The present invention relates to materials and methods in relation to immunosuppression. One of the greatest challenges of modern immunology is to devise strategies for the selective abrogation of harmful immune response, as is the case in IgE-mediated allergies and auto-immune diseases. The ability to induce specific immunological suppression is also of particular importance for the development of novel therapies which depend on the administration of xenogeneic molecules such as proteins. Therapies based on xenogeneic proteins have rapidly expanded in recent years as a result of the increasing number of recombinant proteins, synthesized by new methods of molecular biology or produced by hybridoma technology. Among "these xenogeneic proteins one may cite diverse enzymes, various biological response modifiers (BRMS) such as lymphokines and hormones, urine monoclonal antibodies or recombinant; antibodies directed to specific determinants of human T cells such as CD3, CD4 or IL-2 receptors, or of adhesion molecules. In addition, the use of some of these xenogeneic proteins by themselves, some, typically monoclonal antibodies, (mAbs) have been also used in the form of conjugates with diverse radioactive elements and toxins eg chain A of ricin, for the in vivo diagnostic localization and destruction of undesirable cells such as cancer cells. However, although the xenogeneic proteins referred to above have been shown to be potentially useful additions to the therapeutic armamentarium in a number of diseases, their effectiveness is undermined by their inherent immunogenicity. This is so even where the mAbs have been humanised. Thus, the patient's production of antibodies against the xenogeneic treatment protein eg a murine mAb prevents or reduces its desired effector function. The resulting immune complexes often lead to

untoward pathophysiological effects, including serum sickness and occasionally even anaphylaxis.

UK Patent No. 1,578,348 and US Patent No. 4,261,973 disclose that an allergen (AL) such as ovalbumin and the non-dialyzable constituents of the aqueous extract of ragweed pollen and dog albumin, may be converted to a tolerogen by coupling it to an optimal number (n) of monomethoxy polyethylene glycol (mPEG) molecules. Injection of tolerogenic PEG conjugates of these ALs into rats and mice, prior to administration of the corresponding unmodified ALs led to the abrogation of the capacity of the mice to mount humoral antibody responses to those ALs.

UK Patent No. 2,238,959 followed on from the above work and disclosed that pre-treatment of a recipient with a tolerogen suppresses the immune response not only to the antigen incorporated in the tolerogen, but also to a conjugate of that antigen and at least one additional antigenic moiety which may be a hapten or another unrelated protein. For example, the patent discloses that injection of a tolerogenic conjugate of human IgG into mice, prior to administration of conjugates of human IgG with either dinitrophenyl DNP or DNP-keyhole limpet haemocyanin (KLH) led to the abrogation of the capacity of the mice to mount humoral antibody responses to both human IgG and the conjugated moiety DNP or DNP-KLH. If however DNP-KLH was injected into mice pretolerized with a noncovalent mixture of DNP-KLH and a PEG conjugate of human IgG, the mice mounted normal humoral ,antibody responses to DNP and KLH, but remained suppressed to human IgG.

Other studies in this field have been reported. Katre, N.V. (1990 J. Immunol. 144, 209-213) reports that the immune response of rabbits to the immunogenic recombinant versions of human IL-2 (rhuIL-2) was

suppressed by administration of the tolerogenic mPEG conjugate of rhuIL-2 prior to injection of rhuIL-2 by itself. Most importantly, the administration of diverse xenogeneic proteins (Meyers F.T., et al 1991 Clin. Pharmacol. Ther. 49, 307-313 and Nucci, M. L. , et al. 1991, Adv. Drug Delivery 6, 133-151) including ALs (Dreborg, S., et al 1990 Crit. Rev. Ther. Drug Carrier Syst. 6, 315-365) has been shown to be safe. For example, non-immunogenic mPEG conjugates of bovine adenosine deaminase are being repeatedly injected into children with severe combined immuno-deficiencies (Hershfield M.S. et al, 1987 N. Engl. J. Med. 316, 589- 596). Maiti, P.K., et al 1988 (Int. J. Cancer Suppl. 3, 17-22) report that a single injection into mice of appropriately synthesized mPEG conjugates of monoclonal (myeloma) human IgG (hulgG) resulted in long term (>300 days) suppression (>95%) of the murine anti-hulgG antibody (Ab) response in spite of repeated, subsequent immunizing injections of hulgG. Lang G.M. , et al. 1992 Immunol. Lett. 32, 247-252 reports that mPEG conjugates of murine mAbs to both human and murine CD4 epitopes suppressed antibody responses of outbred Long Evans rats to isotypic and idiotypic determinants of the antibodies. Mosier, D.E., 1990 J. Clin. Immunol. 10, 185-191 and

McCune, J. et al, 1991 Annu. Rev. Immunol. 9, 399-429 report that severe combined immunodeficient (SCID) mice reconstituted with normal human peripheral blood leucocytes (PBL) ie hu-PBL-SCID mice pretreated with tolerogenic mPEG conjugates of murine mAbs resulted in suppression of the human anti-mouse Ig antibody (HAMA) response (including anti-idiotype antibody) to subsequent injections of murine anti-human CD4 mAbs. hu-PBL-SCID mice represent the closest recently developed in vivo model of the human lymphoid system.

The present application concerns discoveries and ideas going on from the prior art reported above and technical applications based thereupon. Firstly the application discloses that treatment of hu-PBL-SCID mice with a tolerogenic covalent conjugate of mPEG and an anti-ovalbumin IgGl murine mAb (Mab-2), suppressed the human anti-mouse Ab responses to both the common (γl,κ) and the idiotypic determinants of Mab-2. Moreover, the Mab-2(mPEG) ₃₆ conjugate suppressed the immune responses of hu-PBL-SCID mice to the common and idiotypic determinants of murine mAbs to the 2, 4-dinitrophenyl residue and to human CD4, consisting also of γl and K chains. Thus it appears that a tolerogenic mPEG conjugate of a murine mAb induces pan-suppression of the human lymphoid system with respect to other murine mAbs that share the isotypic determinants of the original mAb incorporated in the conjugate. Hence it appears that anti-mouse antibody responses to any murine IgG mAb would be suppressed by one of eight mPEG conjugates, each incorporating one of the four subclasses of IgG and one of the two light chains. Generally this shows that quite surprisingly the substance eg mAb, AL, BRM in the mPEG need not be exactly the same as the substance for which immunosuppression is sought, but only needs to have structural elements in common. Therefore the present application provides a tolerogen for administering to a patient in order to reduce the patient' s immune response to a given antigen (eg antigen 'X' ), which tolerogen comprises a water- soluble covalent conjugate of another antigen (eg antigen Y) with one or more non-immunogenic water-soluble polymers, wherein said another antigen although different to said given antigen has structure in common with it.

The given antigen and the another antigen may both comprise at least part of a xenogeneic protein. The xenogeneic proteins may be antibodies. The antibodies

may be rodent. The antibodies may be murine. The antibodies comprising the given antigen and the another antigen may share the same isotypic determinants. The antibodies comprising the given antigen and the another antigen may have different idiotypes.

Also provided is a kit which comprises a panel of eight different tolerogens as described above. There will be first to fourth tolerogens comprising antibodies with kappa light chains and of isotypes IgGl, IgG2, IgG3 and IgG4 respectively and fifth to eight tolerogens comprising antibodies with lambda light chains and of isotypes IgGl, IgG2, IgG3 and IgG4 respectively.

Also provided is a method for making a tolerogen as described above which comprises the steps of selecting another antigen which is different to said given antigen but which has structure in common with it; and conjugating said another antigen to one or more non- immunogenic water-soluble polymers to render it tolerogenic. The water soluble polymer may be selected from the group consisting of poly(alkylene-glycols) , poly(vinylalcohols) , poly(vinylpyrrolidones) poly(acrylamides) homo- and hetero- polymers of amino acids, poly(saccharides) physiologically-acceptable derivatives, mixtures, combinations and functional equivalents thereof. The polymer may be poly(alkylene glycol) or its monomethoxy derivative. The polymer may be poly(ethylene glycol) or its monomethoxy derivative. Where the water soluble polymer is poly(ethylene glycol) it may have a molecular weight in the range of 2000-

35,000. Preferably the molecular weight may be in the range of 3000-6000.

Also provided is a method of treating a patient with a tolerogen as described above in order to reduce the patient's immune response to said given antigen.

The patents may be human.

In the disclosures to date, tolerogens have been disclosed for the purpose of suppressing a patient's immune response against a xenogeneic protein eg a mAb in order that the mAb may attack an undesired pathogen or tissue eg a tumour. However, the applicants have realised that in certain circumstances the body mounts an undesirable immune response against some xenogeneic material for example an allograft. Treatment Abs may for example, be used which are specifically directed against a mediator of the patent' s immune response against an allograft. However and as discussed in the introduction, the body may also mount an immune response against the treatment Ab which are themselves xenogeneic. Therefore a tolerogenic conjugate of the treatment Ab can be used to suppress the immune response against the treatment Ab, which is therefore able to act against the mediator and reduce the patient's immune response against the eg allograft. Thus the present application provides a tolerogen for administering to a patient in order to reduce their immune response to a treatment antibody or antibody binding fragment which is specific for an epitope of a mediator of the patient's immune response against a xenogeneic epitope, which tolerogen comprises a water- soluble covalent conjugate of one or more non-immunogenic water-soluble polymers with said treatment antibody or antibody binding fragment.

The xenogeneic epitope may be presented by an allograft. The allograft may be a heart.

The mediator may be a cytotoxic T lymphocyte. The epitope may be part or all of the T cell receptor.

The treatment antibody or antibody binding fragment may be rodent derived. It may be murine derived. The treatment antibody may be anti-clonotypic.

Also provided are pharmaceuticals comprising tolerogens as described with one ore more excipients.

Tolerogens as described may be used in the preparation of medicaments for the control of immune responses against any xenogeneic materials. For example, for the control of immune responses involved in allograft rejection, or immune responses directed against a treatment substance eg a protein. Tolerogens as described may be used in the treatment of patients, particularly human patients, where it is desired to control immune responses against any xenogeneic material. For example to treat patients to prevent allograft rejection or to treat patients to prevent or limit the immune response mounted against a treatment substance eg a protein.

Finally, it is known that hu-PBL-SCID mice represent an in vivo model of the human immune system. The present applicants have however realised that a SCID mouse may be engrafted with PBL of a particular patient thereby providing that patient with their own personalised in vivo model of their immune system. This means that any tests which involve a monitoring of that particular patients immune response may first be tried out on their personal in vivo model. This is of enormous benefit, not least because it minimises patient trauma. Thus the present application also provides a method for testing a patient's (P) individual immune response to antigen 'X' which comprises the steps of:

(i ) isolating a sample of peripheral blood leucocytes (PBL) from the patient;

(ii) engrafting a mouse having severe combined immunodeficiency (SCID) with said sample in order to create a p-PBL-SCID test mouse having PBLs of the patient; (iii) immunizing said test mouse with antigen 'X';

and

( iv) detecting any patient antibody response in said test mouse.

Between steps ( ii) and (iii) the method may also comprise treating said test mouse with a tolerogen designed for administration to the patient to reduce the patient' s immune response to antigen 'X' . The tolerogen may comprise a water-soluble covalent conjugate of antigen 'X' with one or more non-immunogenic water- soluble polymers. The PBL sample may comprise both T cells and B cells. The PBL sample may be enriched for B cells. The PBL sample may contain approximately 20xl0 ⁶ T cells and 20xl0 ⁶ B and MN cells.

The patient antibody response in the test mouse may be compared to a patient antibody response mounted in a control mouse which comprises a further p-PBL-SCID mouse treated with suitable control material, for example PBS.

The method may comprise detecting a patient antibody response which is specific to said antigen 'X'. The sample of PBL may be substantially free of erythrocytes.

The patient antibody response may be detected by immunoassay. The immunoassay may be an enzyme immunoassay. The antigen 'X' may comprise at least part of a xenogeneic protein. The xenogeneic protein may comprise an antibody. The antibody may be of rodent. The antibody may be murine.

The antibody may be specific for a target antigen in the patient. The target antigen may comprise a part or product of a pathogen. The target antigen may comprise a part or product of the patient's cell. The patient may be human.

Also provided is a SCID mouse for use in a method as described above which has been engrafted with a sample of

PBL from a particular patient.

The ideas presented above have been discussed independently of one another to aid understanding. However the concepts may be combined in various ways. For example the method for testing a patient's individual immune response using a tailorised mouse model of their immune system, can be used to test out an immune response against any treatment material eg mAb and to establish the effect of a tolerogen in reducing that immune response. The treatment mAb could be for control of the patient' s own immune response against a desired but xenogeneic material eg an allograft. As explained earlier, the mAb in the tolerogen may be different to the treatment mAb provided they have structural elements eg isotypic determinants, in common.

Thus the present application also provides a method for testing a patient's (P) individual immune response to an antibody 'X' or a binding fragment thereof, which is specific for an epitope of a mediator of the patient's immune response against an antigen ' Z ' , which method comprises the steps of: (i) isolating a sample of peripheral blood leucocytes (PBL) from the patient; (ii) engrafting a mouse having severe combined immunodeficiency (SCID) with said sample in order to create a p-PBL-SCID test mouse having PBLs of the patient; (iii) treating said mouse with a tolerogen for administration to the patient to reduce the patient's immune response to antibody 'X'; (iv) immunising said test mouse with antibody 'X' ; (v) detecting any human antibody response in said test mouse; wherein the tolerogen comprises a water-soluble covalent conjugate of an antibody 'Y' with one or more non-immunogenic water soluble polymers and wherein said antibody 'Y' although different from antibody 'X' has structure in common with it.

In order that the above is understood more clearly, embodiments and experimental data are described in detail by way of example only, and not by way of limitation. Reference is made to the following figures. Figure 1 shows the relationship between human IgG level and HAMA response in hu-PBL-SCID mice. Twenty-three 3 to 4-week old SCID mice were engrafted i.p. with 20xl0 ⁶ T and 20xl0 ⁶ (B+MN) cells on day 0. Eleven of the mice were injected with the tolerogen, Mab-2(mPEG) ₃₆ , and the remaining 12 mice were injected with PBS prior to immunization with 20 μg of ha-Mab-2 on days 8 and 28. Each hu-PBL-SCID mouse was bled on day 42 and the sera were assayed for total human serum IgG levels and HAMA response; each point on the graph represents the relationship between these two parameters with respect to one of the test and control mice (represented, respectively, by circles and squares).

Figure 2 shows a flowchart for the protocol for testing transferable suppression in hu-PBL-SCID mice; Figure 3 shows a flowchart for the protocol for generation of responder CD4 ^* T cells.

Each SCID mouse received 50xl0 ⁶ PBL of each volunteer on day 0. All mice were immunized 12 hr later with 100 μg of Mab-2 emulsified in FCA. Two weeks later, human leucocytes were isolated from pooled spleens and lymph nodes of each group of 2-4 hu-PBL-SCID mice and treated with anti-H-2 ^d mAbs and RC. The cells were stimulated twice with 100 μg/ml of Mab-2 or 20 μg/ml of PPD at 21 days interval. Two weeks after the last stimulation, the culture cells were treated with 0KT8 plus RC . The residual viable cells were used as responder CD4 ^* T cells. Figure 4 shows suppression of HAMA to the idiotype of murine anti-OVA monoclonal Mab-2 by Mab-2(mPEG) ₃₆ conjugate.

Panels A and B: hu-PBL-SCID mice injected with either 200 μg of Mab-2(mPEG) ₃₆ (circles) or PBS (squares)

were immunized with 1 μg of Mab-2 in FCA 7 and 21 days later, and were bled 21 days after the second immunization (ie, 42 days after receipt of the tolerogen of PBS). All sera were diluted 1/100 in 0.01% normal murine Balb/c serum (NMS) and diluted two-fold in the presence of 0.01% NMS prior to being tested on DNP ₉ -0VA coated plates. Each curve represents the data from the serum of one mouse. Panel A: Each serum was titrated in the presence of a mixture of biotinylated and non- biotinylated Mab-2; Panel B: Each serum titrated in the presence of a mixture of biotinylated and non- biotinylated Hl-DNPγ-109.3; Panel C: Each serum from the PBS treated group was passed through an immunosorbent composed of rabbit anti-human IgG (open squares) or an immunosorbent composed of rabbit anti-mouse Ig (closed squares) prior to titration in the presence of biotinylated and non-biotinylated Mab-2.

Figure 5 shows suppression of HAMA to the idiotype of the murine anti-DNP mAb, Hl-DNPγ-109.3, by Mab- 2(mPEG) ₃₆ conjugate.

Panel A and B: hu-PBL-SCID mice injected with either 200 μg of Mab-2(mPEG) ₃₆ (circles) or PBS (squares) were immunized with 5 μg of Hl-DNP-γ-109.3 in FCA 7 and 21 days later and bled 21 days after the second immunization (ie, 42 days after receipt of the tolerogen or PBS). All sera were diluted 1/100 in 0.01% normal murine Balb/c serum (NMS) and diluted two-fold in the presence of 0.01% NMS prior to being tested on DNP ₉ -0VA coated plates. Each curve represents the data from the serum of one mouse. Panel A: Each serum was titrated in the presence of a mixture of biotinylated and non-biotinylated Mab-2.

Panel B: Each serum was titrated in the presence of a mixture of biotinylated and non-biotinylated Hl-DNPγ- 109.3. Panel C: Each serum from the PBS treated group was

passed through an immunosorbent composed of rabbit anti- human IgG (open squares) or an immunosorbent composed of rabbit anti-mouse Ig (closed squares) prior to titration in the presence of biotinylated and non-biotinylated Mab- 2.

Figure 6 shows suppression of HAMA to the idiotype of the murine anti-human CD4 mAb, Leu3a, by Mab-2(mPEG) ₃₆ conjugate.

Hu-PBL-SCID mice injected with either 100 μg of Mab- 2(mPEG) ₃₆ or PBS received two immunizing injections of 1 x 10 ⁷ T cells coated with Leu3a 7 and 28 days later; these mice were bled 14 days after the second immunizing injection of the Leu3a-coated cells. All sera were diluted 1/100 in 0.5% human cord serum, and 1.0 ng/ml of protein A purified mouse IgG and diluted two-fold in the presence of 0.5% human cord serum and 1.0 ng/ml of protein A purified mouse IgG prior to addition to the"inhibition of proliferation" test system. Each curve represents the data from the serum of one mouse. The data depicted by the empty squares represent the proliferation of OVA-reactive T cells in the presence of OVA and in the absence of anti-CD4 mAb and of sera from hu-PBL-SCID mice. The data depicted by the filled squares represent the background proliferation of OVA- reactive T cells in the absence of OVA, of anti-CD4 mAb and of sera from hu-PBL-SCID mice.

Panel A: Titers obtained in the presence of 0KT4 mAb for individual sera of control mice treated with PBS (in lieu of the tolerogen). Panel B: Titers of individual sera of mice treated with the tolerogen in the presence of 0KT4.

Panel C: Titers of individual sera of mice treated with PBS (in lieu of the tolerogen) in the presence of Leu3a; Panel D: Titers of individual sera of mice, which

had been treated with the tolerogen, in the presence of Leu3a mAb.

Figure 7 shows a schematic illustration of the basic principles underlying the strategy for the protection of the donated heart from destruction by the patient's cellular and humoral immune responses.

Figure 8 is a flowchart of the strategy. The numbers refer to the procedures corresponding to the sections identified under "Specific Aims". Median Survival Time is abbreviated to MST.

Part 1

Mice having the condition of severe combined immunodeficiency (SCID) lack functional T and B leucocytes and are, therefore, unable to reject xenogeneic cells and may be engrafted with a healthy individual's human peripheral blood leucocytes (hu-PBL). Remarkably, very few of the SCID mice exhibit a transient graft versus host reaction. The hu-PBL-SCID mice produce on immunization Ag-specific human Ab responses. Hence the present applicants have realised that the hu-PBL-SCID mouse system represents an ideal model for testing the immune responses of individual patients to a given BRM and the possibility of suppressing this response with tolerogenic conjugates.

Specifically, in this study the applicants, used (as a model BRM) the murine mAb directed to ovalbumin (OVA) , referred to as Mab-2, both as an immunogen and for the synthesis of the corresponding tolerogenic Mab-2(mPEG) ₃₆ conjugate. Utilising this system the applicants demonstrated that (i) treatment of hu-PBL-SCID mice with Mab-2(mPEG) ₃₆ 7 days before immunization with heat-treated Mab-2 (ha-Mab-2) induced tolerance of the human anti- murine Ig Ab (HAMA) response to Mab-2, but not to OVA, and ( ii ) the mechanism underlying the induction of

immunosuppression by tolerogenic mPEG conjugates in hu- PBL-SCID mice appeared to be identical to that responsible for suppression of intact normal mice which are subjected to the same type of tolerogenic protocol, ie, the suppression was found to be due to the generation of human suppressor CD8 ^* T (Ts) cells, which downregulated the CD4 ⁺ helper T (The) cells in an Ag- and MHR class I- specific manner.

MATERIALS AND METHODS Animals

BALB/c scid/scid mice were obtained from Jackson Laboratories (Bar Harbour, ME) and bred by brother-sister mating under sterile conditions in the Central Animal Care Facility of the University of Manitoba. A breeding nucleus of C.B-17 scid/scid mice was a generous gift of Dr. D.E. Mosier (La Jolla, CA). The sera of the offspring of both strains were tested for total mouse Ig at 3- to 4-week intervals after birth by ELISA (detection limit: 50 ng/ml). Mice which had less than 50 ng/ml of mouse serum IgG were selected for breeding purposes and their Ig levels were monitored at 2- to 3-week intervals. The Ig level of pregnant female mice were determined 2-3 days before delivery and that of their offspring at intervals of 3-4 weeks. The mouse serum IgG levels were also monitored throughout the experiments at intervals of 3-4 weeks and animals having mouse serum Ig levels in excess of 500 ng/ml were eliminated. Peripheral Blood Leucocytes (PBL) The PBL were collected by leukaphoresis from three healthy EBV-, HIV- and HBV-seronegative donors and were isolated by discontinuous Ficoll-sodium metrizoate density gradient (d=1.091). The contaminating erythrocytes were lysed by ammonium chloride. The PBL of each donor were partitioned by passing through a nylon

wool column into cell fractions enriched with respect to T cells (CD3 ^* > 90%; CD3 ^" /DR ^* < 5%) and B plus MN cells (CD3 ^* < 10%;CD3 ^" /DR ^* > 85% (Julius, M.H., et al. Eur. J. Immunol. 3, 645-9, 1973). The PBL, as well as the T and (B+MN) cell fractions, which had been suspended in 90% of heat inactivated fetal calf serum (FCS) and 10% DMSO, were frozen at a freezing rate of -l°/min with the aid of a programmed freezer (CryoMe , Mt. Clemens, MI ) and maintained in liquid nitrogen until used. The HLA type of each donor is given in the footnote to Table 3. For some experiments, the PBL were freshly collected from a donor 12 months after immunization with diphtheria- pertussis-toxoid vaccine. Antibodies The hybridoma cell line producing BALB/c anti-DNP mAb (Hi-DNP-109.3; γ _x κ ) (Liu F.E., et al, J. Immunol. 124, 2728-37, 1980) was kindly provided by Dr. H. Yamamoto (National Center of Neurology and Psychiatry, Tokyo, Japan). Hybridoma cell lines producing mAbs to human CD3 (0KT3), human CD4 (OKT4), human CD8 (0KT8 ) , K ^d (H-2K ^d 31- 3-4S' ) and D ^d (H-2D ^d 34-4-20S), were purchased from American Tissue Culture Collection, Bethesda, MS. Biotinylated anti-HLA-DR mouse mAbs and avidin-FITC conjugates were obtained from Beckton Dickinson (Mountain View, CA). The BALB/c anti-OVA mAb (Mab-2; γ^K) was established in Winnipeg and purified from ascites as described (Lang, G.M. et al, 1992 supra). Each of the other mAbs was purified from culture supernatants with the aid of protein A-agarose (Pierce, Rockford, IL) and/or of protein-G fast-flow ( LKB-Pharmacia, Uppsala, Sweden) and the biotinylated mAbs were synthesized with biotin-LC-hydrazide (Piece, IL). hu-PBL-SCID Mice

Three- to 4-week old SCID mice were engrafted i.p. with 40xl0 ⁶ hu-PBL, or a mixture of T and (B+MN) cells

(the cell numbers used for a given experiment are as indicated in the Tables) . Total human serum IgG levels and HAMA responses of each hu-PBL-SCID mouse, which had been engrafted with 20xl0 ⁶ T and 20xl0 ⁶ (B+MN cells, are illustrated in Figure 1, from which it is obvious that the Ig representing the HAMA constituted only a negligible amount of the total human Ig produced by hu- PBL in the SCID mice. Antigens and Tolerogen The OVA and purified protein derivative (PPD) were obtained, respectively from Sigma (St. Louse, MO) and CedarLane (Hornby, ON). Aggregate-free OVA was isolated by gel-filtration and was used as the immunizing Ag. To increase the immunogenicity of Mab-2, this mAb was heated twice at 63°C for 1 h; the resulting ha-Mab-2 was used as the immunizing Ag.

The purified, aggregate-free monomeric Mab-2 was converted to the tolerogenic derivative, Mab-2(mPEG) ₃₆ , by reaction with the "activated intermediate" of mPEG (average Mr = 3,200 Da), as described earlier (Lang G.M. et al, 1992 supra). The tolerogenic fraction of the conjugate preparation was isolated from interfering contaminants (ie, high molecular weight cross-linked products, residual unmodified Mab-2, and unreacted but deactivated mPEG intermediate) by gel-filtration chromatography on Superose 6 (Pharmacia-LKB, Uppsala, Sweden) .

ELISA For determination of total mouse Ig and total human

IgG in the serum of each hu-PBL-SCID mouse, ELISA plates (Corning Inc., Corning, NY) were coated with 100 μg/ml of rabbit anti-mouse Ig (Zymed, San Francisco, CA) or 10 μg/ml of rabbit anti-human Ig Abs which had ben absorbed with immobilized human or mouse serum Ig, respectively.

For measurement of human Abs of Mab-2 mAb, the ELISA plates were coated with Mab-2 (10 μg/ml). After coating, the plates were treated with 10% BSA for 2 h and, as usual, each serum, serially diluted, was added to the wells; the plates were maintained at 37°C for 4 h, or at 4°C overnight. Protein A-purified serum IgG of BALB/c or of C57BL/6 served as standards for determination of murine Ig in sera of BALB/c or C.B-17 scid/scid mice respectively. Human IgG isolated by gel filtration was used as a standard for the determination of the human Ig in sera of SCID mice. After incubation, the ELISA plates were washed with PBS containing 0.01% Tween 20; biotinylated rabbit anti-mouse Ig Abs (Zymed, South San Francisco, CA) or biotinylated mouse anti-human Ig mAb (Zymed; HP) which had been absorbed, respectively, with human and mouse serum, was then added. The plates were incubated at 37°C for 2 h or at 4 ^C C overnight, washed 4 times, and finally the amounts of biotinylated Abs were determined colorimetrically at 405 nm by reaction with alkaline phosphatase-strepavidin conjugates (Zymed), and nitrophenyl phosphate (Sigma) as the substrate.

Treatment of hu-PBL-SCID Mice

Twelve ours after engraftment of PBL, the test and control hu-PBL-SCID mice received an i.p. injection of 200 μg of Mab-2(mPEG) ₃₆ or PBS, respectively. Seven and 21 days later, ie on days 8 and 28, all mice received two i.p. immunizing injections of 20 μg of ha-Mab-2 or 100 μg of OVA. On day 42 the serum of each mouse was assayed by ELISA for total human Ig Abs to Mab-2, and for total human IgG levels.

Cell Transfer Experiments

For the determination of the phenotype of the human Ts cells induced by Mab-2(mPEG) ₃₆ in hu-PBL-SCID mice, the

double cell transfer experiments were performed as illustrated in Figure 2. Briefly, SCID mice were engrafted with 40xl0 ⁶ PBL on day 0 and received i.p. injections of Mab-2(mPEG) ₃₆ and of ha-Mab-2 on days 1 and 8, respectively; these mice are referred to as the "first cell recipients" and serve as the "donor SCID mice". On day 22, human T cells were isolated from spleen and lymph nodes of each hu-PBL-SCID mouse by passage through nylon wool column. The non-adherent cells were treated with a mixture of anti-H-2- ^d mAbs (ie, anti-K ^d and anti-D ^d mAbs), or with a mixture of anti-H-2 ^d mAbs and 0KT4, or a mixture of anti-H-2 ^d mAbs and 0KT8, in the presence of rabbit complement (RC ). The residual viable cells of each of these three batches of cells, obtained from a single mouse, were co-transferred with 15xl0 ⁶ T and 20xl0 ⁶ (B+MN) cells of the original volunteer (the cells had been maintained in liquid nitrogen) into a second recipient SCID mouse. Twelve hours and 28 days after the cell transfer, all second cell recipients were immunized with 20 μg of ha-Mab-2. Fourteen days after the second immunization, the serum of each mouse was assayed for human anti-Mab-2 Ig Abs and for total mouse Ig levels by ELISA

Proliferation Test of Mab-2-specific CD4 ^* Cells

The protocol for this test is illustrated in the flowchart in Figure 3. The culture medium was RPMI-1640, containing 5% human cord serum and 5xl0" ⁵ M of 2- mercaptoethanol. For isolation of responder CD4 ^* T cells, individual

SCID mice were engrafted with 50xl0 ⁶ PBL of each of the three volunteers and immunized s.c. with 100 μg of Mab-2 in Freund's complete adjuvant (FCA). Fourteen days later, the spleen and lymph node cells of each mouse were isolated and pooled, and treated with a mixture of anti-

H-2 ^d mAbs plus RC . The viable cells were stimulated twice with 100 μg/ml of Mab-2 or 20 μg/ml of PPD at 21 days intervals. Two weeks after the last stimulation, the cultured cells were harvested and then treated with 0KT8 plus RC . The residual cells were used as responder CD4 ^* T cells.

For isolation of CD8 ^* T cells, each SCID mouse was engrafted with 50xl0 ⁶ PBL of volunteer "A" and 12 hr later they were injected i.p. with 200 μg of Mab-2(mPEG) ₃₆ or PBS, and all the mice received i.p. an immunizing injection of 20 μg of ha-Mab-2 on day 8. Fourteen days later, the CD8 ^* T cells were isolated from the pooled spleen and lymph node cells of each hu-PBL-SCID mouse by passage through nylon wool column, and treated with a mixture of anti-H-2 ^d and 0KT4 mAbs in the presence of RC . Half a million of the responder CD4 ^* T cells of each volunteer and 5x10" irradiated (2,000 rad) PBL of the same volunteer were co-cultured for 6 days with 100 μg/ml of Mab-2 or 20 μg/ml of PPD in the presence of 5xl0 ⁴ CD8 ^* T cells of either control or tolerized mice. The irradiated hu-PBL served as the source of APC. The cells were pulsed for the last 12 h of culture with ³ H-TdR (0.5μCi/well).

RESULTS

The Ag-specific Tolerogenic Effect of Mab-2(mPEG) ₁₆ on the

Human Lvmphoid System

In the exploratory phase of this study the primary and secondary HAMA responses could not be consistently induced in all hu-PBL-SCID mice which had been engrafted with more than 40xl0 ⁶ fresh PBL and given two i.p. immunizing injections of the highly immunogenic form of

Mab-2, ie, ha-Mab-2, or of Mab-2 emulsified in FCA on days 0 and 28, even though more than 1.0 mg/ml of total human Ig was detected in some of the mice (data not

shown). Since B cells represent only 15-20% of PBL, it was decided to transfer into the SCID mice a cell population enriched in B cells. The results listed in Table 1 (Experiments 1 and 2) demonstrate that this approach was successful, in as much as HAMA responses were induced in all hu-PBL-SCID mice which had been engrafted with a mixture of 20xl0 ⁶ T cells and 10 or 20xl0 ⁶ (B+MN) cells on day 0, and immunized with 20 μg of ha-Mab-2 on days 8 and 28. Moreover, as predicted, the magnitude of the HAMA responses depended on the number of (B+MN) cells. Parenthetically, it ought to be pointed out that although no data are given for the levels of the HAMA responses after a single immunizing injection because they were below the level of detection by ELISA (ELISA titers were in the range of 10-50), the primary HAMA responses could be detected by the micro-double- diffusion test. This discrepancy in detectability of the primary HAMA response is attributed to the dilution of the low titered Abs which is necessitated in ELISA. Furthermore, the hu-PBL-SCID mice treated with the tolerogenic conjugate, Mab-2(mPEG) ₃₆ , 12h after engraftment of the human lymphoid cells and given two immunizing injections of ha-Mab-2, displayed Ag-specific suppression of their secondary HAMA responses. On the other hand, their secondary anti-OVA IgG responses were not affected by pretreatment with this mPEG conjugate. The tolerogenic effect of Mab-2(mPEG) ₃₆ on the HAMA response was confirmed in Experiment 3 utilizing SCID mice engrafted with PBL of two other volunteers, "B" and "C". However, because of the limited availability of these volunteers' PBL, the groups of mice receiving their PBL consisted of only 2-3 animals. In any case, all these results demonstrate that mPEG conjugates of murine Mab-2 suppressed the human Ab response in an Ag-specific manner.

Relationship of HAMA Responses and Serum Human Ig Levels

As reported by many groups, it was confirmed in this study that the serum human Ig levels of hu-PBL-SCID mice varied over a wide range even when the mice were engrafted with PBL of the same donor. Therefore, the

HAMA responses were compared with the human Ig levels in sera of both control and test hu-PBL-SCID mice, ie, mice which had received, respectively, PBS and Mab-2(mPEG) ₃₆ prior to immunization. As illustrated in Figure 1, total human IgG levels of all hu-PBL-SCID mice on day 42 after engraftment with 20xl0 ⁶ T and 20xl0 ⁶ (B+MN) cells of a single donor (ie, volunteer "A"), were in the range of 3- 5 mg/ml. Moreover, no obvious differences in total human IgG levels were observed on comparing these levels for control and tolerized hu-PBL-SCID mice. Similarly, the induced HAMA responses varied among the animals within each group; nevertheless, the results in Table 1 demonstrate that HAMA responses of tolerized mice were consistently lower than those of the test mice. Moreover, it is interesting to note that (i) there was some direct relation between the levels of total human Ig and of HAMA in test hu-PBL-SCID mice, and (ii ) the deviations in both total Ig and HAMA levels were essentially identical for the control and tolerized hu- PBL-SCID mice.

Induction of CD8 ^* Ts Cells in hu-PBL-SCID Mice by Mab-2(mPEG),.

In earlier studies, it has been demonstrated that the Ag-specific tolerogenic effects of Ag(mPEG) _n conjugates in normal mice were due to induction of transferable CD8 ^* Ts cells. Similarly, in the experiment described here, it was shown that treatment of hu-PBL- SCID mice with Mab-2(mPEG) ₃₆ and then with ha-Mab-2 resulted in the induction of human CD8 ^* T cells. As

illustrated in the flow chart in Figure 2 and described under Methods, human CD4 ^* and CD8 ^* T cells were isolated separately from each donor hu-PBL-SCID mouse and co- transferred with the standard dose of 15xl0 ⁶ T cells and 20xl0 ⁶ (B+MN) cells of PBL of the same volunteer (the latter cells had been kept in liquid nitrogen prior to transfer into the second SCID mouse recipient).

It is evident from the results listed in the lower section of Table 2 that the co-transfer of the untreated PBL with the T cell fraction of an immunuosuppressed SCID mouse resulted in a dramatic reduction of the HAMA response, and that this effect was abrogated by treatment of the T cells with 0KT8, but not with 0KT4, in presence of RC . On the other hand, the T cells from hu-PBL-SCID mice which had been immunized with ha-Mab-2 without pretreatment with the mPEG conjugate did not have a lowering effect on the HAMA response induced in the second SCID recipient on day 42 (upper section of Table 2).

Suppression of Proliferation of CD4 ^* T Cells by CD8 ^* T Cells

In an attempt to shed some light on the mechanism of action of the suppressive CD8 ^* Ts cells of hu-PBL-SCID mice, which had been induced by tolerogenic Mab-2(mPEG) ₃₆ , the effect of these cells on proliferation of Ag-specific CD4 ^* T cells was tested. For this purpose, as illustrated in Fig. 3, SCID mice were engrafted with whole PBL of each donor and then immunized s.c. with Mab-2 emulsified in FCA on day 1. Two weeks later, human T cells were isolated from pools of spleen and lymph node cells of hu- PBL-SCID mice and then restimulated in vitro twice with Mab-2 or PPD at 21-day intervals. Two weeks after the last stimulation, CD4 ^* T cells were isolated from the culture, as illustrated in the flowchart in Figure 3 and

described under Methods, and were used as responder CD4 ^* T cells; the CD8 ^* T cells were induced as described previously.

As is evident from the results listed in Table 3, responder T cells of each donor showed significant proliferative activity by stimulation with the relevant Ag in the absence or presence of CD8* T cells of hu-PBL- SCID mice which had not been treated with the conjugate. However, the addition of CD8 ⁺ T cells of tolerized hu-PBL- SCID mice which had been engrafted with hu-PBL of donor "A" caused significant reduction of proliferation activity of responder T cells of donor "A" to Mab-2, but not to PPD. Moreover, the CD8 ^* T cells "A" also suppressed proliferative responses of CD4 ^* T cells of donor "B" who shared class I Ags, ie, A2, A24, B51, cl, c3, and Class II ags, ie, DR4, which those of donor "A". On the other hand, CD8 ^* T cells "A" did not decrease the response of T cells of donor "C" against Mab-2 and PPD, whose class I Ags, cl, and class II Ags, DR4 and DR9, were identical with those of volunteer "A". These results are interpreted to indicate that CD8 ^* T cells suppressed proliferative responses in an Ag-specific and Class I-specific specific manner.

This study establishes the method for the induction of HAMA responses in SCID mice which had been engrafted with appropriate numbers of T and (B+MN) cells, and also shows that the system used is appropriate for evaluation of the tolerogenicity of mPEG conjugates of murine mAbs in relation to their ability to suppress the formation of HAMA responses. Furthermore, the suppression was shown to b due to the generation of human CD8 ^* Ts cells, which downregulated CD4 ^* Th cells in an Ag- and MHC-specific manner, ie in accord with identical conclusions derived from systems involving the suppression of Ag-specific immune responses in intact mice to a variety of Ags by

the corresponding tolerogenic Ag(mPEG) _n conjugates.

In the exploratory phase of this study, utilizing 40-60xl0 ⁶ of freshly harvested hu-PBL from three EBV- seronegative volunteers, for engraftment of SCID mice, the applicants did not consistently induce significant primary and secondary HAMA responses (data not shown) . Thus, only 8 out of 47 SCID mice mounted a weak HAMA response over a broad range of ELISA titers, ie 50 - 430. Moreover, there was no apparent relation between the Ab titers and the total "human" IgG levels. By contrast, the anti-tetanus toxoid (TT) Ab response in SCID mice, engrafted with an identical dose of PBL from the same volunteers, were significantly higher and there seemed to be a direct relation between anti-TT Ab titers and the level of human IgG in these mice.

Other investigators, using the same procedure, have reported similar findings. Thus, one study found that only 1 out of 10 SCID mice, engrafted with unfractionated PBL from healthy volunteers, produced a significant HAMA response and, by contrast, all 10 SCID mice, which had been engrafted with PBL of melanoma patients undergoing immunotherapy with murine anti-melanoma mAb, produced high levels of HAMA, which included high titers of anti- idiotypic human Abs. Taking an overview of all the above results, it is clear that the hu-PBL-SCID mouse model has its limitations, particularly for studies involving the induction of primary and secondary Ab responses with unfractionated PBL of volunteers who had not been sensitized to the Ags being used in these mice. The inability of inducing a consistent and significant de novo response may possibly be due to lower numbers of surviving naive human B and/or T cells than of the corresponding memory cells in SCID mice. In addition it is to be stressed that B cells represent only 15-20% of the cells in PBL, in contrast with their much higher

numbers in lymphoid organ.

In an attempt to circumvent these apparent limitations of the hu-PBL-SCID mouse model, the applicants explored the possibility that a more vigorous and consistence HAMA response could be generated with SCID mice engrafted with distinct populations of human lymphoid cells at appropriate numbers. Indeed, as is evident from the results listed in Table 1, significant HAMA and human anti-OVA Ab responses were induced in SCID mice engrafted with a mixture of 20xl0 ⁶ T cells and 20xl0 ⁶ (B+MN) cells. In addition these results demonstrate that (i ) the HAMA responses were proportional to the number of engrafted (B+MN) cells, and (ii) there was a relation between the levels of total human IgG and the HAMA titers of secondary responses in a given hu-PBL-SCID mouse. The latter finding is in agreement with the recent conclusion of Duchosal et al (Nature 355, p 258- 262) who established that, although there were wide variations in human anti-TT Ab titers produced in different SCID mice engrafted with the same dose of unfractionated PBL of one and the same volunteer, there was a direct relation between Ab titers and human IgG levels produced in these mice. All these results may suggest that the percentage of survival of B cells might determine the level of total human IgG levels in a given hu-PBL-SCID mouse, and that the percentage of surviving B cells and of memory B cells in a given batch of hu-PBL could be identical. Thus, the failure of detection of HAMA responses in hu-PBL-SCID mice, which had been engrafted with unfractionated PBL of healthy volunteers, may be due to a low number of B cells committed to mouse Ig in a given batch to hu-PBL rather than to a limited clonal heterogeneity of the B cells, which had survived in the mice (Saxon, A., et al, J. Clin. Invest. 87, 658- 665, 1991).

More recently, it was reported that human mature T cells in hu-PBL-SCID mice were refractory to simulation by anti-CD3 mAb, but they proliferated in response to exogenous 11-2 (Tary-Lehmann, M. , et al, J. Exp. Med. 175, p 503-516, 1992). Though the present applicants did not analyze the surface expression of CD45RO and HLA-DR Ags on T cells isolated from hu-PBL-SCID mice, their results indicate that the functional CD8 ^* Ts cells were induced in hu-PBL-SCID mice, which had been subjected to the combined activation signals of the tolerogenic Ag(mPEG) _n conjugate and the corresponding free Ag. Moreover, the fact that Ab responses were induced in hu- PBL-SCID mice is considered to indicate that affinity maturation of B cells, which is T cell-dependent, occurred in these mice (Carlsson R. et al, J. Immunol.

148, 1065-1072, 1992). All these results taken together suggest that T cells responsible for a positive or a negative immune response are functional in the environment of the hu-PBL-SCID mice. The reason(s) for the differences between the refractory state of T cells found in the study referred to above (Carlsson R. et al 1992 supra) and their functional capacity in other systems remains to be elucidated.

On administration of murine mAbs to man - even in conjunction with immunosuppressive agents - some individuals produce HAMA directed to the isotypic determinants, others to the idiotypic determinants of mAbs, and yet others would to both determinants, or no detectable HAMA at all. These results reflect that the differences in responsiveness among difference individuals are controlled by their Ir gene. Hence, it is envisaged that utilizing the SCID mouse model with the PBL of a given patient it would be possible to determine not only the immunogenicity of a murine mAb and the tolerogenicity of its corresponding mPEG conjugate, but

also to predict for the same patient the nature and the level of his/her HAMA response when the patient is to be treated with the murine mAb in question.

SUMMARY

Severe combined immunodeficient (SCID) mice were reconstituted with normal human peripheral blood leucocytes (PBL) and were shown to produce a human anti- mouse Ig antibody (HAMA) response on immunization with heat-treated murine monoclonal IgGl antibody (mAb) to ovalbumin (OVA), referred to a ha-Mab-2. The HAMA response was proportional to the number of B cells and mononuclear cells transferred from a given batch of PBL. However, pretreatment of hu-PBL-SCID mice with a tolerogenic covalent conjugate of monomethoxypolyethylene glycol (mPEG) and Mab-2 suppressed the HAMA response on subsequent injections of ha-Mab-2 and this suppression was shown to be antigen-specific, ie, it id not suppress the antibody response to OVA and did not affect the level of production of human Ig of hu-PBL-SCID mice. The suppression was due to the generation of human suppressor CD8 ^* T (Ts) cells, which downregulated CD4 ^* helper T cells in an Ag- and MHC class I-specific manner, ie, these findings were in accord with the previously shown immunosuppressive effect of tolerogenic mPEG conjugates in normal mice.

Table 1 ANTIGEN SPECIFIC SUPPRESSION OF HUMAN ANTIBODY RESPONSES

IN IIU-PBL-SCII) MICE BY MaVι 2(mPEG) ₃₆

Footnote for Table 1 Experiments 1 and 2

Groups of four, 3-4 week old, BALB/c SCID mice received i.p., on day 0, the indicated numbers of T and (B+MN) cells, which had been isolated from the PBL of a healthy volunteer "A" by passage through a nylon wool column. The freshly isolated cells served for Experiment 1. For Experiment 2, cells which had been maintained in liquid nitrogen were used. Twelve hours after cell transfer, all test mice received i.p. 200 μg of Mab-

2(mPEG) ₃₅ and the control mice received PBS instead. All mice received two immunizing injections of either 20 μg of ha-Mab-2 of 100 μg of aggregate-free OVA on days 8 and 21. The mice were bled on day 35 and their sera were assayed by ELISA for human IgG antibodies to Mab-2 and OVA. Experiment 3

Groups of two to three, 3-4 week old, SCID mice received i.p. on day 0 the indicated number of T and (B+MN) cells of two healthy volunteers "B" and "C", which had been kept in liquid nitrogen. Twelve hours later, the test mice received i.p. 200 μg of Mab-2(mPEG) ₃₆ and the control mice received PBS. All mice were immunized with 20 μg of ha-Mab-2 on days 8 and 21. On day 35, the ELISA titers of human IgG antibodies to Mab-2 were individually determined.

Table 2

INDUCTION OF CD8 ⁺ Ts CELLS IN HU-PBL-SCID MICE BY Mab-2(mPEG) ₃₆ *

Treatment of T cells Human anti-Mab-2 IgG ELISA

Treatment of donor of donor mice prior titers of second SCID hu-PBL-SCID mice to transfer to recipients on day 42 second SCID recipient

None 1 ,420 1.640 1,710

a-H-2 ^d + RC 2,090 2.570 2,810

PBS plus a-H-2 ^d 4- OKT4 + RC 1 ,700 1.910 1 ,620 c o ha-Mab-2 a-H-2 ^d + OKT8 + RC 2,420 2.770 2,620

a-H-2 ^d + RC 190 130 290

Mab-2(mPEG) ₃₆ plus a-H-2 ^d + OKT4 + RC 170 260 < 50 ha-Mab-2 a-H-2 ^d + OKT8 + RC 1 ,550 2.180 2,510

* For protocol, see Flowchart 1. The first SCID mice which had received hu-PBL are referred to as "donor SCID mice"; the treatment of these mice and of their T cells are indicated in the left segment of the Flowchart 1.

Table 3 ANTIGEN-SPECIFIC SUPPRESSION OF PROLIFERATION OF CD4 ⁺ T CELLS BY CD8 ⁺ T CELLS OF HU-PBL-SCID MICE TOLERIZED WITH Mab-2(mPEG) ₃₆

Footnote for Table 3 a) The HLA types of volunteers "A", "B", and "C" are

A(2,24) b(51,-) C(cl,c3) DR(4,9); A(2,24) B(51,w52) C(cl,c3) DR(4,-); A(3,-) B(35,-) C(cl,-) DR(4,9), respectively.

(b) Responder CD4 ^* T cells of each volunteer in upper and lower panels were stimulated separately with Mab-2 or PPD, respectively. Two weeks after the last stimulation, the CD4 ^* T cells were isolated as described under Methods.

(c) As the stimulating Ag, 100 μg/ml of Mab-2 or 20 μg/ml of PPD were added.

(d) Each batch of responder CD4 ^* cells ( 5xl0 ⁵ ) and the irradiated PBL (5xl0 ⁴ ) of the same volunteer (as source of APC) were col-cultured for 6 days with

Mab-2 or PPD in the absence, or presence of 5xl0 ⁴ CD8 ^* T cells of primed or tolerized hu-PBL-SCID mice, which had been engrafted with PBL of volunteer "A". All cultures were pulsed for the last 12 hr with H- TdR (0.5 μCi/well). The results represent means and

S.E.M of 6 replicate cultures.

Part 2

In this study the applicants investigates the extent of suppression of anti-idiotypic (a-id) Abs since it has been generally accepted that a-id Abs, produced by patients receiving murine mAbs, interfered with the binding of the mAbs to their cell targets. Moreover, the a-id response was more pronounced when the a-id Abs were directed to tissue Ags. This effect is probably due to the increased immunogenicity of the corresponding anti- idiotopes as a result of their acquiring enhanced rigidity after binding to tissue antigens. Hence, the applicants evaluated the effectiveness of suppressing the a-id responses by tolerogenic mPEG conjugates of mAbs of soluble and tissue antigens. Clearly, for optimizing the tolerogenicity of the mPEG conjugate of a given mAb, it would be advisable to use a conjugate comprising the particular mAb. However, the applicants in this study investigated the use of only the tolerogenic mPEG conjugate of Mab-2 in conjunction with four mAbs directed to different Ags, but possessing the same heavy and light chains. The four mAbs used were: (i) Mab-2 (γ ,κ) directed to one of the epitopes of ovalbumin (OVA), ( ii ) H ₁ -DNPγ=109.3 (γ ₁₍ κ) directed to the 2,4-dinitrophenyl (DNP) group (for the sake of brevity this mAb will be referred to as H _x ) , and ( iii ) 0KT4 (γ _2b ,κ) and Leu-3a(γ _α ,κ) which are mAbs directed to different epitopes of the human CD4 marker of helper T (Th) cells.

The applicants found that treatment of hu-PBL-SCID mice with the tolerogenic Mab-2(mPEG) ₃₆ suppressed the anti-id responses with respect to not only the idiotopes of Mab-2, but also to the idiotopes of the other three mAbs consisting of the same heavy and light chains.

METHODS AND MATERIALS

The methods of purifying Mab-2, H _x , Leu3a and OKT4 and of synthesizing Mab-2(mPEG) ₃₆ have been described elsewhere ( Bitoh, S, et al Hum. Antibod. Hybridomas 1993, vol 4 p 144; Lang, G.M. et al, Immunol, letters 32, 247-252, 1992). The procedures for generating hu-PBL- SCID mice were similar to those developed by Mosier D.E. et al in Nature 335, 256-259, 1988, J. Clin. Immunol. 10, 185-191, 1990; Science 251, 791-794, 1991. For the present study both BALB/c scid/scid and C.B-17 scid/scid mice were used. Each mouse was engrafted with 20xl0 ⁶ T cells and 20xl0 ⁶ of a mixture of B and mononuclear (B+MN) cells, which had been isolated by leukaphoresis from the blood of one healthy, EBV-, HIV- and HBV-negative volunteer.

Treatment of Mice prior to Immunization

The test mice used for the Mab-2 and H _x systems received 12 h after engraftment of the human cells a single i.v. injection of 200 μg of the immunosuppressive Mab-2(mPEG) ₃₆ and the control mice were administered PBS. The mice which were used for the Leu-3a system received 100 μg of Mab-2(mPEG) ₃₆ . For the Mab-2 and H _x systems, four SCID mice were used for each test and control group. In the subsequent experiments with Leu-3a, the number of animals was decreased to three per group; this was predicated by the fact that the number of experiments which could be performed was limited by the availability of suitable blood donors and by the resources available for maintaining the colony of SCID mice.

Detection of Anti-id Abs to Mab-2 and H _x mAbs

Because of the low level of Abs produced, these were not detected by ELISA on plates coated with Mab-2 or _{l r} but were visible on micro-immunodiffusion in agar gel. Hence, to increase the sensitivity of the Ab detection

procedure, the method of "inhibition of ELISA" was developed as described below. For this purpose, the serum of each SCID mouse was diluted 10-fold with PBS containing 1% of normal BALB/c mouse serum for neutralization of HAMAs to determinants common to all murine IgG. This serum will be referred to as "neutralized serum" . The residual anti-Mab-2 or anti-H ₁ Abs in the "neutralized sera" were considered to represent anti-id Abs specific for the respective mAbs. For the detection of anti-Mab-2 id Abs, the

"neutralized serum" was incubated with a mixture of 10 parts of nonbiotinylated and 1 part of biotinylated Mab-2 (referred to hereafter as B.Mab-2) in ELISA plates coated with DNP ₉ -0VA, and the bound antibody was reacted with avidinylated alkaline phosphatase, using nitrophenyl phosphate as the substrate. The optical density (O.D. ) readings at 405 nm, obtained under standardized conditions with the "neutralized serum" of PBS-treated control SCID mice, which contained anti-id Abs, represented the minimal "reference" ELISA titers due to the maximal effect of the inhibition of binding of the B.Mab-2 to the DNP _g -OVA coated plates by the anti-id Abs. By contrast, maximal O.D. readings were obtained with the "neutralized serum" in the absence of anti-id Abs. Clearly, the increase in O.D. determined on dilution of the "neutralized serum" of control mice was due to the decrease in the concentration of the anti-id Abs, which would have reacted with the B.Mab-2 and would have blocked their binding to DNP _g -OVA. For the detection of anti-DNP ₉ -OVA H _x id Abs, the

ELISA was performed with DNP-OVA coated plates, as described in the preceding paragraph, except that a mixture of 10 parts of nonbiotinylated H _x mAbs (referred to hereafter as B.H _x ) were used instead of B.Mab-2.

Detection of Anti-id Abs to Leu3a by CD4 ^* T Cell Proliferation Assay

For induction of these anti-id Abs and for the preparation of the OVA reactive CD4 ^* human T cells, the CD4 ^* T cells of the same donor (whose hu-PBL cells had served for engraftment of the SCID mice) were used. For the production of anti-id Abs these hu-PBL cells were first inactivated by mitomycin C (MMC) and lOxlO ⁷ of these cells were incubated with 1 μg of Leu3a for 45 min at 0°C and washed with RPMI-1640 medium, and the Leu-3a coated CD4 ^* T cells were then used for immunization. (For the treatment with MMC, these cells were maintained in 25 μg of MMC/ml for 60 min at 37°C. )

The CD4 ^* T cells were prepared by stimulating lOxlO ⁶ of the hu-PBL in culture with 100 μg/ml of OVA at 2-3 week intervals. Ten days after the third stimulation the cultured cells were harvested, passed through a Sephadex G-10 column, treated with 0KT8 in the presence of rabbit complement (RC ), and washed thrice with RPMI-1640 medium and centrifuged; the viable cells were used as the OVA- sensitive CD4 ^* T cells.

As in the case of Mab-2 and h _x mAbs, the anti-id Abs induced by Leu3a in control mice were below the level of detection by ELISA on plates coated with Leu3a. Hence, the presence of the anti-id Abs in sera of control and test hu-PBL-SCID mice, which had been treated with the anti-human Cd4 mAb, was established by determining the degree of inhibition by these anti-id Abs of the abrogation of proliferation of OVA-specific CD4 ^* T cells by Leu3a in the presence of OVA.

For the detection of anti-id Abs induced by Leu-3a, the following assay was developed. Each serum of the control and test hu-PBL-SCID mice was diluted 10-fold with RPMI-1640 medium containing 5% of human cord serum as a source of lymphokines, 5xl0 ^-5 M of 2-mercaptoethanol,

and 10 ng/ml of protein A-purified mouse IgG, and maintained at 37°C for 2 h, and finally centrifuged and filtered through a millipore membrane (0.22 μm). The OVA-reactive CD4 ^* T cells ( 5xl0 ⁵ ) were co-cultured for 2 days with lxlO ⁵ irradiated (2000 rad) autologous (B+MN) cells, as a source of APC, in the presence of OVA (100 μg/ml) , and of Leu3a ( 10 ng/ml) and of each serum at different dilutions. Finally, the cells were pulsed with ³ H-TdR (0.5 μCi/well) and the culture was continued for 12 additional hours. At the end of this period, the maximal and minimal proliferation indices of OVA-specific CD4 ^* T cells were determined, respectively, in the presence and absence of anti-CD4 mAb. Thus, the addition of sera of hu-PBL-SCID mice which contained anti-id Abs directed to the idiotypic determinants of Leu3a inhibited the abrogation of proliferation of CD4 ^* T cells by Leu3a and, as would be expected, there was an inverse relationship between this inhibition and the dilution of the anti-id serum.

RESULTS

I. Suppression of Human Anti-id to Mab-2 and H _x MaBS

I-l Mab-2 System

Balb/c SCID mice received two s.c. injections of lμg of Mab-2 in Freund' s complete adjuvant (FCA) on days 7 and 28 after administration of Mab-2(mPEG) ₃₆ or PBS. On day 42 all mice were bled for determination of their anti-id Abs.

As it evident from the results of the inhibition of ELISA, illustrated in Fig. 4A, the sera of SCID mice pretreated with the tolerogenic Mab-2(mPEG) ₃₆ conjugate were devoid of detectable anti-id Abs (empty circles), as contrasted to sera of all four control mice ( filled circles) which contained anti-id Abs (Fig. C). For confirmation that the above inhibition of ELISA titers

was due to "human" anti-id Abs specific for the idiotypic determinants of Mab-2, two variations of the assay, described below, were used:

(i ) The ELISA plates were coated as before with DNP _g -OVA and the sera incubated with B.H _X , instead of B.Mab-2.

The rest of the Methods and the corresponding ELISA titers were plotted in Fig. 4B. From these results, it is evidence that the anti-id Abs specific for Mab-2 which had been present in sera of control SCID mice did not react with idiotope(s) of H _x , ie, no inhibition of ELISA was detected with sera of control or tolerized mice by interaction with H _x . ( ii ) To demonstrate that the ELISA titers were due to "human", and not to murine Abs, the serum of each control SCID mouse was absorbed with immunosorbents consisting of Affi-gel Hz (Pierce, Rockford, IL) to which were coupled rabbit Abs to human Ig or to mouse Ig. The absorbed serum was then assayed for the presence of id Abs utilizing the ELISA conditions used for the results illustrated in fig.

4-A. As is evident from the results plotted in Fig. 4-C, whereas detectable anti-id Abs were measured in the control sera absorbed with Abs to mouse Ig, the sera absorbed with Abs to human ig were depleted f anti-id Abs, ie, the anti-id Abs were human Abs.

All these results support the conclusion that pretolerization of hu-PBL-SCID mice with Mab-2(mPEG) ₃₆ , followed by injections of Mab-2, resulted in suppression of the immune response of the engrafted human lymphoid system to the idiotypic determinants of Mab-2.

1-2. The H _x System

The present applicants have made the new and surprising discovery that SCID mice tolerized by Mab- 2(mPEG) ₃₆ are also anergic to H ₁ . These two mAbs differ

only in their idiotopes and the Ts cells would recognize the same isotypic determinants present on these two mAbs.

Indeed this was found in hu-PBL-C.B-17 scid/scid mice which received first 200μg of Mab-2(mPEG) ₃₆ and two injections of 5 μg of H _x in FCA 7 and 28 days later; the control animals received PBS in lieu of the conjugate. A fortnight after the second immunization, the anti-id HAMA response was determined by the method of inhibition of ELISA, as described earlier, ie, the plates were coated with DNP ₉ -0VA and reacted with B.Mab-2 or B.H _x mAbs in the presence of the "neutralized serum" of each mouse at increasing dilutions. Finally, the bound antibody was detected with avidinylated alkaline phosphatase and nitrophenyl phosphate. Also, as in the previous experiment, to ascertain the human nature of the anti-id Abs produced by the control mice to H _{l f} the sera of all mice were absorbed with the two rabbit immunosorbents specific for human and mouse Ig.

As can be determined from an inspection of the results illustrated in Fig. 5-A, addition of B.Mab-2 to the "neutralized serum" of either tolerized or control mice ( hich have been immunized with H _x ) did not inhibit the ELISA titers since any anti-id Abs specific for the H _x would not be expected to interact with Mab-2. By contrast, as illustrated by the data in Fig. 5-B, on interaction of the "neutralized sera" of control mice with B.H _x mAbs, the presence of anti-H _x id Abs was demonstrated and, these anti-id Abs were not detectable in sera of animals which had been tolerized with Mab- 2(mPEG) ₃₆ . Similarly, as in the previous experiment, the results plotted in Fig. 5-c demonstrate that the anti-id Abs due to immunization with H _x had the antigenic determinants of human Ig, and not of murine ig.

All the results of the experiments (described in I-l and 1-2) support the conclusion that Mab-2(mPEG) ₃₆

immunosuppressed the human lymphoid system to idiotypic determinants present on two different Ig molecules consisting of the same γ-1 and K chains. Clearly, these results open the possibility that administration of tolerogenic mPEG conjugates of a given murine mAb may abrogate the ability of the human lymphoid system to respond to murine mAbs possessing diverse idiotypic determinants, as long as these mAbs have some determinants which are identical to the specific determinants of the Ig (eg, isotypic determinants), incorporated in the mPEG conjugate which is responsible for induction of the Ts cells.

11 Suppression of Human Ab Responses at Murine anti- Human CD4 mAb ( ie Leu3a) by Mab-2(mPEG) ₃₆

In this section are described the experiments for establishing the possibility of suppressing the anti-id responses to a murine mAb directed to a tissue epitope of the human CD4 marker, ie, Leu3a. Three-week old C.B-17 scid/scid mice were engrafted with human leukocytes and

12 h later they received 100 μg of mAb-2(mPEG) ₃₆ or PBS. As stated earlier, in an attempt to simulate as closely as possible the conditions in patients, all the mice received 7 and 28 days later lOxlO ⁶ Th cells from the same volunteer after coating them with Leu3a. Fourteen days after the second immunization all mice were bled and each serum was tested for anti-id Abs by the procedure described in the Methods section.

It is important to note that in these experiments, the dose of the Mab-2(mPEG) ₃₆ conjugate was reduced to 100 μg in spite of the fact that this mAb combines with the CD4 antigens of Th cells, a procedure which may render them even more immunogenic than the mAbs reacting with soluble antigens. (In the preceding "exploratory" experiments, the higher doses of 200 μg of the

tolerogenic conjugate was used; the rationale for this larger does was to minimize in the initial experiments the possibility of not inducing suppression. ) Nonetheless as shown below, this reduction in Mab- 2(mPEG) ₃₆ did not impair the suppressive activity of this conjugate even with respect to the Leu3A anti-id Abs.

From the data plotted in Figure 6, it can be seen, as represented by the empty squares, that, in the absence of anti-CD4 mAb, OVA-reactive T cells proliferated in the presence of OVA, ie, the proliferation indices of these Th cells in the absence of interfering Abs were of the order to 18,489 dpm ± 3,762. By contrast, the background levels of proliferation in the absence of OVA were of the order to 1,210 ± 258 dpm (these are represented by the filled squares). The addition of 0KT4 or Leu3a led to reduction in the proliferation of these cells to background levels. As illustrated in Figure 6-A, sera from control hu-PBL-SCID mice (which had received PBS, in lieu of the tolerogen, and the CD4 ^* T cells coated with Leu-3a) abrogated the inhibitory effect of Leu-3a on the proliferation index, but not of 0KT4 (Figure 6-C), ie, these results are taken as evidence for the presence of anti-Leu-3a Abs in these sera which neutralized the effects of Leu-3a. On the other hand, as illustrated in Figures 6-B and 6-D, respectively, sera from test hu-PBL- SCID mice, ie, mice which had received Mab-2(mPEG) ₃₆ before the injection of the Leu-3a-coated T cells had no detectable inhibitory activity with respect to the capacity of either Leu-3a or 0KT4 to abrogate the proliferation of OVA-reactive CD4 ^* T cells. These results are considered to demonstrate that pretreatment of hu- PBL.SCID mice with the tolerogenic conjugate, Mab- 2(mPEG) ₃₆ , induced suppression of HAMA responses to the epitopes of Leu-3a, including their anti-id response. A variety of mAbs (ie, xenogeneic, chimeric,

humanized, or even "human" mAbs produced by genetic engineering) directed to tissue antigens either by themselves, or as immunoconjugates with toxins, radioactive elements, enzymes, prodrugs, etc., are being used for diverse therapeutic purposes as "magic bullets". However, despite the recent elegant advances leading to the syntheses of tailor-made Abs by genetic engineering, all these Abs induce at least anti-id Abs, and the immunogenicity of their idiotypic determinants has been shown to be highest when they are directed to tissue Ags of the host. Therefore, the effectiveness of therapies involving the administration of these agents is counteracted by the recipient' s Abs not only to isotypic determinants of the mAbs and to the determinants of their "cytotoxic payloads", but also to their idiotopes. The results of studies reported herein provide the uniform conclusion that tolerogenic mPEG conjugates derived from a murine mAb specific to the soluble antigen OVA were able to suppress the HAMA responses in hu-PBL-SCID mice to murine mAbs directed to both soluble and tissue antigens. The applicants attribute this pan- tolerogenicity of Mab-2(mPEG) ₃₆ (at least as observed with respect to the three mAbs used in this study) to the phenomenon of linked immunological suppression which is due to the recognition of common epitopes among these three mAbs by the Ts cells induced by mab-2(mPEG) ₃₆ . Hence, one may envisage that it would be possible to engineer a truly subclass pan-specific tolerogenic mPEG conjugate of a murine mAb, capable of suppressing the HAMA responses to the different idiotypic epitopes of mAbs directed to diverse antigens. For this purpose it will be essential either ( i ) to incorporate the epitopes specifying the different isotypic subclass determinants of the light and heavy chains into the engineered Ig onto which will be grated the requisite number of mPEG

molecules, or (ii) to use different mAbs possessing the same heavy and light chains as the mAb used for the synthesis of the corresponding mPEG conjugate.

Clearly, the present results indicate that the use of tolerogenic mAb(mPEG) _n conjugates would lead to a substantial reduction in, if not to the full, abrogation of HAMA responses and, consequently, make it possible to use mAbs for therapeutic biologically active agents. Thus, the teaching therein paves the way to the development of therapies involving mAbs to a diversity of human markers, such as are present on leucocyte adhesion molecules and other cells which would have to be downregulated or altogether eliminated, eg, CD4 ^* T cells and tumor cells, respectively, in autoimmune conditions and malignancies.

SUMMARY

Severe combined immunodeficient (SCID) mice were reconstituted with normal human peripheral blood leucocytes (hu-PBL). Treatment of these hu-PBL-SCID mice with a tolerogenic covalent conjugate of monomethoxypolyethylene glycol (mPEG) and an anti- ovalbumin, IgGl murine monoclonal antibody (mAb), referred to as Mab-2, suppressed the HAMA responses to both the common (γl,κ) and the idiotypic determinants of Mab-2. Moreover, the Mab-2(mPEG) ₃₆ conjugate suppressed the immune responses of hu-PBL-SCID mice to the common and idiotypic determinants of murine mAbs to the DNP residue and to human CD4, which mAbs consisted also of γl and Kchains. It is concluded that a tolerogenic mPEG conjugate of a murine mAb induces pan-suppression of the human lymphoid system with respect to other murine mAbs which share the isotypic determinants of the original mAb (ie, here the Mab-2) incorporated in the conjugate; hence, it may be anticipated that HAMA responses to any

murine IgG mAb would be suppressed by 1 of 8 mPEG conjugates, each incorporating 1 of the 4 subclasses of IgG and 1 of the 2 light chains.

Part 3

Cardiac transplantation, first performed in 1964, has become an accepted therapy for patients who have suffered end-stage heart diseases. By 1989 over 9,000 orthotopic heart transplants had been performed worldwide. Two of the main causes of death of heart recipients during the first year after transplantation are:

(i) acute immunological rejection mediated by the recipient's leucocytes recognizing the allo-antigens (allo-Ags) of the transplant,

(ii) enhanced susceptibility to infections (and increased incidents of malignancies) due to long-term immunosuppressive therapy with a battery of non-specific immunosuppressive drugs (ISD) eg cyclosporin A, FK506, rapamycin, mizoribine, 15-deoxyspergualin.

The ISDs are often administered in conjunction with mAbs to specific determinants of human T cells, e.g., CD3, CD4 and the IL-2 receptor; these mAbs lead to the downregulation of T-dependent immune responses. However, the use of ISDs is limited by their side effects.

Similar considerations apply to murine mAbs which lead, as a result of their immunogenicity, to the development of HAMA responses and to consequent side effects, i.e., fever, diarrhea, hypertension, malaise, nausea, occasional pulmonary edema, and even fatal anaphylaxis. In an attempt to eliminate complications due to HAMA, chimeric, "humanized" and even human mAbs have been produced by ingenious combinations of techniques of immunology and genetic engineering. However, so far the production of antibodies (Abs) by the patient to the

idiotypic ( Id) determinants of the engineered mAbs has not been overcome.

As for all organ transplants, heart graft survival depends on the degree of cross-matching of the histocompatibility Ags of donor and recipient, i.e., on the closeness of the HLA (human leucocyte antigen) complexes of the two. Although myocytes, in situ, do not normally express class I and class II HLAs, these Ags are present on vascular endothelial cells of the transplanted heart. The expression of class II Ags on the cellular components of the transplanted heart is still a matter of controversy. In any case, it is to be noted that survival rates for 9 grafts with no HLA-B or HLA-DR mismatches were better than those for hundreds of mismatched grafts, a conclusion which has been too well known in relation to kidney transplants. However, in contrast to renal transplantation, the crucial predicament in cardiac transplantation is that hearts are grafted under conditions where selection of a donor with well matched HLAs is difficult. Moreover, failure to prevent acute immunological rejection of heart allografts leads normally to the patient's death.

In spite of gigantic advances made in recent years in surgical procedures and in the development of effective ISDs, it is dispiriting that untoward physiological reactions and transplant rejections still occur. Moreover, in spite of the great potential of mAbs, their limited success in suppression of the host's immune response is due to their immunogenicity. Hence, the development of an immunologic strategy leading to selective suppression of the recipient's immune response, without the induction of HAMA and without unleashing side effects, is urgently required.

In Figure 7, the donor's cells, D, possess a hypothetical mosaic of four HLAs (g _x to g ₄ ) among which g ₂

is the incompatible HLA recognized uniquely by the patient's Th ₂ and/or Tc ₂ cells. The cytotoxic T cell (CTL or TC) has the capacity of destroying the D cells, whereas the Th ₂ cell is essential for activating the patient's Tc ₂ cells. Therefore, administration of the two corresponding Abs ('I-'T, ), which recognize the unique antigenic determinants of Vα ₂ /Vβ ₂ and Vα ₂ ' /Vβ ₂ ' of the TCRs of the respective Th ₂ and Tc ₂ cells, should annul the participation of these T cells in the rejection mechanism; the production of these antibodies, referred to as a-cmAbs, is described below.

This is based on the discovery that some patients, in spite of having received HLA mismatched kidneys, did not reject the kidneys even after withdrawal of ISD therapy. Similar observations have also been made in mice and rats. The retention of these grafts may be attributed to the patients' having produced (i) Abs which blocked the activity of their CTLs directed selectively against the donors' incompatible HLAs, but not against the HLAs of a third party, and (ii) Abs which suppressed the mixed leucocyte reaction (MLR) between their leucocytes and those of the donors' , but not between their leucocytes and those of a third party. Both types of Abs will be referred to as anti-clonotypic Abs (a- cAbs) , since - as illustrated in Fig. 1 - they are directed specifically against the unique "self" antigenic determinants of the variable regions of the a and/or β chains (Vα/Vβ) of the receptors of the recipient's helper T cells (Th) and Tc cells (TCR) which recognize a given mismatched HLA of the transplant; each "self-antigen" will be referred to as a clonotype, since they are specific for a given clone of T cells of the recipient.

The applicant proposes the application of the method of pegylation of protein Ags to the suppression of HAMA responses in prospective heart transplant recipients.

It is important to stress that the strategy should be most effective when the Ag(mPEG) _n conjugates are administered before exposure to the corresponding Ag. Therefore, induction of long-term downregulation of the Ab response to a particular Ag would require two steps:

Step I: Injection of the tolerogenic conjugate, Ag(mPEG) _n , leading to the generation of Ag-specific suppressor T (Ts) cells.

Step II: Administration of the unmodified Ag at least 7 days after Step I. Thereafter, the Ag can be injected repeatedly over extended periods without further injections of Ag(mPEG) _n .

The reason for this interval between Steps I and II is to allow sufficient time for the activation of Ag¬ specific Ts cells; moreover, additional periodic injections of the unmodified Ag are required to maintain the proliferation of these Ts cells for extended periods. These Ag-specific Ts cells have been shown to downregulate the Ag-specific Th cells and to thus prevent the induction of the Ab response to the specific Ag.

This two-step procedure, rather than injection of only the tolerogenic mPEG conjugate, is due to the fact that coupling of the required number of the long chains of mPEG onto a protein for converting it to a tolerogen (i.e., 10±2 mPEG molecules per protein unit of about 50 KDa), results in the masking of some of the epitopes and in structural modifications of the original Ag and, consequently, in loss of its biological and ligand binding activities. As stated earlier, immunosuppressive strategies utilizing xenogeneic mAbs have as yet not yielded the potential therapeutic benefits of these "magic bullets", primarily because of their immunogenicity in man and their inability to discriminate between Th cells uniquely specific for the allo-Ags of the transplant, and the other Th cells of the host which are involved in protective immune responses. Therefore, the applicant proposes using the immunosuppressive method based on tolerogens for the selective elimination, or radical downregulation, of the T cells of the recipient of the graft which recognize the allo-Ags of the donor's heart and are responsible for its rejection.

This is based on (a) the applicant's recent results demonstrating the induction of specific suppression of ( i ) Abs in outbred rats to Id determinants of murine mAbs

directed against rat and human CD4 by the corresponding mAb(mPEG) _n conjugates, and ( ii ) HAMA responses in hu-PBL- SCID mice by appropriate mPEG conjugates of murine mAbs and (b) increasing evidence that patients producing "blocking" Abs did not reject mismatched kidneys.

A fundamental tenet here, illustrated in Fig. 7 and Fig. 8, is that suppression of rejection of organ transplants in man, including heart transplants, can be achieved by generating anti-clonotypic mAbs (a-cmAbs) in mice or rats which will recognize the clonotypic Ags of the T cell receptors (TCRs) of the patient's Th and Tc cells. These a-cmAbs should react with the unique Vα/Vβ determinants of these cells and, thus, either inactivate these T cells and/or block their interaction with the allograft. Moreover, to prevent the induction of Abs by the patient against the passively administered, xenogeneic a-cmAbs, which would undermine the proposed strategy, the patient will be specifically immunosuppressed to a-cmAbs by administration of tolerogenic mPEG conjugates of a-cmAbs 7 days prior to initiation of the continual series of injections of the unmodified a-cmAbs (the reason for this interval was indicated earlier) .

In essence, this strategy may involve the following 2 phases:

Phase I - Performance of the heart transplantation under the umbrella of the currently used ISDs and continuation of this therapy for a period of 5-8 months, which would be required for the production in rats or mice of ( i ) a-cmAbs to the Vα/Vβ of the TCRs of the recipient's Th and Tc cells recognizing the allo-Ags of the graft, and ( ii ) the tolerogenic mPEG derivatives of these mAbs. In principle, this period could be reduced to a few days (or completely eliminated) when a panel of a- cmAbs becomes available for the most frequently expressed HLAs. In spite of the large number of a-cmAbs required for this purpose, this is achievable on an individual scale by the application of currently available methods of molecular biology and immunology. Phase II - This phase consists of (a) injection,

still under the umbrella of ISDs, of the tolerogenic mPEG conjugates of the above a-cmAbs for activating the Ts cells which will suppress the immune response of the patient specifically to the a-cmAbs, and (b) administration of these unmodified a-cmAbs which should eliminate, or block the patient's CTLs from attacking the heart allograft. The repeated injections of these non- immunogenic and truly protective a-cmAbs are expected to replace the administration of ISDs, thus allowing the return of the patient' s immune system to its normal function.

It is to be noted that it has been recently demonstrated in the applicant's laboratory that specific suppression to Ags may be generated by the corresponding Ag(mPEG) _n conjugates and maintained in vivo under the umbrella of an appropriate immunosuppressive drug (this evidence provides support for the proposition that the metabolic pathways of Ts cells are not affected by this ISD, as are those of Th and B cells). The immunological status of the patient will be monitored at varying intervals with respect to the possible re-appearance of the CTLs, which are expected to have been ablated by the a-cmAbs. This strategy should have advantages over the current blunderbuss approach, involving the indiscriminate use of ISDs and of mAbs to the patient's T cell determinants, which leads to unavoidable side effects. The mPEG conjugates may be synthesized according to methods as previously described. Necessary immunological procedures are commonly used in many laboratories.

Clearly, for the elucidation of the mechanism of a multifactorial system, as is involved in graft rejection, it is imperative to use the simplest, and yet a relevant, experimental model. Singer's extensive and elegant studies of the mechanisms underlying tissue allograft rejection, utilizing a mouse model system (Singer, A. et al., 1984 J. Immunol. 132, 2199-2209; Golding, H. and singer, A. 1985 J. Immunol. 135, 1610-1615; Rosenberg, A.S. et al. 1986. Nature 322, 829-831; Mizouchi, T. et al., 1985 J. Exp. Med. 162, 427-443; Mizouchi, T. et al.,

1986 J. Exp. Med. 163, 603-619) was considered most suitable, since, (i) it is limited to the rejection of skin allografts expressing only class I MHC disparities between wild-type C57BL/6 (B6) mice and the K ^b mutants, (ϋ) each K ^b mutant strain differs from the wild-type B6 mice in the expression of a mutant H-2K ^b molecule, (iii) each K ^bm class I molecule is well characterized biochemically, i.e., H-2K ^bra molecules differ from the H-2K molecules by 2-5 amino acid substitutions which reflect the corresponding gene substitutions, and (iv) the various K ^b mutations alter the K ^b molecule sufficiently so that they differ antigenically. Thus, the rates of rejection of skin grafts from female B6 bm mice by female B6 mice followed a distinct hierarchy: bml skin allografts were rejected most rapidly (MST of 16 days), the MST for bm3 and bmlO grafts was 22 days, and bm6 grafts were retained for at least 150 days. It has also been determined that there was a close relationship between the survival times of allografts of hearts and skin. Hence, the applicant considers that in the initial phase of study, it is justifiable to use skin grafts since they are easier to perform than heart grafts. It is also to be noted that Singer et al. demonstrated that (a) acute tissue allograft rejection caused by (i) differences of class I or class II Ags of the murine major histocompatibility complex (MHC) was initiated by CD4 ^* Th or CD8 ^* Th cells, and (ii ) differences in both class I and class II Ags was initiated by CD4 ^* Th cells , and (b) in both cases the rejection was mediated by CD8 ⁺ Tc cells. Hence, for the success of the strategy, a- cmAbs (and their tolerogenic mPEG conjugates) will have to be generated either (i) to both Th and Tc cells, or ( ii ) to only one of these cells.

As illustrated in Fig. 8, the overall strategy incorporates the following aims.

1. Isolation and identification of wild-type B6 anti-K ^bm3 reactive Th/Tc cells infiltrating the allograft, and also residing in the recipient's draining lymph nodes, and the generation of the corresponding cloned Th and Tc cell lines.

2. Analysis of the Vα/Vβ genes of the TCRs of T cells infiltrating allogeneic hearts, implanted in each B6 mouse between a kidney and its capsule, at different stages of rejection. 3. Analysis of Vα/Vβ genes of the established T cell clones.

4. Generation of rat anti-clonotypic monoclonal Abs.

5. Establishment of the "immune prophylaxis" model utilizing rat anti-clono- typic mAbs and their tolerogenic mPEG conjugates.

6. Establishment of the "rescue therapy" utilizing rat mAbs to clonotypic mu- rine CTLS and mPEG conjugates of these mAbs. (This variation of the therapeutic regimen is not included in Fig. 8.)

EXPERIMENTAL SYSTEMS

1. Isolation and identification of wild-type B6 anti-K ^bl ° ³ reactive Th/Tc cells infiltrating the allograft, and also residing in the recipient's draining lymph nodes, and the generation of the corresponding cloned Th and Tc cell lines.

The rejection of tissue allografts may be due to different B6 anti-K ^bn3 reactive cells generated in response to and reacting with the cells of the graft by one or more pathways. Thus, (i) Th cells of the B6 recipient specific to the graft's allo-Ags may be activated at the site of the graft and/or in the appropriate draining lymph nodes, and (ii ) B6 Tc cells will be generated in the draining lymph nodes of the recipient. These Th and/or Tc cells would then infiltrate the graft, and accelerate the further infiltration of Tc cells which lead to the destruction and final rejection of the graft. Therefore, as indicated in Fig. 8, Th and Tc cell lines will be generated from the T cells which had infiltrated the graft and from T cells of the draining lymph nodes.

For this phase of experiment ( i ) skin of bm3 mice is grafted onto tails of B6 mice (Billingham, R.E. et al 1951 J. E. Biol 28, 385-401) and ( ii ) a piece of bm3 heart placed between the kidney and its capsules of a

wild-type B6 mouse. At 2-day intervals after grafting, and after graft rejection, sections of the grafts are subjected to immuno-histological analysis, for the definition of the phenotypes of the infiltrating T cells utilizing several anti-T cells mAbs, i.e., mAbs to Thy- 1.2 (HO-13-4) (Marshak-Rothstein, A et al., 1979 J. Immunol. 122, 2491-2497), CD4 (YTS191.1), CD8 (YTS169.2) (Cobbold, S.P. et al, 1984 Nature 312, 548-551) and Ly6C (HK1.4) (Havran, W.L, et al 1988 J. Immuno 140, 1034- 1042)

For the isolation of the infiltrating T cells, the grafted hearts are treated with collagenase ( lmg/ml) and DNase (80-100μg/ml for 1 h at 37°C); the Thy-1 ^* T cells are isolated from the cell suspension with magnetic Dynabeads (Dynal, Norway) coated with anti-Thy-1.2 mAb. The Vα/Vβ genes of the infiltrating T cells are to be determined (see next Section), and these cells used to generate cloned Th and Tc cell lines

For the generation of Th cell lines, the isolated T cells are cultured by stimulation at 2-week intervals with irradiated (3,300 rad) spleen cells of bm3 mice in the presence or absence of con A-sup (supernatants from rat spleen cells stimulated for 40 hr with 5 μg/ml of concanavalin A as source of T-cell growth factors) according to the method established by Kimoto and Fathman (1980 J. Exp. Med. 152, 759-770) with slight modifications. The cells of the draining lymph nodes are also to be cultured by the use of this method. After the 3rd or 4th stimulation, single cells are to be picked out from the culture by micro-manipulation for the generation of the corresponding B6 anti-bm3 Th cloned cell lines.

As stated earlier, acute rejection caused by differences in class I Ags was shown to be initiated by CD8 ^* and CD4 ^* Th cells, which recognize class I allo-Ags on allogeneic Ag presenting cells (APC), or in association with class II Ags on self APC, respectively. Both CD8 ^* and CD4 ^* Th cells secrete IL-2 on stimulation with allo-Ags (Mizouchi, T., et al., 1986 J. Exp. Med., 163, 603-619); therefore, the generated Th cell clones may be selected on the basis of their ability to secrete

IL-2. For this purpose, the following two batches of stimulator cells are: (i) one batch of cells prepared from spleen cells of bm3 mice by irradiation at 3,300 rad (these cells will be referred to as APC ^* stimulator cells) , and ( ii ) another batch of cells isolated by passage of the irradiated spleen cells through Sephadex G-10 columns (Ly, I.A. et al., 1974 J. Immunol. Meth. 5, 239-247) to remove the adherent APC (these cells will be referred to as APC ^" stimulator) . To obtain cloned Th cell lines, they are to be cultured with APC ^* or APC ^" stimulator cells in the presence of anti-IL-2R mAbs 7D4 (Malek, T.R. et al., 1983 P.N.A.S. 80, 5694-5698, and in the absence or presence of the irradiated B6 spleen cells (as source of self APC). The reason for the addition of anti-IL-2R mAbs to each culture is to minimize, if not completely inhibit, the consumption of IL-2 produced by allo-Ag reactive T cells by the other cells in the culture; this procedure results in increased sensitivity of the assay. After culture for 4 days, the IL-2 content of the culture supernatants may be determined by a standard method utilizing IL-2 dependent HT-2 (Watson, J. 1979 J. Exp. Med. 150, 1510- 1519) or CTLL-2 (Gillis, S. et al., 1977 Nature 268, 154- 156 cell lines. The clones generated are from (i) T cells of draining lymph nodes, and (ii ) T cells infiltrating several batches of transplanted skin and heart tissue.

For the generation of cloned Tc lines, the T cells infiltrating the graft and residing in the draining lymph node may be maintained by an established procedure

(Yamamoto, H. et al., 1987 Eur. J. Immunol. 17, 719-722). Briefly, the cells from these two sources are stimulated with irradiated bm3 spleen cells at intervals of 5-7 days. Two to 4 days after the 2nd stimulation, the cultured cells are diluted and maintained in 24-well culture plates ( lxl0 ⁵ /well ) with RPMI-1640 medium containing 5% fetal bovine serum, 5xl0' ⁵ M 2- mercaptoethanol, 100 mM glutamine, 100 mM sodium pyruvate, and 10-20% con A-sup; fresh medium is provided at intervals of 2-3 days. After further culture for 4-8

weeks, the cells are subjected to single cell manipulation for cloning. The generated clones may be selected on the basis of their ability to lyse ⁵¹ Cr- labelled LPS- and con A-blast cells of bm3 spleen cells. The CTL clones may also generated from several batches of grafted hearts and lymph nodes of different mice.

The surface phenotypes of each T cell clone may be determined by fluorometric analysis with a flow cell sorter (EPICS model 753, Coulter, CA). For this purpose, the cells are first treated with biotinylated mAbs specific to Thy-1.2 (HO.13.4), CD4 (YST191.1), CD8 (YTS169.1), Ly6C (HK1.4) , Vβ8 (F23.1 ) Staers, U.D. et al., 1985 J. Immunol. 134, 3994-4000), Vβ5 (MR9-4) (Reich, E.P., et al., 1989 Nature 341, 326-328), and Vβll (RR3-15) (Utsunomiya, Y. et al., 1989, J. Immunol. 143, 2602-2608), and then with strep-avidin conjugated phycoerythrin (Becton Dickinson, CA).

2. Analysis of the Vα/Vβ genes of the TCRs of T cells infiltrating allogeneic hearts, implanted in B6 mice between their kidneys and kidney capsules, at different stages of rejection.

It is important to determine if the Vα/Vβ genes of the cloned T cell lines correspond to the T cells which actually infiltrate the transplanted heart, since one may visualize that the T cells selected for cloning may represent cells proliferating preferentially under the culture conditions. Hence, a piece of adult bm3 heart is to be implanted between the kidney and the capsule of each B6 mouse (bm3 mice will serve for controls) . At intervals of 2-4 days after implantation, the infiltrated T cells are to be isolated, as described above, and the CD4 ^* and/or CD8 ^* T cells separated from the isolated T cells with Dynabeads coated with anti-CD4 mAb or anti-CD8 mAb, respectively. Total RNA may be extracted from these cells by the standard guanidium thiocyanate-cesium chloride method (Chirgwin, J. M. et al. , 1979 Biochem. 18, 5294-5299) and the first strand of cDNA synthesized using 5 μg of total RNA in a 20 μl reaction mixture consisting of oligo(dT) primer and AMV reverse transcriptase. Both the Vα and Vβ

genes may be amplified by the polymerase chain reaction (PCR) utilizing the Taq polymerase according to the method of Maeda et al . (1991, Diabetes 40, 1580-1585). For amplification of the corresponding Vβ gene, each of the seventeen different Vβ primers may be used in conjunction with a 3 ' -Cβ primer; 5 ' -Cβ and 3 ' -Cβ primers may serve as controls. The synthetic oligonucleotides of all Vβ primers and Cβ primers may be purchased from the DNA Synthesis Laboratory, University of Manitoba. For quantification of Vβ and Cβ transcripts, the amplified PCR products may be hybridized with ³² P-labelled Hinfl- EcoRI fragment of Cβ (86T5) gene (Hedrick, S.M. et al., 1984 Nature 308, 153-158). The intensities of the auto- radiographic bands may be quantified by densitometry. Reference standard curves may be established with cDNAs of splenic CD4 ^* plus CD8 ^* T cells of the same mice (their cDNA's having been subjected to PCR) after appropriate dilution, with each of the Vβ and 3 ' -Cβ primers of 5' -Cβ and 3 ' -Cβ primers as controls. The Vα gene frequencies may also be determined by this procedure utilizing the corresponding Vα, 3 ' -Cα and 5 ' -Cα primers.

On the basis of this gene analysis, it should be possible to establish (i) if the cloned T cell lines are derived from T cells infiltrating the graft, and (ii) the preferential Vα/Vβ genes of these T cells. Furthermore, by a kinetic analysis of the Vα/Vβ gene usage, it should be possible to establish the sequence of infiltration of the graft by different T cells, i.e., to identify the T cells which infiltrate the tissue in the early, middle and end phase of the rejection episode. This information will be useful for defining the T cell types participating in different stages of the rejection episode, i.e., it should be possible to establish if CD4 ^* and CD8 ^* Th cells initiate the graft reaction, and if CD8" and CD4 ^* CTLs play the major role as effector cells in the rejection episode.

3. Analysis of Vα/Vβ genes of the established T cell clones.

The methods used in the previous section may be employed for the definition of the Vα/Vβ genes of each T

cell clone. Each ss DNA may be hybridized with cloned gene fragments of 15 different Vβ and 11 different Vα genes (kindly provide/d by Dr. E. Palmer, Denver, CO). After the determination of the family of Vα and Vβ genes, the Vα and Vβ cDNAs may be amplified utilizing the corresponding Vα primers and 3'-Cβ, and the corresponding Vβ primers and 3'-Cβ, in order to determine the Jα and Dβ-Jβ gene usages. Each amplified product will be cloned into pCRlOOO vector (In Vitrogen, CA) and will be sequenced using T ₇ /M13 primers.

4. Generation of rat anti-clonotypic monoclonal Abs. After identification of the cloned T cells which participate in different stages of the rejection episode, rat a-cmAbs are to be generated to the Vα-Jα/ /Vβ-Dβ-Jβ regions of the TCRs of the CD4 ^* and CD8 ^* T cells involved in the early phase; these a-cmAbs should inhibit the initiation of graft rejection. Similarly, rat a-cmAbs are to be produced against the Vα/Vβ regions of the CD8 ^* and CD4" CTL clones, derived from T cells infiltrating the heart graft during the middle phase of the episode; these a-cmAbs should inhibit the continuing destruction of the transplanted heart.

For the production of these mAbs, the method described by Dialynas et al. may be used (1983 Imm. Rev. 74 29-56). Lewis rats are immunized twice with each of the identified Th and CTL clones at an interval of 2 weeks. Seven days after the second immunization, all rats are bled and each of their sera absorbed with B6 spleen cells to remove antibodies to common murine Ags. After absorption each serum is tested for its capacity to inhibit ( i ) the proliferation of the corresponding Th cell clones in the presence of allo-Ags, or (ii ) the cytolysis of LPS blast cells of bm3 mice by the corresponding CTL lines. The rats producing the desired a-cAbs are immunized once more and 2 days later their spleen cells fused with mouse myeloma cells, P3.8.653.20AR, for the generation of hybridomas producing a-cmAbs, according to standard procedures.

The IgG a-cmAbs to Th and CTL clones may be screened by the inhibition of proliferation of the corresponding

Th cell lines, or by the suppression of cytotoxic activity of the corresponding CTL lines, respectively. Finally, the selected mAbs may be subjected to further selection on the basis of their capacity to inhibit the MLR between polyclonal B6 spleen cells and bm3 cells, or the cytotoxicity of polyclonal B6 CTLs using ⁵¹ Cr-labelled LPS blast cells of bm3 mice. It is envisaged that (i) the former mAbs(s) would abrogate Th cells involved in the initial phase of the rejection episode, i.e., these mAbs would be most useful as the enabling agents of the "immune prophylaxis" treatment, and ( ii) the latter mAbs would block the ongoing rejection process and may prove useful in the "rescue therapy" for reversing the acute rejection. The mAbs directed against the clonotypes of the B6 CD8 ^* and CD4 ^* Th cells will be referred to as mAbs- 8H or mAbs-4H, respectively, and the corresponding mAbs to the clonotypes of the CTLs recognizing the epitopes of the bm3 mutant will be referred to as mAbs-K. 5. Establishment of the "immune prophylaxis" model utilizing rat anti-clonotypic mAbs and their tolerogenic mPEG conjugates.

As mentioned above, rat IgG mAbs to Vα/Vβ regions of cloned Th cells should suppress the MLR between host and donor cells and are, therefore, to be used to explore their efficacy in "immunoprophylaxis" by inhibition of immune responses of the graft recipient, which lead to the graft's rejection. For the success of this immunological intervention, it is essential that the host should not mount an immune response against the appropriate rat mAbs.

A Regimen for the immune prophylaxis is proposed, consisting of two steps utilizing for the exploratory experiments using skin grafts.

Step I: The B6 mice receive the tolerogenic mAbs- 8H(mPEG) _n and/or mAbs-4H(mPEG) _n , for induction of immunological suppression to mAbs-8H and mAbs-4H, respectively.

Step II: Beginning 7 days later, the B6 mice receive i.v. injections of the "prophylactic mAbs", at intervals determined by the half-life of these mAbs. On

the eighth day pieces of skin of bm3 mice are grafted onto the tail of B6 mice for the determination of the efficacy of this therapy. The control mice may reject the allograft within 14 to 30 days (as stated earlier, the MST depends on the bm strain combinations) . A marked prolongation of MST, will lead to testing of this therapy for heart transplantation, i.e., hearts will be grafted heterotopically in the abdominal cavity by the method of Corry et al. (1973 Transplantation 16, 343-350). 6. Establishment of the "rescue therapy" utilizing rat mAbs to clonotypic murine CTLs and mPEG conjugates of these mAbs.

In an attempt to arrest an ongoing graft rejection, rat mAbs to B6 anti-bm3 CD8 ^* CTL clone(s), [i.e., to mAbs- K(s)] are to be used for the "rescue therapy", i.e., these mAbs will be administered to B6 mice undergoing rejection of bm3 skin grafts. Since, in contrast to the "prophylaxis therapy", this clinical intervention has to be applied without any delay, the arrest of the ongoing graft rejection episode may require the injection of mAbs-K prior to tolerization of the host by administration of mAbs-K(mPEG) _n . However, to avoid triggering the anti-mAbs-K antibody response in the host, which would undermine this therapy, these anti-CTL mAbs will have to be administered under the umbrella of an ISD.

This approach may be tested by examining two variants of the strategy for the "rescue therapy", utilizing the simpler transplantation model involving grafting of a piece of bm.3 skin onto the tail of a B6 mouse, both therapies being initiated 7-10 days after grafting, i.e., after initiation of the rejection episode. The first therapeutic modality involves the i.v. injections of mAbs-K at intervals of 2-4 days without ISD. If prolongation of MST is achieved by this procedure, the same regimen is to be tested for heart transplantation by the method of Corry et al. (1973 supra) .

However, since induction of Ag-specific Ts cells by tolerogenic mPEG conjugates requires about 7 days, in the

second therapeutic regimen, to ensure that the onset of rejection does not occur prior to having induced tolerance to mAbs-K, the grafting of the skin is to be done under the umbrella of ISD with simultaneous administration of mAbs-K(mPEG) _n . Thus, it is conceivable that one might be able to prepare the patient for the "rescue therapy" even prior to the onset of the rejection episode and that the production of CTLs may be prevented altogether. It is also envisaged that the proposed therapeutic strategy could be simplified. Instead of using the a- cmAbs in a two-step procedure, in conjunction with their tolerogenic mPEG conjugates, one may use directly the mPEG conjugates of the Ag-binding fragments of the a- cmAbs, i.e., the Fab', or the (Fab' ) ₂ , or the Fav, or even the antigen binding region of the H chains of these a- cmAbs.

The present invention encompasses methods of treatment, compositions for use in methods of treatment, the use of compositions in treatment, the use of compositions in the manufacture of medicaments for use in treatment, and products containing compositions as a combined preparation for simultaneous, separate or sequential use in therapy; as described herein.