Salk, J. et al.

(1993) Science 260:1270; Cease, K.B et al. (1994) Ann. Rev. Immunology. 72:923. Evidence obtained in the last few years suggests that HIV may subvert the immune response through the interaction of viral gρl20 with the CD4 receptor on T cells. Observations from Finkel's laboratory (Finkel et al, supra; Banda et al, supra) and of Newell et al. (1990) Nature 347:286), indicate that crosslinking of CD4 on the

T cell surface may prime T cells for apoptosis, perhaps via the upregulation of the Fas molecule, CD95 (Oyaizu, N. et al. (1994) Blood 84:2622; Desbarats, J. et al. (1996) Proc. Natl. Acad. Sci. USA 95:11014-11018. Even picomolar concentrations of gpl20 could prime T cells for such activation-induced death. Apoptosis in normal, non-infected ("bystander") CD4⁺ T cells may be programmed by (1) allowing gpl20 proteins to bind to CD4 via their natural affinity, and then (2) adding anti-gpl20 antibodies to bind and crosslink the gpl20-CD4 complexes (Finkel et al, supra; Banda et al, supra). When such programmed or "primed" cells are triggered through their TCR, apoptosis follows. This is reminiscent of the increased rate of apoptosis observed in vitro in T cells from HIV- infected subjects (Gougeon, M-L. et al. (1993) Science. 268:1269; Ameisen, supra) and provides one explanation for CD4+ T cell depletion in AIDS: According to this view, concurrent infection by other organisms or any antigenic challenge for that matter would trigger the death of those T cells which bear a TCR recognizing these antigens and in which the CD4 molecules have been crosslinked via gpl20 and anti- gpl20. Such a mechanism may also explain (1) the enhancement of infection brought about by certain anti-HIV antibodies, and (2) the paradox that HIV appears to cause AIDS after the onset of antiviral immunity.

Ex vivo analysis of CD4⁺ T cells (without prior culture) indicates that the increased apoptosis representes a process ongoing in vivo. Varying numbers of CD4⁺ as well as CD8⁺ cells in lymph nodes (LN) of HIV infected subjects are undergoing apoptosis. In infected adults and children, a fall in the CD4/CD8 ratio correlates with increasing apoptosis of CD4⁺ cells which correlates with CD4 depletion and disease severity. In infected humans and monkeys, most apoptosis in the LNs was occurring in "bystander" (uninfected) cells. Among infected cells, those expressing lower levels of the HIV p24 protein showed higher levels of apoptosis than cells expressing higher levels of p24. Thus, the majority of apoptosis appears to takes place in HIV^{^} cells, and the majority of apoptotic cells are HIV- or HIV^l0 .

The mechanism for such T cell apoptosis has been suggested by Pahwa and colleagues (Oyaizu, N. et al. (1993) Blood 52:3392-3400) who examined apoptosis as a mechanism for CD4+ T cell depletion in HIN-1 infection. They showed that

(1) patient blood mononuclear cells underwent marked spontaneous apoptosis;

(2) stimulation of patient and normal T cells resulted in increased apoptosis; and

(3) cross-linking of CD4 molecules was sufficient to induce apoptosis in CD4+ T cells if cross-linking was performed in unfractionated blood mononuclear cells (but not in purified T cells). The accelerated cell death through apoptosis was concluded to play an important role in the pathogenesis of HIV- 1 infection, and crosslinking of CD4 in vivo contributed to this mechanism. Cross-linking of CD4 molecules, induced either by anti-CD4 monoclonal antibody (mAb) or by HIN-1 envelope protein gpl60 (which includes gpl20) upregulates Fas mRΝA and Fas antigen expression in normal lymphocytes (Oyaizu et al. (1994) surpra). Upregulation of Fas antigen closely correlated with apoptotic cell death. CD4 cross-linking resulted in the induction of interferon-γ (IFΝγ) and tumor necrosis factor-α (TΝF-α) in blood cells, both of which cytokines contributed to Fas upregulation. Anti-LFΝ-γ and anti-TΝF-α antibodies blocked crosslinking-induced Fas upregulation and lymphocyte apoptosis. Hence, aberrant cytokine secretion induced by the crosslinking of CD4 and the consequent upregulation of Fas antigen were concluded to play a critical role in triggering peripheral T cell apoptosis and thereby contribute to HIV disease pathogenesis. Cross-linking of the CD3 molecule (present on all T cells) caused an increase in the Fas ligand. This combination of increased expression of Fas and Fas ligand led to apoptosis.

Based on the foregoing, the present inventors have concluded that the antibody response to gp 120 in an infected subject is an important pathway leading to AIDS progression due to the pathogenetic component of CD4⁺ T cell depletion through bystander apoptosis as described above. Therefore, they have developed novel compositions and methods based on their general, flexible approach to the induction and maintenance of epitope-specific tolerance to eliminate virus-specific immune responsiveness. In particular T helper cell and/or antibody responsiveness to one or more epitopes of viral gpl20 is prevented or inhibited through the induction and maintenance of immune tolerance in T helper cells, B cells or both that are specific for one or a number of selected HIV gpl20 epitopes.

Furthermore, the present inventors have extended this approach to the induction of tolerance to any antigen, be it an autoantigen, an antigen of a microorganism or a tumor antigen, against which an undesired antibody response or T helper cell response occurs in a disease setting and is pathogenic or otherwise deleterious to the host.

Citation of the above documents is not intended as an admission that any of the foregoing is pertinent prior art. All statements as to the date or representation as to the contents of these documents is based on the information available to the applicant and does not constitute any admission as to the correctness of the dates or contents of these documents.

SUMMARY OF THE INVENTION The present inventors have devised novel fusion proteins and DNA constructs coding therefor. The fusion protein includes a desired peptide epitope or several epitopes, toward which immune tolerance is to be established, inserted in particular sites of the immunoglobulin ("Ig") heavy ("H") chain. This product is termed a "fusion immunoglobulin" and is abbreviated "fig" herein. A preferred fig includes an epitope or epitopes of HIV- 1, most preferably from the gpl20 glycoprotein of HIV- 1. To make this construct, DNA encoding the targeted epitope or epitopes is inserted "in frame" within a DNA construct encoding the Ig heavy (H) chain. Iftwo or more targeted epitopes are included, they exist as contiguous or non-contiguous sequences in the protein from which they are derived, and may be either linear or conformational epitopes.

This fusion protein construct is then transfected into a cell line, preferably a myeloma or other line of B lymphocyte lineage (such as a human cell line transformed by Epstein-Barr virus) that produces Ig light (L) chains but that cannot produce H chains due either to a spontaneous or induced mutation. When the transfected Ig H chains are synthesized, they combine naturally with the host cell's Ig L chains to form complete immunoglobulin molecules (H₂L₂) which are secreted. This resultant Ig fusion protein contains the desired target epitope (or epitopes) preferably in its N- terminal region and functions as a tolerogen for both B cells and T cells and induces tolerance in vivo. Transgenic mice producing such a fusion protein are highly tolerant immunologically to the epitopes included in the fig. The present inventors have found that Ig fusion proteins such as these can be presented to the immune system in a tolerogenic fashion, either as an fig preparation or in the form of transgenic hemopoietic precursor cells or B cells expressing the fig, to induce both B and T cell tolerance to the targeted HIV-1 gpl20 epitope..

The present inventors have conceived of an approach that is useful in producing improved and effective immunity against a virus, in particular, human immunodeficiency virus (HIV-1, HIV-2) by inducing tolerance to selected nonprotective viral epitopes as discussed above. Thought a peptide comprising the desired epitopes can be chemically attached to an autologous Ig carrier for tolerance induction, the present inventors have created a fig comprising one or more peptide epitopes and the Ig H chain using recombinant methods as described herein. The invention specifically involves:

1. engineered synthesis of peptide immunoglobulin fusion proteins that are highly tolerogenic.

2. cells transduced with DNA encoding such an engineered fig to induce, but more importantly, to maintain, a state of tolerance to the peptide epitopes.

The invention provides polynucleotides encoding the fig in the form of recombinant DNA molecules in vehicles such as plasmid and retroviral vectors, capable of expression in a desired eukaryotic host cell as disclosed herein. The invention also provides hosts transfected or transduced with the fig constructs which are capable of producing in culture or in vivo the fig molecules and secreting them or displaying them on the cell surface.

This invention is useful for the treatment of any disease in which immunologic reactions are pathologic. The best-known examples are in infectious and autoimmune diseases. In many types of infections, where the host response to the organism damages the host. For example in certain arenavirus infections (for example, lymphocyte choriomeningitis virus infection), the T cell response is responsible for as much or more pathology than the virus itself. Antibody responses and the interaction of the antibodies with complement is responsible for the hemorrhagic shock syndrome elicited by flaviviruses (in particular) dengue virus or arenaviruses, such as Junin virus which causes Argentinean hemorrhagic fever. In both the latter cases, an efficient immune response leads to disaster for the host. Other examples of diseases for which the present invention can be used include viral diseases wherein virus- antibody complexes damage the host. For example, infants congenitally infected with cytomegalovirus have such circulating complexes that are deposited in the kidney. Patients with hepatitis B virus infection have circulating complexes that result in arthritis and glomerulonephritis. Antibodies generated against a virus can also act as autoantibodies directed against normal tissues, even tissues not infected directly by the virus. An example of this is the polyendocrinopathies that develop in newborn animals infected with reovirus type 1 in whom antibodies against antigens in pancreatic islets, the anterior pituitary and the gastric mucosa have been observed. Such examples abound in the art and may be found in any comprehensive treatise on microbiology or infectious diseases.

More classical autoimmune diseases are also treatable by the present invention as either cell-mediated or antibody responses to organ-specific antigens or common or cross-reactive antigens are the known pathogenic agents. Specific tolerance induced by an fig of this invention is a promising therapeutic approach to the treatment of many types of autoimmune disease.

The invention provides an individual fig H chain or fig H chain dimers. Also provided by the invention is an fig molecule comprising (i) two different H chains, one of which is a fusion protein having one or more HIV gpl20 epitopes included in the V region, preferably at the N-terminus of a framework region, most preferably of the first framework region, and (ii) native L chains. Preferably, both H chains of the fig molecule are the fused H chains.

Specifically, the present invention is directed to a fusion immunoglobulin (fig) heavy (H) chain protein comprising a mammalian, preferably human, Ig H chain fused in frame after the leader in its N-terminal region to one or more HIV gpl20 epitopes, wherein the fig H chain is tolerogenic in a host with respect to the gpl20 epitopes. The tolerogenic epitope(s) is or are fused to the variable region of the Ig H chain, preferably at the N terminus of a framework region of the variable region. Most preferably the HIV gpl20 epitope or epitopes are fused to the N-terminal amino acid residue of the mammalian Ig H chain such that all amino acids encoding the gpl20 epitope or epitopes are N-terminal to the Ig-encoding amino acids.

Also provided is an intact fig protein comprising two Ig H chains and two Ig L chains, wherein at least one of the H chains is the fig H chain described above. Preferably, both of the H chains are the above fig H chains. A preferred Ig is one which fixes complement and has a longer serum half life. Thus, in a preferred embodiment, the fig H chain is an Ig γ chain, more preferably an Ig γ, γ₂ or γ₃ chain.

Most preferably, the Ig is human IgG and preferred fig iso types are IgG, IgG₂ and

IgG₃. In one embodiment of the above fig, the one or more gpl20 epitopes comprises a full length gpl20 protein In other embodiments, the gpl20 epitopes are one or more peptides selected from the group consisting of the Cl region, the V3 loop and the C5 region. In yet other preferred embodiments, the gpl20 epitope is a B cell epitope comprising a peptide selected from the group consisting of: VPVWKEATTTLFCASDAKAY (SEQ ID NO:2), EVHNVWATHACVPTD (SEQ ID NO:3), YDTEVHNVWA (SEQ ID NO:4), PQEWLVNVT (SEQ ID NO:5), PQEVVLVNVTENFDMWKNDM (SEQ ID NO:6), PNNNTRKSIR (SEQ ID NO:7), NNNTRKRIRIQRGPGR (SEQ ID NO:8), RKSIR (SEQ ID NO:9), IQRGPGRAFV (SEQ ID

NO: 10), GRAFVTIGKI (SEQ ID NO: 11), PGRAFY (SEQ ID NO: 12), NTRKSIRIQRGPGRAFVTIG (SEQ ID NO: 13),

PNNNTRKSIRIQRGPGRAFVTIGKIGNMRQAHC (SEQ ID NO: 14), NNTRKSIRIQRG (SEQ ID NO: 15), NKRKRIHIGPGRAFYTTKNIIGTIC (SEQ ID NO: 16), RKSIRIQRGPGRAFV (SEQ ID NO: 17), IRIQRGPGR (SEQ ID NO: 18),

KRIRIQRGPGRAFVTIG (SEQ ID NO: 19), QRGPGRAF (SEQ ID NO:20), RGPGRAFV (SEQ ID NO:21), RKRIHIGPGRAFYTT (SEQ ID NO:22), RGPGRAFVTIG (SEQ ID NO:23), SISGPGRAFYTG (SEQ ID NO:24), KRIHI (SEQ ID NO:25), KRIHIGP (SEQ ID NO:26), IHIGPGR (SEQ ID NO:27), HIGPGR (SEQ ID NO:28), HIGPGRA (SEQ ID NO:29), HIGP (SEQ ID NO:30), RIHIGPGRAFYTTG (SEQ ID NO:31), RIQRGPGRAF

(SEQ ID NO:32), IQRGPGRAFV (SEQ ID NO: 10), IQRGPGRAF (SEQ ID NO:33), IRIQRGPGRAFVTI (SEQ ID NO:34), RGPGRAFVTIGKIG (SEQ ID NO:35), QRGPGRA (SEQ ID NO:36), LXXGPGRA (SEQ ID NO:37), IGPGR (SEQ ID NO:38), GPGR (SEQ ID NO:39), GPXR (SEQ ID NO:40), GPGRAF (SEQ ID NO:41), RIHIG (SEQ ID NO:42), HIGPGRAF (SEQ ID NO:43), GRAF (SEQ ID NO:44), GGGDMRDNWRSELYKYKVVK

(SEQ ID NO:45), KYKVVKIEPLGVAPTKAKRR (SEQ ID NO:46), LGVAPTKAKR (SEQ ID NO:47), GGDMRDNWRSELYKYKWKI (SEQ ID NO:48), IEPLGVAPTK (SEQ ID NO:49), RRWQRE (SEQ ID NO:50), PTKAKRR (SEQ ID NO:51) and WQREKR (SEQ ID NO:52). In yet other preferred embodiments, the gpl20 epitope is a T helper cell epitope comprising a peptide selected from the group consisting of: EQLWVTVYYGVPV (SEQ ID NO:53), VYYGVPVWKEA (SEQ ID NO:54), GVPVWKEATTLFC (SEQ ID NO:55), AHKVWATHACV (SEQ ID NO:56), NVWATHACVPTD (SEQ ID NO:57), CVPTNPVPQEW (SEQ ID NO:58), VEQMHEDIISLW (SEQ ID NO:59), EQMHEDIISLWDQ (SEQ ID NO:60), EQMHEDIISLWDQSL (SEQ ID NO:61), HEDIISLWDQSLK (SEQ ID NO:62), VTVYYGVPVWKEATTTLFC (SEQ ID NO:63), WLVNVTENFNM (SEQ ID NO:64),

SLKPCVKLTPLCY (SEQ ID NO:65), CTRPNNNTRKSIRIQRGPG(Y) (SEQ ID NO:66), NTRKSIRIQRGPGR (SEQ ID NO:67), EQRGPGRAFVTIGKI (SEQ ID NO:68), RIQRGPGRAFVTIGK (SEQ ID NO:69), RIHIGPGRAFYTTKN (SEQ ID NO:70), GRAFVTIGKIGNMRQ (SEQ ID NO:71), QRGPGRAFVTIGKIGNMRQAH (SEQ ID NO:72), VGKAMYAPPISGQIR (SEQ ID NO:73), GNSNNESEIFRPGGG (SEQ ID NO:74),

FRPGGGDMRDNWRSEL (SEQ ID NO:75), DMRDNWRSELYKYKV (SEQ ID NO:76), RDNWRSELYKYKVVK (SEQ ID NO:77), CKYKVVKIEPLGVAPT (SEQ ID NO:78), YKYKVVKIEPLGVAP (SEQ ID NO: 79), KVVKIEPLGVAPTKAKRRVVQREKRC (SEQ ID NO: 80), ITLPCRIKQIINMWQEVGKAMYAPPISGQIRC (SEQ ID NO: 81), and ELYKYKVVKIEPLGVAPTKAKRRVVQREKR. (SEQ ID NO:82)

The present invention is further directed to a DNA molecule comprising a nucleotide sequence encoding any fusion Ig H chain as described above.

Also provided is an expression vector useful for producing the above fusion Ig product and for inducing and maintaining immunological tolerance to one or more epitopes of HIV gp 120 protein in a subject, preferably a human. The vector preferably comprises (a) a DNA molecule as above, operably linked to (b) transcriptional and translational control regions operable in a hematopoietic cell or lymphoid cell of the subject. The transcriptional and translational control regions provide for constitutive expression of the DNA sequence in a lymphoid cell or a hematopoietic cell. A preferred vector is a retroviral vector. A naked DNA vector may also be used.

The present invention also provides a hemopoietic or lymphoid cell transformed by a vector as above, which cell stably expresses the fig protein. Stable expression is expression which is not transient, and persists for weeks or even months, preferably for the in vivo lifespan of the cell in which the fig is expressed. Such a cell is preferably a human bone marrow cell, a resting B lymphocyte or an activated B lymphocyte which has been activated by a mitogen or other polyclonal B cell activator.

Also included is a method for producing the fusion Ig of the invetnion by culturing the above transformed cell. For in vitro production of the fig, however, any cell type may be used which can express an Ig H chain gene as well as the DNA encoding the fig and secrete it into the culture medium.

The present invention includes a pharmaceutical composition comprising:

(a) a tolerogenic amount of a fusion Ig molecule having a fusion Ig H chain as described above; and

(b) a pharmaceutically acceptable carrier or excipient for parenteral administration.

Preferably, in the pharmaceutical composition, the fig is an isologous IgG molecule. Also provided herein is a method for immunologically tolerizing a subject to one or more HIV gpl20 epitopes comprising administering to the subject an effective amount of a fusion Ig pharmaceutical composition as described above.

A method for immunologically tolerizing a subject to one or more HIV gρl20 epitopes comprising introducing into the subject an effective amount of transformed cells as described above, thereby tolerizing the subject. In another embodiment, the invention is directed to a method for immunologically tolerizing a subject to one or more HIV gpl20 epitopes comprising introducing into the subject an effective amount of transformed cells as above, thereby tolerizing the subject. Prior to introducing the transformed hemopoietic cells into the subject, the subject may be treated to diminish the host's hemopoietic cells, although this may not be necessary in a patient with AIDS. Tolerance may also be achieved by a combination of treatment with transformed cells and a pharmaceutical composition comprising fig as described above. This invention is also directed to a method of (i) inducing and (ii) maintaining immunological tolerance to an epitope or epitopes of HIV gpl20 protein in a subject, comprising:

(a) administering to the subject an effective amount of a fig pharmaceutical composition as described above to induce the tolerance to the epitope or epitopes; and

(b) administering to the subject an effective amount of transformed hemopoietic or lymphoid cells as described above to maintain the tolerance to the epitope or epitopes, thereby inducing and maintaining the tolerance. However, tolerance is also induced and maintained by means of administering the transformed hemopoetic or lymphoid cells without resorting to the fig itself. Thus, expression of the fig by the transformed cells is sufficient to accomplish the induction and the maintaining functions.

Also included is a method for identifying whether a candidate HIV gpl20 epitope or epitopes are tolerogenic in a first subject when presented to the subject in a fusion Ig molecule, comprising the steps of:

(a) producing a expression vector as above, wherein the gpl20 epitope or epitopes in the fusion Ig are the candidate epitopes;

(b) stably transforming a population of autologous or matched allogeneic cells of the subject with the vector;

(c) introducing the transformed cells into the subject; and

(d) determining whether the subject is tolerant to the candidate epitope or epitopes by

(i) immunizing the subject with the candidate epitope or epitopes in immunogenic form and measuring the immune response in vivo or in vitro, and (ii) comparing the response to an immune response in a second control subject similarly immunized which has not been treated with the transformed cells, such that, if the first subject is tolerant, the candidate epitope or epitopes in the fusion immunoglobulin construct is identified as being tolerogenic. In the above method, the subjects are preferably humans and the transformed cells are human cells.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 (sheets 1/1 through 1/3) shows the amino acid sequence of HIV- 1 gpl20 (mature protein). The boldface sequence, SEQ ID NO:83, (with position numbers in the right margin) is the consensus sequence of the protein from subtype B virus (the prevalent HJN-1 subtype in the United States). The following characters are used in the consensus sequence: ( 1 ) single letter code; UPPER CASE letter indicates that the amino acid residue in that position is conserved for all known viral isolates of subtype B;

(2) lower case letter indicates the amino acid residue is conserved in more than 50% of known isolates;

(3) a "?" indicates lack of consensus at that position (no single residue is found in the majority of isolates).

The concensus sequence is read left to right. Shown vertically below each position in the consensus sequence (where appropriate) are alternative amino residues that have been identified at that position in mutants or variants of subtype B. Residues which happen to be adjacent to one another below the consensus sequence line are NOT to be read left to right as they do NOT represent adjacent residues in an actual gpl20 sequence. (Note: all the variant residues below the consensus sequence line are UPPER CASE for clarity only.) All of the sequence information in Fig. 1 (and Fig. 2) was obtained from The Human Retroviruses and AIDS Genetic Sequence 1995 Compendium, published by the Los Alamos National Laboratory: Theoretical Biology and Biophysics Division, Los Alamos, NM.

Figure 2 (sheets 2/1 - 2/4) shows the aligned consensus sequences for the major subtypes or "clades" of HIV- 1 as published in 77 e Human Retroviruses and AIDS Genetic Sequence 1995 Compendium (see Figure 1). HIV subtypes are defined and distinguished based on their nucleotide (and not amino acid) sequences. Certain "signature sequences" are characteristic of a subtype, for example, the GPGR consensus sequence at the tip of the V3 loop of the subtype B that appears as a GPGQ consensus for most other subtypes. The letter/symbol conventions are the same as those used in the consensus sequence in Figure 1. In addition, the presence of a "." at a position indicates that most isolates (and hence, by definition, the consensus sequence for the subtype) lack an amino acid at that position. All sequences are shown relative to the consensus A sequence (SEQ ID NO:84), and a "-"indicates the same residue as in subtype A at that position. The gpl20 sequences in Figure 2 (SEQ ID NOS:84-104) include the signal peptide (N terminal to the mature protein and indicated above the CONSENSUS- A sequence). The mature protein begins to the right of the "/" Other landmarks indicated include the V3 neutralization loop, the C terminus of gpl20 (indicated by a "/" on sheet 2/4) and the N-terminal segment (about 16 residues) of the HIV-1 gp41 protein.

Figure 3 (sheets 3/1 and 3/2) shows the aligned amino acid sequences of gpl20 (including the signal sequence) from several strains or isolates of HJN-1. The top line of each grouping (in boldface) is the subtype B consensus sequence (SEQ ID NO: 105; also appearing in Figures 1 and 2). The footnotes describing each variant or isolate and the markings used in Figure 3 are as follows:

(1) The first 27 - 30 amino acids left of the "//" mark comprise the signal sequence of gpl20. The mature gp 120 protein begins to the right of the "//". In general, a space appears after each 10 residues. To preserve alignment, spaces have sometimes been omitted and for the consensus sequence, additional residues have been placed above the main sequence line.

(2) CON-B is the consensus sequence for gpl20 of subtype B (SEQ ID NO:105). UPPER or lower case letters are as described for Figs. 1 and 2. The presence of single letter amino acid codes or "?" above the consensus sequence line indicates the existence of additional residues in some subtype B isolates at approximately the positions indicated. In some locations, arrows appear in the sequence line as place indicators for such additional residues. Each arrow is not intended to correspond to a single residue and points to the known residues (usually "?") that may occupy that region in various isolates.. (3) BH10 isolate (SEQ ID NO: 106): Ratner, L. et al. Nature 3/3:277-284(1985) (Genbank

SWISS PROT Accession No. P03375)

(4) LAV-BRU isolate (SEQ ID NO:107): Wain-Hobson, S. et al, Cell 40:9-17(1985) (Genbank SWISS PROT Accession No. P03377 )

(5) ARV2/SF2 isolate (SEQ ID NO: 108): Sanchez-Pescador, R., et al. Science 227:484- 492(1985) (Genbank SWISS PROT Accession No. P03378) (6) MN isolate (SEQ ID NO:109): Gurgo, C. et al, Virology 7 4:531-536(1988) (Genbank SWISS PROT Accession No. P05877)

(7) 92US712.4 isolate (SEQ ID NO:l 10). This sample was part of a set of sequences generated through the NIAID/NIH DAIDS HIV variation program. The virus was derived from an asymptomatic individual from Baltimore, thought to be infected by parenteral i.v. drug user contact. The env sequence clustered with HIV-1 B subtype sequences. Gao, F. et al., J. Virol. 70:1651-1667 (1996) (Genbank SWISS PROT Accession Number U08449). This sequence was randomly chosen as a subtype B isolate for illustrative purposes and for comparison with the more common variant sequences.

Figures 4A and 4B illustrate a preferred engineering strategy for inserting a foreign epitope at the N-terminus of an IgG γ chain. Figure 4A depicts the incoφoration of an oligonucleotide (SEQ ID NO:l 11) encoding the λ phage Cl repressor peptide 12-26 (SEQ ID NO: 112) as described in Examples. This fig was expressed in murine J558 myeloma cells. The present invention introduces an oligonucleotide or polynucleotide encoding one or more native or synthetic gpl20 peptide epitopes into an Ig H chain, preferably a human γ chain (Figure 4B).

Figures 5 A and 5B show strategy for the construction, expression, and epitope recognition of a fusion Ig gene by inserting a foreign epitope into a V_H gene. Fig. 5 A presents a scheme for constructing the fig. A modified 12-26 nucleotide sequence was ligated into a Pstl site of a 1.3-kb murine V_H (LVDJ) chain fragment. The Pstl site appears at the coding sequence of the fifth amino acid of the FRl; therefore, a repeat of the first five FRl amino acids was designed to follow the coding sequence of the 15 amino acids of 12-26, so as not to perturb proper framework region folding after insertion. SDS/10% polyacrylamide gel electrophoresis of purified H-chain transfected immunoglobulins demonstrated proper assembly of H chains with L chains. Fig. 5B is a gel pattern showing recognition of epitopes by immunoblotting. Purified control IgG (P6) of 12-26-IgG (Q3) samples were electrophoresed on SDS/10% polyacrylamide gels, transferred onto nitrocellulose, and probed with antimouse IgGl (left lanes) or with biotinylated anti- 12-26 mAb B3.11 (right lanes) plus AP-conjugated secondary reagents.

Figure 6 shows in vivo effects of 12-26-IgG pretreatment on peptide-specific humoral immune responses. Male BALB/c mice were injected i.v. with a single 1-mg dose of deaggregated protein G-purified P6 ( ▲ ) or Q3 ( / ) IgG. Mice were immunized and total or isotypic anti-peptide IgG titers were quantitated by ELISA 8 days after a secondary antigenic boost. Isotyped anti-peptide titrations (IgG, and IgG_2b) represent mean absoφtion values from assays of three individual mice in each group. 0 , Preimmunization sera

Figure 7 shows in vivo effects of 12-26-IgG pretreatment on peptide-specific cellular immune responses. Tertiary cytokine (IL-2 and IL-4) responses of enriched splenic T cells (3 x 10⁶ cells per ml) from mice displaying tolerized humoral immune secondary responses are shown. IL-2 and IL-4 production in supernatants was determined in triplicate by CTLL and CT.4S assay, respectively. "Medium only" backgrounds were subtracted; these values ranged from 1 to 4 units/ml in all assays. Figure 8 shows structure and genomic Southern blotting of transgenic mice expressing 12-26-IgG specifically in the B-lymphocyte lineage. A murine IgG,^b H chain construct containing endogenous immunoglobulin promoter and enhancer (E_H) sequences was modified to express 12-26 peptide and a repeat of perturbed framework region sequence (FRl) at the N-terminus. Fertilized embryos were injected with this linearized construct and transgenic mice were generated via standard procedures. Genomic DNA from tail biopsies was digested with BamYil and EcoRI to release a 1.3 kb V_H fragment, fractionated on 0.8% agarose/TBΕ, and transferred onto nylon membranes via alkaline Southern transfer . Southern blots were probed with random- primed ³²P-labeled DNA sequence containing 3 tandem repeats of 12-26 nucleotide sequence. Densitometry studies using known amounts of purified, linearized transgene DNA was used to estimate that there are 2-3 integrated copies in Lines 5 and 17. Figures 9A and 9B show profound peptide-specific cellular and humoral immune tolerance in 12-26-IgG-expressing transgenic mouse lines. Fig. 9 A presents titers of total anti-peptide IgG (open symbols), or IgG, isotype (closed symbols) for Line 5 transgenic mice measured after peptide immunizations and secondary boosts. Fig. 9B presents splenic T cell cytokine responses from tolerant Line 5 trangenic (Tg) mice determined by CTLL assay. Error bars signify the standard error of the mean for 3-4 mice per group.

Figure 10 shows profound peptide-specific cellular and humoral immune tolerance in transgenic bone marrow chimeras expressing 12-26-IgG. Chimeras were prepared with 1:1 mixtures of Line 17 Tg and non- transgenic (NTg) bone marrow (/).

Antibody responses to peptide are shown. Anti-HEL specificity controls showed no differences between groups. Nonirradiated mice injected with saline (0) displayed immune responses similar to control chimeras reconstituted with 100% NTg bone marrow (0). Error bars signify standard error of the mean of 2-3 mice per group.

Figure 11 shows the induction of peptide-specific humoral immune tolerance in normal immunocompetent adults by intravenous injection of various preparations of 12-26-IgG-expressing lymphoid tissues. Normal, nonirradiated BALB/c males were injected iv with 4xl0⁷ sex-matched splenocytes, Percoll^®-gradient-purified (60- 10% fraction) resting B cells, 48-hour activated LPS blasts, or crude unfractionated bone marrow cells from Line 17 transgenic mice. Recipients were rested for 7-10 days before immunization with 50 μg peptide in CFA (SC base of tail). Mice were boosted with an additional 50 μg in saline 2 weeks later and serum antibody titers determined 8 days later.

Figures 12A and 12B present an analysis of B-cell tolerance induction in tolerized transgenic or normal adult subjects. Fig. 12A: Nontransgenic ( 0 ), Line 5 transgenic ( V ), or line 17 transgenic ( / ) mice were immunized intraperitoneally with 50 μg 12-26-HEL conjugate in CFA, and boosted with the same in saline 2 weeks later. Anti-peptide and anti-HEL (all >10⁵, not shown) titers were determined by ELISA as described in the text. Fig. 12B: Serum titers from adoptively transferred recipients boosted with 50 μg 12-26-HEL conjugate in IFA were similarly determined: BALB/c recipients were irradiated with 400 rads, and injected with 5 x 10⁷ splenocytes from Line 17 Tg-tolerized donors (closed circles, various sources of lymphoid tissue) or non-transgenic injected, non-tolerized donors (open diamonds). Splenic donors had been previously primed and boosted with 12-26 peptide and HEL (at different subcutaneous locations), and had previously displayed tolerance (experiment from Figure 11).

Figures 13A. 13B,and 13C summarize studies showing the induction of tolerance in previously-primed adult recipients by either resting, B cells, B cell blasts or chemically fixed B cells. BALB/c mice were immunized SC with 50 μg 12-26 peptide in CFA 1-2 weeks before iv injection of 4x107 Line 17 transgenic (!) or nontransgenic control ( 0 ) purified resting B cells (Fig. 13 A), LPS-activated B cell blasts (Fig. 13B), or carbodiimide- fixed B cells (Fig. 13C). The mice were challenged IP with 50 μg soluble peptide 1-2 weeks following tolerizing injections, and antibody titers (ELIS A) determined 8 days later. The graphs show peptide-specific total IgG or two IgG isotypes (IgG, and IgG_2b),

Figures 14 and 15 show B cell expression, epitope recognition, and direct antigenic presentation of retro virally-synthesized peptide-IgG. Fig. 14 shows the structure and proviral integration of murine Moloney leukemia retroviral construct MBAE.BAK. Ten μg of genomic DNA from transduced, G418-resistant (+) or control (-) A20 cells was digested with Sac I, fractionated on 0.8% agarose, Southern- blotted, and probed with 32p_ι_abeled DNA probe containing three tandem copies of 12-26 sequence. Sac I digestion releases an -5.1 kb proviral genome. Fig. 15 shows tissue expression of 12-26 mRNA in long-term (~3 months) recipients of gene- transferred ( + ) or mock-transduced ( — ) BM progenitors. RNA from bone marrow

(B), thymus (T), or spleen (S) was assayed by 12-26 sequence RT-PCR (25). One- tenth of each PCR reaction (except for A20 controls: 1/100th) was Southern-blotted and probed with a non-complementary 32p_ι_abeled 12-26-specific oligonucleotide.

Figure 16 shows the induction of peptide-specific cellular immune tolerance in adult bone marrow chimeras infused with peptide-Ig-expressing progenitor cells.

BALB/c mice were sublethally irradiated (600 rads) and injected iv with 1-2x10^ gene-modified or mock-transduced BM. Recipients were immunized with peptide in CFA 2 months post-infusion and draining LN cells were restimulated in vitro with dilutions of synthetic peptide and 25-50 μg/ml purified protein derivative (PPD, Connaught) in RPMI 1640 with 0.5% heat-inactivated autologous serum. Stimulation indices (SI) represent ratios of proliferation to medium alone backgrounds (5,609+2,271 cpm). IL-2 and IFN-γ were quantitated by CTLL and ELIS A assays, respectively (Gilbert, K.M., et al. (1994) J. Exp. Med. 779:249-258). Additional experiments also revealed a diminution of peptide-specific IL-4 release in LN of tolerized recipients. Error bars signify standard error of the mean for 3 individual mice per group. This experiment was done at least twice with 3-4 mice per group with similar results.

Figures 17 A, 17B and 17C show the induction of peptide-specific humoral immune tolerance in adult bone marrow chimeras infused with peptide-Ig-expressing progenitor cells. BALB/c mice were sublethally irradiated with either (A) 200 rads

(Fig. 17A) , or 600 rads (Fig. 17B,C) and infused with 1-2x10^ gene-transferred (triangles) or mock-transduced (circles) BM cells. Mice were primed and boosted for humoral responses either (Fig. 17 A) one month, or (Fig. 17B,C) 2 months post- infusion with synthetic 12-26 peptide, and HEL as a specificity control. Non- manipulated, immunized BALB/c always produced titers similar to recipients infused with mock-transduced BM cells (Fig. 17A, diamonds). Both total peptide-specific IgG (open symbols), or the main isotype IgGl (closed symbols) were diminished in all experiments. Normal recipients in Fig. 19B and 19C received either 5-FU- pretreated normal BALB/c BM or SCID/BALB/c BM cells. Flow cytometric analysis at the one month sacrifice time (of mice from Fig. 17 A) revealed comparable levels of CD4+ and Ig+ splenocytes in normal BALB/c reconstituted with either normal or SCID BM: (CD4+: 18-25%; Ig+: 40-65%). All experiments were done at least twice with 3-5 mice per group with similar results. Figures 18-19 show peripheral tolerance induction in immunocompetent adults with gene-transferred peripheral B cells expressing engineered peptide-Ig. Figure 18 shows humoral and cellular tolerance induction. Unirradiated mice were injected with

>lxlθ LPS blasts co-cultured with retro virus-producing F6P (+) or mock- transducing ψ-2 (-). One week later, mice were primed and boosted for humoral responses, and sacrificed 3 months later for analysis of splenic memory T cell responses. Cytokine release in individual splenic cultures was determined at 24 hours (IL-2) or 48 hours (IL-4); medium alone background values were less than 1-2 U/ml and were subtracted for clarity (ΔU/ml). Figure 19 shows persistence of gene- transferred B cells. Hybridomas generated from spleens of tolerized mice by PEG fusion of A20 cells with LPS-activated splenocytes (48 hours, 50 μg/ml LPS).

Hybridomas were selected in 1 mg/ml G418 and tested for their ability to activate T- cell hybrid 9C127 as above. Eight representative A20 hybridomas from each recipient (Mice #1-3) are shown.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The nucleotide and amino acid sequences of most known isolates of HIV- 1 are published in "The Human Retroviruses and AIDS Genetic Sequence 1995 Compendium" (hereinafter, "Compendium"), as well as earlier editions of the Compendium). This document is available in paper or electronic form from its publisher, the Los Alamos National Laboratory: Theoretical Biology and Biophysics

Division, Los Alamos, NM. The complete nucleotide and amino acid sequences of individual HIV isolates as well as consensus sequences for recognized HIV-1 subtypes or clades (A, B, C, D, E, F, G, O U, CPZ) have been published in this database for a number of years. The HIV Molecular Immunology Database 1995, Editors: B. Korber et al, Los Alamos National Laboratory: Theoretical Biology and

Biophysics, Los Alamos, NM, 1995, (referred to herein as the "HIVMID") provides T-cell epitope maps, alignments, and annotation (for T helpe epitopes and for CTL epitopes) , as well as a summary and map of linear B cell epitopes and monoclonal antibodies recognizing such epitopes. This application incoφorates by reference the latest Compendium and HIVMID, but is also intended to include updates containing sequences of additional viral isolates as they are added and published. The compendium (database) and HIVMID are publicly available on the World Wide Web at the address "http://hiv-web.lanl.gov".

In particular, the HF -1 env gene (which encodes the gpl60 precursor protein of both gp!20 and gp41 envelope proteins) consensus nucleotide sequences for 10 viral subtypes appear at pages I-A-358 to I-A-364 of the 1995 Compendium (NOV 1995). The env nucleotide sequences of individual viral isolates, grouped by subtypes and appearing along with the consensus sequence for the subtype, are at pages I-A- 174 to I-A-357 of the 1995 compendium. Crosslinking of CD4 molecules on human T cells either by (a) HIV-1 virions bound to CD4 via viral gpl20, or (b) anti-gpl20 antibodies crosslinking of soluble gpl20 bound to CD4, primes or programs the T cells for apoptosis, as described herein. Thus, an infected subject's antibody response to HIV-1, particularly to one or more epitopes of gpl20, contributes to the pathogenetic process by targeting bystander T cells to self-destruct. The inventors have discovered an approach to modulate these responses by inducing selective immunological tolerance either at the level of B cells, T helper cells or both, resulting in diminished antibody responses to one or more gpl20 epitopes. Inducing and maintaining such tolerance in a subject provides a therapeutic approach for treating HIV infection. Furthermore, such tolerance induction to selected epitopes can be used as part of a therapeutic or prophylactic vaccine approach. Thus, an improved HIV vaccine may include in addition to an HIV immunogenic preparation, a fig in accordance with this invention to reduce or prevent undesired antibodies.

The term "tolerant" or "tolerance" as used herein is defined functionally in terms of the immune response to an immunogenic challenge with an antigen. A subject is tolerant if his response to an immunogenic challenge is reduced by at least about 50%, more preferably at least about 80% relative to a non-tolerant control subject.

Tolerance may be manifest by reduced reactivity in vivo such as antibody formation or in vitro, for example, by reduced lymphocyte proliferation.

A "tolerogen" is a form of antigen which, when it encounters the immune system, induces a state of immunological tolerance or hyporesponsiveness or anergy in the host. Such a state is tested by subsequent immunization or challenge of lymphocytes in vitro with the specific antigen in immunogenic form. The term "immunogenic" with reference to an antigen or epitope is also a functional term which is dependent on the nature, form, dose and route of administration of the antigen (epitope) such that it has immunogenic properties, i.e., it induces immune reactivity resulting in antibodies or cellular immunity. Thus the same molecule, e.g., a protein, can be immunogenic or tolerogenic depending on it form (e.g., aggregated or deaggregated) dose or route of administration, all of which is well-known in the art. As described herein, antigens, including low molecular weight haptens, can be rendered non-immunogenic and even tolerogenic by coupling them to homologous immunoglobulin molecules. In fact, a key observation underlying this invention is that such "coupling" can be achieved by recombinant techniques in the form of a fig wherein a peptide epitope (or "antigenic determinant" or minimal antigenic structure) for which tolerance is desired is made part of the fig using methods described herein.

Because certain tolerogenic fig preparations, or cells expressing such an fig, can induce hyporesponsiveness in an already primed or immunized subject (see below), the present invention is useful as a therapeutic tolerogen, to curtail an ongoing immune response to a selected gρl20 epitope or epitopes during the course of HIV disease. In fact, that may be the more significant clinical utility of this invention. Alternatively, the fig tolerogen is to modulate the response to an HIV vaccine such that the subject immunized with the vaccine and treated with the tolerogen responds to particular desired viral epitopes (expressed by the vaccine) and is prevented (or suppressed) in his response to other selected epitopes (expressed by the tolerogen) By judicious selection of these HIV epitopes, most preferably gpl20 epitopes, it is possible to render a host selectively tolerant at the level of T helper cells, B cells, or both. In the case of HIV infection, this would inhibit or prevent the production of antibodies that are of no benefit (e.g., non-neutralizing), and more importantly, are harmful via mechanisms such as bystander apoptosis or enhancing antibodies which promote infection of host cells. On its face, it might appear counterintuitive to inhibit an immune response to a virus which one wishes to eradicate. However, given the differences between epitopes of HIV recognized by antibodies and by cytotoxic T lymphocytes (CTL) (see HIVMID). B cell and/or T helper cell tolerance to one or more (even all) epitopes of gpl20 molecule may still permit an effective CTL response against other (non-tolerizing) gpl20 epitopes or, importantly, other non- envelope HIV proteins which are known to be immunogenic. HIV T helper epitopes and CTL epitopes have been described in a number of publication, for example,

Berzofsky, J.A. (1995) Ann N Y Acad Sci, 754:161-168; Meister GE et al. (1995) Vaccine, 75:581-591; Cease KB et al. (1994) Annu Rev Immunol, 72:923-989; Shirai M; et al. (1994) J Immunol, 152:549-556; Ahlers JD et al. (1993) J Immunol, 150 5647-5665; Berzofsky JA, (1991) Biotechnol Ther, 2:123-135; De Groot AS, et al. (1991) J Infect Dis, 754:1058-1065; Berzofsky JA et al. (1991) J Clin Invest,

55:876-884; Clerici M et al. (1991) Eur J Immunol, 27:1345-1349 Palker TJ et al. (1989) J Immunol, 742:3612-3619, which references are hereby incoφorated by reference.

The present inventors have developed a flexible fusion protein approach for induction of unresponsiveness to defined B-cell and T-cell epitopes in vivo and in vitro. See, for example, Scott and Zambidis, co-pending application U.S.S.N. 08/195,874, PCT Publication WO 95/21926 and Zambidis, E.T. et al, (1996) Proc. Natl. Acad. Sci. USA 95:5019-5024, which references are hereby incoφorated by reference in their entirety. As described herein, this approach originally set forth for other antigens, is adapted for the production of compositions and methods useful for inducing unresponsiveness to one or more HIV gpl20 epitopes.

Epitope-specific tolerance is used to ablate undesired antibody responses while maintaining protective CTL responses. Hence, by inducing B cell tolerance and T helper cell tolerance to all gpl20 epitopes, either by use of a fig into which a complete gpl20 sequence or one or more partial gpl20 sequences have been inserted, or by using a mixture of fusion Ig's each including a subset of gpl20 epitopes, anti-gpl20 antibody responsiveness can be prevented or diminished. Because the CD8 arm of the immune response is not affected, protective antiviral cell-mediated immunity, in particular CTL responses to HIV epitopes, remains intact. IgG-gρl20 peptide fusion proteins are effective tolerogens which modulate anti-gpl20 responses. Furthermore, human hematopoietic precursor cells, whether from BM or other tissues, and their progeny B cells which express the IgG-gpl20 peptide fusion proteins are themselves tolerogenic agents which deliver or present on their surface the selected HIV peptides in tolerogenic form for induction and/or maintenance of the tolerant state. Thus, in one embodiment, the ongoing maintenance of tolerance is achieved by first transfecting bone marrow (BM) cells or peripheral hematopoietic stem cells from any tissue (for example, CD34⁺ peripheral blood stem cells in the human) with a DNA vector which includes a DNA sequence encoding a IgG-gpl20 fusion protein of the present invention. In another embodiment, the tolerogen is presented expressed in a myeloid cell (as determined in studies using SCID mouse BM). The B cell expressing the tolerogenic fig may be a resting B cell, an activated B cell or B cell blast, or a transformed B cell (e.g., leukemia or lymphoma) which has been appropriately attenuated to ablate its oncogenic potential for use in human subjects. Long-lasting, even permanent tolerance can be induced by grafting transfected BM stem cells or peripheral stem cells. This approach is described in more detail in Example IV.

Tolerogen Presentation

B cells are known to be capable of inducing tolerance by presentation of appropriate surface molecules in a tolerogenic fashion (Eynon, E.E. et al. (1992) J.

Exp. Med. 775:131, using human IgM and IgD; Fuchs, E. et al. (1992) Science

255:1156, for the H-Y antigen).

The present inventors discovered that resting B cells expressing a fig, after injection into a recipient subject, induce tolerance for natural epitope included in the fig, such as the phage λ 12-26 epitope. Larger blast cells induced by stimulating such

B cells with bacterial hpopolysaccharide (LPS) (termed "LPS blasts") also tolerize for this peptide. Activated B cells are better tolerogenic vehicles in primed recipients than resting B cells. This is in contrast to the observations of Yuschenkoff et al.

(supra) who found that activated B cells from mice transgenic for and expressing human μ chains lost the ability to tolerize. Transgenic lymphoma cells activated in this way did not induce tolerance but rather appeared to induce an immune response for the same epitope.

In murine studies, splenic B cells stimulated with LPS are infected with a retro virus construct containing the desired epitope. In a model system, the 12-26 IgG fig has been used successfully in this way. Such LPS blasts are tolerogenic for that epitope. Hybridomas produced from the splenic B cells expressing the fig also express the fig transgene. When transgenic BM expressing 12-26 fig or normal (control) BM is injected into recipient mice irradiated with 200R, and the animals are immunized with the peptide in immunogenic form (in adjuvant), the following results have been obtained:

(1) T cells in recipients of transgenic BM are tolerant, measured by T cell proliferation and production of cytokines (IL2, IL4, IFN-γ,. etc.).

(2) recipients are tolerant as far as making IgG antibodies to the peptide. Tolerance to a desired HIV gpl20 peptide epitope included in an fig construct is achieved using as a source of B cells expressing the fig on their surface any population of lymphocytes known to contain B cells or to differentiate into B cells. This may include an unfractionated population, a cell preparation enriched in B cells or their precursors, or a purified B cell population. Any conventional method for enriching or purifying B cells may be employed. Examples of tissue sources for B cells include BM, spleen, LN, peripheral blood or lymph. B cells may be resting or preferably are activated, for example, LPS blasts.

As described in the Examples, when normal murine spleen cells were first stimulated by LPS followed by infection with a retroviral vector carrying the fig transgene and then infused into normal recipients, followed by immunization, the following results were obtained: The T cell responsiveness to the 12-26 peptide showed decreased IL-2 and IL-4 production. Animals had a decreased antibody response to the peptide. The effect on the antibody response can be explained by the T helper cell compartment being tolerant, the B cells being tolerant, or more likely, both. Table I summarizes results in a primed subject: TABLE I

Fixed cells are treated with a carbodiimide. ~ ND=not determined ³ Recipients injected 10 days earlier with 12-26 in complete Freund's adjuvant. Without wishing to be bound by any mechanism, the present inventors propose two possible mechanisms to explain these results:

(1) Antigen presentation without "signal 2" (R. Schwartz (1989) Cell 57: 1073- 1081) results in anergy.

(2) LPS blasts may induced "propriocidal" cell death.. Lenardo and colleagues (Boehme S.A. et al. (1993) Eur. J. Immunol.

25:1552-1560; Boehme S.A. et al. (1993) Leukemia 7 (Suppl 2 :S45-S49; Critchfield, J.M. et al. (1995) Cell. Immunol. 750:71-78; Pelfrey, CM. et al. (1995) J. Immunol. 754:6191-6202) found that stimulated T cells (or T cell hybridomas) produced IL-2 but also underwent suicide termed "propriocidal death." This response is thought to be important for regulating an ongoing immune response wherein suicide of responding cells serves to bring the response to a timely termination.

Evaluation of Potentially Tolerogenic gp!20 Epitopes for Use in fig

Human γ globulin (HGG) (American Red Cross), a model tolerogenic carrier, is used as a carrier in these evaluations of a given peptide ("PEP") corresponding to one or a combination of epitopes of gpl20. MBS (w-maleimodobenzoyl-N- hydroxysuccinimide ester) is a preferred coupling agent because of ease of use and thiol-cleavability (i.e., to prepare control peptide-conjugates, as well as to determine conjugation ratios). A known antigen (hapten) may be used as a specificity control for tolerance, e.g., FITC-coupled HGG. In a standard protocol based upon long-term experience in the present inventors' laboratory (Scott, D.W. et al. (1979) Immunol. Rev. 43:241; Warner, G.L. et al. (1991) J. Immunol. 146:2185), murine spleen cells are cultured for 24 hours with increasing concentrations of PEP-HGG, FITC-HGG or anti-μ (positive control for tolerance); these cells are washed and then challenged with LPS in microculture for 4 days. ELIS As for IgM and IgG anti-PEP, anti-gpl20, anti- HGG and anti-FITC are then performed by standard methodology. This protocol allows for polyclonal stimulation that elicits measurable responses to all of these epitopes Once established, the evaluation can be performed in PEP-primed subjects to verify that tolerance induction can be achieved in secondary B cells (Linton PJ, et al. (1991) J. Immunol. 146:4099). It is also helpful to perform dose response studies using PEP-HGG conjugates, as well as free peptide, administered intravenously. For example, groups of 4-5 mice are injected intravenously with 0.1, 0.3 or 1 mg of PEP alone, PEP-HGG, or FITC- HGG as a specificity control. Four to seven days later, mice are challenged with gpl20 in complete Freund's adjuvant (CFA). Mice are bled on day -7 (before tolerance) and at 10 and 20 days after challenge; mice can then be boosted on day 20 and bled 7 days later to evaluate secondary IgG responsiveness. Heterologous IgG's are known to be tolerogenic in vivo at <10^"8 M (-0.1-1 mg/mouse). Peptides for inducing T cell tolerance are commonly administered at higher concentrations (approximately 10^"7M). It may also be advantageous to establish epitope density requirements for tolerance. Typically, hapten-protein ratios of 5-10 are used with Ig conjugates. It would be desirable to control coupling reactions to achieve molar ratios (PEP:HGG) of 2,4,8, and 16. Because the MBS cross-linker is cleavable, it is possible to quantitate ratios and create peptide-linker only controls. Primed recipients may require tolerogens with a higher epitope density. In the fig embodiment, higher epitope density is translated into inclusion of more copies of the DNA encoding the epitope, for example 2-10 copies, in the fig DNA construct if this is required to overcome a state of preexisting immunity in the subject.

Achievement of tolerance using the above chemical-coupling approaches along with determinations of optimal dose-response relationships and epitope densities will provide the basis and indicate the efficacy of the epitopes to be inserted into the recombinant fig fusion proteins for use as tolerogens.

Choice of HIV gp!20 Epitopes for B Cell and T Helper Cell Tolerance Induction

The tolerogenic IgG-gp-120 peptide fusion proteins may include one or more peptides of gpl20, including the full-length gpl20 protein. If more than one peptide epitope is present, the different peptides may be arranged in the fusion protein in the same order and in contiguous form as they are in the native gpl20 protein. Alternatively, the peptides may be "reshuffled" in the fusion protein. Furthermore, one or more of the gρl20 peptides may be present in the fusion protein in two or more copies, either alone or with another gpl20 peptide. In a preferred embodiment, the one or more epitopes selected for use in the tolerogenic fig is a linear epitope. However, as conformational epitopes become better defined, it will be possible to construct a fig having one or more epitopes which, in combination, yield the conformational determinant in the expressed fig. It is advantageous to use the largest fragment of the native gpl20 protein that

(a) can be fused with the Ig H chain while maintaining the required tertiary structure of the Ig portion of the fusion protein for tolerogenic activity and (b) can be accommodated by the vector used to transfer the fig-encoding DNA. The advantage lies in the fact that the appropriate epitopes of such a fig are selected by the host MHC proteins (of antigen-presenting cells or, in this case, tolerogen-presenting cells) for presentation and tolerance induction. In humans, this would obviate the need to select a priori those epitopes of gpl20 which would interact with the HLA-DR molecules of a given subject to yield an active tolerogen for that subject. As more information relating various HL-A types with HIV gpl20 epitopes becomes available, it will become easier to tailor smaller tolerogenic epitopes for a given subject. Approaches to accomplish this for T helper cells epitopes are already available through various computer based algorithms which are discussed in much more detail below.

Expression of the epitope on the fig can be tested using a conventional immunoassay with an antibody specific for the epitope (if it is a B cell epitope) or with lymphocyte proliferation or cytokine secretion assay (for a T helper cell epitope). Antibodies recognizing such epitopes are available, and T cells can be prepared in vitro or long-term T cells lines of the appropriate specificity are available or can be prepared using conventional methods. The Compendium, and in particular the HIVMID lists antibodies specific for each of the epitopes of Table II, for example. The antibodies may be rodent mAbs, human polyclonal or mAbs or hybrid antibodies generated from such human or rodent mAbs. Alternatively or additionally, the soluble fig can be administered in adjuvant to a host and tested for generation of peptide- specific T-cell responses in vivo, due to processing and presentation by endogenous APC, even in the context of an Ig scaffold (see Examples). A gpl20 epitope of the present invention, in particular a linear or "sequential" epitope, is preferably one comprising a "natural" sequence, defined as the sequence as it occurs in a consensus gpl20 sequence of a particular HIV subtype or a naturally occurring mutant thereof which has been isolated and characterized.

However, the epitope sequence may also be a variant of a natural sequence defined here as a sequence in which one or more amino acid residues has been replaced by a different residue, including substitutions not known to occur in natural viral isolates. The only condition is that the variant sequence maintain the secondary and tertiary structure needed to create the desired the tolerogenic epitope when expressed in a fig protein either in solution or on a cell surface. Hence, it is preferred that any variant maintain (a) the structure of the peptide backbone in the area of the substitution, for example, as a sheet or helical conformation, (b) the charge or hydrophobicity of the molecule at the substitution site, or (c) the bulk of the side chain.

For a detailed description of protein chemistry and structure, see Schulz, G.E. et al, Principles of Protein Structure, Springer- Verlag, New York, 1978, and

Creighton, T.E., Proteins: Structure and Molecular Properties, W.H. Freeman & Co., San Francisco, 1983, which are hereby incoφorated by reference. The types of substitutions which may be made in the gpl20 protein or peptide molecule of the present invention may be based on analysis of the frequencies of amino acid changes between a homologous protein of different species (e.g., Table 1-2 of Schulz et al. (supra) and Figure 3-9 of Creighton (supra). Base on such analysis, conservative substitutions are defined as exchanges within one of the following five groups:

1. Small aliphatic, nonpolar or slightly polar residues: Ala, Ser, Thr (Pro, Gly);

2. Polar, negatively charged residues and their amides: Asp, Asn, Glu, Gin; 3. Polar, positively charged residues: His, Arg, Lys;

4. Large aliphatic, nonpolar residues: Met, Leu, He, Val (Cys); and

5. Large aromatic residues: Phe, Tyr, Tφ.

The three amino acid residues in parentheses above have special roles in protein architecture. Gly is the only residue lacking a side chain and imparts flexibility to the chain. Pro, because of its unusual geometry, tightly constrains the chain. Cys participates in disulfide bond formation which is important in protein folding. Tyr, because of its hydrogen bonding potential, has some kinship with Ser, Thr, etc.

Most deletions and insertions, and substitutions according to the present invention are those which do not produce radical changes in the structural or immunological characteristics of the gpl20 protein or peptide molecule when expressed as part of a fig. However, when it is difficult to predict the exact effect of the substitution, deletion, or insertion in advance of doing so, one skilled in the art will appreciate that the effect will be evaluated by routine screening assays. For example, a variant typically is made by chemical synthesis or site-specific mutagenesis of the peptide-encoding nucleic acid, expression of the variant nucleic acid in an fig construct in recombinant cell culture, and, optionally, purification from the cell culture, for example, by immunoaffinity chromatography using an immobilized antibody specific for the natural (non-variant) epitope. The presence of the desired epitope can be readily ascertained by one skilled in the art using an antibody, for example in an immunoassay using a mAb the binding of which defines the epitope. A standard assay, for example immunofluorescence of flow cytometry may be used to detect the variant epitope on the surface of a cell. Alternatively, the presence of the desired epitope can be detected using a cellular assay, for example an assay which measures the stimulation of T lymphocytes to proliferate or to secrete cytokines. Such assays are well known in the art and are described in detail in the Examples below.

In preferred embodiments, a tolerogenic gpl20-IgG fusion Ig protein modulates the responsiveness of B and/or T cells to non-neutralizing gρl20 epitopes, for example in the Cl region or in the C5 region which contains HLA-cross-reactive

C5 epitopes.

The Cl region of gpl20 is noteworthy for its dominance in being a target for immune reactivity. The V3 loop, in particular the V3 region of the loop (see Tables II and III), is noteworthy as a target for neutralizing antibodies. The C5 region is noteworthy for its cross reactivity with HL-A molecules and its stimulation of autoimmunity. Hence one or more epitopes of one of more of these regions would be useful as tolerogenic epitopes. In particular any gpl20 epitope which stimulates autoimmunity or against which an autoimmune host response is directed (irrespective of mechanism of induction) is a preferred epitope for use in a tolerogenic fig of this invention.

Because the host immune system will process any administered fig for presentation to T lymphocytes in conjunction with the host's MHC glycoproteins on the antigen-presenting cells ("APC")(or more appropriately, "tolerogen-presenting" cells), it is preferably to include a peptide of sufficient length for binding to host MHC molecules and subsequent presentation. As shown below, in particular for T helper cell epitopes, the gpl20 peptide may be as short as about 6 amino acids. Generally larger peptides are preferred, including those with more than one gpl20 epitope in the fig. For example, about 10-20 amino acids, preferably about 10-40 amino acids, more preferably about 10-60 amino acids are included in the fig. This will allow the host cells to select the epitopes appropriate for the particular MHC type. Alternatively, as discussed below, T helper epitopes may be identified and selected using various published computer-based algorithms. It is preferable, though not required, to exclude cysteine from the tolerogenic fig because of the constraints this amino acid imposes on uncontrolled secondary structure. The present inventors have developed a model system which utilizes mice made transgenic for human CD4 which is used to screen fig constructs for their efficacy and utility in humans. Administration to these mice of gpl20 and anti-gpl20 or HIV virions or crosslinked gpl20 leads to sensitization for subsequent apoptosis. Once the animals have been primed in this way, apoptosis occurs "spontaneously" in response to environmental exposure to antigens which engage the TCR and trigger the apoptotic process. Specific antigens including peptides with defined epitopes are administered to more precisely activate T cell apoptosis. This can be evaluated by testing the animals for epitope-specific T cell unresponsiveness or hyporesponsiveness. The induction of B-cell tolerance and T helper cell tolerance to selected gpl20 epitopes can readily be tested in this model for its effect on the pathway of T-cell apoptosis.

The present inventors have utilized a peptide that contains both a T-cell and a B-cell epitope, created a fusion protein of this peptide with an IgG molecule serving as a "carrier" and have used it to induce epitope-specific T-cell and B-cell tolerance

(see Examples). This success indicates the utility of inducing epitope-specific B-cell tolerance to HIV peptides. Such tolerance is exploited to counteract the non- protective or even harmful antibody responses to certain gpl20 peptides (Homsy, J. et al. (1989) Science 244:1357; Finkel, T. et al. (1994) Curr. Opin. Immunol. 6:605; Fust, G. et al. (1995) Immunol. Today, 76:167), as described below.

In the Cl region of gpl20, both T-cell and B-cell epitopes have been shown to be immunodominant (Abacioglu, Y. et al. (1994) AIDS Res. Human Retrovirus. 70:371). Within Cl, these investigators defined boundaries, termed Cla and Clb. The Cla peptide epitope FNMWKND (corresponding to residues 63-69 of Figure 1, SEQ ID NO: 83) can be detected with monoclonal antibody B10. A preferred peptide has the structural motif similar to that used earlier to form a 14-mer with a C-terminal cysteine for coupling ease using MBS ( -maleimodobenzoyl-N-hydroxysuccinimide ester) AAAFNMWKNDGGGC (SEQ ID NO:l 13). This peptide can be chemically conjugated to HGG for evaluation in vivo and in vitro for tolerogenicity. 37

Table IV, above provides amino acids sequences of T helper cell epitopes of gpl20 that have been identified using either human or murine test systems and have been entered in the HIVMID published on the Los Alamos National Laboratory World Wide Web Site. Preferred fig constructs include one or more of the epitopes presented in Table IV linked to the N-terminus of an Ig H chain as was described above for B cell epitopes (e.g., those in Table III). However, the present inventors do not intend to be limited by this listing of sequences which are specifically based on the amino acids sequences of HIV subtype B viruses. The art permits identification of other epitopic sequences derived from other HIV subtypes (discussed above) as well as viral isolates or "quasi species" thereof.

TABLE II

Defined Regions Of _Spl20 For Use As Tolerogenic Epitopes

Region of gp 120 Residues (approximate)¹

Cl 1-95

Cl (subregion) 60-90

VI loop 101-127

V2 loop 144-166

C2 168-269-

C2 223-238

V3 loop 266-301

V3 region 274-288

C3 298-360

C3 (subregion) 331-348

V4 loop 355-388

C4 383-427

V5 428-442

C5 443-481 (C-terminus)

¹ See Figure 1 for consensus amino acid sequence of HIV- 1 subtype B gpl20 (and selected individual amino acid substitution variants in individual viral isolates).

See Figure 2 for consensus amino acid sequence of other HIV-1 subtypes 38

TABLE III HIV gp!20 B Cell Epitopes Identified by mAbs

Region or Virus Amino Acid SEQ

Positions Strain* Residues ID NO:

30-51 LAI ATEKLWNTVYYGVPVWKEATTT 114

31-50 LAI TEKLWVTVYYGVPVWKEATT 115

31-50 LAI GVPVWKEATT 116

42-61 LAI VPVWKEATTTLFCASDAKAY 2

64-78 IIIB EVHΝNWATHACVPTD 3

51-70 LAI YDTEVHΝNWA 4

81-90 LAI PQEVVLVΝNT 5

81-100 LAI PQEVVLVNVTENFDMWKNDM 6

89-103 IIIB PNPQEVVLVNVTENF 117

91-100 LAI ENFDMWKNDM 118

93-96/94-97 LAI/BH10 FNMW 119

94-99 BH10 FNMWKN 120

91-100 LAI ENFDMWKNDM 118

101-110 LAI VEQMHEDIIS 121

101-120 LAI VEQMHEDIISLWDQSLKPCV 122

311-321 HXB10 EQMHEDIISLWDQSLKPCVK 123

101-120 LAI LWDQSLKPCV 124

102-121 LAI EQMHEDIISLWDQSLKPCVK 123

114-123 IIIB MHEDIISLWD 125

122-141 LAI LTPLCVSLKCTDLKNDTNTN 126

162-169 HXB2 STSIRGKV 127

162-171 V2 BH10 STSIRGKVQ 128

170-180 BH10 QKEYAFFYKLD 129 or IIIB

172-191 HXB2 EYAFFYKLDIIPIDNDTTSY 130

162-181 BH10 STSIRGKVQKEYAFFYKLDI 131

172-181 HXB2 EYAFFYKLDI 132

221-220 LAI EPIPIHYCAPA 133

211-230 LAI EPIPIHYCAPAGFAILKCNN 134

222-231 LAI GFAILKCNNK 135

242-261 LAI RPVVSTQLLL 136

252-271 LAI RPVVSTQLLLNGSLAEEEVV 137

257-262 BH10 TQLLLN 138

257-263 BH10 TQLLLNG 139

262-281 LAI NGSLAEEEVVIRSVNFTDNA 140 087

39

Table III, cont.

Region or Virus Amino Acid SEQ

Positions Strain* Residues ID

NO:

261-280 LAI VIRSVNFTDN 141

299-304 IILB LNCTRP 142

299-304 IILB SVEΓNCTRPNNNTRKSI 143

299-308 IIIB PNNNTRKSIR 7

300-315 HXB10 NNNTRKRIRIQRGPGR 8

304-308 IIIB RKSIR 9

309-318/329-338

IIIB IQRGPGRAFV / AHCNISRAKW 144

314-323/494-503 infec GRAFVTIGKI / LGVAPTKAKR 145

316-322 infec PGRAFY 12

302-321 BH10 NTRKSIRIQRGPGRAFVTIG 13

306-338 BH10 PNNNTRKSIRIQRGPGRAFVTIGKIGNMRQAHC 14

307-318 IIIB NNTRKSIRIQRG 15

308-313 MN NKRKRIHIGPGRAFYTTKNIIGTIC 16

308-313 MN V3 tip

304-318 LAI RKSIRIQRGPGRAFV 17

299-304 IILB IRIQRGPGR 18

299-304 IIIB KRIRIQRGPGRAFVTIG 19

308-328 BRU QRGPGRAF 20

V3 BRU RGPGRAFV 21

V3 MN RKRIHIGPGRAFYTT 22

V3 infec -I — G-FY-T 146

311-321 HXB10 RGPGRAFVTIG 23

V3 infec SISGPGRAFYTG 24

V3 MN KRIHI 25

V3 infec IXIGPGR 147

V3 MN KRIHIGP 26

V3 MN IHIGPGR 27 (or infec)

V3 MN HIGPGR 28

V3 infec HIGPGRA 29

V3 infec RKRIHIGPGRAFYTT 22

V3 ? HIGP 30

311-324 MN RIHIGPGRAFYTTG 31

312-318 MN IXIGPGR 147

307-316 IIIB RIQRGPGRAF 32

307-316 IIIB IQRGPGRAFV 10 Table III, cont.

Region or Virus Amino Acid SEQ

Positions Strain* Residues ID

NO: 308-316 IIIB IQRGPGRAF 33

V3 IILB IRIQRGPGRAFVTI 34

V3 316-330 HXB2 RGPGRAFVTIGKIG 35

V3 ? QRGPGRA 36

V3 IIIB LXXGPGRA 37 V3 infec IGPGR 38

V3 MN GPGR 39

V3 MN GPXR 40

308-313 MN GPGRAF 41

V3 MN RIHIG 42 V3 MN HIGPGRAF 43

V3 IIIB GRAF 44

V3 IIIB RAF 148

361-380 LAI IFKQSSGGDPEIVTHSFNCGG 149

362-381 LAI FKQSSGGDPEIVTHSFNCGGE 150 380-393 LAI GEFFYCNSTQLFNS 151

C3 HIV2ROD HYQ [core] 152

C3 HIV2ROD RNISFKA 153

C3 HIV2ROD APGK[core] 154

395-400 BH10 WFNSTW 155 423-437 IIIB ILNMWQKVGKAMYAP 156

429-443 EVGKAMYAPPISGQI 157

429-438 BRU EVGKAMYAPP 158

429-438 BRU GKAMYAPPIS 159

CD4 bs* IIIB AMYAPPI 160 CD4 bs IIIB AMYAPPISGQ 161

425-441 IIIB NMWQEVGKAMYAPPISG 162

412-453 MN GKAMYAPPIS 159

451-470 LAI SNNESELFRL 163

471-490 LAI GGGDMRDNWRSELYKYKVVK 45 490-508 IILB KYKVVKIEPLGVAPTKAKRR 46

314-323 and 494-503 GRAFVTIGKI and LGVAPTKAKR 11 and 47

472-491 LAI GGDMRDNWRSELYKYKVVKI 48

491-500 LAI IEPLGVAPTK 49 503-509 infec RRVVQRE 50

C terminus infec PTKAKRR 51

C terminus infec VVQREKR 52 TABLE IV HIV gp!20 T Helper Cell Epitopes

Virus SEQ

Position No. Strain Amino Acids Ref. ID NO.

39-51 EQLWVTVYYGVPV 1 53

45-55 VYYGVPVWKEA 1 54

48-61 GVPVWKEATTLFC 1 55

72-82 AHKVWATHACV 1 56

74-85 LAI NVWATHACVPTD 2 57

81-92 CVPTNPVPQEVV 1 58

108-119 LAI VEQMHEDIISLW 2 59

109-121 EQMHEDIISLWDQ 1 60

109-123 IIIB EQMHEDIISLWDQSL 3 61

112-124 IIIB, BH10 HEDIISLWDQSLK 3-9 62

115-126 LAI IISLWDQSLKPC 2 164

204-216 SVITQACSKVSFE 1 165

215-228 FEPIPIHYCAFPGF 1 166

233-244 LAI AGFAILKCNNKT 2 167

269-283 IIIB B10 EVVIRSANFTDNAKT 10 168

274-288 IIIB B10 SANFTDNAKTIIVQL 10 169

296-312 LAI IIVQLNQSVE 2 170

292-300 SF2 NESVAINCT 11 171

MN ESVQLN 12 172

303-321 IILB CTRPNNNTRKSIRIQRGPG(Y) 13 66

307-322 IIIB NTRKSIRIQRGPGR 14 67

309-323 IIIB B10 EQRGPGRAFVTIGKI 10 68

315-329 IIIB RIQRGPGRAFVTIGK 6-8,15,16 69

MN analog RIHIGPGRAFYTTKN 16 70

314-328 IIIB B10 GRAFVTIGKIGNMRQ 10 71

324-338 IIIB FVTIGKIGNMRQAHC 3 173

327-341 HXB2 RQAHCNISRAKWNNT 17 174

342-356 IIIB RAKWNNTLKQICSKL 3 175

346-359 QIVKKLREQFGNNK 18 176

364-378 IIIB B10 SSGGKPEIVTHSFNC 10 177

368-377 LAI NKTIIFKQSS 2 178

369-383 IIIB B10 PEIVTHSFNCGGEFF 10 179

394-408 IIIB B10 TWFNSTWSTKGSNNT 10 180

399-413 IIIB B10 TWSTKGSNNTEGSDT 10 181

410-429 PV22 GSDTITLPCRIKQFLNMWQE 19,20 182

424-438 IIIB B10 LNMWQEVGKAMYAPP 10 183

428-443 IIIB B10 KQIINMWQEVGKAMYA 3-8,12,13,16,21,22 184

432-446 IIIB NMWQEVGKAMYAPPI 3 185 Table IV, cont.

Virus SEQ

Positions Strain Amino Acids Ref. ID NO.

437-451 IIIB VGKAMYAPPISGQIR 3 73

459-473 IIIB BIO GNSNNESEIFRPGGG 10 74

466-481 FRPGGGDMRDNWRSEL 18 75

474-488 IIIB BIO DMRDNWRSELYKYKV 10 76

483-497 IIIB RDNWRSELYKYKVVK 3 77

C492-506 IIIB CKYKVVKIEPLGVAPT 3 78

484-498 IIIB BIO YKYKVVKIEPLGVAP 10 79

494-518 IIIB KVVKIEPLGVAPTKAKRRVVQREK ER.CC 1144 80

References: 1. K.J. Sastry et al, AIDS, 1991 5:699-707; 2. R.D. Schπer et al., J. Immunol. 1989 142:1166-1176; 3. P.M. Hale et al, Int'l. Immunology, 1989 1 :4:409-415; 4. K.B. Cease et al, Proc Natl. Acad. Sci. USA, 1987 84:4249-4253; 5. J.A. Berzofsky et al, Nature, 1988 334:706-708; 6. M. Cleπci et al , Nature, 1989 339:383-385; 7. M. Cleπci et al., J. Immunol, 1991 146:2214-2219; 8. M. Cleπcι e. α/., Eur. J. Immunol, 1991 21 :1345-1349; 9. A. Hosmalm et al, J. Immunol , 1991 146:1667-1673; 10. B. Wahren et al, Vaccines, 1989 89:89-93; 11. P. Botarel et al., J. Immunol, 1991 147:3128-3132; 12. F.D.M. Veronese et al, J. Mol. Biol, 1994 243:167-172; 13. T. J. Palker et al., J Immunol , 1989 142:3612-3619; 14. G. Goodman-Smtkoff et al. , Vaccine, 1990 8:257-262; 15. H. Takahashι e/ α/., J. Exp. Med., 1990 171 :571-576; 16. M. Cleπci et al, J. Inf. Dis., 1992 165:1012-1019; 17. A. P. Warren et al, AIDS Res. Hum. Retrovir., 1995 8:559-564; 18. J. Krowka et al , J. Immunol, 1990 144:2535-2540; 19. K. M. Callahan et al, J. Immunol, 1990 144:3341- 3346; 20. M. Polydefkis et al, J. Exp. Med., 1990 171 :875-887; 21. B.F. Haynes et al, J. Exp. Med., 1993 177:717-727; 22. D. M. Klinman et al, AIDS Res. Hum. Retrovir., 1995 11:97-105

Preferred approaches to identification and selection of T cell epitopes, in particular T helper cell epitopes, for inclusion in a fig as described herein, utilize computer-based algorithms. Several computer-driven algorithms have been devised in the art which exploit the alphabetic representation of amino acid sequence information to search for T cell epitopes by searching the amino acid sequence of a given protein for characteristics believed to be common to immunogenic peptides, and thereby locating regions that are likely to induce cellular immune response in vitro. With the rapid expansion of sequence data on geographic subtypes (clades) of HIV and individual HIV quasi-species, the application of these algorithms to HIV proteins can significantly reduce the number of regions which would require in vitro testing for the desired property (generally immunogenicity) although as envisioned by the present inventors, the desired property is tolerogenicity when presented to the immune system as an fig. Computer-driven algorithms can identify regions of HIV proteins that contain epitopes and are less variable among geographic HIV isolates; alternatively, computer-driven algorithms can rapidly identify regions of each geographic isolate's more variable proteins that should be included in a multi-clade tolerogenic fig. Well-known and conventional ways to identify T cell epitopes within protein antigens, which may be used for the present invention, employ a variety of methods, including the use of whole and fragmented native or recombinant antigenic protein, and the "overlapping peptide" method. This approach involves synthesis of overlapping peptides which span the entire sequence of a given protein antigen, in the present case, gpl20. These overlapping peptides are then tested for their capacity to stimulate the relevant T cell responses in vitro, for example T cell proliferative responses (Vordermeier, H.M. et al. (1993) Immunology 80:6-12; Ashbridge, K.R. et al. (1992) J. Immunol. 148:2248-2255). While the overlapping peptide method is thorough, it is both cost- and labor-intensive. The computer based algorithm methods minimize the cost and labor of the overlapping peptide method and avoid the potential omission of sites between overlapping fragments. Such computer-based algorithms designed to predict T cell epitopes from the amino acid sequences of proteins include AMPHI. AMPHI searches a protein's primary structure for peptides with a high probability of folding as amphipathic structures (Margalit, H. et al. (1987) J. Immunol. 138:2213-2229;

Cornette, J.L. et al. In: The Amphipathic Helix (Ed. Epand, R.M.), CRC Press, Boca Raton, 1993). Seventy percent of published epitopes were found to contain sequences that would have been predicted by AMPHI (Margalit et al, supra; Spouge, J.L. et al. (1987) J. Immunol. 138:204-212). The number of known T cell epitopes has quadrupled since the design of AMPHI, and of these, 65% are amphipathic, such that the correlation remains highly significant (Cornette et al, supra). Other epitope prediction algorithms which analyze protein sequences for specific secondary structural or sequence characteristics (Stille, C.J. et al. (1987) Mol. Immunol. 24:1021-1027; Rothbard, J.B. et al. (1988) EMBOJ. 7:93-100; Salomon, M. et al, (1993) Vaccine 11:1067-1073) generally search for a spacing of hydrophobic residues similar to that searched for by the AMPHI algorithm.

DeGroot and colleagues (Meister, G.E. et al. (1995) Vaccine, 13581-591) developed two computer-based algorithms for T cell epitope prediction, OptiMer and EpiMer, which incoφorate current knowledge of MHC-binding motifs. OptiMer locates amphipathic segments of protein antigens with a high density of MHC-binding motifs. EpiMer identifies peptides with a high density of MHC-binding motifs alone. These algorithms exploit the tendency for MHC-binding motifs to cluster within short segments of each protein. Epitopes predicted by these algorithms contain motifs corresponding to many different MHC alleles, and may contain both class I and class

II motifs, features thought to be ideal for the peptide components of synthetic subunit vaccines. Use of these two algorithms provide sensitive and efficient means for the prediction of promiscuous T cell epitopes that may be used to development preparations such as epitope-specific vaccines, or, for the present application, specific tolerogenic epitopes to be used in an fig.

OptiMer examines known amino acid sequences of proteins and generates a list of peptides that contain these motifs; the algorithm then identifies peptides that would be amphipathic if folded as a helix or twisted as a beta-strand, using the AMPHI algorithm. These potentially amphipathic peptides are compared to the list of MHC-binding motif matches. OptiMer extends the predicted amphipathic peptides, to maximize the density of MHC-binding motif matches per length of protein region.

The EpiMer algorithm searches protein amino acids sequences for MHC- binding motif matches, generating a list of matches for each protein. The relative density of these motif matches is determined along the length of the antigen, resulting in the generation of a motif-density histogram. Finally, the algorithm identifies protein regions in this histogram with a motif match density above an algorithm- defined cutoff density value, and produces a list of subsequences representing these clustered, or motif-rich regions. The regions selected by EpiMer may be more likely to act as multi-determinant binding peptides than randomly chosen peptides from the same antigen, due to their concentration of MHC-binding motif matches. OptiMer and EpiMer, have been used to predict putative epitopes in five Mycobacterium tuberculosis (Mtb) protein antigens (14 kDa, 16 kDa, 19 kDa, 38 kDa, and 65 kDa) and three human immunodeficiency virus (HIV) protein antigens (nef, gpl60 which is the precursor of gpl20 and gp41, and reverse transcriptase (RT). To evaluate the new algorithms' predictive power, Meister et al. compared OptiMer- and

EpiMer-predicted epitopes, AMPHI-predicted epitopes, and peptides that would have been synthesized using the "overlapping peptide" method, to a selection of published T cell epitopes for the above proteins. These algorithms were used to predict T cell epitopes from within the published sequences of three HIV protein antigens. Epitopes published for the HIV protein antigens nef and gpl60 were almost exclusively class I

MHC-restricted, while epitopes published for RT were both class I- and class II- restricted.

A version of either OptiMer or EpiMer based on the list of class I-restricted MHC-binding motifs was used to predict putative epitopes for nef and gpl60, while versions of both algorithms based on the combined list of class I- and class II- restricted motifs were employed to predict putative epitopes for the HIV protein antigen RT. In all, 29 putative epitopes were generated by the class I-specific version of OptiMer (totaling 661 amino acids in length); 30 putative epitopes were generated by EpiMer, totaling 614 amino acids in length. AMPHI generated 36 putative epitopes (totaling 666 amino acid residues), and 104 peptides (totaling over two thousand residues in length) would have been required by the overlapping peptide method. For these two HIV protein antigens, the class I-restricted implementations of both OptiMer and EpiMer identified published epitopes with an efficiency comparable to that of AMPHI, and greater than that of the overlapping peptide method. EpiMer's sensitivity per amino acid exceeds that of either OptiMer or AMPHI. For RT, the combined class I/class II implementation of OptiMer generated 18 putative epitopes (totaling 422 amino acids); the same implementation of EpiMer generated 22 putative epitopes (totaling 361 amino acids in length). These values compare with 23 putative epitopes generated by AMPHI (totaling 433 amino acids) and 55 peptides (totaling over one thousand residues) required by the overlapping peptide method. OptiMer and EpiMer predict published T cell epitopes for the HIV protein RT with both efficiency and sensitivity comparable to that of the AMPHI algorithm. EpiMer again attains the highest sensitivity per amino acid of these three algorithms. In a recent comparison of EpiMer predictions to published HLV protein T cell epitopes, the EpiMer algorithm was shown to be 2.4-fold more sensitive (per amino acid residue) than the overlapping peptide method for detecting published T cell epitopes for four HIV proteins, gpl60, nef, tat, and gag. In contrast, AMPHI was somewhat less sensitive (1.6-fold) (Roberts, C.G.P. et al. (1996) AIDS Res. Human Retrovir. 12:593- 607). A summary of comparisons of the overlapping peptide method with the AMPHI and EpiMer prediction method is provided by Roberts et al. (supra).

The above approach to HIV epitopes has been embodied in an algorithm recently named EpiMatrix/HIV which predicts the sequences most likely to bind to MHC molecules when given a number of primary HIV protein sequences and which was developed by A.S. De Groot at Brown University and implemented for the Internet by AVX Design Inc., Providence, Rhode Island. Both a website and an online tool, EpiMatrix is located on the Internet at http://www.epimatrix.com hiv as of November 1, 1996. Use of this algorithm in accordance with the present invention allows selection of peptides that are highly likely to bind to a particular subject's MHC, thereby enabling identification of T-helper epitopes (as well as cytotoxic T-cell epitopes for vaccine development). The EpiMatrix algorithm yields a score for each peptide in a 10-mer frame. Scoring is a quantitative estimate of the likelihood (relative to other sequences) that a peptide will bind to a given HLA molecule. Two scoring methods are used: single-allele predictions score for specific HLA alleles and clustered predictions score peptides by the prevalence of MHC alleles in selected populations Matrices for all of the major (greater than 10% population prevalence)

MHC alleles representing world populations are included in the algorithm (B.M. Jesdale et al, Vaccines '97, Cold Spring Harbor Laboratory Press). EpiMatrix reduces the total number of regions of HIV proteins to be evaluated in vitro, permitting more rapid identification of desired epitopes. (See, also AIDSWEEKLY Plus, 18 November 1996 issue). Additional MHC binding motif-based algorithms have been described by K.C. Parker et al. (J. Immunol. (1994) 752:163-175) and Y. Altuvia et al. (1995) J. Mol Biol. 249:244-250). In these algorithms, binding to a given MHC molecule is predicted by a linear function of the residues at each position, based on empirically defined parameters, and in the case of Altuvia et al, known crystallographic structures are also taken into consideration, j. Hammer et al. (J. Exp. Med. (1994) 750:2353- 2358) described a technique known as "peptide side chain scanning" which is used to predict binding peptides for an MHC allele. The EpiMer/EpiMatrix algorithm predicted putative T cell epitopes from protein sequences for HIV-1 nef, gpl60, gag p55, and tat that required fewer peptides and therefore fewer amino acid residues to be synthesized than either AMPHI-predicted peptides or overlapping peptides. For the four HIV-1 proteins, EpiMer predicted 43 peptide epitopes, AMPHI predicted 68 peptides , and the overlapping peptide method (20 amino acid long peptides overlapping by 10 amino acids) would have required 161 peptides. Details (amino acid start and stop, number of MHC binding motifs) of the predicted proteins are available36. Regions of HIV proteins that contain as many as 20 to 30 MHC binding motifs can be identified using this algorithm.

The various known methods for epitope prediction are not mutually exclusive. As the contributions of side chains and tertiary peptide structure to peptide-MHC binding are better quantified, the development of a computer algorithm that predicts T cell epitopes based on a matrix of side chain information such as one described by J. Hammer (1995) Curr. Opin. Immunol. 7:263-269) will become available. The identification of novel structural features which are able to independently predict peptide recognition and their subsequent synthesis into a combined algorithm with statistically verifiable predictive capacity, allows a dramatic reduction in the time and effort required to synthesize and test potential T cell antigenic sites for HIV proteins, by allowing the prediction of sites with a high concentration of antigenic features.

HIV protein regions that contain multiple overlapping class-II restricted epitopes, also known as "multi-determinant" or multi-determinant peptides, have been identified in mice and humans. Such regions might be important to include in the synthesis of an fig having multiple tolerogenic T helper cell epitopes as described herein. This is particularly useful if a multi-determinant T cell epitope is involved in stimulating antibody responses (i.e., to B cell epitopes).

Table V, below presents a list of epitopes of gpl20 (and several N-terminal epitopes of gp41) which were identified using EpiMer (Roberts et al, supra). These sequences are from the BH10 strain of HIV- 1. The amino acid sequence of this HIV strain was obtained from the SWISS-PROT protein sequence data bank , Accession No. P03375 (EMBL Data Library, Heidelberg, Germany). The residue numbers shown in Table V are from this sequence bank. Those residues beyond amino acid 511 are part of gp41, not gpl20. In a preferred embodiment, the present invention provides a tolerogenic fig H chain or intact fig molecule which includes at the N- terminus of the H chain one or more of the HIV peptide epitopes listed in Table V.

TABLE V

T Cell Epitopes of Hiv gp!20 Identified by Epimer Algorithm

AMINO AMINO

ACIDS SEQUENCE ACIDS SEQUENCE

19-34 TMLLGMLMICSATEKL 168 - 185 KVQKEYAFFYKLDIIPID

(SEQ ID NO: 186) (SEQ ID NO: 187)

20-28 MLLGMLMIC 168-176 KVQKEYAFF

21-29 LLGMLMICS 169-177 VQKEYAFFY

21-30 LLGMLMICSA 169-178 VQKEYAFFYK

22-29 LGMLMICS 171-179 KEYAFFYKL

22-30 LGMLMICSA 173-181 YAFFYKLDI

24-32 MLMICSATE 173-182 YAFFYKLDII

24-33 MLMICSATEK 174-181 AFFYKLDI

25-33 LMICSATEK 174-182 AFFYKLDII

26-34 MICSATEKL 175-182 FFYKLDII

175-183 FFYKLDIIP

36-54 VTVYYGVPVWKEATTTLFC 175-184 FFYKLDIIPI

(SEQ ID NO:63) 176-184 FYKLDIIPI

36-44 VTVYYGVPV

36-45 VTVYYGVPVW 198-214 TSVITQACPKVSFEPIP

37-46 TVYYGVPVWK (SEQ ID NO: 188)

38-46 VYYGVPVWK 198-207 TSVITQACPK

39-46 YYGVPVWK 199-207 SVITQACPK

42-51 VPVWKEATTT 199-208 SVITQACPKV

44-52 VWKEATTTL 200 - 208 VITQACPKV

44-53 VWKEATTTLF 202-210 TQACPKVSF

45-53 WKEATTTLF 204-213 ACPKVSFEPI

45-54 WKEATTTLFC 205-213 CPKVSFEPI

84-95 WLVNVTENFNM 249 - 261 HGIRPVVSTQLLL

(SEQIDNO:64) (SEQ ID NO: 189)

85-93 VLVNVTENF 249 - 256 HGIRPVVS

85-94 VLVNVTENFN 251-259 IRPVVSTQL

87-95 VNVTENFNM 251-260 IRPVVSTQLL

252 - 260 RPVVSTQLL

115-127 SLKPCVKLTPLCY

(SEQ ID NO:65) 284 - 296 IIVQLNQSVEINC

116-124 LKPCVKLTP (SEQ ID NO: 190)

116-125 LKPCVKLTPL 284 - 292 IIVQLNQSV

117-125 KPCVKLTPL 285 - 293 IVQLNQSVE

119-127 CVKLTPLCV 286 - 294 VQLNQSVEI

286 - 295 VQLNQSVELN

288 - 296 LNQSVEINC TABLE V, cont.

AMINO AMINO

ACIDS SEQUENCE ACIDS SEQUENCE

310- 330 QRGPGRAFVTIGKIGN 482 - 527 ELYKYKWKIEPLGVA MRQAH (SEQ ID NO: 191) PTKAKRRWQREKRAV

312- ^■320 GPGRAFVTI GIGALFLGFLGAAG

314- ^■322 GRAFVTIGK (SEQ ID NO: 194)

315- ^■322 RAFVTIGK 482 - 490 ELYKYKVVK

315 - ^■323 RAFVTIGKI 483 - 491 LYKYKVVKI

317- ^■325 FVTIGKJGN 483 - 492 LYKYKVVKIE

318- ^■326 VTIGKIGNM 486 - 494 YKVVKIEPL

318- ^■327 VTIGKIGNMR 486 - 495 YKVVKIEPLG

319 - ^■327 TIGKIGNMR 488 - 496 VVKIEPLGV

320- ^■328 IGKIGNMRQ 488 - 497 VVKIEPLGVA

320- ^■329 IGKIGNMRQA 489 - 497 VKIEPLGVA 491 -499 IEPLGVAPT

351- ■363 EQFGNNKTIIFKQ 493 - 502 PLGVAPTKAK (SEQ ID NO: 192) 494 - 502 LGVAPTKAK

351 - -359 EQFGNNKTI 495 - 503 GVAPTKAKR

353 - ^■361 FGNNKTIIF 495 - 504 GVAPTKAKRR

353 - ■362 FGNNKTIIFK 496 - 504 VAPTKAKRR

496 - 505 VAPTKAKRRV

381- ■392 EFFYCNSTQLFN 497 - 505 APTKAKRRV

(SEQ ID NO: 193) 500 - 508 KAKRRVVQR

382- ^■390 FFYCNSTQL 503-511 RRVVQREKR

382- -391 FFYCNSTQLF 505-513 VVQREKR/AV^*

383 ^■ -391 FYCNSTQLF 506-514 VQREKR/AVG

383 ■ -392 FYCNSTQLFN 506-515 VQREKR/AVGI

414 - 445 ITLPCRIKQIINMWQEV 507-515 QREKRAVGI

1 GKAMYAPPISGQIRC 510-518 KR/AVGIGAL

1 (SEQIDNO:81) 511-519 R/AVGIGALF

414- -422 ITLPCRIKQ

414. -423 ITLPCRIKQI

416. -424 LPCRIKQII

416. -426 LPCRIKQILNM

418 -426 CRIKQIINM

420. -428 IKQILNMWQ

420. -428 IKQIINMWQ

420 -429 IKQILNMWQE

424. -433 INMWQEVGKA

426 -435 MWQEVGKAMY * "/" follows the C terminal residue of

427 -435 WQEVGKAMY gpl20

428 -436 QEVGKAMYA Application of the above approach to general HIV tolerogenic preparations may be restricted by the amount of sequence variation in individual quasi-species, HIV strains, and HIV subtypes as well as by the MHC background of the subject population. For example, the region of amino acids at about 130 to 160 (depending upon which subtype or isolate), has a great deal of inter-strain variation and may therefore best be avoided in designing a tolerogenic fig which has the broadest range of applicability across virus variants and subtypes. HIV peptide epitopes which contain multiple MHC binding motifs, either conserved across HIV strains or derived from several different HIV strains, may be ideal candidates for targeting for T helper cell-directed tolerance induction, as it is assumed that the tolerogen will be presented in vivo by host MHC molecules. Thus, epitopes with multiple MHC binding motifs or having an MHC binding motif present in the highest frequency in the subject population (race, ethnic group, e.) would be preferably selected for inclusion in a tolerogenic fig. The EpiMer algorithm is particularly well suited for identifying and selecting such epitopes.

Preparation of Recombinant fig and its Transfer

The present invention provides polynucleotides encoding the fig in the form of recombinant DNA molecules in vehicles such as plasmid and retroviral vectors, capable of expression in a desired eukaryotic host cell as disclosed herein. The invention also provides hosts transfected or transduced with the fig constructs which are capable of producing in culture or in vivo the fig molecules and secreting them or displaying them on the cell surface.

A preferred engineering strategy for inserting a foreign epitope at the N- terminus of an IgG γ chain is shown in Figure 4 A and 4B. Figure 4 A depicts the incoφoration of an oligonucleotide, in this example encoding the λ phage Cl repressor peptide 12-26. However, the present invention exploits the same general scheme wherein a native or synthetic gpl20 peptide epitope is inserted in place of the 12-26 peptide. This is illustrated in Figure 4B. Any Ig gene construct may be used for insertion of the tolerogenic epitope or epitopes. A preferred Ig gene encodes human Ig, more preferably an Ig comprising a human γ chain. The general principles of recombinant DNA technology are utilized, as described for example, in Sambrook, J. et al. , Molecular Cloning: A Laboratory Manual, 2nd Edition, Cold Spring Harbor Press, Cold Spring Harbor, NY, 1989; Ausubel, F.M. et al. Current Protocols in Molecular Biology, Vol. 2, Wiley- Interscience, New York, 1987; Lewin, B.M., Genes IV, Oxford University Press,

Oxford, (1990); Watson, J.D. et al, Recombinant DNA, Second Edition, Scientific American Books, New York, 1992, which references are hereby incoφorated by reference in their entirety.

The DNA construct encodes an individual fig H chain, although the protein products of this invention include both the fig H chain and a complete assembled Ig molecules comprising the fig H chain having one or more HIV gpl 20 epitopes in combination with a native human Ig L chain. The fig may also comprise two different H chains, one of which is a fusion protein having one or more HIV gpl 20 epitopes added to or included in the V region. Genetic sequences, especially cDNA sequences, encoding either a complete fig H chains, the fig V regions or a human Ig C region of any Ig isotype, most preferably, an IgG isotype (i.e., a human Cγ chain) are also provided herein.

The invention also provides a genetic sequence, especially a cDNA sequence encoding an Ig V region fusion protein in which the V region encoding DNA has been combined in frame with one or more HIV gpl 20 epitopes. Though, genomic DNA sequences may also be used, cDNA sequences are particularly preferred.

One non-limiting approach to producing the fig comprises the steps of: 1. Selection of one or more gpl 20 epitopes as described below for which tolerance is desired; 2. Preparation of DNA encoding the epitope or epitopes selected above; this can be done by isolating HIV RNA and cloning an preparing cDNA corresponding to all or part of gpl 20, by isolating and cloning DNA from HIV-infected cell, or if the DNA is sufficiently short, synthesizing an oligonucleotide having the desired coding sequence. The latter synthetic approach permits construction of artificial combinations of two or more gpl 20 epitopes or which are not contiguous in the native protein.

3. Introduction of appropriate restriction enzyme recognition sites in the gpl 20 DNA to permit ligation to Ig H chain encoding DNA, preferably human γ chain- encoding DNA; this can be done by PCR, site-directed mutagenesis or synthetically;

4. Selection of endogenous restriction sites of the Ig-encoding DNA or modification of the DNA as above to introduce restriction sites corresponding to those in the gpl 20 DNA such that they can form cohesive ends and be ligated 5. Ligation of the gpl 20 DNA to the Ig H chain DNA using conventional methods.

6. Expression and production of the fig H chains or intact Ig molecules (H₂L₂) in a selected host, preferably human lymphoid or hematopoietic cells.

Oligonucleotides which can be used as primers for introducing useful restriction sites into the gpl 20 and human Ig DNA for subsequent linkage are well known in the art. See, for example, Sambrook et al, supra.

In an alternate embodiment, rather than using DNA encoding an entire Ig H chain, the gpl 20 DNA is linked to an Ig V gene cassette. Because the antibody specificity of the fig is not important, any V region DNA can be selected. A preferred V gene is one which encodes a protein which, after fusion of a gpl 20 epitope or epitopes, even a full length gpl 20 protein, still maintains its ability to fold properly in an full Ig molecule (H₂L₂).

The variable (V) domain of an Ig chain includes hypervariable (HV) regions which are also known as complementarity-determining regions (CDRs) because they are important in "determining" the structure of the antibody combining site that is complementary the epitope bound. Each H and L chain V region has three HVs or

CDRs. The segments on either side of each HV region which are relatively invariant are termed "framework regions" (FRs). Thus, the order of these regions in a V domain (from the N-terminus)is as follows: FR1-HV1-FR2-HV2-FR3-HV3-FR4. For example, the three HV regions are roughly from residues 28-35, 49-59 and 92- 103, respectively. The framework regions form the β sheets that provide th structural framework of the domain, with the HV sequences corresponding to three loops at one edge o each sheet that are juxtaposed in the folded protein. The HV loops from the V_H and V_L domains are brought together, creating a single HV site at the tip of the Fab fragment which forms the antigen binding site. (See, for example, Janeway, C.A., Jr. et al,

IMMUNOBIOLOGY, 2n ed., Garland Publishing Inc., New York, 1996, chapter 3).

The first framework region (FRl) is the most N-terminal of the V region. Eisen, H.N., GENERAL IMMUNOLOGY, (J. Lippincott Co., Philadelphia, 1990) at pages 57-59, in particular Figure 14-19 at page 58, shows the amino acid sequences of the first framework region of 5 different human H chains. The first framework region includes the 30 N-terminal amino acids at which point the HV1 region follows. A framework region of nine different human K L chains belonging to three different groups Vκl, VKII and VKIII are shown in this textbook figure.. Again, the FRs are about 30 residues, with a number of positions in each group serving as "framework residues" which serve to characterize each Vt group. In the present invention, the heterologous epitope of the fig is preferably inserted immediately N terminal to the first framework region. In other embodiments, it may be fused "deeper" into the Ig sequence within the V region.

A spacer comprising between about 1 and 10 amino acids, preferably about 3- 5 residues, can be present between the C terminal residue of the heterologous epitope(s), preferably a gpl 20 epitope(s) and the N terminal residue of the Ig V region, provided that the protein can fold properly to present the gpl 20 epitope while maintaining its tolerogenic properties. In a preferred arrangement, as exemplified below, a repeat of the 5 N-terminal amino acids of the Ig H chain is inserted N- terminal from the added the gpl 20 peptide (or peptides) such that this pentapeptide sequence is repeated on either end of the inserted gpl 20 sequence. If more than one gpl 20 peptide is included, a spacer as described herein may be linked to one or more of the added gpl 20 peptides. A major puφose of the spacer is to permit unimpeded folding and proteolytic processing of the fig as if it were an normal Ig protein. This assures proper surface expression of the fig and association with MHC proteins on the surface of a tolerogen-presenting cell.

Advantage can be taken of the natural antigen-binding property of the V region, as is described below for the model murine constructs wherein the V region was specific for the NIP hapten. A complete Ig H chain is constructed by combining the now altered V gene construct containing additional gpl 20 DNA with a C gene construct encoding a desired human C region, preferably a human Cγ protein. The most preferred C region would be that encoding the γ₃ isotype.

Ig H chain (or V_H) cDNA vectors are typically prepared from human cells and modified by site-directed mutagenesis to place a restriction site at the position in the human sequence in which the gpl 20 DNA is to be grafted. Preferably this is 5' to the nucleotide encoding the N-terminus of the Ig H chain or the V_H protein.

Two coding DNA sequences are said to be "operably linked" if the linkage results in a continuously translatable sequence without alteration or interruption of the triplet reading frame. A DNA coding sequence is operably linked to a gene expression element if the linkage results in the proper function of that gene expression element to result in expression of the coding sequence.

Expression vehicles include plasmids or other vectors, such as retroviral vectors. A preferred vehicle carries a functionally complete human V_H and C_H having appropriate restriction sites engineered so that any gpl20-encoding nucleotide sequence with appropriate cohesive ends can be conveniently ligated thereto. These vehicles can be used as intermediates for propagation of DNA encoding any desired H chain (V_HC_H) ready to receive a gpl 20 DNA sequence, and for the expression of the complete fig (gp!20-V_HC_H). Preferred hosts are mammalian cells, most preferably human cells, grown in vitro for prolonged periods, or taken from a host, cultured in vitro for puφoses of transfection and then reintroduced into the host. Mammalian cells provide post-translational modifications to the Ig protein molecules including leader peptide removal, folding and assembly of H and L chains, glycosylation of the protein chains and secretion of the complete functional fig protein. Mammalian cells which may be useful as hosts for the production of fig proteins include cells of lymphoid origin, such as the hybridoma Sp2/O-Agl4 (ATCC CRL 1581) or the myeloma P3X63Ag8 (ATCC TIB 9), also abbreviated as P3, and its derivatives. A preferred murine cell line for expressing the fig of this invention is J558L. Any cell line which allows for efficient expression and secretion of the fig constructs of the present invention and which promotes proper folding of the fig is preferred. Known human lymphoid or hematopoietic cell lines may be used, including B lymphoblastoid lines, lymphomas, hybridomas or heterohybridomas. Examples of cell lines and approaches for expression of recombinant or chimeric or hybrid or modified Ig genes are described in Shin, S.U. et al, (1993) Int. Rev. Immunol. 70:177-186; Wright, A. et al, (1992) Crit.

Rev. Immunol. 72:125-168; Shin, S.U. et al. (1992) Immunol. Rev. 750:87-107; Morrison, S.L., (1992) Annu. Rev. Immunol. 70:239-265; Morrison, S.L. et al, (1989) Adv. Immunol. 44:65-92; Weidle et al, (1987) Gene 57:21; Whittle et al, (1987) Protein Engineering 7:499; Morrison, S.L., (1985) Science 229:1202-1201; Morrison S.L. et al, (1984) Annu. Rev. Immunol. 2:239-256, all of which references are incoφorated by reference in their entirety. In a preferred embodiment, human hematopoietic cells obtained from the intended recipient or those histocompatible with the recipient are transfected with the fig DNA construct.

Many vector systems are available for the expression of cloned Ig H and L chain genes in mammalian cells (see Glover, D.M., ed.(1985) DNA Cloning, Vol. II, pp 143-238, IRL Press). Different approaches can be followed to obtain complete H₂L₂ antibodies. It is possible to co-express H and L chains in the same cells to achieve intracellular association and linkage of H and L chains into complete tetrameric H₂L₂ antibodies. The co-expression can occur by using either the same or different plasmids in the same host. Genes for both H and L chains can be placed into the same plasmid, which is then transfected into cells, thereby selecting directly for cells that express both chains. Alternatively, cells may be transfected first with a plasmid encoding one chain, for example the L chain, followed by transfection of the resulting cell line with an H chain plasmid containing a second selectable marker. Cell lines producing H₂L₂ molecules via either route could be transfected with plasmids encoding additional copies of H, L, or H plus L chains in conjunction with additional selectable markers to generate cell lines with enhanced properties, such as higher production of assembled H₂L₂ antibody molecules or enhanced stability of the transfected cell lines. One particular strategy for inserting an HIV peptide sequence at or near the N- terminus of an Ig H chain is related to that described in Hebell, T. et al (1991) Science 254:102-105 and Ballard, D.W. et al (1993) Proc. Natl Acad. Sci. USA 53:9626-9630. A first plasmid is constructed which preferably includes .a full genomic sequence of the Ig H chain and selectable markers, for example, neomycin and or/ampicillin resistance genes. The source DNA encoding the HIV gpl 20 epitope or epitopes PCR is amplified to create the DNA encoding the desired single or multiple epitopes. Appropriate restriction sites are included on the primers so that the epitope-encoding DNA can be spliced into the Ig gene-containing vector. The gpl 20 epitope sequence is subcloned into a site, preferably the V_Hsite of the first plasmid. Recombinant clones are analyzed for proper orientation and polymerase induced errors by double stranded DNA sequencing methods (e.g., Sequenase® kit from U.S. Biochemical).

The promoter sequences useful for the DNA constructs of the of the present invention are any promoters which allow efficient expression of the fig DNA of the invention in a target cell of choice, for example a hematopoietic progenitor cell or a lymphoid cells, more preferably a B cell. Preferred promoters are the promoters of the Ig gene into which the foreign epitope-encoding DNA is being inserted. However, other known promoters of either eukaryotic or viral origin may be used. Suitable promoters are inducible or repressible or, more preferably, constitutive. Examples of useful eukaryotic/viral promoters include the promoter of the mouse metallothionein I gene (Hamer, D., et al. (1982) J. Mol. Appl. Gen. 7:273-288); the TK promoter of Heφes virus (McKnight, S. (1982) Cell 57:355-365); the SV40 early promoter (Benoist, C, et al. (1981) Nature 290:304-310); and the yeast gal4 gene promoter (Johnston, S.A., et al. (1982) Proc. Natl. Acad. Sci. (USA) 79:6911-6915; Silver, P.A., et al. (1984) Proc. Natl. Acad. Sci. (USA) 57:5951-5955). Strong promoters are most preferred.

The fig construct into which the gpl 20 epitope(s) has been inserted is introduced ("gene transfer") into the appropriate target cells by conventional methods, e.g. , direct physical transfer of plasmid DNA, or preferably, by virus-mediated transfer, for example using a retroviral vector, as discussed below.

A number of means for transferring genes are known in the art and may be used herein, including, for example, electroporation and lipofection. A preferred, and relatively efficient means for achieving transfer of genes is by retrovirus-mediated gene transfer (Gilboa, E. (1987) Bio-Essays 5:252-258; Williams, D.A. et al. (1984)

Nature 570:476-480; Weiss, R.A. et al, RNA Tumor Viruses, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1985). One class of retroviruses, recombinant amphotropic retroviruses have been used as vectors for the transfer of genes into human cells (Cone, R.D. et al. (1984) Proc. Natl. Acad. Sci. USA 57:6349- 6353; Danos, O. et al. (1988) Proc. Natl Acad. Sci. USA 55:6460-6464.

When the targets for gene therapy are bone marrow or blood stem cells, for example, it may be advantageous to manipulate the cells in vitro with cytokines and then to infect them with the vector bearing the fig gene (Wilson, J.M. et al. (1990) Proc. Natl. Acad. Sci. USA 57:8437-8441). Recombinant amphotropic retroviruses have been recognized as useful vectors for transferring genes efficiently into human cells, for example to correct enzyme deficiencies (Cone, R.D. et al. (1984) Proc. Natl. Acad. Sci. USA 57:6349-6353; Danos, O. et al, (1988) Proc. Natl. Acad. Sci. USA 55:6460-6464). For safety reasons, it is important that a retroviral vector used for gene therapy be capable of infecting only desired cells and not cause generalized infection of cells throughout the body of the individual being treated. In the past, this has generally been accomplished by using helper-defective virus preparations, or mutants lacking the ψ packaging sequence, etc.

Another viral vector system useful for this invention is the recombinant adeno associated viral (AAV) transduction system (Lebkowski, J.S., et al. (1988) Mol. Cell. Biol. 5:3988-3996). AAV DNA integrates into cellular DNA as one to several tandem copies joined to cellular DNA through inverted terminal repeats (ITRs) of the viral DNA. (Kotin, R.M., et al. (1990) Proc. Natl Acad. Sci. USA 87:2211-2215). The transgene DNA size limitation and packaging properties are the same as with any other DNA viral vectors. AAV is a linear single stranded DNA parvo virus, and requires co-infection by a second unrelated virus in order to achieve productive infection. AAV carries two sets of functional genes: rep genes, which are necessary for viral replication, and structural capsid protein genes (Hermonat, P.L., et al. (1984) J. Virol. 57:329-339). The rep and capsid genes of AAV can be replaced by a desired DNA fragment to generate AAV plasmid DNA. Transcomplementation of rep and capsid genes are required to create a recombinant virus stock. Upon transduction using such virus stock, one recombinant virus uncoats in the nucleus and integrates into the host genome by its molecular ends.

Liposomes may be used to encapsulate and deliver a variety of materials to cells, including nucleic acids and viral particles (Faller, D.V. et al. (1984) J. Virol. 49:269-212). Preformed liposomes that contain synthetic cationic lipids form stable complexes with polyanionic DNA (Feigner, P.L., et al. (1987) Proc. Natl. Acad. Sci. USA 54:7413-7417). Cationic liposomes, liposomes comprising some cationic lipid, that contained a membrane fusion-promoting lipid dioctadecyldimethyl-ammonium- bromide (DDAB) efficiently transfer heterologous genes into eukaryotic cells (Rose, J.K., et al. (1991) Biotechniques 70:520-525). Cationic liposomes can mediate high level cellular expression of transgenes, or mRNA, by delivering them into cultured cell lines (Malone, R., et al. (1989) Proc. Natl. Acad. Sci. USA 5(5:60776081).

Ecotropic and amphotropic packaged retroviral vectors infect cultured cells in the presence of cationic liposomes, such as Lipofectin (BRL, Gaithersburg, MD), and in the absence of specific receptors (Innes, CL. et al. (1990) J. Virol. 64:951-961).

Physical means well-known in the art can be used for direct gene transfer, including administration of plasmid DNA (Wolff et al, 1990, supra) and particle- bombardment mediated gene transfer, originally described in the transformation of plant tissue (Klein, T.M. et gl. (1987) Nature 327:10; Christou, P. et al. (1990) Trends Biotechnol. 6:145) but also applicable to mammalian tissues in vivo, ex vivo or in vitro (Yang, N.-S., et al. (1990) Proc. Natl. Acad. Sci. USA 57:9568; Williams, R.S. et al. (1991) Proc. Natl. Acad. Sci. USA 55:2726; Zelenin, AN. et al. (1991) FEBSLett. 280:94; Zelenin, AN. et al. (1989) FEBSLett. 244:65; Johnston, S.A. et al. (1991) In Vitro Cell. Dev. Biol. 27:11). Furthermore, electroporation, a well-known means to transfer genes into cell in vitro, can be used to transfer DΝA molecules according to the present invention to tissues in vivo (Titomirov, AN. et al. (1991) Biochim. Biophys. Acta 7055:131).

Gene transfer can also be achieved using "carrier mediated gene transfer" (Wu, CH. et al (1989) J. Biol. Chem. 264:16985; Wu, G.Y. et al. (1988) J. Biol. Chem. 263:14621 ; Soriano, P. et al. (1983) Proc. Natl. Acad. Sci. USA 50:7128; Wang, C-Y. et al (1982) Proc. Natl. Acad. Sci. USA 54:7851; Wilson, J.M. et al. (1992) J. Biol. Chem. 267:963). Preferred carriers are targeted liposomes (Νicolau, C. et al. (1983) Proc. Natl. Acad. Sci. USA 50:1068; Soriano et al, supra) such as immunoliposomes, which can incoφorate acylated monoclonal antibodies into the lipid bilayer (Wang et al. , supra), or polycations such as asialoglycoprotein/polylysine (Wu et al. , 1989, supra).

In general, improved efficiency of gene transfer is attained by the use of promoter enhancer elements in the plasmid DΝA constructs (Philip, R., et al. (1993) J. Biol. Chem. 268:16087-16090). The disclosure provided herein focuses on the gpl20-derived amino acid sequence that is present in the final fig product. The nucleotide sequences encoding the desired peptide epitopes are not specifically listed here but are evident to those skilled in the art. First, the full native sequences for HIV gpol20 (consensus for each subtype as well as individual viral isolates reported to date) are provided in the Compendium cited above. One skilled in the art will know how to utilize alternate coding sequences for expressing the desired native or synthetic gpl 20 peptides which are to be included in the tolerogenic fig. Any nucleotide sequence which encodes a chosen peptide epitope or series of epitopes may be used. Distinct gpl 20 epitopes may be combined in any order or combination provided that the coding nucleic acids provide an in-frame sequence both with respect to the gpl 20 epitopes and with respect to the Ig H gene utilized to construct the fig.

Uses of the Invention

Treatment of an individual infected with HIV using the tolerogenic fig of this invention comprises parenterally administering a single or multiple doses of the fig to a subject, preferably a human. The fig is preferably an isologous Ig, that is, of the same species as the subject. A most preferred fig is fusion IgG molecule. An effective tolerogenic dose is a function of the size and number of particular HIV gpl 20 epitopes included in a particular fig construct, the patient and his clinical status, and can vary from about 0.01 mg/kg body weight to about 1 g kg body weight.

A subject can be given this amount in a single dose or in multiple repeated doses.

Doses of hematopoietic cells or B cells expressing the fig are preferably administered at a dose between about 10⁶ and 10¹⁰ cells on one or several occasions.

The route of administration may include intravenous (iv) , subcutaneous (SC), intramuscular, intrapulmonary, intraperitoneal or other known routes. The preferred route for administration of fig proteins or cells for tolerogenesis is by iv injection.

The fig of this invention may be advantageously utilized in combination with other therapeutic agents useful in the treatment or prevention of HIV disease, including prophylactic or therapeutic vaccine preparations, antiviral chemotherapeutic agents, immune response modulators including cytokines and hematopoietic growth factors, protective antibody reagents, etc.

Having now generally described the invention, the same will be more readily understood through reference to the following examples which are provided by way of illustration, and are not intended to be limiting of the present invention, unless specified.

EXAMPLE I

Epitope-Specific Tolerance Induction by Gene Transfer of an Engineered Peptide-

Immunoglbulin Fusion Protein

In this study, the present inventors took advantage of the IgG molecule as a tolerogenic carrier, and created an engineered tolerogen with a grafted epitope at the N-terminus of an IgG heavy chain. This engineered IgG was recognized by the immune system in a tolerogenic manner. The model epitope chosen for this initial analysis is the well-characterized class-II MHC-restricted peptide sequence from the cl λ repressor protein (pi -102), residues 12-26. This peptide contains both a B- and T- cell epitope. and is the immunodominant determinant in H-2^d mice immunized with the entire protein (26-30). It was thus possible to measure tolerance induction to a single determinant at both the B-cell and T-cell levels. Furthermore, tolerogenic Ig- peptide constructs could be expressed in adoptively-transferred hematopoietic tissue for the permanent modulation of epitope-specific immune responses in mature adults. These studies show that the model fig, 12-26-IgG, is an efficient tolerogen in adult animals and serves as the basis for expansion of this approach to other epitopes of clinical utility as described below. MATERIALS AND METHODS Mice. Male and female BALB/cByJ (H-2^d) and CAF1 (H-2^d/a) mice were obtained from the Jackson Laboratory (Bar Harbor, ME) and were used at 6- 10 weeks of age.

Medium: RPMI 1640 medium (GIBCO-BRL, Gaithersburg, MD) was supplemented with 5% FCS (Hyclone, Logan, UT), 2-ME, L-glutamine, penicillin, streptomycin, MEM nonessential amino acids, and sodium pyruvate. Antibodies: Hybridoma B3.11, which produces a monoclonal IgG, specific for the 12-26 peptide was a kind gift of Drs. Tom Briner and Malcolm Gefter (Immulogic,

Waltham, MA). B3.11 was affinity purified with goat anti-mouse IgG sepharose columns and biotinylated, or used as a neat culture supernatant. All alkaline- phosphatase (AP)-conjugated reagents were purchased from Southern Biotechnology Assoc. (Birmingham, AL). Synthetic peptide: The 12-26 15-mer LEDARRLKAIYEKKK (SEQ ID NO:l 12) was prepared with a solid-phase method and purified to >92% homogeneity using standard HPLC methods. Peptide was conjugated to bovine albumin serum (BSA) rabbit gamma globulin (RGG), or keyhole limpet hemocyanin (KLH) as described (Roy, S. et al (1989) Science. 244:515-515). Oligonucleotides: The following complementary synthetic oligonucleotides encoding the 12-26 sequence were designed with BamHI/Clal restriction ends, phosphorylated with T4 kinase and ATP, and cloned into the hypervariable region of flagellin construct pPX 1647: DWS1 : (SEQ ID NO: 195)

5'-CGA TCT GGA GGA CGC GCG GCG GCT GAA GGC GAT ATA CGA GAA GAA GAA GG-3'

DWS2: (SEQ ID NO: 196)

5'-GAT CCC TTC TTC TTC TCG TA T ATC GCC TTC AGC CGC CGC GCG TCC TCC AGA T-3'

PCR primers were also designed to amplify a modified 12-26 sequence from the chimeric 12-26-flagellin construct. This sequence includes 5' FRl V_H sequence and

PstI restriction sites at each flanking ends:

Ig-one: 5'-TGATCTACTGCAGCTGGAGGACGCGCGGCG G-3' (SEQ ID NO:197)

Ig-two:

5'- CGACCTCCTGCAGTTGGACCTGCTTCTTCTTCTCGTATAT-3' SEQ ID NO:198) ELISA : To determine the specificity of binding of our peptide-specific mAb B3.11 to

12-26-fusion proteins, competitive inhibition ELISA' s were conducted as follows: biotinylated B3.l l was incubated 1 :1 (vol/vol) with decreasing amounts of inhibitor in ELISA binding buffer (0.25% BSA, 0.05% Tween 20 in saline). Mixtures were then incubated on peptide-coated (10 μg/ml) ELISA plates (Immulon 4 Dynatech), and subsequently streptavidin-AP was added as a secondary reagent. Percent inhibition of binding (A₄₀₅) was calculated as: [(average binding of antibody alone minus average binding of antibody incubated in presence of inhibitor)/average binding of antibody alone] xlOO. ELISA determinations of serum peptide-specific IgG responses were done by coating ELISA plates with 50 μg/ml synthetic peptide. Antigen-coated plates were blocked with 1% gelatin/0.05% Tween 20 buffer, and duplicate serial dilutions of serum were incubated and probed with goat anti-mouse IgG isotype-specific secondary reagents. Titers are expressed as the geometric mean of the reciprocal dilution required to bring A₄₉₀ readings to prebleed levels or <0.08 O.D. Protein Engineering Design:

Preparation of murine H chain IgG] construct encoding the 12-26 sequence at the N- terminus

Our strategy for inserting a foreign peptide sequence at the N-terminus of an IgG H chain is similar to what has been described (Hebell, T. et al. (1991) Science

254:102-105; Dal Porto, J. et al. (1993) Proc. Natl. Acad. Sci. USA 90:6671-6675).

Plasmid pSNR (Ballard, D.W. et al. (1986) Proc. Natl. Acad. Sci. USA. 55:9626-9630), which contains neo and amp resistance genes, as well as the full genomic sequence for a IgG,⁰ H chain specific for the NP hapten, was obtained from Dr. Douglas Fearon (Cambridge University) and modified. A modified 12-26 sequence was created via PCR amplification of this sequence from the chimeric flagellin construct A29 (described in WO95/21926) utilizing PCR primers "Ig-one" and "Ig-two". The modified 12-26 sequence was subcloned into the V_H site of pSNR and recombinant clones were analyzed for proper orientation and Taq polymerase mutational errors by double-stranded DNA sequencing methods (USB Sequenase 2.0 kit). Expression, purification, and quantitation of transfected IgG:

Construct ρQ3.EZl (Q3), as well as the control pSNR IgG, construct (P6) were electroporated into J558L myeloma cells (which produce only a λ light chain) as described by Hebell et al, supra, Ballard et al, supra, and Dal Porto et al, supra.

Stably transfected clones were isolated in 1 mg/ml G418 (GIBCO-BRL), subcloned, and transfected IgG's from selected clones were purified from bulk supernatants or ascites with anti-mouse IgG-Sepharose or protein G columns. Since the original H chain binds with high affinity to the NIP (5-iodo-4-hydroxy-3-nitrophenylacetyl) hapten, purified or serum transfectoma IgG was quantitated using a modified NIP- gelatin binding ELISA , using anti-mouse IgG,-AP as a secondary reagent. In vitro and in vivo tolerance induction and immunization protocols:

Peptide-specific tolerance induction in adult recipients was accomplished by intravenous ("iv") injection (in the lateral tail vein) of either 1 mg purified, deaggregated, chimeric (Q3) or control IgG (P6) diluted in saline, or by 3 repeated injections of mitomycin C-treated (50 μg/ml, SIGMA) P6- or Q3-secreting transfectomas. For measurement of humoral immune responses 10 days following iv tolerization, animals were immunized subcutaneously ("SC") at the base of the tail and intraperitoneally ("ip"), with 50 μg synthetic 12-26 peptide and 25 μg hen egg lysozyme (HEL, SIGMA) emulsified in Freund's complete adjuvant (CFA, SIGMA).

Mice received an additional antigenic boost of 50 μg peptide and 25 μg HEL injected ip in saline 2 weeks after initial priming. Mice were bled to assess serum anti-peptide antibody responses 8 days after this boost. Splenic memory T cell ("tertiary") responses were analyzed in culture 8 weeks following secondary boosts. Splenic T cells were enriched by panning on anti-Ig coated plates, and restimulated (3x10⁶/ml) with dilutions of peptide and irradiated splenic APC (2500 rads, lOVml). For analysis of secondary (LN) responses following iv tolerization, animals were immunized in hind footpads with 20 μg peptide emulsified in CFA, and draining popliteal LNs were harvested 9 days later and restimulated in culture with dilutions of peptide and 50 μg/ml purified protein derivative (PPD, Connaught, Swiftwater, PA). IL-2 and IL-4 secreted into the medium were determined from culture supernatants at 24 and 48 hours, respectively, in LN or splenic T-cell cultures using recombinant cytokines as standards.

In vitro B-cell tolerance induction experiments were done on enriched splenic B cells essentially as described by before (Waldschmidt, T.J. et al. (1983) J. Immunol.

757:2204-2209; Phillips, N.E. et al, (1983) J. Immunol. 750:602-606; Warner, G.L. et al, (1991) J. Immunol. 746:2185-2191). Supernatants from 3-4 day cultures were assayed for IgM production by ELISA by coating wells with peptide- or FITC-BSA conjugates, and probing with goat anti-mouse IgM-AP. Results represent experiments repeated 2-3 times; individual points represent the arithmetic mean of triplicate or quadruplicate values with standard deviations generally less than 15% (omitted for clarity). RESULTS

Preparation and characterization of a murine IgG, self-carrier containing the λ cl repressor 12-26 peptide at the V_H N-terminus

The 12-26-IgG construct was prepared by modifying plasmid pSNR, which contains the genomic sequence encoding a murine IgG,^b H chain. Isologous IgG, was chosen because of its documented activity as a tolerogenic "carrier" of potency equal to IgG₂ and greater than other Ig isotypes or other serum proteins. We chose to insert a foreign epitope at the N-terminus of the V_H region (Figure 5 A), because insertions at this location have been shown not to alter normal immunoglobulin folding and structure (Hebell et al, supra; DaPorta et al, supra). Analysis of transfected, purified chimeric 12-26-IgG (Q3) or control pSNR IgG (P6) by SDS-PAGE showed that H chains can successfully pair with J558L light chains (L). The chimeric H chain (Q3) containing the additional 12-26 sequence was about 1.8 kDa larger than the control IgG (P6). Purified transfected IgGs expressed the 12-26 epitope as shown by western blotting and ELISA utilizing peptide-specific mAb B3.11 (Figure 5B). Furthermore, in competitive inhibition ELISA, chimeric 12-26-IgG effectively competed with free synthetic peptide or a chemical conjugate of 12-26 with rabbit IgG for binding to mAb B3.11. These results suggest that the inserted peptide is recognized efficiently by epitope-specific antibodies/B cells on the exterior surface of the recombinant IgG, without significantly perturbing H chain tertiary structure.

Additionally, the recombinant 12-26-IgG chimera is immunogenic and capable of priming 12-26-specific T and B cells in vivo. Mice immunized with Q3 emulsified in CFA were able to prime 12-26-specifιc T cells comparable to the response elicited with synthetic peptide. In vitro restimulation of LN cultures with synthetic peptide resulted in T-cell proliferation as well as IL-2 and IL-4 production in peptide- and Q3-primed, but not P6-primed LN cells. Immunization also led to a high serum anti- 12-26 IgG antibody titer detectable by peptide-specific ELISA. 12-26-IgG stimulated IL-2 production in an I-A^d-restricted 12-26 specific T-cell hybridoma (9C127). These results suggest that the confirmation of the inserted foreign epitope is not only recognized by specific antibodies, but the peptide (or one like it ) can also be processed and presented to T cells in a physiologically relevant manner by APC even in the context of a self-IgG scaffold.

In vivo induction of peptide-specific immune self tolerance with soluble purified engineered 12-26-IgG.

To test the efficacy of genetically engineered tolerogens, we analyzed both humoral and cellular immune responsiveness after iv administration of high doses of soluble, deaggregated 12-26-IgG fusion Ig. Mice were injected with 1 mg of either chimeric Q3 or control P6 immunoglobulins, and challenged 10 days later with a mixture of 12-26 peptide and HEL (as a specificity control) emulsified 1 :1 in CFA. Secondary humoral immune responses were analyzed one week after an additional boost. Figure 6 shows that mice receiving pretreatments of Q3, but not control P6, were dramatically unresponsive to peptide challenge as assessed by ELISA of anti- peptide IgG, whereas control anti-HEL antibody titers were unaffected. Although the predominant Ig isotype in this anti-peptide response in Balb/c mice is IgG„ antibodies of all isotypes including IgG_2b were consistently diminished by the tolerogenic treatment with 12-26-IgG (Figure 6).

To test the potential of inducing unresponsiveness with peptide-Ig-transfected cells as a model for gene-therapy-based tolerogenesis, Balb/c mice received 3

(consecutive weekly iv injections of transfectomas secreting Q3 (or P6 control IgG); the cells had first been treated with mitomycin C. This protocol resulted in transient appearance in serum of the transfected IgG's at levels reaching at least 10-500 ng/ml (assessed by NIP-gelatin ELISA). This type of treatment resulted in diminution of peptide-specific humoral immune responses as well as reduction of LN cell proliferative responses.

Since unresponsiveness as measured by serum antibodies may results from tolerance of B-cells, T-cells or both, the cellular basis of the observed tolerance was analyzed by measuring T-helper (Th) cytokine responses 8 weeks after immunogenic challenge. Restimulation of splenic memory T cells (Figure 7) revealed that both

Thl-type (IL-2) and Th2-type (IL-4) responses were absent in tolerized mice, a result consistent with the observed lack of anti-peptide IgG_2b and IgG, antibodies(Figure 6), which are dependent on these Th cell subsets. The T cell response to peptide was diminished in 12-26-IgG pretreated animals when measured as short-term LN restimulation assays . Mice tolerized with 1 mg of 12-26-IgG 10 days previous to peptide challenge had reduced LN IL-2 responses, but unaffected recall proliferative responses to the antigen PPD compared to control P6-injected animals. These results indicate efficient induction of Th cell tolerance to the immunodominant peptide which results in an inability to prime any subset of T cell response to the peptide.

Thus, a foreign immunogenic peptide genetically engrafted into an Ig scaffold can be very efficiently presented to the immune system in a tolerogenic manner when administered by the appropriate route and method. Thus pretreatment with peptide-Ig chimeras delivered either as single high doses or via slow release by transfected autologous B cells have utility in achieving efficient epitope-specific manipulation of undesired T-cell responses. Analysis of a novel 12-26-flagellin immunogen for testing the efficacy of 12-26-IgG on B-cell tolerance.

To test the efficacy of the fig 12-26-IgG as a B-cell tolerogen, it was necessary to challenge B cells with an immunogenic, T-independent form of the 12-26 epitope.

Since polymerized flagellin is a well-characterized T-independent antigen, we constructed a 12-26 flagellin fusion protein with a strategy previously described

(Newton et al, supra). Western blotting and ELISA analyses of purified WT (pPX) and 12-26-flagellin (A29) showed that although flagellin epitopes are readily expressed in both recombinant flagellins, the inserted epitope was detectable only in chimeric flagellin A29. Polymerized 12-26-flagellin stimulated splenic B cells to secrete anti- 12-26 IgM antibodies. The stimulatory effect was comparable to that of the polyclonal B cell mitogen, bacterial Hpopolysaccharide (LPS). A concentration of 0.1 μg/ml was found to be minimally mitogenic (as assessed by anti-fluorescein [FITC] IgM ELISA's) and used for subsequent experiments. These results broaden the context in which the inserted epitope can be recognized: IgG and the polymerized flagellin molecule. In the latter context, the epitope readily stimulate B cells to produce epitope-specific IgM antibodies. We also tested the ability of 12-26-IgG to induce specific B-cell unresponsiveness. Enriched B cell populations were incubated in vitro with various doses of Q3 or P6 control IgG's, washed, and then cultured with either mitogenic LPS or 12-26-flagellin. Alternatively, BALB/c mice were injected iv with 1 mg of each protein, and splenic B cells were harvested and challenged in vitro 10 days later.

Supernatants from 3 day cultures were assayed for 12-26-specific IgM antibody or anti-FITC antibody as a specificity control. Pretreatment with Q3, but not P6, either in vitro or in vivo, markedly suppressed the anti- 12-26 IgM response, whereas anti- FITC control IgM responses were unaffected. Thus, in addition to inducing potent Th tolerance, the fig construct is independently can induce epitope-specific unresponsiveness in B cells. The magnitude of B cell tolerance was more modest in vivo than T cell tolerance, possibly reflecting either a requirement for higher epitope valency (the fig provides only a bivalent epitope, one on each arm of the H chain), or a higher dose requirement. A reduction of antibody responsiveness of similar magnitude was observed after adoptive transfer of in v/vo-tolerized B cells, admixed with nontolerized naive T cells, into secondary immunodeficient recipients which were then challenged with the peptide in CFA. DISCUSSION

The development and maintenance of the unresponsive state in newly emerging lymphocytes is a lifelong process requiring the persistence of antigen.

Exposure of mature B and T cells to antigen in an adult immune system may lead to either activation or tolerance depending on the route and method of exposure, as well as the availability of costimulatory signals from specialized APC. Since a major goal in clinical therapy in a variety of conditions (e.g., infection, autoimmunity, allergy, transplantation) is the induction of specific immune unresponsiveness in adult mature lymphocytes, a variety of approaches have exploited these pathways of exposure. Of these approaches, experimental tolerance induction with gamma-globulin carriers has been most extensively described. IV administration of soluble, deaggregated IgG's in the absence of adjuvants, induces both antigen-specific B-cell and T-cell tolerance even in the absence of a thymic environment. Mechanisms of specific clonal anergy/inactivation and deletion have been implicated in this type of experimental model.

In the foregoing studies, the present inventors described for the first time, the tolerogenic capability of an engineered self-IgG expressing a model class II MHC- restricted immunodominant peptide. This novel epitope, deliberately expressed at the

N-terminus of an IgG heavy chain construct, was tolerogenic in vivo and in vitro. Conceptually similar approaches have been utilized to express immunogenic (rather than tolerogenic) malarial or viral peptides in the CDR3 loop of Ig H chains for the induction of enhanced anti-peptide immune responses, as described above. As with other such epitopes, the 12-26-IgG protein could act as an efficient immunogen when administered in an immunogenic manner (i.e., emulsified in CFA). Zaghouani et al, 1993, supra showed that T-cell activation (for a class Il-restricted epitope) was enhanced 100-1000 fold when the epitope was part of an Ig-chimera, presented in vitro by stimulatory dendritic cells as APC. The present results similarly show that an approximately 100-fold lower molar quantity of 12-26-IgG (as compared to free peptide) stimulated similar numbers of peptide-specific LN T cells from immunized mice.

The increased efficacy of the fig's of the present invention, both as immune activators and as tolerance inducers, may indicate that common pathways are utilized. The increased efficacy may directly result from (a) an increased half-life and (b) an

Fc-receptor mediated uptake of the "carrier" portion of the Ig molecule (Stockinger, B. (1992) Eur. J. Immunol. 22:1271-1278) leading to improved presentation of the grafted foreign epitope(s). In the absence of adjuvants (which act in part by mobilizing APC having efficient costimulatory capability), high doses of soluble, deaggregated serum protein may be preferentially taken up by "non-professional"

APC, such as resting B cells, via the process of Fc receptor-mediated endocytosis or phagocytosis, and subsequently presented by these "non-professional" APC (Parker, D.C et al. (1991) FASEB J. 5:2777-2784; Eynon, E.E. et al. (1992) J. Exp. Med. 775:131-138; Fuchs, E.J. et al. (1992) Science 255:1156-1159). Furthermore, IgG carriers can induce efficient B-cell unresponsiveness by mechanisms involving the crosslinking of surface IgM to Fc receptors. One or more of the above mechanisms may be responsible for the enhanced tolerogenic efficiency of Ig carriers. In contrast, the mere iv injection of soluble, deaggregated peptides can suffice to induce effective Th cell tolerance (Scherer, M.T. et al. (1989) Cold Spring Harbor Symp. Quant. Biol. 54:497-504), but is insufficient to induce specific B cell unresponsiveness.

The present inventors findings are summarized as follows. A fig, specifically the 12-26-IgG fusion protein, can present an epitope in a tolerogenic fashion and induce both B- and T-cell tolerance. A convenient property of this epitope allows simultaneous study of both cellular and humoral immune responses to a single immunodominant peptide. The 12-26 peptide can induce a vigorous antibody response which is predominantly of the IgG, isotype, and can prime Th cells of both the Thl and Th2 phenotype. Tolerance induction with 12-26-IgG was globally effective in suppressing every type of immune response which can be elicited by this immunodominant peptide. The inventors have therefore provided a powerful approach to determining the efficacy of inducing specific unresponsiveness to a defined antigens, particularly peptide antigens, for the modulation of undesired immune responses. The present approach has advantages of that inserting heterologous epitopes into the H chain CDR3 because the N-terminus insertion does not restrict the size of the epitope or epitopes fused to the tolerogenic IgG carrier. Therefore, not only short peptides, but also larger, more complex foreign antigens may be fused in an fig construct for tolerogenic presentation.

Finally, because this approach provides what may be envisioned as a genetically transmissible "hapten-carrier" complex, these tolerogenic figs when expressed as a transgene-transferred into hematopoietic tissues or cells, can be used to both induce and maintain tolerance for the long term. Such studies are reported in Example III, below. Recipients of BM stem cells which have been transduced with a retroviral vector for the long-term expression of fig cDNA constructs. The application of most immediate interest for the present invention is the use of this approach to block and ineffective and potentially harmful antibody responses which occur during HIV infection (Clerici, M. et al, (1993) Immunol. Today 74:107-110).

EXAMPLE II Tolerance to HIV gpl 20 Epitopes from the C5 Region: Detection

The study described above using a phage λ epitope was extended to two gpl 20 epitopes:

(1) a C5 peptide KYKVVKIEPLGVAPTKAKRRVVQREKR (SEQ ID NO: 199) positions 485-511 of gpl20 from the BH10 strain (see Figure 3) which is cross- reactive with HLA-C monomoφhic determinants (DeSantis, C. et al. (1993) J. Infectious Dis. 765:1396; Palker, T.J. et al. (1987) Proc. Nat'l Acad. Sci. USA 54:2479; and

(2) a Cl immunodominant peptide (such as residues 90-120 of the BH10 isolate, above, that contains distinct B- and T-cell epitopes (Abacioglu et al. (supra).

Many antibody responses to HIV can be non-protective, and can enhance viral uptake by monocytes or promote T-cell apoptosis (Finkel et al. , supra; Banda et al. , supra; Kliks, S.C. et al (1996) Proc. Nat'l Acad. Sci. USA 90:11518. Initially, the inventors focused on defining peptides smaller than the original 35-mer in C5 defined by Beretta and colleagues (DeSantis, C et al. (1993) J. Infectious Dis. 765:1396.

Using mAbs (Robinson, W.E. Jr. et al. (1990) Proc. Nat. Acad. Sci. USA 57:3185, peptides were examined that contain either the KYK or KAKRR (SEQ ID NO:200)motifs that have been defined for HLA cross-reactivity. At least two 15-18- mers were identified by an ELISA inhibition assay. Results discussed in Example I, above, indicated that, in contrast to free peptide, λ 12-26-IgG was tolerogenic for B cells as a bivalent molecule, although the mechanism of this unresponsiveness was not determined. Design of shortened peptides containing major C5 epitopes:

As a model epitope, the inventors initially chose the C-terminal peptide KYKVVKIEPLGVAPTKAKRRVVQREKR (SEQ ID NO: 199) (residues 485-51 in the BH10 variant (Figure 3) and which ; corresponds approximately to positions 455-481 of the consensus sequence in Figure 1 ). This is in the conserved C5 C-terminal region of gpl20. This peptide contains the B-cell epitope consisting of the KYK KAKRR (SEQ ID NO:200)motifs that are recognized by the M38 murine mAb (DeSantis et al, supra; Palker et al, supra). The epitope recognized by M38 has been noted to be KYKVVKEIPLGVAPTKAKRR of SEQ ID NO: 199. MAb M38 also binds to the C-terminus of gpl20, in a gp41 binding region. M38 also reacts with a common motif in the HLA-C heavy chain al region (KYKRQAQADRVNLRKLR; SEQ ID NO:201) that is mimicked in this C5 peptide. HIV-infected individuals have HLA class I-gpl20 cross-reactive antibodies.

The inventors first established that a 35-mer containing this M38-defined epitope was tolerogenic in vivo when chemically coupled to heterologous rabbit IgG. Since the C5 peptide was relatively large and not readily available. Shorter peptides containing the KYK and KAKRR sequences with different spacer residues and with a C-terminal cysteine for more controlled coupling to IgG carriers can be designed.

Although the residues between the two M38 epitopes (IEPLGVAPT; SEQ ID NO:202) are not recognized in seropositive individuals (Scott, D.W. et al. (1993) Adv. in Molec. and Cell. Immunol. 7:119, it was important to determine the contribution of these amino acids to the epitope conformation. The mAbs described in Robinson et al. , supra, were used to analyze reactivity to these new peptides as well as the requirement for the intervening sequences. The peptides designed were:

1. AAKYKGGGGGKAKRRGGC (SEQ ID NO.203)

2. AAKYKGGGPTKAKRRGGC (SEQ IDNO.204)

3. AAKYKGVAPTKAKRRGGC (SEQ IDNO:205) Control peptides (for example, available from the National Institute of Allergy and Infectious Diseases) encompassed the KYK, KAKRR (SEQ ID NO:200) motifs or the entire IEPTGVAPTKAKRR (SEQ ID NO:206)sequence recognized by the human mAbs. Using a competitive ELISA assay, the inventors found that peptide #3, above ("P3"), was similar in activity to the full sequence and that peptides containing only the KYK motif were noninhibitory, as expected, with these human anti-C5 mAbs. Treatment with P3-HGG induced unresponsiveness for an anti-P3 response (and an anti-HGG response). These results suggest that VAPT motif, while not immunoreactive, contributes to the structural integrity of this epitope. Peptide 3 (P3) contains both motifs that have been reported to be recognized by M38 and are important in anti-HLA-C recognition, for which tolerance induction is one goal.

Importantly, these results mean that shorter peptides still express B-cell epitopes.

It is expected that anti-Cl and V3 loop antibody responses would not be affected by the above peptides because the stimulus is a polyclonal mitogen. It is expected that anti-HLA crossreactivity will be eliminated if KYK-specific unresponsiveness is induced. In order to achieve tolerance to Cl region or V3 loop epitopes, the present invention would require that the fig include one or more epitopes from these regions.

EXAMPLE III

Resting and Activated B Lymphocytes Expressing fig are Tolerogenic Vehicles Since antigen-presenting B-lymphocytes are known to either augment or downregulate T-cell dependent immunity , it should be possible to modulate the immune response to a selected antigen (such as an autoantigen, a viral antigen or a tumor antigen) via gene-transfer of exogenous genes and constitutive expression in vivo by autologous APC. Such an approach would be advantageous for the induction of unresponsiveness, since tolerance to foreign antigens could be maintained indefinitely in vivo, especially if gene-transfer into long-lived lymphoid progenitors is achieved.

Previous models have led to apparently divergent results and have shown that B cells can be either essential (Ron, Y. et al (1981) Eur. J. Immunol. 77:964-968; Janeway, C.J. et al. (1987) J. Immunol. 755:1051-1055; Constant, S. et al. (1995) J.

Immunol. 755:3734-3741; Morris, S.C et al. (1994) J. Immunol. 752:3777-3785) or nonessential (Sunshine, G.H. et al (1991) J. Exp. Med. 774:1653-1656; Ronchese, F. et al (1993) J. Exp. Med. 777:679-690; Epstein, MM et al. (1995) J. Exp. Med. 752:915-922) for T cell priming, and can be critical for either activating or tolerizing (Eynon, E.E. et al. (1991) Transplant. Proc. 23:129-130; Eynon, E.E. et al. (1992) J. Exp. Med. 775:131-138; Fuchs, E.J. et α/. (1992) Science. 255:1156-1159; Buhlmann, J.E. et al. (1995) Immunity. 2:645-653) naive T cells, and even previously activated T- cell clones (Gilbert, K.M. et al. (1994) J. Exp. Med. 179:249-258). The state of activation of the collaborating B cells and T cells as well as the antigen-specificity for the interactions have appeared to be important for the different outcomes.

To analyze the ability of antigen-presenting B cell to serve as a modulator of the immune response, the present inventors generated a unique transgenic mouse system (see Example I) in which a foreign class Il-restricted immunodominant epitope is expressed as a self antigen specifically in the B cell compartment. The foreign epitope, residues 12-26 from λ cl repressor protein was grafted in-frame at the N- terminus of a murine IgG, heavy chain and is made endogenously as a transgene in the B-lymphocyte lineage. The tolerogenic capabilities of this soluble engineered immunoglobulin in immunocompetent adult mice is described above. The present study describes the tolerogenic nature of transgenic hematopoietic tissue expressing such a fig molecule. This approach takes advantage of the efficiency of the immunoglobulin secretory and endocytic pathways to synthesize and present an exogenous "neo" self-peptide, and provides a model for inducing peripheral tolerance to undesirable humoral and cellular immune responses using gene therapy strategies. I. Materials and Methods Mice and Reagents. Male and female B6D2 (H-2^b/d) and B ALB/cByJ (H-2^d) mice were purchased from the Jackson Laboratories (Bar Harbor, ME) at 3-8 weeks of age, and housed in pathogen-free, microisolater cages. RPMI 1640 medium (GIBCO- BRL, Gaithersburg, MD) was supplemented with either heat-inactivated 5% FCS (Hyclone, Logan, UT), or heat-inactivated 0.5% autologous mouse serum (Jackson Immunochemicals), 2-ME, L-glutamine, penicillin, streptomycin, MEM nonessential amino acids, and sodium pyruvate. Hybridoma B3.11 , which produces an IgG, specific for the 12-26 peptide was from Drs. T. Briner and M.Gefter (Immulogic Coφ., Waltham, MA), and was originally derived by fusion with splenocytes from peptide-immunized BALB/c mice. Monoclonal antibody (mAb) B3.11 was affinity purified from bulk-cultured supernatants with goat anti-mouse IgG sepharose columns and biotinylated. All alkaline-phosphatase (AP)-conjugated secondary reagents were purchased from Southern Biotechnology Assoc. (Birmingham, AL). The 12-26 15- mer LEDARRLKAIYEKKK (SEQ ID NO:l 12), or an N-terminal cysteine-modified 16-mer was prepared with a solid-phase method and purified to >92-95% homogeneity using standard HPLC methods. The cysteine-modified 12-26 peptide was covalently conjugated to hen egg white lysozyme (HEL) with Sulfo-MBS (Pierce, Rockford, Illinois), a sulfhydryl-specific crosslinking reagent.

Generation of Transgenic (Tg) Mice. The preparation of a chimeric murine IgG,^b H chain construct, specific for the NP hapten and engineered to express the 12- 26 peptide at the N-terminus, is described in detail above. The entire ~10 kb genomic construct containing the original endogenous immunoglobulin promoter, enhancer, and polyadenylation sequences was shuttled into pBluescript KS+/- (Stratagene, La Jolla, CA) and excised as a XhoI/NotI fragment. The linearized transgene was purified over a continuous 10-40% (wt vol) sucrose gradient and dialyzed against 5 mM Tris-HCI/0.15 mM EDTA (ph 7.5). Tg mice were derived by pronuclear injection of fertilized B6D2 eggs, and implantation into pseudopregnant females as described by Hogan et al. (Hogan, B. et al. (1986) Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor Lab. Press, Plainview, N.Y. pp. 81-141, incoφorated by reference). Three original Tg founders were identified by genomic Southern blotting of tail DNA wid a ³²P-labeled probe containing 3 cloned, tandem copies of the 12-26 cDNA sequence. Two of these founders (Line 5 and Line 17) were selected for further analysis, bred onto the BALB/c background for at least 5-10 generations, and confirmed for H-2^d homozygosity via RFLP Southern blot analysis before use in BALB/c adoptive transfer experiments. Lines 5 and 17 were also rederived by Cesarean section (Taconic Labs) and thereafter housed in sterilized microisolater units at the Holland Laboratory to ensure healthy microorganism-free strains of Tg mice. Tg offspring obtained via BALB/c matings were, in general, heterozygous for their transgene and distinguished from their nontransgenic (NTg) littermates by either 12-26 sequence Southern blotting of tail DNA, or serum NIP- binding IgG, ELISA. Preparation of Bone Marrow Chimeras. Eight week-old BALB/c recipients were sublethally irradiated (650 rads) with a ¹³⁷Cs source and injected iv with 10⁷ cells consisting of a 1 :1 pooled mixture of non-Tg/Line 17 Tg bone marrow (BM) cells that had been depleted of erythrocytes. Control mice were injected with NTg littermate BM cells(after 650 rads) or saline only (with no irradiation). All Tg/NTg donor BM was completely sex-matched and syngeneic with BALB/c recipients. Adoptively- transferred mice were rested for 7-8 weeks before immunization studies.

Preparation of Lymphoid Cells for Tolerance Induction. 12-26 peptide- specific tolerance induction in normal adult (6-10 week old) BALB/c mice was accomplished by iv injection of Line 17 Tg hematopoietic tissue. Unconditioned recipients were generally injected with 2-4 x 10⁷ cells from preparations of purified resting B cells, LPS-activated B cell blasts, unfractionated splenocytes, or crude BM cells from Tg or control NTg donors. Ten days following such injections, recipients were immunized with antigens SC as described below. Bone marrow (from both femurs and tibiae) or spleen tissue was prepared in serum-free complete RPMI and depleted of erythrocytes. Splenic B cells were obtained by depleting splenocytes of T cells by treatment with anti-T cell cocktail plus baby rabbit complement. Resting B cells were harvested by further fractionation on Percoll gradients and collecting the 60-70% layers as previously described (29). For preparation of activated B cell populations, purified B cells (4xl0⁶/ml) were incubated for 48 hrs in complete RPMI (5% FCS) in the presence of 50 μg/ml LPS (Sigma, St. Louis, MO), and washed 3 times before further use. For chemical fixation, purified B cells were treated with carbodiimide (ECDI, Sigma) by incubating 10⁸ cells in 0.5 ml of 75 mM ECDI (in saline) for 1 hour, on ice. All cells were washed extensively prior to iv injection.

Measurement of Peptide-Specific Cellular and Humoral Immune Responses. All protocols are essentially as described previously (Lai, M-Z et al. (1987) J. Immunol. 759:3973-3980; Scherer, M.T. et al. (1989) Cold Spring Harbor Symp. Quant. Biol. 54:491-504; Soloway, P. et al. (1991) J. Exp. Med. 774:847-858). Ten days following iv tolerization, animals were immunized to induce cellular or humoral immune responses. For measurement of humoral immunity, animals were injected SC at the base of the tail with 50 μg synthetic 12-26 peptide emulsified 1:1 in complete Freund's adjuvant (CFA). In some experiments, animals were also injected with 20 μg hen egg lysozyme (HEL) in CFA, intraperitoneally (ip). Two weeks later, mice received an additional antigenic boost of 50 μg peptide and 10 μg HEL in saline, injected ip. Antibody titers were determined from serum obtained 8 days after secondary boosts. Splenic memory T cell responses were measured in vitro 6-8 weeks following these secondary challenges.

For more detailed analysis of cellular immune responses, animals were immunized SC in the hind footpads with 20-50 μg peptide in CFA, and draining popliteal and inguinal lymph nodes (LN) were harvested 9 days later. LN cells were restimulated in vitro with synthetic peptide or 25-50 μg/ml purified protein derivative (PPD, Connaught, Swiftwater, PA) in complete RPMI with 0.5% heat-inactivated autologous mouse serum (Jackson Immunochemicals, West Grove PA). On day 3, cultures were pulsed with 1 μCi/well of [³H]thymidine and incubated an additional 14-

20 hours, for the determination of proliferative responses. Cells were then harvested on glass fiber filters and incoφorated ³H was detected using a direct beta counter (Packard, Matrix 9600). IL-2 and IL-4 cytokine production was quantitated using CTLL and CT.4S bioassays, respectively, testing serial dilutions of culture supernatants. Recombinant IL-2 (Genzyme, Cambridge, MA) and IL-4 (from Dr.

William Paul, NIH) were used to generate standard curves. Dilutions of anti-IL-2 mAb S4B6 and anti-IL-4 mAb 11B11 (ATCC, Rockville, MD.) were included in the assays to establish cytokine specificity.

For determination of B-cell tolerance induction, Tg mice or adoptively transferred recipients were immunized with a chemical conjugate of cysteine-modified

12-26 and HEL (12-26-HEL). Tg or NTg control mice were immunized ip with 50 μg 12-26-HEL emulsified 1 :1 in CFA and then boosted with 10 μg of the same conjugate in saline 2 weeks later. Titers of IgG antibodies specific for the peptide- or HEL were determined by ELISA 8 days following this boost. Irradiated (400 rad) BALB/c recipients were adoptively transferred (iv) with 5 x 10⁷ splenocytes from previously tolerized BALB/c, and boosted ip with 100 μg 12-26-HEL conjugate in incomplete Freund's adjuvant (IF A). Serum bleeds were collected 8 days following this boost, and antibody titers determined by ELISA.

Antigen-Presentation Studies. The ability of Tg B cells to directly present endogenous 12-26 peptide was assessed with T-cell hybridoma 9C127 which recognizes 12-26 peptide in the context of 1 A^d. Tg or control littermate B cell APC were purified as described above, and recultured in varying numbers in 200 μl microcultures with 10⁴ 9C127 cells/well in complete RPMI with 5% FCS. Supernatants were harvested 48 hours later, and multiple dilutions were assayed for IL-2 production as above.

Immunologic Methods. ELISA determinations of serum peptide-specific or HEL-specific IgG responses were performed by coating plates with 50 μg/ml synthetic peptide or 5 μg/ml HEL and following standard ELISA protocols. Briefly, antigen-coated plates were blocked with 1% gelatin/0.05% Tween 20 buffer, and duplicate serial dilutions of serum were incubated and probed with goat anti-mouse

IgG isotype-specific secondary reagents conjugated to alkaline phosphatase. Titers are expressed as the geometric mean of the reciprocal dilution required to bring A₄₉₀ readings to prebleed levels or <0.09 O.D.

12-26-IgG H chain protein was detected in serum of Tg mice via its ability to bind to the NIP (5-iodo-4-hydroxy-3-nitrophenylacetyl) hapten using a modified NIP- binding ELISA (Grosschedl, R. et al. (1984) Cell. 55:647-658). Dilutions of sera from Tg mice were incubated on ELISA plates coated with NIP-gelatin or NIP-BSA conjugates (10 μg/ml), and subsequently probed with goat anti-mouse IgG,-AP as a secondary reagent. Detection of the 12-26 epitope in Tg sera could be demonstrated by similarly using NlP-sepharose beads (from Dr. T. Imanishi-Kari, Tufts University) to immunoprecipitate 12-26-IgG. Samples were boiled in 2X SDS loading buffer, electrophoresed on 10% SDS-PAGE, and transferred onto nitrocellulose filter in a buffer with 25 mM Tris, 192 mM glycine, and 20% methanol, pH 9.0 at 20°C overnight. Blots were blocked in 2% BSA in TBST (50 mM Tris, 200 mM NaCl, pH 7.5, 0.05% Tween 20), and probed with biotinylated mAb B3.11 (anti-12-26 epitope) plus streptavidin-AP as a secondary reagent.

FACS analysis. Cells were stained for surface antigens and analyzed by flow cytometry using standard methods. The following conjugated antibodies were from commercial sources and used with appropriate fluorochrome-labeled isotype/species matched or secondary reagent controls: RA3-6B2, rat anti-mouse Ly5 (B220)-PE

(Caltag, San Francisco, CA), YTS 191.1, rat anti-mouse L3/T4 (CD4)-PE (Caltag),

YTS 169.4, rat anti-mouse Ly-2 (CD8)-FITC (Caltag), goat anti-mouse IgGl

(adsorbed)-PE (Caltag), goat anti-mouse IgM (H+L)- FITC (Hyclone). Data was acquired on a Becton Dickinson FACScan and analyzed with LYSIS II software.

II. RESULTS

Generation of Transgenic Mouse Lines Expressing a Novel Peptide-IgG, Construct Specifically in the B-Lymphocyte Compartment

The unique tolerogenic properties of an engineered peptide-IgG fusion protein expressing residues 12-26 of λ cl repressor protein at the N-terminus of a murine H chain specific for the NP hapten is described above. The engrafted epitope is recognized on the exterior surface of assembled IgG by a peptide-specific mAb (B3.11). More importantly, soluble fusion protein administered in adjuvant can efficiently generate peptide-specific T-cell responses in vivo, suggesting that 12-26 (or an extremely similar peptide) is processed and presented by endogenous APC, even in the context of an Ig scaffold. To further characterize the potential for expressing tolerogenic IgG fusion proteins in vivo, we generated Tg mice expressing the engineered genomic (rearranged) H chain construct driven by its endogenous immunoglobulin promoter/enhancer sequences (Example I). Tg founders possessing 2-3 integrated copies were identified via Southern blotting of genomic tail biopsy

DNA using a cDNA probe containing 12-26 sequence (Figure 8). Two lines (5 and 17) were bred onto the BALB/c (H-2^d) strain and further analyzed for expression of engineered IgG. Unlike Tg mice expressing rearranged IgM constructs, both lines had no apparent suppression of endogenous Ig rearrangements and expressed amounts of surface IgM, serum levels of IgM and IgG, as well as B220, CD4, and CD8 markers that were comparable to NTg littermates. These results are in agreement with previous observations for IgG H chain Tg experiments (Storb, U. (1987) Ann. Rev. Immunol. 5:151 -174; Yamamura, K. et al. (1986) Proc. Natl. Acad. Sci. USA. 55:2152-2156.; Tsao, B.P. et al. (1992) J. Immunol. 749:350-358; Radic, M.Z. et al. (1995) J. Immunol. 755:3213-3222).

Serum expression of the NP-binding Tg H chain was detected as described by Grosschedl et al, supra. Since the Tg V_H region binds with high affinity to the NIP hapten in combination with λl light chains, functional 12-26-IgG was detected indirectly with a NIP-binding IgG, ELISA Although probably representing a fraction (only λ light chain-associated) of expressed Tg serum protein, NIP-binding IgG, assays revealed that Line 17 and 5 expressed between 1000-25000 ng/ml and 50-1000 ng/ml, respectively. The higher serum expression for Line 17 mice correlated with increased expression of surface IgG, in splenocytes as compared to Line 5 or NTg littermates. Direct presentation of endogenously synthesized 12-26 by Tg B cells was demonstrated with a peptide-specific hybridoma (9C127). More importantly, 12-26 peptide expression was demonstrated directly via immunoprecipitation of Tg serum with NlP-sepharose beads and immunoblot analysis with mAb B3.11. B-lymphoid expression of the transgene mediated by a specific Ig promoter/enhancer was demonstrated as significantly increased expression following activation by bacterial

LPS, but not by Con A stimulation of splenocytes. These results collectively show that the Tg 12-26-IgG, H chain readily combines with endogenously-synthesized light chains to be expressed specifically as a self molecule by a large fraction of B cells without perturbing expression of endogenous IgM rearrangements. Thus, the 12-26 peptide is secreted as well as processed and presented via the Ig endocytic pathway as an endogenous B cell self-epitope.

Profound Peptide-Specific Cellular and Humoral Immune Tolerance in Mice Expressing 12-26-IgG, During Development or in Adult Bone Marrow Chimeras

Tg mice expressing foreign "neo" self-antigens have firmly established that tolerance induction can readily occur for membrane-bound and soluble proteins which are expressed ubiquitously, or in a tissue-specific manner during normal development (Goodnow, CC (1992) Ann. Rev. Immunol. 70:489-518; Miller, J.F.A.P. et al (1992) Ann. Rev. Immunol 70:51-69). In our initial experiments, we wished to establish that Tg animals expressing a model immunodominant epitope as part of serum IgG protein were similarly unresponsive to an immunogenic challenge with the epitope. Since the

12-26 peptide contains both a T-cell and a B-cell epitope, we could measure both cellular and humoral immune responses to this relatively simple determinant with immunization assays. Draining LN cells from subjects who received SC injection of synthetic peptide in CFA and which were subsequently restimulated with antigen displayed a profound proliferative unresponsiveness and IL-2 production.

Furthermore, NTg mice (H-2^d) primed with peptide in adjuvant and followed by a subsequent boost of peptide in saline 2 weeks later, developed an extremely high titer serum antibody response dominated by antibodies of the (Th2 -mediated) IgG, isotype (Soloway. et al, supra). Profound humoral unresponsiveness was observed in Tg animals immunized in this manner (Fig. 9A). This could not be due to immune- complex binding with circulating serum fig (12-26-IgG) since these tolerant animals had diminished splenic memory T cell responses to 12-26 peptide (Figure 9B). The extent of cellular and humoral unresponsiveness was comparable for both Line 5 and Line 17 suggesting that even lower levels of expression (Line 5) efficiently satisfied antigenic thresholds for tolerance induction.

In the immunocompetent adult, solid induction of tolerance to foreign transplantation antigens or viral CTL epitopes has previously been shown to be most effective in subjects in whom hematopoietic or lymphoid ablation is followed by reconstitution with antigen-expressing BM-derived APC (Ildstad, S. T. et al. (1984) Nature 507:168-170; Cobbold, S.P. et al. (1984) Nature 572:548-551; Roberts, J.L. et al. (1990) J. Exp. Med. 777:935-940; Oehen, S.V. et al. (1994) Cellular Immunol. 755:342-352; Nemazee, D. et al. (1989) Proc. Natl. Acad. Sci. USA. 56:8039-8043). To ascertain whether tolerance to a class Il-restricted T cell and B cell epitope could similarly be achieved, we constructed Tg BM chimeras. Sublethally irradiated adult BALB/c mice were injected with 1 :1 mixtures of Line 17 Tg and NTg littermate BM, and the recipient's immune system was allowed to redevelop for 7-8 weeks in the presence of fig (12-26-IgG)-producing lymphoid tissue. Such treatment followed by immunogenic challenge with synthetic peptide revealed profoundly suppressed cellular and humoral (Figure 10) peptide-specific immunity in these normal adult recipients.

The magnitude of tolerance as well as serum NIP-binding IgG, levels observed for these BM chimeras was comparable to that observed in Tg animals expressing 12-

26 IgG continuously during ontogeny. More interestingly, even injection of nonirradiated subjects with large numbers of crude syngeneic Line 17 BM resulted in high levels of serum NIP-binding IgG, which could be detected as long as one year post-infusion.

Tolerogenicity of fig-Expressing Lymphoid Tissue Transferred to Unmanipulated Adult Subjects Although tolerance in the neonate or hematopoietically-ablated adult recipient is believed to involve thymic participation (central tolerance), we analyzed the potency of peptide-flg-expressing lymphoid tissue for the induction of tolerance in mature peripheral lymphocytes. To induce peripheral tolerance in an unmanipulated immune-competent adult, we injected iv various preparations of Line 17 Tg hematopoietic tissue into normal adult BALB/c subjects. We first compared the in vivo tolerogenic efficacy of injecting large numbers of resting vs. activated B cells. Resting B cells are known to be competent in antigen processing and presentation functions, but have been described to possess defective costimulatory ability, in contrast to LPS- or surface Ig-activated B cell blasts which express abundant B7-1, B7-2, and CD40. Suφrisingly, injection of a variety of different 12-26-IgG- expressing lymphoid preparations, including Percoll gradient-purified resting B cells, LPS-activated blasts, crude BM , and even crude splenocyte preparations, were all highly effective in diminishing humoral (Figure 11) and cellular immune responses to the 12-26 peptide in adult recipients. Analysis of B-Cell Tolerance Induction in Transgenic or Normal Tolerized Recipients

Although processed 12-26 peptide induced efficient tolerance in T cells, the bivalent epitope-containing IgG molecule could potentially induce B cell tolerance (Chiller, J.M. et al. (1970) Proc. Natl. Acad. Sci. USA. 65:551-556; Parks, D.E. et al. (1980) J. Immunol. 724:1230-1236; Tighe, H. et al. (1995) J. Exp. Med. 757:599-

606). To test for the potency of such an effect, line 5 and 17 Tg mice were challenged with peptide conjugated to a different carrier, hen egg lysozyme (HEL), as a source of T cell help for potentially tolerized B cells. Potentially self-reactive anti-peptide B cells can receive foreign-reactive T cell help from HEL-specific T cells to produce autoantibodies. Immunization with 12-26-HEL in adjuvant, followed by a boost, revealed that both Tg lines displayed B-cell tolerance, manifest as a reduction in anti- peptide IgG titers (Figure 12 A). In contrast, all animals expressed similar levels of anti-HEL IgG titters (>10⁵). Interestingly, although the lower expressing Line 5 was solidly unresponsive to 12-26 when challenged with peptide alone (Figure 12B), immunization with the conjugate revealed low but significant anti-peptide responses not evident in the higher expressing Line 17. Thus, although both lines are solidly tolerant at the T cell level, a more potent B cell tolerance appeared to have been associated with higher self-peptide concentrations in Line 17 mice.

We also assessed the effect on B-cell tolerance of transferring 12-26-IgG- expressing lymphoid tissue to normal, unconditioned recipients. Normal immunocompetent subjects that were rendered tolerant by various preparations of Line 17 lymphoid cells, displayed similar levels of humoral immune tolerance (Figure 11). Since BM cells, resting B cells, and activated LPS B cell blasts secrete or express on their surface varying amounts of the fig, it is conceivable that, although they are all effective in inducing T-cell tolerance (via similar tolerogen presentation pathways), they may have differing effects on B cell tolerance induction. Thus, we adoptively transferred splenocytes from previously-tolerized animals immunized with 12-26 peptide and HEL (as a specificity control), and which had previously displayed unresponsiveness (Figure 11), into irradiated BALB/c recipients. These recipients were boosted with peptide-HEL (to stimulate T-cell help for potentially tolerized B cells). Anti-peptide IgG titers were measured. Suφrisingly, although all animals displayed relatively lower titers than non-tolerant controls, recipients of splenocytes from LPS-blast tolerized donors appeared to have a more profound B cell tolerance (Figure 12B). Splenocytes from tolerized animals were also restimulated in vitro with peptide and analyzed for IL-2/IL-4 cytokine responses. This experiment confirmed that the various lymphoid treatments resulted in similarly decreased cytokine responses , and thus similar T-cell tolerance induction as was suggested by results described above. Induction of Tolerance in Previously Primed Adult Subjects

Various experimental models of tolerance induction have established that it is possible to diminish specific immunity in a naive, antigen-inexperienced recipient (Eynon, E.E. et l. (1991) Transplant. Proc. 25:729-730; Eynon, E.E. et al. (1992) J. Exp. Med. 775:131-138; Fuchs, E.J. et al. (1992). Science. 255:1156-1159). In contrast, inducing unresponsiveness in an antigen-primed (immunized) adult has been more difficult (Fuchs et al, supra; Eynon, E.E. et al. (1993) J. Immunol. 757:2958- 2964).

We tested the ability of tolerogen-synthesizing B cells to modulate an ongoing immune response by SC immunization of recipients with peptide in adjuvant and waited 1-2 weeks before injecting of Line 17 Tg B cell preparations as tolerogen.

Primed recipients received one of four preparations: (1) Percoll purified resting B cells, (2) crude BM cells, (3) LPS-activated B cell blasts, or (4) chemically-fixed B cells. One week later, they were boosted with peptide in saline, and humoral immune responses were subsequently determined. Although both resting B cells and crude BM cells produce specific unresponsiveness in antigen-naive recipients, both were ineffective in diminishing peptide-specific humoral immunity in previously primed subjects (Figure 13 A). LPS- activated Tg B cells completely reversed the ongoing immune response (Figure 16B). A significant reduction in anti-peptide antibody titers was also produced by treatment with fixed Tg B cells (Figure 13C). Thus, the most potent tolerogenic treatment for already immune subjects, which caused them to become as unresponsive as antigen- naive subjects, was the infusion of activated fig-synthesizing B cells. Furthermore, diminution of anti-peptide antibody responses by activated or fixed fig-expressing Line 17 B cells was observed not only for total IgG levels. Rather antibodies of the IgG, isotype, a Thl -dependent response, and antibodies of the IgG_2b isotype, a Th2 dependent response, were also diminished, thus ruling out a possible "immune- deviation" or class-switching effect (Asherson, G.L. et al. (1965) Immunology. 9:206- 215).

III. DISCUSSION A variety of protein engineering strategies have established the efficacy of expressing heterologous epitopes in immunoglobulin frameworks for the enhancement of specific immunity (Billetta, R. et l. (1991) Proc. Natl. Acad. Sci. USA. 55:4713- 4717.; Zaghouani, H. et al. (1993) Science. 259:224-221). The present inventors have shown (see Example I) that similarly constructed peptide-Ig molecules can induce specific tolerance to a foreign immunodominant epitope. The above studies extend this approach to a novel strategy in which antigen-presenting B cell may be engineered to express immunoglobulins which contain within their structure tolerogens that can be employed to manipulate an undesired immune responses. Expression of otherwise immunogenic determinants in an IgG fusion protein which is synthesized, secreted, and also directly presented by lymphoid tissue, was shown to be a highly efficient tool for the induction of immune self-tolerance in mature, immunocompetent subjects.

In the model described herein , iv injection of any of a variety of crude or purified lymphoid cell preparations were effective for tolerance induction. We were suφrised to find that even unfractionated splenocytes were effective, since such a population includes dendritic cells which have been reported to prevent induction of unresponsiveness (to the H-Y antigen) (Fuchs et al. supra). This is most likely due to the restricted expression of the self-antigen in the present system to the B-lymphoid cell compartment whereas H-Y is expressed by B cells and other APC. Transgenic peptide-Ig chimeric molecules have the potential to be presented directly or secreted and re-presented, making it likely that tolerance induction by injected of peptide-expressing lymphoid tissue occurs via multiple pathways. This may also explain the potency of the fig tolerogens. Secretion of the fig tolerogen by activated transgenic B cells and re-presentation by non-transgenic APC may provide an additional tolerogenic pathway. This is supported by our observations that high doses of soluble peptide-IgG, or very low doses of fig from secreting transfected cells, upon injection d in vivo, are sufficient for inducing tolerance (Zambidis et al, supra). T-cell clonal deletion has been described in other transgenic models in which soluble self-Ig antigenic determinants were presented in the periphery or in the thymus (53,

64). Thus, although direct presentation of self-antigens by B cells may be sufficient for peripheral tolerance induction, additional pathways using other APC such as macrophages (especially for soluble IgG antigens) (Phillipis, J.A. et al. (1996) J. Exp. Med. 755:1339-1344) may also tolerize independently. BM-derived B cells or purified resting B cells are deficient in costimulatory function. Hence, for these cells, direct presentation of endogenously synthesized 12- 26 peptide (signal 1) is the most likely primary tolerogenic pathway. Such cells express very little membrane or secreted IgG,. Thus, a relatively low level of production of soluble and/or membrane fig capable of interacting with surface IgM molecules specific for the foreign epitope may explain why such cells were relatively less tolerogenic for the B cell compartment. In contrast, activated B cells with increased secretion of fig were more efficient B cell tolerogens, and were the only preparation we tested which could shut down an ongoing immune response.

Recently, high doses of antigen given to naive or even primed recipients was shown to cause clonal deletion of both peripheral T and B lymphocytes via mechanisms of programmed cell death, or apoptosis. Peripheral deletion of mature lymphocytes resulted from an exhaustive immune response ("propriocidal regulation") which was IL-2-dependent and mediated by the apoptosis-regulating surface molecules Fas and Fas ligand (Crispe, LN. (1994) Immunity. 7:347-349; Critchfield, J.M. et al. (1994) Science. 265:1139-1143; Singer, G.G. et al. (1994) Immunity. 7:365-371; Pulendran, B. et al. (1995) Nature 575:331-334; Shokat, K.M. et al. (1995) Nature 575:334-338; Lenardo, M.J. (1991) Nature 555:858-861).

One inteφretation of the present findings is that prolonged presentation of peptide/MHC complexes provided by live transgenic B cells, combined with sustained production and re-presentation of secreted fig activates those mechanisms responsible for high dose tolerance. The existence of an activation-driven apoptosis mechanism may explain our observation of tolerance induction in subjects during an ongoing immune response when treated with transgenic LPS blasts expressing the fig.

Abundant evidence exists that transmittal of signal 1 alone (for example by using fixed APC) may be important in tolerance induction in vitro or in vivo of preactivated T cell clones, or tolerance induction in a naive immune system. We now add the notion that resting or fixed (costimulation-deficient) transgenic B cells, which are not expected to secrete significant amounts of Ig, may evoke a different tolerogenic pathway than does a large dose of activated B cell APC injected iv. In the present studies, resting B cells were less efficient at curtailing an ongoing immune response.

It is concluded that the expression of a fig construct comprising a selected foreign epitope or epitopes in peripheral B cells using gene therapy strategies has great practical utility for modulating humoral and cellular immune responses. In comparison to currently used methods of high dose tolerance or oral tolerance, genetic transfer and expression of tolerogens in lymphoid APC requires only knowledge of the DNA sequence encoding the target epitope towards which tolerance is desired. The present method avoids the cumbersome antigen purification synthesis steps . More importantly, since clinically useful tolerance would require that the antigen (tolerogen) persist, its genetic expression in long-lived APC or pluripotential hematopoietic stem cell precursors provides a means for achieving the requisite persistence. The present inventors have also induced peptide-specific tolerance by expression a fig construct in peripheral B cells or hematopoietic stem cells using retroviral-mediated gene transfer. EXAMPLE IV

Genetically-Transferred Central and Peripheral Immune Tolerance Via Retroviral- Mediated Expression of Immunogenic Epitopes in Hematopoietic Progenitors or

Peripheral B Lymphocytes One potential strategy for the induction of clinically relevant tolerance is indirectly related to the original demonstration by Medawar's group of tolerance induction to foreign MHC antigens via injection of allogeneic hematopoietic cells into neonates (Billingham et al, supra). In adults, attempts to induce tolerance to foreign grafts by injecting accessory-cell depleted splenocytes (Ryan, J.J., et al. (1984) J. Immunology 755:2343-2350; Hori, S., et al. (1989) J. Immunology 745:1447-1452) or syngeneic transfected cells (Madsen, J.C, et al. (1988) Nature 552:161-164) has met with, at best, limited success. The advent of efficient methods for gene transfer into hematopoietic cells may, in theory, enable the expression of foreign antigens for the induction of tolerance via pathways similarly used for tolerance induction to naturally-expressed epitopes (e.g. MHC, Mis antigens) in bone marrow chimeras

(Ramsdell, F., et al. (1989) Science 246:1038-1041; Roberts, J.L., et al. (1990) J. Exp. Med. 552:161-164; Gao, E-K, et al. (1990) J. Exp. Med. 777:1101-1121 ; Sachs, D.H., et al. (1993) Transplantation Proc. 25:348-349). Since such approaches have usually required some degree of myeloablation, a more desirable approach would be to adoptively transfer genetically modified peripheral APC (Sutkowski, N., et al. (1994)

Proc. Natl. Acad. Sci. USA 97:8875-8879), which engraft more efficiently in an unconditioned host. An excellent candidate for such a strategy would be the peripheral B lymphocyte which has been described to possess immune-modulating characteristics. Previous studies have shown that antigen-presenting B cells are capable in some circumstances of inducing peripheral tolerance of (a) mature, naive T cells in vivo (Webb, S., et al. (1990) Cell 65:1249-1256; Eynon, E.E., et al. (1992) J. Exp. Med. 775:131-138; Fuchs, E.J., et al. (1992) Science 255:1156-1159; Buhlmann, J.E., et al. (1995) Immunity 2:645-653), or (b) previously activated T cell clones in vitro (Gilbert, K.M., et al. (1994) J. Exp. Med. 779:249-258). To test the potential for gene-transfer of a target antigen into autologous APC for the induction of specific immune tolerance, we created a recombinant, replication- defective retroviral vector for the expression of a foreign class II MHC-restricted immunodominant model epitope, 12-26 (Soloway, P., et al. (1991) J. Exp. Med. 774:847-858; Lai, M-Z, et al. (1987) J. Immunology 759:3973-3980), fused at the N- terminus of a murine IgG H chain, as described above. Engineered Ig expressing heterologous epitopes has been described for the potentiation of peptide-specific immunity (Zaghouani, H., et al. (1993) Science 259:224-227), and the Examples above expanded this approach by describing the tolerogenic properties of a soluble engineered 12-26-IgG fusion protein. The genetic transfer and expression of immunogenic epitopes, or whole complex antigens by appropriate "non-professional" APC has great utility for the specific elimination of undesirable immunity associated with HIV infection, as describe herein, autoimmune states (Tisch, R. et al. supra; Higgins, P.J. et al. supra; Critchfield, J.M., et al, supra), recombinant clotting factor administration (Allain, J.P. et al. (1976) Blood 47:973), and gene therapy protocols (Yang, Y., et al. (1995) J. Virol. 69:2004-2015; Tripafhy, S.K. et al. (1996) Nature Medicine 2:545-550).

A. Materials and Methods

1. Replication-defective retroviral vectors and gene-transfer protocols

12-26-IgGι H chain cDNA was derived by RT-PCR from J558L myeloma cells, transfected with the rearranged genomic construct (Examples I-III; Zambidis et al, supra) and subcloned into retroviral vector MBAE (Kang, J., et al. (1990) Proc.

Natl. Acad. Sci. USA 57:9803-9087) containing long terminal repeats (LTR), ψ+ packaging signals, a neomycin resistance gene, and cloned human b-actin promoter sequences. PCR primers encoded 5¹ Ig H chain leader and 3' IgGj as well as Sal I restriction site sequences: VH 5* primer: TGGACTAAGTCGACACCATGGGATGGAGC (SEQ ID NO:207)

Gl 3' primer: TCGGAAGGGTCGACGGATCATTTACCAGGAGA (SEQ ID NO:208)

A high titer (10^-10^ neomycin-resistant NIH 3T3 CFU/ml) ψ-2 packaging line (F6P) was prepared with recombinant plasmid MBAE.BAK, and assayed for helper virus via horizontal spread of neomycin resistance with NIH 3T3 cells. Ecotropic F6P was prepared by "ping-pong" amplification using amphotropic line PA317. Producer lines were stored in liquid nitrogen and freshly thawed cells were utilized for each individual experiment.

B cell lines CH31, A20, J558L, and NS-1 (ATCC, Rockville, MD) were transduced with recombinant retrovirus via co-culture with adherent F6P cells for 24- 48 hours in the presence of 6 μg/ml polybrene (Sigma). Cells in suspension were washed and recultured in 1 mg/ml G418 for selection of stable transductants prior to genomic Southern blot, RT-PCR, ELISA, or antigen-presentation studies. Infection of BM progenitors and quantitation of G418-resistant colony-forming cells (CFC) has been described (Keller, G., et al. (1985) Nature 318:149-154; Bodine, D.M., et al. (1989) Proc. Natl. Acad. Sci. USA 86:8897-8901). BM was harvested from femurs and tibiae of 6-8 week old BALB/c donors injected IV with 150 mg/kg 5-fluorouracil

3-4 days previously. Erythrocyte-depleted BM was co-cultured (5xlθ6/ml) with irradiated (2000 rads) F6P or ψ-2 parental cells (mock transduction). Ten ml cultures in complete RPMI 1640 with 15% FCS were incubated at 37° C, 5% CO2 for 48 hours, and included 200 U/ml each of IL-3, IL-6, and IL-7 (Genzyme). 4 μg/ml polybrene was added to co-culture during the last 24 hours of infection.

Splenic B cells were similarly infected in vitro via co-culture with viral- producing F6P or parental ψ-2 (mock transduction). Peripheral B cells were purified with anti-T cell antibody cocktail plus complement and Percoll density gradients (60- 70% layers). Purified B cells were pre-stimulated with 50 μg/ml bacterial

Hpopolysaccharide (LPS, E. coli 055:B5, Sigma) overnight, and recultured

(3xlθ /ml, 5 ml cultures) with irradiated F6P in the presence of 4 μg/ml polybrene and 50 μg/ml LPS for an additional 24 hours.

2. Tolerance induction and measurement of peptide-specific cellular and humoral immunity

Adult (6-8 week old) BALB/c recipients (Jackson Labs) were sublethally irradiated (200-600 rads total body irradiation) and injected intravenously (lateral tail vein) with 1-2x10^ gene-transferred or mock-transduced BM progenitor cells.

Unconditioned, normal BALB/c were similarly injected with >lxl0⁷ gene-transferred LPS blasts. All cells were washed extensively in serum-free medium and injected IV in a volume of 500 μl. Recipients were analyzed for expression of recombinant retrovirus as well as immune tolerance to 12-26 peptide 2-12 weeks later. Tolerized recipients were immunized for cellular or humoral immune responses as described above (see also: Soloway, P. et al. (1991) J. Exp. Med. 774:847-858; Zambidis et al, supra). Animals were injected SC with 50 μg synthetic 12-26 peptide emulsified 1:1

CFA. and in some experiments, also ip with 10 μg hen egg lysozyme (HEL) in CFA as a specificity control. Two weeks later, mice received an additional ip boost of 50 μg peptide and 10 μg HEL in saline. Antibody titers were determined from serum bleeds 8 days after secondary boosts. Splenic memory T cell responses were measured in vitro 6-8 weeks following these secondary challenges by reculturing purified T cells (3xl0^/ml) with irradiated (2500 rads) BALB/c splenocytes

(lxlO^/ml) and dilutions of synthetic peptide. Serum peptide-specific or HEL- specific IgG responses were determined by ELISA as described (supra). Cellular responses from draining popliteal and inguinal LN cells were assayed 9 days after SC immunization with 20 μg peptide in CFA. Cultures were pulsed with [^Hjthymidine, harvested and counted as described above) IL-2 and IL-4 cytokine production was quantitated as above. Dilutions of anti-IL-2 mAb S4B6 and anti-IL-4 mAb 11B11 (ATCC) were included to confirm specificity. IFN-γ was measured using a commercial ELISA kit (Intertest-γ, Genzyme). 3. RT-PCR and immunologic methods

Detection of 12-26-IgG transcripts in transduced cell lines or hematopoietic tissue from BM-injected mice was performed with a 12-26 sequence RNA-PCR assay. Primers were designed to amplify 5' immunoglobulin leader sequence, "Vjj 5' primer", as above, and 3' 12-26 sequence, ("3' pep primer"): GGC AAC AGA AGC TTT CAC TTC TTC TTC TCG TAT (SEQ ID NO:209).

Briefly, 1-5 μg of total RNA from various tissue was reverse-transcribed (2 rounds) with AMV reverse transcriptase, dNTP's, and oligo dT and random hexamer primers

(Invitrogen cDNA cycle kit) at 42° C The resultant cDNA was amplified with 5' and 3' primers described above and Taq polymerase (Perkin-Elmer Cetus). PCR conditions were 45 seconds at 93° C, 2 min at 47° C, and 2 min at 72° C for 35-40 cycles. Amplified DNA products were loaded (1/10-1/100 sample) onto 2% agarose/TBE gels and subsequently transferred onto nylon membranes for Southern blot analysis. 12-26 sequences were confirmed in RT-PCR amplified DNA samples with a g-32p-labeled oligonucleotide ("oligo Ig-one") encoding 12-26 but which does not overlap with the 3' PCR primer: TGATCTACTGCAGCTGGAGGACGCGCGGCGG (SEQ ID NO:210). Tissue RNA samples were compared via b-actin RT-PCR using commercially available primers (Stratagene).

12-26-IgG H chain protein was detected in culture supernatants of transduced cell lines, or in sera of mice injected with gene-transferred cells, via its ability to bind to the NIP hapten using a modified NIP-binding ELISA as above. Briefly, dilutions of culture supernatants or sera were incubated on ELISA plates coated with NIP- gelatin conjugate (and subsequently probed with goat anti-mouse IgGi -AP. Standard curves with affinity-purified 12-26-IgG from supernatants of transfected J558L were used for quantitation.

B. Results and Discussion

A recombinant retroviral vector (Kang, J., et al, supra ) was modified by inserting a PCR-derived cDNA encoding the 12-26-1 gG H chain sequence (Figure 14), and a high titer ecotropic packaging line (F6P) was generated for the in vitro infection of cell lines and hematopoietic tissue via co-culture methods (Keller, G., et al. (1985)

Nature 575:149-154; Dick, J.E., et al. (1985) Cell 42:71-79; Bodine, D.M., et al. (1989) Proc. Natl. Acad. Sci. USA 56:8897-8901). For initial studies, a variety of B cell lines at various stages of differentiation were transduced, including CH31 (immature), A20 (mature, activated), NS-1, and J558L (plasmacytomas). Intact proviral integration in transduced G418-resistant A20 cells could be verified by genomic Southern blotting using a DNA probe specific for 12-26 sequence (Figure 14). More importantly, 12-26-IgG H chain can assemble with endogenous light chains in transduced B cell lines, to be expressed as a membrane surface protein, or secreted into cultured supernatants (50-80 ng/ml) inNS-1 and J558L myelomas. Immunoprecipitation of secreted 12-26-IgG and immunoblot analysis with a peptide- specific monoclonal antibody (B3.l l) could directly demonstrate the expression of 12-26 peptide.

Although retrovirally-synthesized gene products are expected to give rise primarily to processed peptides presented by MHC class I molecules, endogenously- derived peptides can also be routed to endocytic class II MHC compartments in some cases (Weiss, S., et al. (1991) Cell 64:161-116). Such a pathway should be enhanced for retrovirally-encoded 12-26-IgG H chain due to the efficient nature of the Ig secretory pathway in targeting the endosomal compartment. To test for direct presentation of this model immunodominant class Il-restricted peptide, we tested the ability of 12-26-IgG-transduced A20 cells to directly activate peptide-specific T-cell hybrid 9C127, which recognizes 12-26 in the context of I-A^ (Lai, M-Z, et al, supra). Efficient presentation of endogenously synthesized peptide was demonstrated, and this effect was blocked with antibodies to CD4 or class II MHC molecules. These results predict that, in vivo, the 12-26 peptide could be recognized directly by T cells from a variety of gene-transferred APC (both lymphoid and non-lymphoid). Additionally, synthesis, L chain assembly and secretion of 12-26-1 gG H chain by B cells can potentially result in re-presentation of the molecules by endogenous host APC.

The next experiments tested the potential of genetically modified BM cells to specifically tolerize a regenerating immune repertoire ("central tolerance" induction). BM chimeras were produced in sublethally irradiated (200-650 rads) BALB/c mice by infusing 5-fluorouracil (FU)-pretreated donor BM which had been co-cultured with F6P. This protocol leads to newly developing lymphocytes and APC (lymphoid and non-lymphoid) that are derived from both the host, as well as the transplanted BM progenitors expressing 12-26-IgG. Mice were immunized 4-12 weeks post-infusion and specific immune responses were measured. Analysis of hematopoietic tissue

(collected at sacrifice) using a peptide sequence-specific RT-PCR assay indicated that transcripts were expressed consistently and reproducibly in the BM of all gene- transferred recipients (Figure 15), although variably in the thymus or spleen. A very sensitive NIP-binding IgG_j ELISA was used that detects secreted 12-26-IgG in the serum which is a result of pairing of the fusion protein H chain with endogenous λ L chains in B cells derived from the gene-transferred stem cells. Although this assay detects only a fraction of H chain secreted by B cells (λ light chain-paired), expression could be detected in approximately one-third of all gene-transferred BM recipients (Table V). Despite these variable serum expression patterns, dramatic and reproducible peptide-specific tolerance was observed in all 22 studied recipients of 12-

26-expressing progenitor cells, which tolerance was demonstrated following immunization with synthetic peptide in adjuvant. Specific T-cell proliferative and cytokine responses of draining LN cells were significantly diminished (Figure 16), as were anti-peptide antibody levels following priming and boosting (Figure 17A,B). Since the present BM transduction protocol was designed to effect viral integration into early hematopoietic progenitors, and both the 5' viral LTR and β-actin promoters are non-specific as to cell lineage, a variety of differentiated cells with APC capacity may directly tolerize developing lymphocytes. We predicted that non-lymphoid APC derived from transduced stem cells (e.g. monocytes, macrophages, dendritic cells), which cannot synthesize Ig L chains, and thus cannot secrete the fig, may nevertheless play a critical role in direct presentation of transgenic peptide to developing lymphocytes. To test this, syngeneic BM from SCID mice was gene-transferred the tolerogenic activity of myeloid APC was analyzed. Although hematopoietic tissue from SCID mice is deficient in developing mature lymphoid cells, the APC function of cells of the myeloid (non-lymphoid) lineage remains intact (Dorshkind, K. et al.

(1984) J. Immunology 752:1804-1808). 5-FU-pretreated SCID or normal BALB/c donor BM cells were co-cultured with F6P, and stem cells were injected into sublethally irradiated normal BALB/c recipients. (With SCID BM donors, normal lymphocytes can regenerate only from the recipients' stem cells). BM chimeras were rested over 2 months and subsequently immunized with 12-26 peptide for measurement of humoral immune tolerance.

Peptide-specific tolerance was comparable in recipients of either normal or lymphoid-deficient BM. In contrast recipients of mock-transduced BM (not expressing the 12-26 peptide from the fig) had high-titer antibody responses (Figure 17A,B). Analysis of G418-resistant hematopoietic colony forming cells (CFC) from recipients of transduced normal BM or transduced SCID BM showed that the proportion of BM-derived myeloid stem cells expressing the fig construct was similar (1-5%), Table V). These results indicate that in addition to tolerogenesis by lymphoid APC, transduced myeloid BM-derived APC can also share this activity. This result may explain the consistently solid tolerance observed in all recipients of transduced normal BM, regardless of detectable levels of 12-26-IgG (the B-lymphocyte-derived transgene) in the subjects' serum.

Although efficient induction of tolerance in newly arising lymphocytes was readily and reproducibly accomplished with genetically-modified BM, a more clinically practical approach for gene-transfer tolerogenesis would be the induction of peptide-specific peripheral tolerance in a mature immune repertoire without the need for prior myeloablation. It is known that potent tolerance in normal, immunocompetent subjects can be induced by injecting large numbers of mature B lymphocytes expressing a "foreign" antigen (e.g., H-Y or Mis). Therefore, the present inventors tested the tolerogenicity of fig -transduced peripheral B cells.

The approach comprised stimulating Percoll® gradient-purified splenic B cells to proliferate with bacterial LPS, brief co-culture with F6P, and subsequent iv injection into normal, immunocompetent (non-irradiated) BALB/c recipients. This treatment resulted in an efficient suppression of peptide-specific humoral immunity comparable to that observed in the BM chimera experiments described above.

TABLE V

Expression of Serum 12-26-IgG and G418-resistant BM Progenitors in Genetically-Tolerized Mice

Legend to Table V: Recipients of F6P or mock-infected 5-FU-treated BM progenitors were assayed for transgene expression in serum or BM at indicated times. BM CFC were assayed at sacrifice time in 0.3% semisolid agar cultures in long-term recipients. BM cells were cultured at

10⁶ /well in complete IMDM plus 15% FCS, 200 U/ml IL-3, and 10% ORIGEN conditioned medium (GIBCO, BRL) containing IL-1, G-CSF, GM-CSF, M-CSF, and IL-6. Erythro-myeloid colonies were grown with and without 1 mg/ml G418, and the percentage which were viable and neomycin-resistant (G418^) were counted after 7-10 days. BM recipients were conditioned with either 200 rads (Expt. 1) or 600 rads (Expts. 2-4). Detailed experimental results for Expt. 1 and 4 are presented in Figures 17A and 17B, respectively. * NT: not tested Splenic memory T cell responses measured approximately 3 months after immunization in these subjects were markedly reduced (IL-2 and IL-4 cytokine reduction), indicative of effective tolerization in both T helper cell compartments (Thl and Th2). See Figure 18). Furthermore, G418-resistant hybridomas could be generated from LPS-activated spleen cells of these tolerized mice by fusing the spleen cells with A20 lymphoma cells in the presence of PEG. These hybrids stimulated peptide-specific 9C127 T cells directly (Figure 19), thus proving that gene-transferred peripheral B cells are capable of persisting and presenting 12-26-IgG self-antigen for prolonged periods in a way which results in induction and maintenance of peptide- specific tolerance. Such long-term persistence (>3-6 months) is consistent with results obtained after injection of either normal peripheral B cells (Sprent, J. et al (1991) J. Exp. Med. 174:111-12%) or genetically-modified LPS-activated peripheral B cells (Sutkowski, N., et al, supra).

The foregoing results show that the present inventors have in hand a novel, efficient strategy for delivery of a foreign peptide, which would otherwise be an immunogen, to an adult immune system in a tolerogenic manner in the form of a soluble fig protein expressed in hematopoietic tissue. Genetic transfer of a selected target single or multi-epitope sequence into a multipotential stem cell or into a peripheral B cell permits the induction of, and more importantly, the long-term maintenance of, specific immune self-tolerance in the autologous host. The choice of a model immunodominant peptide, 12-26, capable of inducing both high titer IgG antibody responses (Th2 -mediated), as well as vigorous cellular (Thl -mediated) responses underscores the versatility of the present method. Although gene-transfer of BM expressing class I MHC-restricted CTL epitopes efficiently induced tolerance (Ally, B.A., et al. (1995) J. Immunology 755:5404-5408), the present approach of fusing an antigenic sequence to an Ig molecule allows for the efficient presentation of a retrovirally-synthesized class II MHC-restricted epitope. Furthermore, in addition to effective Th tolerance induction, the bivalent nature of the secreted form of the tolerogenic epitope on the two H chains of the Ig-molecule can independently mediate effective peptide-specific B cell tolerance, probably via Fc-mediated antibody feedback mechanisms (Zambidis et al, supra). Thus, the potency of tolerance induction using the present invention can exploit multiple pathways in the immune mechanism.

The use of tolerogenic peptide-Ig constructs facilitates "tailor-designing" the immune response to a whole antigen by selectively inducing immunity (Zaghouani,

H., et al, supra ) or tolerance to selected epitopes be they immunodominant or cryptic. In contrast to expressing a heterologous epitopes in the CDR3 region of the Ig H chains, fusing an foreign antigenic sequences at the N-terminus is not limited by size restrictions, and can thus be adapted for expressing large multi-epitope antigens, for example, autoantigenic proteins such as factor VIII (Allain et al, supra), myelin basic protein (Higgins et al, supra; Critchfield et al, supra ), or glutamic acid decarboxylase (Tisch et al, supra). Delivery of the tolerogen as a gene sequence has many advantages over present tolerance induction protocols, since only the cDNA sequence of the target antigen, for example, one or more HIV gpl 20 epitopes, needs to be known. This avoids the need for a protein purification strategy. More importantly, since experimentally acquired tolerance eventually wanes, expression and persistence of the tolerogen in long-lived or multipotential hematopoietic tissue has the potential to modulate permanently a specific immune response.

As described herein, an important application of the genetic tolerogenesis method of the present invention is to help eliminate genetically-altered cells encountered in gene therapy protocols. Autologous cells genetically modified with adenoviral and retroviral vectors are known to induce immunity in a competent recipient due to immune recognition of vector-encoded products leading to subsequent elimination of transduced cells via both cellular and humoral immunity (Yang et al, supra). Although, for example, immunity to low-level expression of viral proteins of first-generation, El -deleted adenovirus can undoubtedly be reduced with further genetic manipulation of the vectors, rejection of the foreign transgenes expressed by such vectors remains an even more significant obstacle (Tripathy et al, supra). The present results suggest that tolerogenic pretreatment of immunocompetent recipients with vector-transduced autologous APC expressing viral or foreign transgenes will allow the prolonged expression and multiple administration of therapeutic transgenes in immunocompetent recipients without need for generalized immunosuppressive drugs

Although solid evidence exists for the tolerogenic role of lymphohematopoietic APC in irradiated bone marrow chimeras,

The foregoing studies exploited the ability of mature peripheral B cells to induce efficient peripheral tolerance in unconditioned adults. These findings represent the first example of using transduced LPS blasts as tolerogenic vehicles. Others have reported tolerance induction via presentation of antigen by resting B cells to naive T and have implicated the poor expression of costimulatory molecules such as B7-1 and

B7-2 (Hathcock, K.S., et al. (1994) J. Exp. Med. 750:631-640). Antigen-presentation by resting B cells has thus far been successful in inducting tolerance in naive recipients, but has proven ineffective in primed (Fuchs et al, supra) or allo-MHC- reactive recipients (Buhlmann et al, supra) unless an anti-gp39 (CD40-ligand) antibody was simultaneously injected to prevent upregulation of B cell costimulatory function. Paradoxically, costimulation-competent LPS blasts, as in the present experiments, could serve as efficient tolerogenic APC in vivo in antigen-naive recipients, or could induce tolerance in vitro in previously activated T cell clones (Gilbert et al, supra). In studies with transgenic mice expressing the fig construct specifically in the

B cell compartment, both purified resting fig-expressing B cells or their LPS-activated counteφarts were highly tolerogenic in normal, antigen-naive adults. In contrast, only the activated transgenic B cells were effective in tolerizing an ongoing response in a previously-immunized recipient. EXAMPLE V

HIV gpl 20 Crosslinking In Vivo Induces apoptosis of T Cells Studies were performed in mice transgenic for the human CD4 gene. Normal BALB/c mice or mice transgenic mice for human CD4 ("CD4hu") were immunized with 20 μg of gpl 20 in complete Freund's adjuvant, boosted with gpl 20 in incomplete adjuvant and then injected intravenously with 1 μg of gpl20 in PBS. Peripheral blood lymphocytes were harvested at various times after the last injection (of soluble gpl 20). Total number of T cells in the peripheral blood were evaluated using flow cytometry to enumerate CD3⁺ cells. Table VI, below shows the results as percent of total blood lymphocytes which are CD3⁺

Table VI

HIV gpl 20 crosslinking induces apoptosis of CD3 cells in vivo in human CD4 transgenic mice

The only significant reduction in numbers of T cells was observed in the transgenic mice which had been immunized, that is, mice expressing the human CD4 molecule which can bind gpl 20 or gpl20-anti-gpl20 complexes.

Experiments were performed to study the apoptosis resulting from the ligation of T cell receptors. Mice were immunized as above. Spleens were harvested 9 days after this last injection and were cultured with medium or with anti-CD3 mAb (145.2C1; 50 μg/ml coated wells) for 24 hours; cells were then harvested, fixed and assayed for apoptosis by propidium iodide uptake and flow cytometry. The percent of hypodiploid, apoptotic cells at 24 or 48 hours with anti-CD3 and at 24 hours with anti-IgM are shown in Table VII. Table VII

Induction of Apoptosis After T Cell Receptor Ligation

"Bkgrnd" = Background values of apoptosis of freshly isolated cells (as opposed to cells cultured 24 or 48 hours).

The results show that apoptosis is increased significantly only in T cells of immunized CD4 transgenic mice whose T cells receptors have been ligated polyclonally with anti-CD3 mAb. This effect is most dramatic after 48 hours .of culture. Collectively, the foregoing results prove that gpl 20 epitopes have the capacity to prime T cells for apoptosis shown as direct observation of apoptosis in vitro and as a loss of T cells in vivo which occurred selectively only in subjects who were both immunized with gpl 20 and who had the appropriate target T cells (bearing transgenic human CD4) to which complexes of gpl20 and anti-gpl20 antibodies could bind,. These results further support the findings of Finkel's group cited above and provide an even stronger basis for the utility of inducing epitope specific tolerance as described herein.

EXAMPLE VI

Ongoing Immune Responses to HIV gpl 20 in Human CD4 Transgenic Mice Contributes to T Cell Decline upon IN. Administration of gp!20

(most references are cited in this section as numbers and appear in a list at the end of the section)

The HIV retrovirus interacts with the host immune system in a puzzling way. Virtually everyone infected with the virus synthesizes antibodies directed against a number of the viral envelope epitopes. However, much of this humoral response has little if any protective value over the course of HIV pathogenesis [5-16]. Titers of neutralizing antibodies in AIDS patients are low [6], and the antibodies might cross- react with self-components due to molecular mimicry and structural/genetic similarities [7-9]. Furthermore, crosslinking CD4 by anti-CD4 antibodies or gpl 20 and anti-gpl20 antibodies can upregulate Fas expression and prime Th cells for activation- induced apoptosis [1, 10-16].

Crosslinking of human CD4 coreceptor (huCD4), by HIV gpl 20 and anti- gpl20 antibody in vitro as well as in vivo, upregulate Fas (CD95) expression and prime T cells for activation-induced apoptosis (Banda et al. , supra; Wang et al. , supra). Based on these observations, we hypothesized that an ongoing humoral immune response to gpl 20 might not serve the host in a protective or virus- neutralizing manner upon exposure to HIV gpl 20. Rather , the response might sensitize even noninfected cells for apoptosis. Immunization mice transgenic for huCD4 ("huCD4 Tg") and control mice with 20 μg gpl 20 in CFA led to titers exceeding 1 :10⁵ within three weeks. We injected i.v. 1 μg of rgpl20_SF2 into huCD4 Tg and non-transgenic BALB/c and BALB/c x C57B1/6 F, mice (CB6 FI) mice that had been immunized with rgpl20_SF2. The same amount of rgpl20 was also administered i.v. to unprimed huCD4 Tg mice and nontransgenic controls. Boosting gpl20-primed control mice with gpl 20 gave rise to increased numbers of T and B cell as well as in the antibody titers. In sharp contrast, boosting the primed huCD4 Tg mice (which express huCD4 on both T and B cells) with soluble gpl 20 resulted in lower secondary antibody titers than in controls. The response to an irrelevant antigen, HEL, was also reduced in the gpl 20 -primed and boosted huCD4 Tg mice. Furthermore, on day 6 after a single bolus of gpl 20, the number of peripheral T cells and B cells in immunized huCD4 Tg decreased to 50% of the control levels. Moreover, compared to the control groups, the splenocytes from gpl20-pretreated immunized huCD4 Tg had a lower level of CD3+ T cells and underwent extensive apoptosis after anti-CD3 treatment. These in vivo results were consistent with the in vitro findings: Crosslinking of huCD4 on the spleen lymphocytes of huCD4 Tg mice using rgpl20_SF2 and anti-gpl20 antibody not only sensitized T cells for apoptosis, but also induced apoptosis er se. Thus, precautions should be taken when employing HIV envelope gpl 20 as one of the HIV vaccine components. Tolerogenic therapies should be considered when treating HIV infected subjects in this manner. Materials and Methods

Mice — BALB/cByJ and CB6 FI mice were purchased from the Jackson Laboratories (Bar Harbor, ME) at 6-10 weeks of age, and housed in pathogen-free, microisolater cages. Line 313 huCD4 Tg mice were obtained from Dr. Terri Finkel, Denver, CO). These transgenics were originally produced by Dr. Richard Flavell by injecting a huCD4 transgene into fertilized eggs and were maintained by repeated backcrosses on the C57B1/6 background [17]. The F, offspring between huCD4 Tg mice and BALB/cByJ are produced in our animal facility by crossbreeding female BALB/cByJ with male huCD4 Tg mice to yield huCD4 expressing mice histocompatible with CB6 FI mice. The huCD4 molecule was shown to be functional in calcium signal transduction and in overcoming the block in positive selection induced by in vivo injection of mAbs to the endogenous mouse CD4 [17]. Since in these transgenic mice, expression of huCD4 is driven by CD2 regulatory elements, both B and T cells express huCD4. Approximately, 85% of splenic cells from Line 313 huCD4 transgenic mice and more than 50%) of spleen cells from the F, offspring between huCD4 Tg mice and BALB/cByJ expressed the huCD4 receptor on their surface. Recombinant wild type gpl 20 (rgpl20_SF2), from Dr. K. Steimer (Chiron Coφoration, Emeryville, CA), binds to the huCD4 molecules in a dose-dependent manner and competes with huCD4 mAb (Leu-3a; Becton Dickinson, Mountain View, CA) for huCD4 binding. Cell Culture

RPMI 1640 medium (GIBCO-BRL, Gaithersburg, MD) supplemented with heat-inactivated 5% fetal calf serum (Hyclone, Logan, UT), 50 μM 2-mercaptoefhanol, 2 mM L-glutamine, 100 U/ml penicillin, 100 U/ml streptomycin, MEM nonessential amino acids, and 1 mM sodium pyruvate was used. Antibodies

The following antibodies were purchased from PharMingen ( San Diego, CA): FITC- labeled anti-hamster IgG, PE-labeled anti-mouse CD3, biotin-labeled anti -mouse CD 19 and anti-Fas antibody Jo2) FITC- labeled mouse anti-human CD4 mAb (FITC- Leu-3a) was obtained from Becton Dickinson (View Mountain, CA). The following biotin- or FITC- labeled antibodies were purified in our lab by standard protocols: anti-mouse CD3 (145.2C11), anti-mouse CD4 (GK1.5), anti-mouse CD8 (53-6.72) and anti-mouse CD45R (RA36B2, B220). Anti-gpl20 antibodies used for huCD4 crosslinking were obtained as follows: human monoclonal anti-gpl20 antibodies directed against gpl20 C-terminal peptide, (450-30D 100,100,1,1; abbreviated '^SO-

SO")) and against the V3 loop (694/98-D 100,100,10,1; abbreviated "694/98") were from Dr. Susan Zolla-Pazner (VA Medical Center, NY); sheep anti-gpl20 antibody (6205) directed against the 15 C-terminal residues was obtained from International Enzymes, Inc. (Fallbrook, CA); human HIV-1 gpl20 monoclonal antibodies (F105) [Dr. M. Posner, refs. 18-21 ] and (AD3) [Drs. K. Ugen and D. Weiner, ref. 22], as well as goat polyclonal HIV-1SF2 gpl 20 antibodies raised against glycosylated and non- glycosylated gpl 20 [Dr. K. Steimer, refs. 23-27] were obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH. Mouse polyclonal HIV-1_SF2 gpl 20 antisera were produced in our lab by immunization of BALB/c J mice with HIV-1_SF2 gpl 20 in CFA intradermally and boosting twice with

HIV-1_SF2 gpl 20 in IF A intramuscularly after 2 weeks and 1 month of primary immunization, respectively. The antibody titer was >1/10⁵ as determined by ELISA using rgpl20 coated plates (1 μg/ml in Tris coating buffer, pH 9.0). In vitro crosslinking huCD4 by gp!20 and anti-gp!20 followed by anti-CD3 activation Viable splenocytes from Line 313 huCD4 Tg were incubated (5 x lOVml) with rgpl20_SF2 (20 (g/ml) on ice for 30 min., washed twice, and reincubated with various anti-gpl20 antibodies (2 (g/ml or 1 : 1000 dilution) at 37°C for 45 min. While an aliquot of the cells was checked for surface levels of Fas expression by flow cytometry, 1 x 106 cells were transferred onto anti-CD3 antibody (145.2C11)- precoated 96 well plates in a 200 (1 volume and incubated at 37°C, 5% CO₂ for further 24-72 hr. The cells were then harvested and assayed for apoptosis by DNA content analysis, as described below [28].

Immunization protocol and intravenous administration of gpl 20

BALB/cJ mice and Line 313 huCD4 transgenic mice were immunized intradermally with 20 μg HIV-1_SF2 rgpl20 emulsified in CFA and boosted intramuscularly with 20 μg rgpl20 in IF A 9-12 days later. Ten days after boosting, when high titers of anti gpl 20 sera were detectable, a single dose of 1 μg of rgpl20 was administered intravenously into the immunized animals, as well as animals that had not received rgpl20 immunization. The percentages of human CD4+, mouse CD3+, CD8+ and B220+ cells in the peripheral blood were followed using dual color flow cytometry described below. In some experiments, the spleens were harvested within 9 days after gpl 20 iv injection and the percentages of splenic CD3+ T cells were determined by flow cytometry. The apoptotic cells in freshly harvested spleens were assessed by in situ terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling staining (TUNEL). Spontaneous apoptosis and anti-CD3 activation- induced apoptosis was measured after anti-CD3 in vitro treatment for 24-72 hr. To search for depletion of antigen-specific T cells by repeated gpl 20 iv injection, the F, offspring between huCD4 Tg and BALB/cByJ mice, as well as CB6 FI control mice, were immunized intradermally with 20 μg HIV-1_SF2 rgpl20 and 20 μg hen egg-white lysozyme (HEL, Sigma Chemical Co., St. Louis, MO) emulsified in CFA. Three weeks after primary immunization, the animals were boosted intravenously with 1 μg rgpl20_SF2 and 1 μg HEL in PBS, and the iv injection was repeated three times at 10 day intervals. One week after each iv injection, the CD3+ T cells were determined by flow cytometric analysis and the gpl 20 and HEL specific IgG responses were measured by ELISA by coating plates with 1 μg/ml gpl 20 or 1 μg/ml HEL, respectively. Antibody titers were determined using CA-Cricket Graph software and were expressed as the serum dilution that would bring the OD to pre-immunization levels (OD₄₀₅ s. 0.04), assuming parallelism of curves. In other experiments, gpl 20- primed mice were sacrificed on day 1, 4, 7, 11, and 20 after a single bolus of gpl 20 iv, and lymph nodes and spleens were harvested,and determinations made of cell phenotypes, spontaneous apoptosis and anti-CD3 stimulation index. Flow cytometric analysis

The surface level of Fas expression on the splenocytes from Line 313 huCD4 transgenic mice after huCD4 crosslinking was measured by staining cells with hamster anti-Fas antibody (Jo2), followed by FITC-labeled anti-hamster IgG. To determine the percentage of peripheral CD3+ T cells, blood was removed from the retroorbital plexus. White blood cells were prepared by lysing red blood cells with Tris-buffered ammonium chloride buffer (pH 7.2). After washing with PBS, the cells were stained with FITC-labeled Leu-3a, or anti-B220, or anti-CD8 antibody plus biotin-labeled anti-mouse CD3 antibody, followed by Streptavidin-PE staining. The same procedure was applied to spleens for measurement of CD3+ T cells. To measure Annexin V positive T cells, lymph node and spleen cells were washed twice with PBS and resuspended in binding buffer (10 mM Hepes/NaOH, pH 7.4,140 mM NaCl, 2.5 mM CaCl₂) [41]. To the cell resuspension was added fluorescein labeled Annexin V

(R&D SYSTEMS, Inc., Minneapolis, MN) and biotin-labeled anti-CD3 antibody, followed by Streptavidin-PE staining. The FITC and PE fluorescence analysis were performed by CELLQuest software in FACScan flow cytometry (Becton Dickinson). Apoptosis analysis The percentage of cells undergoing apoptosis was quantitated by a flow cytometric method described earlier [28]. Briefly, cells (1 x 10⁶) were fixed in 70% ethanol for 1 hour at 4 C. The cells were then washed and resuspended in 1 ml PBS, to which 1 μl RNase solution (10 mg/ml in PBS) was added and incubated at 37°C for 1 hour. Following the addition of 5 μl of propidium iodide (PI, 10 mg/ml in PBS), the PI fluorescence of individual cells was measured using flow cytometry. Cell debris and clumps were excluded by gating for single cells by forward and side light scatter and by FL-2 area vs. FL-2 width. A distinct cell cycle region of apoptosis (AQ) could be identified below the G_Q/G, diploid peak and the percentage of cells in the A₀ region was quantitated. Statistical Analysis

The difference among different treatment groups was analyzed by one-way ANOVA (StatView, BrainPower, Inc., Calabasas, CA). RESULTS In vitro induction of apoptosis by gpl 20 crosslinking ofhuCD4

We hypothesized that gpl 20 crosslinking of huCD4 on splenic cells from huCD4 transgenic mice would induce apoptosis or prime for apoptosis despite the fact that these cells cannot be infected by HIV [34] due to the lack of a necessary cofactor (such as fusin) in mice [35]. We cultured splenocytes from Line 313 huCD4 transgenic mice with rgpl20_SF2, and then crosslinked huCD4 by the addition of murine hyperimmune anti-gpl20 antibody, and finally stimulated aliquots of these cells with anti-CD3 antibody. Data from three sets of 24 hr in vitro apoptosis induction experiments showed that crosslinking huCD4 by gpl 20 and anti-gpl20 antibody prior to TCR ligation primed huCD4 Tg splenocytes for anti-CD3 activation-induced apoptosis. In contrast, anti-CD3 antibody alone reduced spontaneous apoptotic cell death. This phenomenon of reduction of apoptosis in unprimed T cells presumably reflects cell cycle entry induced by anti-CD3 activation. Furthermore, crosslinking huCD4 by gpl 20 and anti-gpl20 antibody er se induced apoptosis in the huCD4 Tg splenocytes, though to a lesser extent than that with additional anti-CD3 treatment. Neither gpl 20 alone nor anti-gpl20 antibody alone had any effect on the priming and apoptosis induction, suggesting that anti-gpl20 antibody was required for huCD4 crosslinking-mediated apoptosis induction in vitro.

Crosslinking huCD4 on spleencells of these transgenic mice via gpl 20 and anti-gpl20 antibody modestly upregulated surface levels of Fas expression (Table VIII), as shown by others [10, 13, 14]. Furthermore, the induction of apoptosis and upregulation of Fas expression by huCD4 crosslinking is not dependent on the specificity of anti-gpl20 antibody. Thus, we tested a variety of anti-gpl20 antibodies for huCD4 crosslinking, either mAbs raised against a number of gpl 20 antigenic domains or polyclonal antibodies. The results on induction of apoptosis and upregulation of Fas expression using different kinds of anti-gpl20 antibodies for CD4 crosslinking are summarized in Table VIII. gpl 20 Induces Acute Loss ofCD3+ T cells in Immunized huCD4 Transgenic Mice A paradox in HIV pathogenesis is that the virus appears to cause AIDS after the onset of antiviral immunity [11]. Wang et al [15] reported that injection into huCD4 transgenic mice of HIV-gpl20 and subsequent gpl20-specific antibodies from AIDS patients (passive transfer of immunity) induced massive long-lasting T lymphocyte deletion. To investigate if an ongoing immune response to gpl 20 (active immunity) could crosslink huCD4 and lead to T cell depletion upon exposure to HIV- gpl20, we administered gpl20_SF2 intravenously into gpl20_SF2 immunized huCD4 transgenic mice.

The time course of CD3+ T cell loss in an in vivo experiment was analyzed Though the number of peripheral CD3+ T cells varied in individual mice, there was a drop of peripheral CD3+ T cells only in immunized huCD4 Tg mice after a single i.v. injection of gpl 20. None of the other three control groups showed a drop of CD3+ T cells. This CD3+ T cell loss occurred acutely, reached its peak (50% of the pre- injection level) on day 6 after i.v. gpl 20 injection and then gradually recovered. Although we did observe CD3+ T cell depletion for more than two weeks in individual animals , the CD3 + T cell drop in the majority returned to its preinjection level within two weeks.

Splenocytes from huCD4 transgenic mice were treated and analyzed as described. Apoptosis induction with anti-CD3 (%) refers to the percentages of apoptotic cells after crosslinking of huCD4 crosslinking with anti gpl 20 antibodies and after 24 hr of treatment with anti-CD3. Fas upregulation by huCD4 crosslinking for 45 min. was calculated as the % increase in median fluorescent channel over medium control.

The mean percentages of peripheral blood CD3+ T cells on day 0 and day 6 after i.v. gpl20 injection from all in vivo experiments were compared. On day 6, injection of gpl 20 induced significant peripheral CD3+ T cell loss in immunized huCD4 Tg, compared to the other three groups.

The first bolus of gpl 20 induced a significant, though transient, loss of peripheral CD3+ T cells in huCD4 transgenic mice, but not in CB6 FI control mice. Suφrisingly, repeated gpl 20 iv injections afterwards were not able to produce a state of long-lasting T cell loss, neither to induce the T cells to decline again after recovery from the first drop, though a slight lower level of CD3+ T cells were constantly observed after gpl 20 injections in the huCD4 transgenic mice than in the controls.

Interestingly, gpl20 i.v. injection also resulted in a loss of peripheral blood CD19+ B cells which express huCD4 driven by CD2 promoter but not the CD3-/CD 19-cells in the gpl20-immunized huCD4 transgenic mice, in the same pattern as the loss of

CD3+ T cells.

T cell depletion via apoptosis is not restricted to peripheral blood, but also occurs in spleen and lymph nodes To investigate if T cells in spleen and lymph node are also deleted in immunized huCD4 transgenic mice receiving gpl 20 iv injection, we harvested spleens and lymph nodes 1 - 20 days after the first gpl20 i.v. injection and measured CD3+ T cells by flow cytometry. The numbers of splenic CD3+ T cells in all gpl20- immunized and -pretreated huCD4 Tg were decreased to 50-75% of those in control groups. Moreover, lower levels of CD3+ T cells were also observed in the lymph nodes from gpl 20 - immunized and -pretreated huCD4 Tg. The decrease in CD3+ T cells in spleens and lymph nodes suggested that the T cell drop was not restricted to peripheral blood. Furthermore, as measured in a TUNEL assay, while no difference from control groups was noted in some gpl20-pretreated immunized huCD4 Tg, a slightl increase (by 3-5%) in apoptotic cell death was observed in freshly harvested spleens from other gpl20-pretreated immunized huCD4 Tg, as shown by others [16]. In addition, when fluorescein labeled Annexin V was employed to detect phosphatidylserine expression on early apoptotic cells [41], a higher percentage of Annexin V positive T cells was observed in the spleens and the lymph nodes from gpl20-immunized huCD4 transgenic mice than in the organs of control mice, indicating that gpl20-induced apoptosis occurs in peripheral lymphoid organs as well as blood. Taken together, these results suggest that a higher number of T cells in the gpl20-immunized huCD4 transgenic mice undergo apoptosis after receiving gpl 20 iv injection. gpl 20 injections lead to a lower secondary antibody titers in primarily immunized huCD4 transgenic mice, presumably by depletion of antigen-specific T and B cells

To further confirm that crosslinking in vivo of huCD4 by gpl20/anti-gpl20 binding causes depletion and/or sensitizes huCD4-expressing T and/or B cells for activation-induced apoptosis, we injected gpl20 and an irrelevant antigen (HEL) i.v. into huCD4 Tg mice and CB6 FI mice that had received primarily immunization with gpl 20 and HEL in CFA. We then measured the secondary antibody titers against gpl 20 and HEL. Boosting with gpl 20 not only boosts specific T and B cells for secondary responses, but the gpl 20 can also bind to huCD4 receptors to prime for apoptosis on all huCD4+ cells. We used the response to HEL as an irrelevant control response although HEL-specific T and B cells would be expected to bind gpl 20 to their huCD4 receptors like gpl 20 specific cells. The results showed that huCD4 Tg mice primed with soluble gpl 20 had lower secondary titers than did controls, and the response to an irrelevant antigen, HEL, was also reduced in the gpl 20 - primed/boosted huCD4 Tg mice. These results indirectly support our hypothesis that gpl20 injection depletes a population of antigen-specific T cells and/or B cells.

TCR ligation induces further apoptosis in the spleens of immunized huCD4 transgenic mice

As mentioned above, in immunized huCD4 Tg mice, the total CD3+ T cell population was depleted by 50%. Numbers of CD3+ cells then recovered to pre- injection levels within two weeks after a bolus of gpl 20 injection. To test if the undeleted / recovered CD3+ T cells were primed for apoptosis by in vivo huCD4 crosslinking, we harvested spleens after a single bolus of gpl 20 and assayed in vitro the apoptosis stimulated by anti-CD3 antibody. Anti-CD3 treatment resulted in lower stimulation in spleen cells from gpl 20 - immunized and -pretreated huCD4 Tg compared to CB6 FI control mice. Anti-CD3 treatment also increased the percentage of apoptotic cells compared to the medium control treatment only in immunized huCD4 Tg mice receiving a bolus injection of gpl 20 . These ex vivo results indicated that the undeleted / recovered CD3+ T cells in the gpl20-pretreated immunized huCD4 Tg mice were also primed for apoptosis by crosslinking of huCD4 in vivo . DISCUSSION

Efforts are currently underway to elucidate the mechanisms responsible for AIDS pathogenesis and to establish a protective vaccine for HIV. The precursor envelope glycoprotein of HIV, gpl 60, and mature proteins, gpl 20 and gp41, have been considered to be important if not essential as vaccine components, because epitopes of these proteins induce both antibody and cytotoxic T cell responses in man. Although usually viewed as potentially protective, the role of humoral immune responses to viral envelope gpl 20 in HIV pathogenesis was investigated in the current study. Our findings , as well as those of others [11, 14-16, 29, 30], suggest that endogenous anti-gpl20 antibody in AIDS patients' serum may actually promote rather than neutralizing and inhibitng HIV pathogenesis. Therefore, to avoid provoking even greater T cell depletion in AIDS, tolerogenic therapies should be considered in HIV vaccine design . This concept is supported by the results of Finkel et al. [16], who used huCD4 and HIV gpl 20 double transgenic mice to address the role of anti- gpl 20 antibody in T cell depletion. Their findings, that antibodies crosslinking huCD4/gpl20 complexes are a determinant for the outcome of the T cell responses to stimuli in vitro and in vivo, are consistent with the present results.

Here, we followed the numbers of CD3+ T cells in human CD4 transgenic mice in vivo after immunization and a single intravenous exposure to soluble gpl 20. In such mice, a bolus of gp 120 led to rapid depletion of CD3+ T cells in the periphery and in the spleen of gpl 20 immunized huCD4 transgenic mice; non- immune transgenic or immunized non-transgenic mice, were unaffected by this treatment. We measured CD3 T cells instead of human CD4 expressing cells since, in preliminary studies, we found that membrane expression of huCD4 was either downregulated or blocked by gpl20:anti-gpl20 binding. Expression of huCD4 is driven in these mice by the CD2 promoter, which results in expression in both T and B cells (over 80% of splenocytes). Our results indicate that huCD4-expressing B cells may also be depleted in vivo. In humans, a very small number of B cells (0.1-1%) express CD4 molecules annd their function in human is still unclear, as is their fate in HIV infection. More importantly, the present results suggest that huCD4 crosslinking- transmitted death signal might not necessarily require association with TCR/CD3 signaling pathway and that, under appropriate circumstances, huCD4 crosslinking is enough to send the death signal and induce the cells to die.

Because we cannot exclude Fas involvement in huCD4 crosslinking-mediated apoptosis, it is important to know the threshold of Fas expression in apoptosis induction. While there is a higher percentage of Annexin V positive T cells in the lymph nodes and spleens from gpl20-pretreated immunized huCD4 transgenic mice than those in the control mice, we have not detected a significant increase in apoptotic cells in vivo in these tissues using in situ TUNEL staining. Annexin V binds to phosphatidylserine, which is translocated from the inner side of the plasma membrane to the outer layer and becomes exposed at the external surface on early apoptotic cells [41]. This event occurs well before the DNA fragmentation as measured by in situ TUNEL staining, therefore, it is possible that rapid phagocytosis of early stage apoptotic T cells may be occurring so that an increased level of apoptotic cells may be evanescent.

Experiments in which repeated gpl 20 injections are given to the immunized huCD4 transgenic mice suggest that at least some antigen specific T cells and B cells have been depleted by gpl20/anti-gpl20 antibody crosslinking huCD4 via apoptosis. However, to our sinprise, multiple gpl 20 injections did not induce a long-lasting T cell loss from the periphery, though the splenocytes from these mice were still primed for TCR activation-induced apoptosis in vitro. Given the fact that huCD4 transgenic mice produced a large amount of anti-gpl20 antibodies after the first bolus of gpl 20, circulating in the bloodstream, gpl 20 injected thereafter may not be able to compete for binding to huCD4 molecules, tthe relatively rapid recovery of CD3 T cells in vivo may reflect the small amount of available gpl 20 when delivered to the bloodstream of immunized mice as a bolus, in contrast to the small but steady production of viral gpl 20 by HIV in infected individuals. Recent studies [16] with non-tolerant gpl 20 transgenic mice are encouraging for the validity of this model. Transgenic mice in which production of gpl 20 can be controlled by an inducible promoter (to avoid partial tolerance seen in other work [16]) would be helpful in this direction. In summary, our studies demonstrate that an ongoing humoral immune response to gpl 20 in huCD4 transgenic mice can participate in priming T cells to die upon exposure to gpl 20. While transient, these primed T cells remain hypersensitive to activation-induced death in vitro by antiCD3 crosslinking, and undergo apoptosis in vivo when exposed to a specific antigen that can bind to their TCR. We believe this system will be useful for evaluating one of the fundamental processes underlying HIV pathogenesis. Using the present invention, it will be possible to test potential immune modulation therapies to reverse this sensitivity to apoptosis by inducing tolerance to gpl 20 epitopes [40]. References

(1) Ameisen, J. C. et al, Immunol. Today 1991. 12:102; (2) Starcich, B. R. et al, Cell 1986. 45:637; (3) Abacioglu et al, supra; (4) Modrow, S.et al, J. Virol. 1987. 61 :570; (5) Willey, R. L.et al, Proc. Natl. Acad. Sci. USA 1986. 83:5038; (6) Weiss, R. A.et al, Nature 1985. 316:69; (7) Katz, D. H., AIDS Res. Hum. Retro. 1993. 9:489; (8) Golding, H. et α/., J. Exp. Med. 1988. 167:914; (9) Kion, T. A. et al, Science 1991. 253: 1 138. (10)

Desbarats et al. supra; (1 1) Banda et al, supra fl. Exp. Med. 1992. 176: 1099; (12) Newell et al, supra . [Nature 1990. 347:286; (13) Oyaizu et al, \994, supra [Blood 1994. 84:2622; (14/15) Wang et α/., 5wpra; (16) Finco, O. et al, Eur. J. Immunol. 1997. 27:1319; (17) Paterson, R. K. et al, J. Immunol. 1994. 153:3491; (18) Posner, M. R. et al, Acquired Immune Defic. Syndr. 1993. 6:7; (19) Cavacini, L. A. et al, J. Acquired Immune Defic.

Syndr. 1993. 6:353; (20) Posner, M. R. et al, Hybrido a 1987. 6:61 1 ; (21) Posner, M. R. et al, J. Immunol. 1991. 146:4325; (22) Ugen, K. E. et al, In: Vacines 93, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. 1993. p. 215; (23) Haigwood, N. L. et al, AIDS Res. Hum. Retro. 1990. 6:855; (24) Steimer, K. S. et al, In: Vacines 88, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. 1988. p. 347; (25) Levy, J. A. et al, Science

1984. 225:840; (26) Sanchez-Pescador et al, supra; (27) Haigwood, N. L. et al,, In: Vacines 90, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. 1990. p. 313; (28) Ezhevsky, S. A. et al, Mol. Biol. Cell 1996. 7:553; (29) Goronzy, J.et al, J. Exp. Med. 1986. 164:911; (30) Jamali, l.et al, J. Immunol. 1992. 148:1613; (31) Muro-Cacho, C. A. et al, J. Immunol. 1995. 154:5555; (32) Chun, T.-W. et al, Nature 1997. 387: 183; (33)

Haφer, M. E. et al, Proc. Natl. Acad. Sci. USA 1986. 83:772; (34) Lores, P. et al, AIDS Res. Hum. Retro. 1992. 8:2063; (35) Feng, Y. et al, Science 1996. 272:872; (36) Scott, D. W. et al, J. Immunol. 1996. 156:2352; (37) Lanzavecchia, A., J. Exp. Med. 1995. 181 : 1945; (38) Salemi, S. et al, J. Exp. Med. 1995. 181 :2253; (39) Siliciano, R. F. et al, Cell 1988. 54:561; (40) Zambidis, E. T. et al, Mol. Med. 1997. 3:212; (41) Vermes, I. et al, J.

Immulnol. Meth. 1995. 184:39. EXAMPLE VII

Induction of Immune Tolerance to Foreign Immunogenic Epitopes via Retroviral-

Mediated Expression of Foreign Protein: the IgG Scaffold is Important for Induction and Maintenance of Humoral Immunological Tolerance Results described above showed that recipients of bone marrow or of LPS- stimulated B-cell blasts, both of which were retrovirally gene-transferred with an immunodominant peptide in-frame to the V region of a murine IgG H chain, were rendered profoundly unresponsive to that epitope. To further investigate whether tolerance to larger molecules can be achieved via this approach and whether the IgG scaffold is important for induction and maintenance of immunological tolerance, we engineered two retroviral constructs (MBAE-1-102 and MBAE-l-102-IgG). The first of these included the DNA encoding the pi -102 peptide of bacteriophage λ. The second had DNA encoding that peptide fused to the murine Ig γ chain DNA such that the peptide was expressed at the N-terminus of the H chain. These vectors were used for gene transfer.

Specificity of pi -102 humoral tolerance in genetically tolerized bone marrow recipients was examined. CB6 F, mice were sublethally irradiated with 400 rads and injected with mock-transduced or 1-102 -IgG gene-transduced bone marrow cells. Mice were primed and boosted with pi -102 and HEL in CFA. Antibody responses were measured in ELISA by coating plates with 50 μg/ml synthetic peptides (peptides

12-26, 73-88 or 55-69). The titers were determined by using CA-Cricket graph software and expressed as the dilution which brings the OD₄₀₅ to the pre-immune level. Each experiment had 3-4 mice per group. The efficacy in induction and maintenance of tolerance by pi -102 and pl-102-IgG gene transfer in bone marrow recipients was also examined. CB6 F, mice were sublethally irradiated as above and injected with mock-transduced or pi -102 transduced or pi -102 -IgG gene-transduced bone marrow cells. Mice were later primed and boosted with pl-102 and HEL in CFA. Antibody responses were assayed and analyzed as above.

The results showed that recipients of bone marrow cells or peripheral B cells that had been gene-transferred with MB AE- 1 - 102-IgG were specifically hyporesponsive to the pi -102 protein (and this occurred in a strain-specific manner). That is, Balb/c and C57B1/6 mice recognize epitopes contained in residues 12-26 and 73-88, respectively, whereas F, hybrid mice between these strains recognize epitopes at both sites. Gene-transfer produced tolerance in Fi mice to the whole pi -102 protein, as well as to the major determinants. No "epitope spreading" to minor epitopes was observed. The results suggest that the self IgG scaffold is necessary for long-lasting unresponsiveness because recipients of pl-102-IgG fusion protein construct remained tolerant to secondary challenge whereas controls given cells transfected with pi -102 construct (not fused to the IgG) regained responsiveness. These results demonstrated that the host can then present the relevant epitopes in a

MHC-haplotype-specific manner to the immune system and induce profound tolerance. This results is directly applicable to treatment of autoimmune diseases, as well as for creating a receptive environment for foreign or otherwise immunologically "unacceptable" proteins to be administered in the context of gene therapy.

The references cited above are all incoφorated by reference herein, whether specifically incoφorated or not.

Having now fully described this invention, it will be appreciated by those skilled in the art that the same can be performed within a wide range of equivalent parameters, concentrations, and conditions without departing from the spirit and scope of the invention and without undue experimentation.

While this invention has been described in connection with specific embodiments thereof, it will be understood that it is capable of further modifications. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains and as may be applied to the essential features hereinbefore set forth as follows in the scope of the appended claims

Claims

WHAT IS CLAIMED IS:

1. A fusion immunoglobulin (Ig) heavy (H) chain protein comprising a mammalian Ig H chain fused in frame in its N-terminus to one or more HIV gpl 20 epitopes, wherein said fusion Ig H chain is tolerogenic in a host with respect to said gpl 20 epitopes.

2. A fusion Ig H chain protein according to claim 1, wherein said gpl 20 epitope is fused to said mammalian Ig H chain at its N-terminal region.

3. A fusion Ig protein comprising two Ig H chains and two Ig light (L) chains, wherein at least of said H chains is a fusion Ig H chain according to claim 1.

4. A fusion Ig protein according to claim 3, wherein both of said H chains are Ig H chains according to claim 1.

5. A fusion Ig protein according to claim 1 , wherein said H chain is a Ig ╬│ chain.

6. A fusion Ig protein according to claim 5, wherein said H chain is a human Ig ╬│₃ chain

7. A fusion Ig H chain according to claim 1 , wherein said one or more gpl 20 epitopes comprises a full length gpl 20 protein.

8. A fusion Ig H chain protein according to claim 1, wherein said mammalian Ig is human Ig.

9. A fusion Ig H chain according to claim 1 , wherein said one or more gpl 20 epitopes comprises a Cl region peptide, a V3 loop peptide or a C5 region peptide.

10. A fusion Ig H chain according to claim 1 , wherein said one or more gpl 20 epitopes is a B cell epitope which is a peptide selected from the group consisting of VPVWKEATTTLFCASDAKAY (SEQ ID NO:2), EVHNVWATHACVPTD (SEQ ID NO:3), YDTEVHNVWA (SEQ ID NO:4), PQEVVLVNVT (SEQ ID NO:5), PQEVVLVNVTENFDMWKNDM (SEQ ID NO:6), PNNNTRKSIR (SEQ ID NO:7), NNNTRKRIRIQRGPGR (SEQ ID NO:8), RKSIR (SEQ ID NO:9), IQRGPGRAFV (SEQ ID NO: 10), GRAFVTIGKI (SEQ ID NO: l 1), PGRAFY (SEQ ID NO: 12),

NTRKSIRIQRGPGRAFVTIG (SEQ ID NO: 13),

PNNNTRKSIRIQRGPGRAFVTIGKIGNMRQAHC (SEQ ID NO: 14), NNTRKSIRIQRG (SEQ ID NO: 15), NKRKRIHIGPGRAFYTTKNIIGTIC (SEQ ID NO: 16), RKSIRIQRGPGRAFV (SEQ ID NO: 17), IRIQRGPGR (SEQ ID NO: 18), KRIRIQRGPGRAFVTIG (SEQ ID NO: 19), QRGPGRAF (SEQ ID NO:20), RGPGRAFV

(SEQ ID NO:21), RKRIHIGPGRAFYTT (SEQ ID NO:22), RGPGRAFVTIG (SEQ ID NO:23), SISGPGRAFYTG (SEQ ID NO:24), KRIHI (SEQ ID NO:25), KRIHIGP (SEQ ID NO:26), IHIGPGR (SEQ ID NO:27), HIGPGR (SEQ ID NO:28), HIGPGRA (SEQ ID NO:29), HIGP (SEQ ID NO:30), RIHIGPGRAFYTTG (SEQ ID NO:31), RIQRGPGRAF (SEQ ID NO:32), IQRGPGRAFV (SEQ ID NO: 10), IQRGPGRAF (SEQ ID NO:33),

IRIQRGPGRAFVTI (SEQ ID NO:34), RGPGRAFVTIGKIG (SEQ ID NO:35), QRGPGRA (SEQ ID NO:36), IXXGPGRA (SEQ ID NO:37), IGPGR (SEQ ID NO:38), GPGR (SEQ ID NO:39), GPXR (SEQ ID NO:40), GPGRAF (SEQ ID NO:41), RIHIG (SEQ ID NO:42), HIGPGRAF (SEQ ID NO:43), GRAF (SEQ ID NO:44), GGGDMRDNWRSELYKYKVVK (SEQ ID NO:45), KYKVVKIEPLGVAPTKAKRR (SEQ ID NO:46), LGVAPTKAKR (SEQ

ID NO:47), GGDMRDNWRSELYKYKVVKI (SEQ ID NO:48), IEPLGVAPTK (SEQ ID NO:49), RRVVQRE (SEQ ID NO:50), PTKAKRR (SEQ ID NO:51) and WQREKR (SEQ ID NO:52).

11. A fusion Ig H chain according to claim 1 , wherein said one or more gpl 20 epitopes comprises a T cell epitope which is a peptide selected from the group consisting of EQLWVTWYGVPV (SEQ ID NO:53), VYYGVPVWKEA (SEQ ID NO:54), GVPVWKEATTLFC (SEQ ID NO:55), AHKVWATHACV (SEQ ID NO:56), NVWATHACVPTD (SEQ ID NO:57), CVPTNPVPQEVV (SEQ ID NO:58), VEQMHEDIISLW (SEQ ID NO:59), EQMHEDIISLWDQ (SEQ ID NO:60), EQMHEDIISLWDQSL (SEQ ID NO:61), HEDIISLWDQSLK (SEQ ID NO:62),

VTVYYGVPVWKEATTTLFC (SEQ ID NO:63), WLVNVTENFNM (SEQ ID NO:64), SLKPCVKLTPLCY (SEQ ID NO:65), CTRPNNNTRKSIRIQRGPGY (SEQ ID NO:66), NTRKSIRIQRGPGR (SEQ ID NO:67), EQRGPGRAFVTIGKI (SEQ ID NO:68), RIQRGPGRAFVTIGK (SEQ ID NO:69), RIHIGPGRAFYTTKN (SEQ ID NO:70), GRAFVTIGKIGNMRQ (SEQ ID NO:71), QRGPGRAFVTIGKIGNMRQAH (SEQ ID NO:72), VGKAMYAPPISGQIR (SEQ ID NO:73), GNSNNESEIFRPGGG (SEQ ID NO:74), FRPGGGDMRDNWRSEL (SEQ ID NO: 75), DMRDNWRSELYKYKV (SEQ ID NO:76),

RDNWRSELYKYKWK (SEQ ID NO:77), CKYKWKIEPLGVAPT (SEQ ID NO:78), YKYKVVKIEPLGVAP (SEQ ID NO: 79), KVVKIEPLGVAPTKAKRRVVQREKRC (SEQ ID NO:80), ITLPCRIKQIINMWQEVGKAMYAPPISGQIRC (SEQ ID NO:81), and ELYKYKWKIEPLGVAPTKAKRRWQREKR (SEQ ID NO:82).

12. A DNA molecule comprising a nucleotide sequence encoding a fusion

Ig H chain according to claim 1.

13. A DNA molecule comprising a nucleotide sequence encoding a fusion Ig H chain according to any of claims 5-10.

14. An expression vector which expresses a product useful for inducing and maintaining immunological tolerance to one or more epitopes of HIV gpl 20 protein in a subject, comprising:

(a) a DNA molecule according to claim 1 , operably linked to

(b) transcriptional and translational control regions operable in a hematopoietic cell or lymphoid cell of said subject.

15. An expression vector which expresses a product useful for inducing and maintaining immunological tolerance to one or more epitopes of HIV gpl 20 protein in a subject, comprising:

(a) a DNA molecule according to any of claims 2 or 5-10, operably linked to (b) transcriptional and translational control regions operable in a hematopoietic cell or lymphoid cell of said subject.

16. A vector according to claim 13 which is a retroviral vector.

17. A vector according to claim 13, wherein the transcriptional and translational control regions provide for constitutive expression of the DNA sequence in lymphoid cells.

18. A hemopoietic or lymphoid cell transformed by a vector according to claim 13, which cell stably expresses said fusion Ig.

19. A hemopoietic or lymphoid cell transformed by a vector according to claim 14, which cell stably expresses said fusion Ig.

20. A hemopoietic or lymphoid cell transformed by a vector according to claim 15, which cell stably expresses said fusion Ig.

21. A human bone marrow cell transformed by a vector according to claim

13, which cell stably expresses said fusion Ig.

22. A cell according to claim 17 which is a B lymphocyte.

23. A B lymphocyte according to claim 21 which has been activated in culture.

24. A pharmaceutical composition comprising:

(a) a tolerogenic amount of a fusion Ig molecule having a fusion Ig H chain according to claim 1 ; and (b)a pharmaceutically acceptable carrier or excipient for parenteral administration.

25. A pharmaceutical composition according to claim 21 , wherein the Ig is an isologous IgG molecule.

26. A pharmaceutical composition comprising:

(a) a tolerogenic amount of a fusion Ig molecule having a fusion Ig H chain according to any of claims 2, or 5-10; and

(b) a pharmaceutically acceptable carrier or excipient for parenteral administration.

27. A pharmaceutical composition comprising:

(a) a tolerogenic amount of a fusion Ig molecule according to claim 3 or 4; and

(b) a pharmaceutically acceptable carrier or excipient for parenteral administration.

28. A method for immunologically tolerizing a subject to one or more HIV gpl 20 epitopes comprising administering to said subject an effective amount of a pharmaceutical composition according to claim 23.

29. A method for immunologically tolerizing a subject to one or more HIV gpl 20 epitopes comprising administering to said subject an effective amount of a pharmaceutical composition according to claim 24.

30. A method for immunologically tolerizing a subject to one or more HIV gpl 20 epitopes comprising administering to said subject an effective amount of a pharmaceutical composition according to claim 25.

31. A method for immunologically tolerizing a subject to one or more HIV gpl 20 epitopes comprising administering to said subject an effective amount of a pharmaceutical composition according to claim 26.

32. A method for immunologically tolerizing a subject to one or more HIV gpl 20 epitopes comprising introducing into said subject an effective amount of transformed cells according to claim 17, thereby tolerizing said subject.

33. A method for immunologically tolerizing a subject to one or more HIV gpl 20 epitopes comprising introducing into said subject an effective amount of transformed cells according to claim 18, thereby tolerizing said subject.

34. A method for immunologically tolerizing a subject to an HIV gpl 20 epitope comprising introducing into said subject an effective amount of transformed cells according to any of claims 19-22, thereby tolerizing said subject.

35. A method of (i) inducing and (ii) maintaining immunological tolerance to an epitope or epitopes of HIV gpl20 protein in a subject, comprising:

(a) administering to said subject an effective amount of a pharmaceutical composition according to claim 23 to induce said tolerance to said epitope or epitopes; and

(b) administering to said subject an effective amount of transformed hemopoietic or lymphoid cells according to claim 17 to maintain said tolerance to said epitope or epitopes, thereby inducing and maintaining said tolerance.

36. A method for identifying whether a candidate HIV gpl 20 epitope or epitopes are tolerogenic in a first subject when presented to said subject in a fusion Ig molecule, comprising the steps of:

(a) stably transforming a population of hematopoietic or lymphoid cells of said subject with a vector according to claim 13 in which the gpl 20 epitope or epitopes in the fusion Ig encoded y said vector are said candidate epitope or epitopes;

(b) introducing said transformed cells into said subject; and

(c) determining whether said subject is tolerant to said candidate epitope or epitopes by measuring the reduction of an ongoing antibody response and comparing with a similar response in a second control subject who is (i) untreated or (ii) treated with control untransfected cells or (Hi) treated with cells transfected with a control Ig lacking said candidate epitopes; such that, if said first subject is tolerant, said candidate epitope or epitopes in said fusion Ig is identified as being tolerogenic.

37. A method according to claim 36, wherein said first and second subjects are humans.