WO1993002696A1 - T-cell membrane protein for b cell ig class switching - Google Patents

Abstract

Purified protein, εRMP, defined by the partial internal sequence (I) characterized in that it is a serine protease, and ε receptor modulating protein that modulates the low affinity Fc receptor for IgE on lymphocytes CD23, has a molecular weight of approximately 17 kDa, decreases the binding avidity of IgE of CD23-bearing B cells and does not diminish the cell surface expression of CD23 and purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgG1 synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNη are disclosed.

Description

T-CELL MEMBRANE PROTEIN FOR B CELL IG CLASS SWITCHING

Cross- Reference to Related Application This is a continuation-in-part of Serial No. 07/741,671, Filed August 7, 1991, to which priority is claimed. Government Rights

The government has certain rights in the invention described herein in that this work was supported by NIH Grant AI-24526 and in part by a contract from QUIDEL Corporation, San Diego, CA.

Field of the Invention This invention relates to mimunochemistry and immunology and to methods and materials for modifying immune responses.

The suppressive factor of allergy (SFA) was initially characterized as a crude protein preparation which has both IgE-suppressive and CD23-modu- lating activities (Tung, A.S., Chiorazzi, N. and Katz, D.H. (1978) J. Immunol. 120,2050-2057; Katz, D. H. and Tung, AS. (1978) 7. Immunol.. 120,2060-1066; Katz, D.H. (1979) J.Exp.Med. 149,539-543; Katz, D.H. (1978) ImmunoLRev 41,77-73). This invention resulted from studies of the present inventors resulting in the discovery that crude SFA can be purified into two functionally and biochemically distinct protein molecules. One protein (30 Kda) suppresses IgE biosynthesis ( ._., purified SFA), which was shown to be distinct from IFN. The second protein (17 Kda) modulates CD23, and is thus designated ε receptor modulating protein (εRMP). εRMP has a unique internal amino acid sequence,

Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala 1 5 10 15

Lys Lys (SEQ ID NO:2). Further studies on the effects of εRMP on IgE biosynthesis showed that 1) εRMP exists not only as 17-Kda soluble form but also as 36-Kda T cell membrane form; and 2) εRMP singularly induces gennline IgE heavy chain transcripts and enhances IgE synthesis by B cells in the presence of IL-4. Soluble εRMP can be used as a therapeutic for altering the immunological characteristics of humans and other mammals and both soluble and membrane forms of εRMP can be used in tests and assays in aid of diagnosis, for separating or concentrating immunological species, and in research on the immune system of mammals. Background of the Invention

The following abbreviations are used throughout this patent: ACNacetonitrile;

BLT N,α-benzyloxycarbonyl-L-lysine thiobenzyl ester; DMSOdimethyl sulfoxide; DNPdinitrophenyl; εRMP ε receptor modulating protein; εRMPmε receptor modulating protein, membrane-form εRMPs ε receptor modulating protein, soluble-form

IL-4 interleukin-4 INFy interferon-y;

HPLC high performance liquid chromatography

Kda kilodalton

KHL keyhole limpet hemocyanin

LPS lipopolysaccharide; Mab monoclonal antibody

MHC major histocompatibility complex mw molecular weight;

OPA o-phthalaldehyde;

PMSF phenylmethylsulfonyl fluoride; PBS phosphate buffered saline

PVDF polyvinylidene difluoride

RP-HPLC reversed-phase high performance liquid chromatography; T_H T-helper cell

TFA trifluoroacetic acid; TNP trinitrophenyl.

SFA suppressive factor of allergy; A factor termed suppressive factor of allergy (SFA) which exerted selective regulatory activity on the production of murine IgE antibodies has been described (Tung, AS., Chiorazzi, N. and Katz, D.H. (1978) J. Immunol. 120,2050-2057; Katz, D. H. and Tung, A.S. (1978) /. Immunol.. 120,2060-1066; Katz, D.H. (1979) J.Exp.Med. 149,539-543; Katz, D.H. (1978) ImmunoLRev 41,77-73). This activity was initially identified as components in serum (or ascites fluids) obtained from mice which had been repeatedly injected with mycobacteria-containing complete Freund's adjuvant. Administration of crude SFA to rodents selectively suppresses IgE antibody production, hence its nomenclature (Katz, D. H. (1984) Allergy 39,81-100; Ishizaka, K. (1984) in PROGRESS IN IMMUNOLOGY V., eds. Yamamura, Y. and Tada, T., Academic Press, Tokyo, pp. 455-460; Katz, D.H. (1986) /.Allergy ClinΛmmunol 78,980-985). In addition to selectively regulating IgE antibody synthesis, SFA suppresses functional expression of the CD23. This rosette inhibition results from SFA acting on CD4⁺ T cells to release an IgE-binding factor (Ishizaka, K. (1988) Ann.Rev.Immunol. 6,513-525) termed suppressive effector molecule SEM; (Marcelletti, J.F., del Guercio, P. and Katz, D.H. (1985) Ann.Inst.Pasteur/- Immunol. 136C.387-394; Marcelletti, J.F. and Katz, D.H. (1984) J. Immunol. 133,2829-2835; Marcelletti, J.F. and Katz, D.H. (1984) /. Immunol.. 133,2845-2852; Marcelletti, J.F. and Katz, D.H. (1986) J. Immunol. 137,2599-2606; Marcelletti, J.F. and Katz, D.H. (1984) J. Immunol. 133,2837-2845). Another lymphokine, interleukin-4 (1L-4), has the opposite activity, /._. it up-regulates CD23 expression (Coffman, R.L., Ohara, J., Bond, M.W., Carty, J., Zlotnik, A. and Paul, W.E. (1986) J.Immunol 136,4538-4542; Conrad, D.H., Keegan, A.D., Kallil, K.R., Dusen, R.V., Rao, M. and Levine, AD. (1988) /. Immunol. 141,1091-1098; Conrad, D.H., Waldschmidt, T.J., Lee, W.T., Rao, M., Keegan, A.D., Noelle, R.J., Lynch, R.G. and Kehry, M.R. (1987) /. Immunol. 139,2290- 2295; Finkelman, F.D., Katona, I., Urban, J.F., Snapper, CM., Ohara, J. and Paul, W.E. (1986) Proc.Nall.Acad.Sci.USA 83,9675-9680) and monoclonal antibody to IL-4 blocks in vivo IgE biosynthesis in mice (Finkelman, F.D., Katona, , Urban, J.F., Snapper, CM., Ohara, J. and Paul, W.E. (1986) Proc.NatlAcad.Sci.USA 83,9675-9680).

A murine T cell hybridoma line that produces SFA has been constructed (Marcelletti, J.F. and Katz, D.H. (1989) Celllmmunol 12,314- 320). The present invention relates to the purification of native murine SFA derived from this hybridoma that revealed the existence of two biochemically and functionally distinct proteins, one of which modulates CD23 molecules and the other of which suppresses IgE biosynthesis. The CD23-modulating protein decreases the avidity of IgE binding to CD23- bearing B cells without decreasing CD23 expression on the cell surface, and has thus been designated "ε receptor modulating protein (εRMP)." The IgE suppressive protein (Le. SFA) is shown to be distinct from IFNy or IFNα.

Summary of the Invention It has now been discovered that the murine suppressive factor of allergy (SFA), which was originally described in sera or ascites of mice repeatedly injected with complete Freund's adjuvant, consists of two functionally and biochemically distinct protein molecules. One protein modulates the low affinity Fc receptor for IgE on lymphocytes, Le. CD23; the second protein suppresses IgE biosynthesis. Each protein was purified from culture fluids of the T cell hybridoma line MBI-1.15 which was adsorbed onto and eluted from methyl-silica and then subjected to reversed- phase high performance liquid chromatography. The purified CD23- modulating protein (mw of 17 Kda) decreases the binding avidity of IgE of CD23-bearing B cells without affecting quantitative expression of CD23; it is thus designated ε receptor modulating protein (εRMP). The purified IgE-suppressive protein (SFA; mw of 30 Kda) suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes. The IgE suppressive activity of SFA is not blocked by anti-IFNy monoclonal antibody and is thus distinct from IFN. Collectively, the data presented indicate that εRMP and SFA are novel protein molecules involved in modulating the CD23 molecule and IgE antibody synthesis, respectively.

Having identified and purified a murine 17-Kda protein which diminishes the avidity of binding between IgE and CD23 (low affinity Fc receptor for IgE) without decreasing the quantitative expression of the CD23, designated ε receptor modulating protein (εRMP), this protein was further characterized. It was discovered that: 1) εRMP is inactivated by phenylmethylsulfonyl fluoride (PMSF) and decomposes N,α-benzyloxycarb- onyl-L-lysine thiobenzyl ester (BLT), as well asN-succinyl-Ala-Ala-Pro-Phe- p-nitroanilide; 2) εRMP does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells; and 3) the partial internal amino acid sequence of εRMP, obtained by using in situ cyanogen bromide cleavage on polyvinylidene difluoride (PVDF) membrane is unique. These data thus clearly demonstrate that εRMP is a novel serine protease controlling the functional expression of CD23 through the participation of CD4⁺ T cells.

Briefly, one facet of the invention is embodied in a composition of mater consisting essentially of purified εRMP characterized in that it is a serine protease, an ε receptor modulating protein that modulates the low affinity Fc receptor for IgE on lymphocytes CD23, has a molecular weight of approximately 17 Kda, decreases the binding avidity of IgE of CD23- bearing B cells and does not diminish the cell surface expression of CD23 and purified IgE-suppressive protein having a molecular weight of approximately 30 Kda that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFΝ and methods of preparing and using the same in research and in the treatment of allergic reactions, e.g. inflammation, in mammals.

Further studies showed that, in addition to modulating CD23, εRMP costimulates with IL-4 the de novo synthesis and secretion of IgE by cultured B cells. Since such costimulating activity is reminiscent of a similar synergism with IL-4 previously observed with cell membranes from activated T cells, we examined isolated membranes from the εRMP-producing MBI- 1.15 T cell line for comparable activity; indeed, as shown herein, MBI-1.15 cell membranes do exhibit this synergism. Furthermore, we show that a monoclonal antibody (Mab), 2E5B, specific for the 17-Kda soluble form of εRMP, blocks the costimulating activities of both the soluble εRMP and MBI-1.15 T cell membranes for 1L-4 induced de novo synthesis of IgE by cultured B cells. This anti-εRMP Mab also detects a 36-Kda membrane- bound protein species which appears to be related to soluble εRMP by immunochemical criteria. Finally, the membrane-bound proteins, present on MBI-1.15 T cells, induce germ-line IgE heavy chain transcripts in 1-29 B cells independently of IL-4, and this inductive event is also specifically blocked by the 2E5B anti-εRMP Mab. These findings suggest that T cell membrane-bound εRMP molecules are crucial proteins involved in B cell class switching in the course of IgE biosynthesis.

Finally, both IL-4 and εRMP induce Iε transcription, but neither molecule by itself can induce class switching to IgE synthesis. This clearly suggests that both εRMP and IL-4 each have another important molecular effect (which may or may not be identical) on B cells, that is essential for class-switching, but only when both molecules are present simultaneously is the complete mechanism of class-switching manifested. Brief Description of the Drawings Figure 1 depicts in graphical form data from the reversed phase high performance liquid chromatograph (RP-HPLC) of SFA. Figure 2 depicts in graphical form data from SDS-PAGE test for

CD23-modulating activity of the molecules described.

Figure 3 depicts in graphical form data from molecular sieving HPLC and SDS-PAGE of IgE-suppressive activity.

Figure 4 depicts in graphical form data showing that purified CD23- modulating activity (17 Kda protein) does not suppress quantitative expression CD23 but decreases binding of IgE onto CD23-bearing B cells. Figure 5 depicts in graphical form data showing that anti-IFNy antibody does not block the effect of SFA on IgE synthesis by LPS+IL-4- stimulated lymphocytes.

Figure 6 depicts in graphical form data showing that εRMP is inhibited by phenylmethylsulfonyl fluoride (PMSF).

Figure 7 depicts in graphical form data showing that purified εRMP requires CD4⁺ T cells to functionally suppress CD23.

Figure 8 depicts the results of immunoprecipitation and Western blot analyses with 2E5B anti-εRMP Mab. (A) Radioiodinated MBI-1.15 cell- surface proteins were immunoprecipitated with 2E5B anti-εRMP Mab- bound Sepharose and analyzed on 13.5%SDS-PAGE. Lane 1, non- reducing; Lane 2, reducing conditions. (B) MBI-1.15 cell lysates separated on 13.5% SDS-PAGE in non-reducing (lane 1) or reducing (lane 2) conditions, transferred onto nitrocellulose membrane, and incubated with radioiodinated 2E5B anti-εRMP Mab.

Figure 9 depicts data demonstrating the induction of Iε by T cell membranes is specifically blocked by 2E5B anti-εRMP Mab. 1-29 B cells (1 x 105/ml) were cocultured with MBI-1.15 cell membranes (1 x 10⁷ equivalent/ml) or IL-4 (1000 U/ml) in the presence or absence of 2E5B anti-εRMP Mab (1:20 or 11B11 anti-IL-4 Mab (50 μg/ml) for 3 days, and poly (A)⁺ RNA was prepared. Each lane was loaded with 0.5 μ.g of poly (A)⁺ RNA for separation on 1% agarose-formaldehyde gel using minigel apparatus, and Northern blot analysis was performed with nick- translated Iε probe. +, added to culture; -, not added to culture. Description of the Preferred Embodiments

T cell regulation of immunoglobulin (Ig) synthesis and secretion by B cells has been one of the most intensely scrutinized processes in modern immunology. The complexity of events and the various molecules involved in intercellular T-B interactions have been largely, but not completely, unraveled. For example, the precise membrane-bound molecules on T cells which participate in cognate and/or non-cognate signal transduction for Ig heavy chain transcription in B cells are still unknown. The studies described here have focused on these processes in the IgE antibody system.

Interleukin-4 (IL-4)³ has been described as an IgE class-switch factor. However, IL-4 alone fails to induce synthesis of IgE by human or murine purified B cells. Thus, IL-4 induction of IgE synthesis by purified B cells requires the presence of a costimulant, which in humans can be provided by T cells, anti-CD40 antibody, hydrocortisone, Epstein-Barr virus infection, or phorbol ester-activated EL-4 murine thymoma cells, and in mice can be provided by T cells (Coffman, R.L., Seymour, B.W.P., Lebman, D.A., Hiraki, D.D., Christiansen, J.A., Shrader, B., Cherwinski, H.M., Savelkoul, H.F.J., Finkelman, F.D., Bond, M.W., & Mosmann, T.R. 1988. The role of helper T cell products in mouse B cell differentiation and isotype regula¬ tion. Immunol Rev. 102:5) or lipopolysaccharide (LPS). Amongst these, T cells most likely provide the important physiological role. Sorting out the stepwise events involved, Vercelli, et al. (Vercelli, D.,

Jabara, H.H., Arai, K-I, & Geha, R.S. 1989. Induction of human IgE synthesis requires interleukin 4 and T/B cell interactions involving the T cell receptor/CD23 complex and MHC class II antigens. J.Exp.Med. 169:1295) demonstrated that cognate interactions involving T cell receptors and class II major histocompatibility (MHC)/antigen complexes, while required to activate helper T (T_H) cells, are not necessarily involved in later events. Thus, when T and B cells are cultured in adjoining chambers separated by a membrane, IL-4 fails to induce IgE synthesis. IL-4 induction does occur when T and B cells are cultured together, and kinetic studies showed that physical contact with T cells is a prerequisite early step to enable B cells to respond to IL-4 signals. Thereafter, activated T cells use other signals, not requiring cognate physical contact, to render B cells susceptible to IL-4- induced class-switching (Vercelli, D., Jabara, H.H., Arai, K-I, & Geha, R.S. 1989. Induction of human IgE synthesis requires interleukin 4 and T/B cell interactions involving the T cell receptor/CD23 complex and MHC class II antigens. J.ExpMed. 169:1295). These findings suggest that IgE synthesis is dependent on two main signals provided by already activated T_u cells: the first mediated by non- cognate, but contact-dependent, physical interactions between T and B cells (Parronchi, P., Tiri, A., Macchia, D., De Carli, M., Biswas, P., Simonelli, C, Maggi, E., Del Prete, G., Ricci, M., and Romagnani, S. 1990. Noncognate contact-dependent B cell activation can promote IL-4-dependent in vitro human IgE synthesis. /. Immunol. 144:2102); and the second by T cell- derived (or exogenously added) IL-4. Of particular importance in this regard is that plasma membranes prepared from activated T_H cell clones are minimally required for IL-4-dependent IgE synthesis by B cells (Hodgkin, P.D., Yamashita, L.C., Coffman, R.L., & Kehry, M.R. 1990. Separation of events mediating B cell proliferation and Ig production by using T cell membranes and lympliokines. /. Immunol. 145:2025), Le., significant IgE synthesis is observed when B cells are cocultured with IL-4 plus activated T_H cell membranes. Not only T_H2-but also T_πl-cell membranes exhibited this activity, whereas CD8⁺ T cell membranes did not (Hodgkin, P.D., Yamashita, L.C, Coffman, R.L., & Kehry, M.R. 1990. Separation of events mediating B cell proliferation and Ig production by using T cell membranes and lymphokines. /. Immunol. 145:2025, Hodgkin, P.D., Yamashita, L.C., Seymour, B., Coffman, R.L., and Kehry, M.R. 1991. Membranes from both T_H1 and T_H2 T cell clones stimulate B cell proliferation and prepare B cells for lymphokine-induced differentiation to secrete Ig. /. Immunol. 147:3696). At molecular levels, T_H cell membranes induced germ-line IgE heavy chain transcripts in human B cells (Gauchat, J.-F., Lebman, D.A. Coffman, R.L. Gascan, H., & de Vries, J.E. 1990. Structure and expression of germline ε transcripts in human B cells induced by interleukin 4 to switch to IgE production. /. Exp. Med. 172:463), possibly increasing accessibility of recombinase to class-switch regions and therefore facilitating IL-4- induced class switching to IgE. Another study showed that Mabs to human B cell CD40 enabled IL-4 to induce B cell class switching to IgE (Jabara, H.H., Fu, S.M., H\Geha, R.S., & Vercelli, D. 1990. CD40 and IgE: synergism between anti-CD40 monoclonal antibody and interleukin 4 in the induction of IgE synthesis by highly purified human B cells. J.Exp.Med. 172:1861). However, the signals transmitted to B cells by anti-CD40 Mabs are probably different from those emanating from T_H cell membranes since anti-CD40 Mabs were incapable of inducing germ-line IgE heavy chain transcription in human B cells (Gascan, H., Gauchat, J.-F., Aversa, G., Van Vlasselaer, P., & de Vries, J.E. 1991. Anti-CD40 monoclonal antibodies or CD4⁺ T cell clones and IL-4 induce IgG4 and IgE switching in purified human B cell via different signaling pathways. /. Immunol. 148:8). Thus, the nature of the molecules on TH cell membranes which mediate physiological TH-B cell interactions for IL-4 induced IgE synthesis is still unknown. CD23 is the low affinity IgE Fc receptor, and its interaction with IgE is in part regulated by a T cell-derived CD23- modulating protein (denoted ε receptor-modulating protein, or εRMP). εRMP is a 17-Kda novel serine protease identified in and isolated from the supernatant fluids of the murine T cell hybridoma, MBI-1.15 (Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and IgE- suppressive proteins. Proc. Natl Acad. Sci. USA. 88:4718, Matsushita, S., & Katz, D.H., 1991. The murine ε receptor modulating protein: A novel serine protease which modulates CD23 binding of IgE. Cell Immunol 137:252).

The T cell hybridoma, MBI-1.15, and a B cell hybridoma clone, designated 2E5B, are being deposited contemporaneously herewith with the American Type Culture Collection (ATCC), Rockville, Maryland, USA, an approved depository under the Budapest Treaty, under conditions as follows:

The deposited cell line are identified respectively as "MBI-1.15" identified in the ATCC by the access number ATCC and "MBI-2E5B" identified in the ATCC by the access number ATCC

During the pendency of this application, access to both deposits will be afforded to the Commissioner upon request; All restrictions upon the availability to the public of the deposited biological materials will be irrevocably removed upon granting of a patent on this application;

Said deposits will be maintained in said depository, or another depository acceptable to the Commissioner, for a period of at least thirty years from the date of deposit or for the enforceable life of the patent or for a period of five years after the date of the most recent request for the furnishing of a sample of the deposited cell line, which ever is longest; and

Said deposit(s) will be replaced if either deposit should become nonviable or non-replicable. In the course of studying the possible function(s) of εRMP in the regulation of IgE biosynthesis, we have identified a T cell membrane-bound form of εRMP which appears to play a crucial role in inducing Ig class- switching in B cells to IgE. Herein, we describe the function of εRMP in IgE biosynthesis using 1) murine in vitro IgE synthesis induced by IL-4 in the presence of soluble εRMP (εRMPs) or T cell membrane-bound εRMP; 2) T cell membrane-induced germ-line IgE heavy chain transcription by 1-29 B cells; and 3) the effects of anti-εRMP Mab on these assay systems. MATERIALS AND METHODS

Proteins and Chemicals. The following reagents were purchased: 2,4,6-trinitrobenzenesulfonic acid (TNBS; ICN Pharmaceuticals, Cleveland, OH); trifluoroacetic acid (TFA; Pierce, Rockford, IL); acetonitrile (ACN; Baxter, Muskegon, MI); Rapid-AG-Stain (ICN Radiochemicals, Irvine, CA): and low molecular weight markers for SDS-PAGE (Bio-Rad, Richmond, CA). For quantitation of IFNy, murine IFNy ELISA kit was used (Amgen Biologicals, Thousand Oaks, CA). Monoclonal Antibodies and Complement. DNP-specific monoclonal IgE antibody was obtained from the hybridoma previously reported (Liu, F.- T., Bohn, J.W., Ferry, EX. Yamamoto, H., Moliuaro, C.A., Sherman, L.A., Klimnan, N.R. and Katz, D.H. (1980)/. Immunol. 124,2728-2736), and was isolated by affinity chromatography on DNP-BSA-coupled Sepharose 4B. Monoclonal anti-mouse CD23 (B3B4) was generously provided by Dr. Daniel Conrad (Rao, M., Lee, W.T. and Conrad, D.H. (1987) /. Immunol. 138,1845-1852), Medical College of Virginia (Richmond, VA). Monoclonal anti-Thy 1.2 derived from the F7D5e7 hybridoma clone (Lake, P., Clark, E., Khorshidi, M. and Sunshine G., Eur. J. Immunol.. 1979. 9:875), monoclonal anti-L3T4 obtained from the clone GK1.5 (Dialynas, D.P., Quan, Z.S., Wall, K.A., Pierras, A., Quintans, J., Liken, M.R., Pierres, M. and Fitch F.W., /. Immunol. 1983. 131:2445), and monoclonal anti-Lyt-2.2 (purchases from Accurate Chemical and Scientific Corp., Westbury, NY) were used for depletion of T cells, CD4⁺ T cells, and CD8⁺ T cells, respectively. Guinea pig complement (C) was obtained from Pel-Freeze Biologicals, Rogers, AR, and was extensively absorbed at 4°C with BALB/c spleen cells before use. PMSF, dimethyl sulfoxide (DMSO) and all the serine proteases employed were purchased from Sigma, St. Louis, MO. PVDF membrane was purchased from Millipore Corporation, Bedford, MA.

Cell Culture of MBI-1. Murine T cell hybridoma MB1-1 was produced by fusion of allostimulated T cell blasts from DBA/2 donors and AKR-derived T cell line BW5147 as described (Marcelletti, J.F. and Katz, D.H. (1984) /. Immunol. 133,2837-2845; Marcelletti, J.F. and Katz, D.H. (1989) CelLImmunol. 12,314-320). A cloned subline of MBI-1 (termed MBI-1.15) that produces high titer of CD23-modulating activity (measured by rosette assay) was isolated by limiting dilution and maintained in Dulbecco's Modified Eagle's Medium (DMEM) plus 10% fetal calf serum (FCS, JR Scientific, Woodland, CA), L-glutamine (2 mM), and antibiotic supplements. To prepare FCS-containing MBI-1.15 supernatant, 5 x 10⁴ MBI-1.15 cells were cultured in 0.5 ml of 10% FCS/DMEM for 48 hours using 24-well culture plates (Costar, Cambridge, MA). To obtain FCS-free MBI-1.15 supernatant, 1 x 10⁹ MBI-1.15 cells were washed twice with phenol red-free DMEM (Sigma, St. Louis, MO), transferred to 3 liters of the same media without FCS, and cultured for 24 hours in a spinner culture flask. As a negative control for biological assays, mock FCS/DMEM supernatant was processed in precisely the same way as hybridoma fluid and used at the same protein concentration (11 μg/ml).

Murine εRMPs was purified from serum-free culture supernatants of T cell hybridoma MBI-1.15 by C18 reversed-phase high-performance liquid chromatography (HPLC) as described previously (Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and IgE- suppressive proteins. Proc. Natl Acad. Sci. USA. 88:4718). Recombinant IL-4 was generously provided by Dr. J. Ohara (University of Colorado). Oligopeptide-amide with 19 amino acid residues with cysteine inserted into its N-terminus was synthesized based on internal amino acid sequence of purified εRMP as described (see also Matsushita, S., & Katz, D.H., 1991, The murine ε receptor modulating protein: A novel serine protease which modulates CD23 binding of IgE. Cell. Immunol. 137:252).), (Multiple Peptide Systems Corp., San Diego, CA),

Cys Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala 1 5 10 15 Ala Lys Lys (SEQ ID NO:l)

(CAKPAPKKEEKKKKAAAKK-amide). The synthetic oligopeptide was conjugated to keyhole limpet hemocyanin (KLH) using the bifunctional crosslinker, m-maleimidobenzoyl-N-hydroxysuccinimide ester (Lerner, R.A., Green, N., Alexander, H., Liu, F.-T., Sutcliffe, J.G., & Shinnick, T.M. 1981. Chemically synthesized peptide predicted from the nucleotide sequence of the hepatitis B virus genome elicit antibodies reactive with the native envelope protein of Dane particles. Proc. Natl, Acad, Sci. USA 78:3403). Lewis rats (8-week old female, Harlan Sprague Dawley, Indianapolis) were immunized with the oligopeptide-KLH conjugate emulsified in complete Freund's adjuvant (lOOμg/injection/rat), twice subcutaneously and once intrapertioneally at 2-week intervals. Spleen cells of a rat with the highest antibody titer were fused with the non-secreting mouse B cell myeloma, FO. Standard techniques of hybridoma selection and propagation was used and the clones producing the relevant antibodies were screened by ELISA. A B cell hybridoma clone, designated 2E5B, secreted antibody that reacted with not only the immunogeα but also native εRMPs purified to homogeneity (Matsushita, S., Marcelleti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and IgE-suppressive proteins. Proc. Natl Acad. ScL USA. 88:4718). The 2E5B Mab was a rat IgM as determined by rat Mab typing kit (The Binding Site, Inc., San Diego, CA). Biological Assays.

In Vitro Culture for Enhanced Expression of CD23 on Murine

Lymphoid Cells. In an exemplary procedure, either unfractionated or subpopulation-depleted spleen cells of 8 to 10-week old female BALB/c mice (Jackson Labs, Bar Harbor, ME) were cultured at a final density of 5 x 10⁶ nucleated cells/0.5 ml/well in 24-well macroplates in the presence of monoclonal IgE (Liu, F.-T., Bolin, J.W., Ferry, E.L., Yamamoto, H.,

Molinaro, CA., Sherman, L.A., Klimnan, N.R. and Katz, D.H., /. Immunol.

1980. 124:2728), effectors such as εRMP, SFA IFNy or medium (-

Dulbecco's Modified Eagle's Medium plus 10% FCS, 2 mM of L-glutamine, and antibiotic supplements) only. In some experiments, 1 xg/ml of εRMP was treated either with 1 mM PMSF/DMSO or with DMSO for 15 min at

37°C and co-cultured with spleen cells at a final concentration of 1 pg/ml.

The cultures were incubated in a 10% CO, atmosphere for 16 to 20 hr, after which the non-adherent cells were harvested, transferred to test tubes and analyzed by the rosette assay for frequencies of CD23-positive cells.

Most of the adherent cells were removed when cells were transferred from culture plate to test tubes. Cells cultured in the absence of IgE revealed background expression of CD23. To prepare subpopulation-depleted spleen cells, cell suspensions were incubated with pre-determined optimal concentrations of various Mabs (Marcelletti, J.F., Katz, D.H., Cell. Immunol. 1989. 120:314) for 30 min at 4°C. washed twice, then incubated with pre-absorbed C solution. After incubation for 45 min at 37 "C cells were washed twice and resuspended in culture medium.

Rosette Assay for the Detection of CD23⁺ Cells. After coculture with IgE in the culture plate, non-adherent cells were transferred to test tubes and washed twice with 2% FCS/Minimum Essential Medium (MEM) and incubated with 10 μ/ml of anti-DNP IgE for 60 min in the presence of 0.02% azide at 4°G This procedure saturates cell-surface CD23 with anti- DNP IgE and little de novo induction of CD23 occurs (Richards, M.L., Marcelletti, J.F. and Katz, D.H. (1988) J.Exp.Med. 168,571-580). After washing, 50 microliters of lymphocytes (2.5 X 10⁶ cells), prepared as described, were added to 100 μl of TNP-coated sheep red blood cells (Chen, S.-S., Bohn, J.W., Liu, F.-T. and Katz, D.H. (1981) /. Immunol. 127,166-174) in 12 x 75 mm Falcon tubes. The cell mixtures were centrifuged at 50 x g for 7 mins. After unperturbed incubation for 2 hours at 4°C the cell pellets were re-suspended gently and trypan blue was added to each tube. A rosette-forming cell is defined as a small lymphocyte that is tightly surrounded by at least 3 indicator cells (most rosettes have 5 or more indicator cells). The specificity of rosette-formation has been confirmed previously (Marcelletti, J.F. and Katz, D.H. (1986) /. Immunol. 137,2599-2606. ;Chen, S.-S., Bohn, J.W., Liu, F.-T. and Katz, D.H. (1981)

/. Immunol. 127,166-174). All assays were performed in triplicate or duplicate. In cases where the data are expressed as % rosette suppression, the following formula was used to calculate the percentage:

%CD23⁺ Cells in Positive Control — %CD23⁺ Cells in Test Culture % = x lOO

%CD23⁺ Cells in Positive Control — CD23⁺ Cells in Background Test Culture = those IgE-stimulated cultures containing modulating proteins; Positive Control = those IgE-stimulated cultures containing only IgE as inductive stimulus.

Radioimmunoassay Using Radioiodinated IgE and Anti-CD23. Radioiodinated probes were also used to detect CD23 expressed on lymphocytes. After overnight culture of spleen cells in the presence of IgE as described in the rosette assay system, cells were incubated with either ¹²⁵I-IgE or ¹²⁵I-anti-CD23 monoclonal antibody in 0.1% azide-containing PBS at 4°C for 2 hr (Richards, M.L., Marcelletti, J.F. and Katz, D.H. (1988) J.Exp.Med. 168,571-580). Cell suspensions were then overlayed onto 15% mineral oil/n-butyl phthalate in microcentrifuge tubes, centrifuged, and the cell pellets were counted on a gamma counter.

Cell-Surface Radioiodination and Immunoprecipitation. MBI-1.15 T cells were externally labeled with ¹²⁵I using lactoperoxidase (Phillips, D.R., & Morrison, M. 1971. Exposed protein on the intact human erythrocyte. Biochemistry 10:1766) and solubilized in Tris buffer containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and 0.5% Nonidet-P40 (NP40). Ascites-containing 2E5B Mabs and control ascites prepared from parental B cell hybridoma line, FO, were semi-purified by 50%-saturated ammonium sulfate precipitation and coupled onto cyanogen bromide-activated Sepharose 4B (Pharmacia). In immunoprecipitation, cell lysates were pre- cleared extensively with control ascites-bound Sepharose and rocked with 2E5B-Sepharose overnight at 4°C Sepharose was then washed successively with 1) washing buffer (0.5% NP40, 5 mM KI, 0.02% azide, 1 mg ml of ovalburain in PBS), 2) 0.5 M NaCl-containing washing buffer, 3) 0.1% SDS- containing washing buffer, and 4) washing buffer (each wash was done in new microcentrifuge tubes to reduce background), and eluted with SDS- PAGE sample buffer at 90°C for 3 min. The eluate was analyzed on 13.5% homogeneous SDS-PAGE in non-reducing or reducing conditions followed by autoradiography. In Vitro Murine IgE Antibody Responses. An in vitro murine IgE antibody production system was used (Marcelletti, J. F. and Katz, D. H., Elicitation of antigen-induced and secondary murine IgE antibody responses in vitro, CelLImmunol. 135:471, 1991) in which 5 x 10⁵/ml whole spleen cells, 20 μg/ml LPS (Difco Inc., Detroit, MI), and 30-3000 U/ml IL-4 (generously provided by Dr. J. Ohara, University of Colorado, Denver, CO) were cocultured in 96-well flat-bottomed plates in the presence of various effectors for 7 days at 37°C in 10% C0₂/humidified atmosphere. Quantities of IgE, IgGl, IgG2a, IgG2b, IgG3, IgM and IgA proteins in supernatant fluids were determined by ELISA using biotinylated anti-Ig antibody for IgE or alkaline phosphatase-labeled anti-Ig antibodies for other isotypes (Fisher Scientific Co., Los Angeles, CA). Anti-mouse IgE antibody was generated using monoclonal anti-DNP IgE as immunogen and purified using monoclonal anti-ragweed antigen-E IgE for adsorption. DNP-specific IgM, IgGl, IgG2b and IgE purified from hybridoma supernatant fluids as described (Liu, F.-T., Bohn, J.W., Ferry, E.L. Yamamoto, H., Molinaro, CA., Sherman, L.A., Klinman, N.R. and Katz, D.H. (1980) /. Immunol. 124,2728-2736), and affinity-purified mouse myeloma IgG2a, IgG3, and IgA purchased from ICN Immunobiologicals (Lisle, IL) were used as standards for ELISA.

B cells were prepared by treating spleen cells from CAF1 mice (8-12 week-old female, Jackson Laboratories), with anti-Thy 1.2 Mab (Lake, P., Clark, E., Khorshidi, M., & Sunshine, G. 1979. Production and character- iztion of cytotoxic Thy-1 antibody-secreting hybrid cell lines. Detection of T cell subsets. Eur. J. Immunol 9:875) and guinea pig complement followed by passage through Sephadex G-10 column (Pharmacia). Preliminary dose titration experiments showed that MBI-1.15 cell membranes exhibited maximum IgE-inducing activity at 1 - 3 x 10⁷ cell equivalent/ml. In setting up the cultures, mixtures of all components except viable cells were incubated for 30 min at 37°G to allow 2E5B Mab to react with εRMP first, then B cells were added. Quantities of IgE in supernatant fluids were determined by ELISA as described previously (Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and IgE-suppressive proteins. Proc. Natl Acad, ScL USA. 88:4718). Batch Purification. SFA was purified from a single 3-liter pool of

FCS-free MBI-1.15 supernatant fluid. The supernatant fluid, acidified with TFA to pH 4.5, was gently agitated with 20 g of pre-wetted methyl-silica (BONDESIL-C1; Analytichem International, Harbor City, CA) overnight at 4°C The silica was then collected by gravity sedimentation, washed extensively with 0.1% TFA, eluted with 500 ml of 30% ACN/0.1% TFA (shown in pilot studies to be optimal for this elution). The eluate was lyophilized and then reconstituted with 2 ml of water.

Purification by HPLC. The reconstituted methyl-silica eluate was loaded onto a 4.5 x 250 mm C18 reversed-phase HPLC column (Vydac, Hesperia, CA) equilibrated with 0.1% TFA. The column was eluted at a flow rate of 1.0 ml/min at room temperature in a gradient of increasing ACN concentration in 0.1% TFA with the OD at 280 nm continuously monitored. Fractions with SFA activity were concentrated with Speed Vac (Savant, France), loaded onto a 7.5 x 600 mm molecular sieving HPLC column (Bio-Sil TSK-250, Bio-Rad) equilibrated with PBS, and then eluted in PBS at 1.0 ml/min.

SDS-PAGE of the Purified Proteins and Elution from the Gel. The HPLC fractions were vacuum dried with Speed Vac and analyzed by gel electrophoresis in 13.5% homogeneous polyacrylamide with sodium dodecyl sulfate (SDS-PAGE) with or without 2-mercaptoefhanol. Samples were treated with SDS for 5 min at 50^UC right before analysis. The electrode buffer was 25 mM Tris, 21 mM boric acid, 0.1% SDS/water. To elute proteins from the gel, gel fragments were incubated at 37°C with 0.5 ml of 0.1% SDS/PBS for 2 hours. The eluted proteins in the fluid were added to culture at the final dilution of 1:10⁶ to 1:10⁸. Preliminary studies document- ed that residual SDS had no effect in these biological assays at dilutions greater than 1:10⁵.

Quantitation of serine protease activity Ten microliters of purified εRMP, -.-chymotrypsin or trypsin (Sigma, St. Louis, MO), at designated concentrations, were incubated with 90 μl of 0.2mM BLT (Calbiochem, La Jolla, CA), 0.22 mM 5,5-dithiobis-2-nitrobenzoic acid (Calbiochem, La Jolla, CA) at pH 8.0 in 0.1 M Tris-Hcl buffer for 60 min at room temperature (RT) in 96-well microtiter plates, after which OD₄₀₅ was determined (Pasternack, M.S. and Eisen, H.N., Nature 1985. 314:743). OD₄₀₅ of εRMP solution was the same as the background level.

Protein sequencing The εRMP purified by reversed-phase high performance liquid chromatography (RP-HPLC; Matsushita, S., Marcelletti, J.F., Katz, L.R. and Katz D.H., Proc. Natl Acad. Sci. USA, 88:4718, 1991) was evaporated by vacuum centrifugation and reconstituted with a small amount of water. After treatment with SDS for 5 min at 50 °C the sample was loaded onto a mini-gel of 13.5% homogeneous polyacrylamide as a separating gel and 5% homogeneous polyacrylamide as a stacking gel. Electrophoresis was performed using ininigel apparatus, at 100 V constant voltage with an electrode buffer of 25 mM Tris, 21 mM Boric acid, 0.1% SDS. To minimize artificial N-terminal blocking, sodium thioglycolic acid was added to the electrode buffer at a final concentration of 0.1 mM, and the separating gel was made 1-2 days before analysis. After electrophoresis, the protein was electrophoretically transferred onto PVDF membrane, stained with Coomassie Brilliant Blue R-250, de-stained, and rinsed with water. εRMP was cut out with a clean razor and sequenced on an Applied Biosystems model 470 sequenator equipped with on-line PTH analysis using the regular program 03RPTH, in the Protein Structure Laboratory of the University of California, Davis (Davis, CA). The PTH-derivatives were separated by RP-HPLC over a Brownlee C-18 column (Matsudaira, P., /. Biol Chem. 1987. 262:10035; Yuen, S.W., Chui, A.H., Wilson, K.J. and Yuan, P.M., BioTechniques 1989. 7:74). For internal amino acid sequencing, the combination of in situ cyanogen bromide cleavage and o-phthalaldehyde (OPA) was used as described previously elsewhere (Hewick, R.M., Hunkapillar, M.W., Hood, L.E. and Dreyer, W.J., /. Biol Chem. 1981; 256:7990; Gross, J., Brauer, A., Bringhurst, R., Borbett, C and Margolies, M.N., Proc. Nail Acad. ScL USA 1985. 82-5627).

Preparation of Plasma Membranes. Sterile plasma membrane preparations were obtained by the SEAT buffer method (Shimizu, Y., & Shimizu, N. 1986. Rat hepatoma cell variants resistant to insulin-diphtheria toxin A fragment conjugates. /. Biol Chem. 261:7342) with some modifica- tions. Briefly, cells were washed three time with saline, once with 0.25 M sucrose, and resuspended in 1 ml of 0.25 M sucrose, 10 mM acetic acid, 10 mM triethanolamine, pH 7.4. After two freeze/thaw cycles, cells were pulverized in a tissue grinder on ice for 3 min with rotating motion, and centrifuged at 5,000 rpm for 15 min in a Sorvall SS-34 rotor. The post- nuclear supernatant thus obtained was then centrifuged at 16,000 rpm for 60 min, washed twice in saline, and resuspended in serum-free culture medium at 4 x 10⁸ cell equivalent/ml using 27-G needles. Use of EDTA was avoided since preliminary experiments showed that the IgE-inducing activity of plasma membranes was significantly decreased in the presence of EDTA. All the reagents were handled sterilely to avoid contamination with bacterial LPS.

Western Blot Analysis. The IgM fraction of 2E5B Mab was purified from 50%-saturated ammonium sulfate precipitates of ascites, by molecular sieving HPLC (Bio-Sil TSK-250, Bio-Rad; 14), and radioiodinated. MBI- 1.15 cell lysates were separated on 13.5% homogeneous SDS-PAGE in non- reducing or reducing conditions and transferred onto nitrocellulose filters using Tris-glycine buffer containing 10% methanol. After blocking the filter with PBS containing 0.5% T een 20 and 1% bovine serum albumin for 1 hr, the filter was incubated in the presence of radioiodinated 2E5B Mab for 1 hr at room temperature. The filter was washed with Tween/PBS containing 1.5 M NaCl, followed by Tween-PBS extensively, and subjected to autoradiography.

Northern Blot Analysis for Germ-Line IgE Heavy Chain Transcripts. Total RNA was prepared by the guanidinium thiocyanate method (Chirgin, J.M., Pryzybyla, A.E., MacDonald, R.J., & Rutter, W.J. 1979. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18:5294) from 1-29 B cells (1 x 10⁵/ml) cultured for 3 days with MBI-1.15 cell membranes (1 x 10⁷ equivalent/ml) or IL-4 (1000 U/ml), in the presence or absence of 2E5B anti-εRMP Mab (1:20) or 11B11 anti-IL-4 Mab 50 μg/ml; (Ohara, J., & Paul, W.E. 1985. Production of a monoclonal antibody to and molecular characterization of B-cell stimulatory factor-1. Nature 315:333). Poly (A⁺) RNA was purified from these total RNA preparations using PolyATract MRNA isolation system (Promega), separated on 1% agarose gels containing formaldehyde, and transferred onto nitrocellulose filter. The 0.5-kb Pst I-Hinc II genomic DNA fragment which includes a sequence specific to germ-line IgE heavy chain transcripts (Iε) but does not contain Cε sequence (generously provided by Dr. Fredrick Alt, College of Physicians and Surgeons of Columbia University; (Rothman, P., Chen Y-Y, Lutzker, S., Li, S.C, Stewart, V., Coffman, R., & Alt, F.W. 1990. Structure and expression of germ line immunoglobulin heavy-chain ε transcripts: Interleukin-4 plus lipopolysaccharide-directed switching to Cε. Mol Cell. Biol. 10:1672.), was used as a probe. The filter was hybridized overnight at 42°C with the nick-translated probe in dextran sulfate, formamide, salmon sperm DNA and Denhardt's solution-containing 4xSSC- Tris buffer, washed in 0.1%SDS/0.2xSSC at 42°C, and exposed for autoradiography. RESULTS

T Cell Hybridoma MBI-1.15 Produces CD23-Modulatin Activity as Well as IgE Suppressive Activity. Culture supernatant fluid of T cell hybridoma MBI-1.15 suppresses functional expression of CD23 detected by rosette assay (CD23-modulating activity). TABLE I SUPERNATANT OF T CELL HYBRIDOMA MBI-1.15 SUPPRESSES FUNCTIONAL EXPRESSION OF CD23 Culture I E⁵" MBI-1.15 Su ernatant- I E % RFC- = % Su ression

* Cultured at a final concentration of 20 μg/ml.

-f- 5 x 10* MBI-1.15 cells/0.5 ml of 10% FCS/DMEM were cultured for 48 hrs and supernatant collected was tested in rosette assay system at designated final dilutions.

4= Expressed as the geometric mean ± standard deviation of triplicate assays. P values were calculated against group 2. Statistically significant differences existed between group 2 and groups 3 and 4 (p<0.001 in both cases). A shown in Table I, the percentage of CD23⁺ cells in murine whole spleen cells increased from 16% to 36% after overnight coculture with IgE (p<0.001) (Lee, W.T., Rao, M. and conrad, D.H. (1987) /. Immunol. 139,1191-1196; Conrad, D.H. (1990) Ann.Rev.Immunol 8,623-630). This induction was inhibited in a dose-related manner by supernatant of MBI-1.15 cells added to the culture. The supernatant did not directly affect background expression of CD23. The CD23-modulating activity was observed with fluids FCS- containing the FCS-free cultures of MBI-1.15 cells. The supernatant was also tested for suppressive activity against IL-4- induced IgE synthesis by LPS-activated murine spleen cells as shown in Table II.

TABLE II

SUPERNATANT OF T CELL HYBRIDOMA MBI-1.15

SUPPRESSES IL-4-INDUCED IgE SYNTHESIS BY

LPS-ACTIVATED MURINE SPLEEN CELLS*

Whole spleen cells of BALB/c mice were cultured in triplicate for 7 days in the presence or absence of LPS (20 μg ml), IL-4 (1000 U/ml), and supernatant of T cell hybridoma MBI-1.15 at designated final dilutions. FCS-free MBI-1.15 supernatant and mock solution were concentrated 100-fold with saturated ammonium sulfate precipitation and dialyzed extensively against DMEM after addition of FCS. The designated final dilutions correspond to original supernatant, i.e., 1:200 of 100-fold supernatant is expressed as 1:2.

+ Total IgE concentrations in the culture supernatants at the end of the culture were determined by ELISA and expressed as the geometric mean ± standard deviation. _P values were calculated against group 2, and differences were found to be statistically significant in the cases of groups 3 (-<0.002), 4 (p<0.005) and 5 (p<0.02).

Murine spleen cells incubated with either LPS only or IL-4 only did not synthesize detectable IgE, and addition of MBI-1.15 supernatant did not enhance IgE synthesis. On the other hand, spleen cells incubated with LPS with IL-4 displayed significant IgE synthesis (culture 2; 140 ng/ml). Addition of MBI-1.15 supernatant, in dilutions of 1:2 to 1:8, to this culture system significantly suppressed IgE synthesis, whereas mock supernatant fluid did not. These data thus demonstrate that T cell hybridoma MBI-1.15 produces suppressive activity against both functional expression of CD23 molecules and LPS+IL-4-stimulated IgE synthesis.

Purification of Suppressive Activity by Reversed-Phase HPLC. The FCS-free MBI-1.15 culture fluid was batch-purified using methyl-silica and the 30% ACN/0.1% TFA eluate was tested and confirmed to contain both CD23-modulating and IgE-suppressive activities. The active eluate from methyl-silica was applied to C18 reversed-phase HPLC column equilibrated with 0.1% TFA and eluted in increasing gradient of ACN: 0.1% TFA. Two milliliters of concentrated methyl-silica eluate were applied to a C18 RP- HPLC column and eluted with a linear gradient of 0 to 54% acetonitrile over 80 min at 1.0 ml/min. Samples were assayed for activity suppressing functional expression of CD23 (rosette assay) and IL-4-induced IgE synthesis by LPS-activated spleen cells. Results in Figure 1 are presented in suppression units per milliliter with 1 unit per milliliter defined as the lowest concentration causing at least 50% suppression in each assay. As shown in Figure 1, the CD23-modulating activity (expressed as rosette suppression units/ml) co-eluted .with a major peak of OD^ at 37-min retention. Most of the OD reading of this high peak is due to non-protein substance in the media, because the fraction had veiy little protein (0.5 μg) on SDS-PAGE and plain DMEM treated in the same way showed the same peak on HPLC Unexpectedly, no IgE-suppressive activity was detected in this fraction, but was observed in a later fraction (at 68-min retention), which was conversely devoid of any CD23-modulating activity. Both of these activities of the HPLC eluates were abolished by treatment with trypsin. Taken collectively, these data indicate that the two distinct activities ascribed to SFA are mediated by two distinct proteins.

SDS-PAGE Profile of the Purified Protein with CD23-Modulating Activity. The HPLC fraction with CD23-Modulating activity was analyzed on SDS-PAGE. The sample applied to one lane corresponds to one liter of culture supernatant starting material. A fraction with the highest CD23- modulating activity from the RP-HPLC chromatography was concentrated, divided into two aliquots and electrophoresed on a 13.5% SDS-PAGE using two lanes in non-reducing conditions. One of the two lanes was stained by Coomassie Brilliant Blue. Gel slices from an unstained section were then incubated in 0.1% SDS-PBS at 37°C for 2 hr, tested for CD23-modulating activity by rosette assay at a final dilution of 1:10⁸, and reported as mean values of duplicate cultures. As shown in Figure 2, after Coomassie Brilliant Blue staining the gel revealed a major band at 17 Kda, and this mw of the protein was retained even under reducing conditions. A parallel unstained lane run in non-reducing conditions was sliced horizontally into 16 pieces based on mw eluted with SDS/PBS, and tested in the rosette assay. Results summarized in Figure 2 show that the activity was associated with a single band with mw of 17 Kda. The eluate from a parallel blank lane had no suppressive activity at this dilution. The dense band at the dye front is suggestive of small peptide fragments, but no activity was eluted from this area.

Further Purification and SDS-PAGE Profile of the Purified Protein with IgE-Suppressive Activity. Since preliminary studies revealed that IgE- suppressive activity was substantially lost during either SDS/PBS elution of electroelution, the C18 HPLC fraction with IgE-suppressive activity was analyzed on a molecular sieving HPLC column. Fractions of IgE-suppressive activity on RP-HPLC was further purified with molecular sieving HPLC and tested regarding IgE-suppressive activity. Active fraction was electrophoresed on a 13.5% SDS-PAGE without (lane 1) or with (lane 2) 2-mercaptoethanol. As shown in Figure 3, fractions were tested for IgE-suppressive activity and the mw was estimated from the elution pattern of standard proteins (shown at the top of the figure). The IgE-suppressive activity was eluted at approximately 17-min retention time, corresponding to 32-39 Kda. SDS- PAGE followed by silver staining of this fraction revealed a single bank at 30 kDa in non-reducing (lane 1) and 33 kDa in reducing (lane 2) conditions, respectively. Although the same amount of protein was used for lanes 1 and 2, reduced protein was much more sensitive than non-reduced protein to silver staining. The faint band observed at 15 kDa was a carryover from the fractions around 20 min.

TABLE HI

SUMMARY OF PURIFICATION OF CD23-MODULATING

AND IgE-SUPPRESSIVE ACTIVITIES

Total Total Protein Activity-}- Purification

Activity Step (μg) -00 — Factor

CD23- Crude Supernatant 3.3 x 10⁴ 3 x l0⁸ 1

Modulating Methyl-silica 7.5 x 10³ 1.5 x 10⁸ 4.6

C-18 HPLC fr 0.5 2 x 10^u 4.4 x 10⁷

IgE- Crude Supernatant 3.3 x 10* 1.2 x 10⁴ 1 Suppressive Methyl-silica 4.5 x 10³ 2.0 x 10³ 1.2

C-18 HPLC fr 10 2.0 x 10³ 5 x 10²

Gel filtration fr 1.0 x 10³ 1.4 x 10³ *

Protein concentration was determined by the Bradford method (crude supernatant and methyl-silica eluate), and by staining of the gels as shown in Figures 2 and 3, with chymotrypsinogen A as a standard

(HPLC fraction and gel filtration fraction). -j- Rosette suppression unit and IgE suppression unit were used for each activity.

Table III summarizes titration of the active fractions of CD23- modulating and IgE-suppressive activities, respectively, at each step. One unit per ml was defined as the lowest concentration that shows more than 50% suppression in each assay. The total yields of CD23-modulating activity increased after HPLC; Le., approximately one thousand times higher than the starting material. As control for the possible addition of non-specific activity by HPLC purification, plain medium was treated and separated on the HPLC column in the same way; the 37-min retention fraction of such control material did not show any CD23-modulating activity.

Purified CD23-Modulating Activity Does Not Change CD23 Expression but Decreases Binding of IgE to CD23-Bearing B Cells. As shown in Figure 4, whole spleen cells were incubated in the presence of IgE (20 μg/ml), with medium (group I), with 100 U/ml of purified CD23- modulating activity (group II), and with 100 U/ml of IFNy (group III). After overnight incubation, aliquots of cells were used for functional expression of CD23 (rosette assay, hatched column), binding of ¹²⁵I-IgE on the cell surface (closed column), and RIA for cell-surface CD23 (shaded column). The IgE- induced control values against which % suppression was derived in the medium-only control cultures were as follows: 1) 19.0% IgE rosettes; 2) 23,000 cpm binding of ¹²⁵I-IgE; and 3) 72,000 cpm binding of ¹²⁵I-anti-CD23. Although purified 17 kDa protein showed CD23-modulating activity measured by the rosette assay (hatched column) and binding of ¹²⁵l-IgE (closed column), it did not inhibit cell surface CD23 detected by radioiodinated anti-CD23 monoclonal antibody B3B4 (shaded column). On the other hand, IFNy inhibited CD23 expression as measured by all 3 of the assays. These data indicate that purified 17 kDa protein decreases avidity of binding between IgE and CD23-bearing B cells without suppressing the amount of CD23 expressed. ["Avidity" refers to the sum of cooperative binding by a bivalent receptor (Le., CD23) to its ligand]. In other words, the 17 kDa protein does not suppress quantitative expression of CD23, but suppresses functional expression of CD23. We therefore designate this protein "ε Receptor Modulating Protein (εRMP)."

Purified IgE-Suppressive Activity (SFA) Suppresses IgE and IgGl but the Effect Is Not Reversed by Anti-IFNy. The purified IgE-suppressive activity (SFA) was tested for its isotype selectivity.

TABLE IV

SFA SUPPRESSES IgE AND IgGl BUT NOT IgM* Immunoglobulins (ng ml)

Stimulation IαE IαG1 IαG2a IαG2b IαG3 IdA Iα

LPS+IL-4 92*20 20000*4500 30*12 34±8 650*210 <206500*1100

LPS+IL-4

+SFA 20*6 800*340 24*9 28*13 380*140 <20 5600*1300

* Whole spleen cells were cocultured with LPS (20 μg/ml), IL-

4 (1000 U/ml), with or without SFA (10 U/ml) for 7 days. Immunoglobulins of all the isotypes were quantitated using ELISA as described in Material and Methods. Results are expressed as the geometric mean ± standard deviation of triplicate cultures. P values were calculated between culture without SFA and culture with SFA. Statistically significant differences existed for IgE and IgGl (p< 0.002 in both cases). As shown in Table IV, IgE, IgGl, IgG3, and IgM were highly synthesized in this culture system. IgG2a and IgG2b were little induced and IgA was undetectable. Purified SFA added in the culture suppressed IgE and IgGl but did not significantly suppress IgM synthesis. IgG3 was partially suppressed. Other poorly induced isotypes were not affected by the addition of SFA. In contrast to a property of IFNy, enhancement of IgG2a was not observed even with lower doses of SFA.

To confirm that SFA is distinct from IFNy, the effect of monoclonal anti-IFNy antibody was tested as shown in Figure 5. The suppressive effect of IFNy on IL-4-induced IgE synthesis by LPS-activated lymphocytes was reversed by anti-IFNy (group II). On the other hand, SFA suppressive activity was not at all affected by anti-INFy (group III). Anti-IFNy itself directly enhanced IgE synthesis slightly (0.05</?<0.1, group I). This might reflect internally produced IFNy during the culture period. In addition, 2000x concentrated MBI-1.15 supernatants and 10 μg/ml of purified SFA were measured by ELISA for their IFNy content. Both preparations showed less than 0.1 ng/ml IFNy concentration. εRMP is a serine protease. The initial purification of εRMP, which diminishes the avidity of IgE binding to CD23 detected by the rosette assay and ¹²⁵I-lgE binding assay (Matsushita, S., Marcelletti, J.F., Katz L.R., and Katz, D.H., Proc. Natl Acad. Sci. USA, 88,4718, 1991) was performed in the presence of a serine protease inhibitor (PMSF) to avoid proteolytic breakdown of εRMP in the supernatants. However, the CD23-modulating activity purified from PMSF-treated supernatants was always lower than should be expected based on the original specific activity observed prior to addition of PMSF. Thus, the purification was repeated in the absence of PMSF and the effect of PMSF on the purified εRMP was tested. As shown in Figure 6, CD23-modulating activity of purified εRMP was abrogated by PMSF treatment, but not by mock treatment. The absence of any activity with PMSF-treated control medium verified the lack of any direct effect of PMSF on the assay system. Purified εRMP was then tested for direct serine protease activity using BLT and N-succinyl-Ala-Ala-Pro-Phe /7-nitroanilide as artificial substrates. Purified εRMP directly decomposed both substrates as -.-chymotrypsin did, exhibiting approximately the same level of the specific activity as α-chymotrypsin. Furthermore, N,α.-/?-tosyl-L-lysine chloromethyl ketone failed to inactivate the activity of εRMP, indicating εRMPs being chymotryptic rather than being tryptic.

TABLE V

Purified εRMP Decomposes BLT and iV-Succinyl-ALA-ALA-PRO-PHE /7-Nitroanilide

OD '4.„05 with

Enzyme BLT N-Succinyl-Ala-Ala-Pro-Phe- p-nitroanilide

0.126

0.127 0.133 0.129 0.131

0.582 0.575 0.277 0.186

0.568 0.552

0.439 Trypsin, α-chymotrypsin, or purified εRMP were incubated with 0.2 mM BLT or N-Succinyl-ALA-ALA-PRO-PHE p-Nitroanilide, in the presence 0.22 mM 5,5-dithiobis-2-nitrobenzoic acid at pH 8.0 in 0.1 M tris-HCl buffer. After a 60 min incubation at RT, OD₄₀₅ was determined. Results are expressed as the geometric means of duplicate determinations.

These observations raise the possibility that εRMP exerts direct proteolytic action on B cells to decrease the binding avidity of CD23 for IgE, without affecting an epitope recognized by anti-CD23 mAb (Matsushita, S., Marcelletti, J.F., Katz, L.R. and Katz D.H., Proc. Natl. Acad. Sci. U SA , 88:4718, 1991). However, this appears very unlikely as shown by results obtained when the effects of εRMP on T cell-depleted populations were tested. Whole spleen cells treated with niAbs and/or C were co-cultured with εRMP and IgE to determine the CD23-modulating activity of εRMP. Results of the rosette assays depicted in Figure 7 are expressed as the mean of % suppression of duplicate cultures in Exp. 1, and as the geometric mean of % suppression ± standard deviation of triplicate cultures in Exp. 2. Thus as shown in Figure 7, εRMP unexpectedly failed to modulate functional expression of CD23 on T cell-depleted populations (Exp. 1); more specifical¬ ly, it appears that the activity of εRMP requires participation of CD4⁺, but not CD8⁺, T cells (Exp. 2). These observations clearly demonstrate that εRMP does not work directly on CD23 of B cells since it requires CD4⁺ T cells to decrease the receptor (substrate)-binding avidity of CD23.

Some, but not all other, serine proteases have εRMP-like CD23-

^' modulating activities We next questioned whether other known serine proteases have εRMP-like activity. Some of the commercially available serine proteases (trypsin, α-chymotrypsin, kallikrein, elastase, and urokinase) were compared to εRMP with regard to CD23-modulating activity by co-culture with whole spleen cells at final concentrations of 1 ng to 1 fg/ml. TABLE VI SOME SERINE PROTEASES SHOW εRMP-LIKE ACTIVITYa)

%Suppression at a final concentration of:

Serine proteases added to culture 1 ng/ml 1 pg/ml 1 fg/ml εRMP 62 ± 13 82 ± 18 43 ± 15 α-chymotrypsin 69 ± 8 59 ± 8 12 ± 12 kallikrein 48 ± 5 54 ± 11 13 ± 17 elastase 38 ± 12 63 ± 4 7 ± 2

trypsin 21 ± 13 -1 ± 4 6 ± 9

Urokinase 0 ± 6 26 ± 13 27 ± 8

*^} After overnight incubation of spleen cells with or without IgE in the presence of various serine proteases at designated concentrations, functional expression of CD23 was determined by the rosette assay.

Background expression of CD23 (Culture without IgE) was 17.2 ± 3.2%, and enhanced expression CD23 (culture with only IgE) was 39.6

± 1.5%. The % suppression was calculated as explained in Materials and Methods. Results are expressed as the geometric mean of %

suppression ± standard deviation of triplicate cultures.

As shown in Table VI, α-chymotrypsin, kallikrein, and elastase exhibited εRMP-like activity, whereas trypsin and urokinase did not. Although εRMP seems to have the highest specific activity among the serine proteases tested, this finding indicates that serine protease-receptor

(substrate) specificity is not very strict for revealing effects on functional expression of CD23. These serine proteases, when tested without IgE, did not affect background expression of CD23. εRMP has a unique internal amino acid sequence The question then

arises of whether εRMP is a novel serine protease or just a known protein expressed in a specific T cell population at a particular activation stage. To

answer this question we determined a partial amino acid sequence of εRMP.

When the initial experiments showed that the purified εRMP was completely

blocked in its N-terminus, we next tried to obtain an internal amino acid

sequence by the combination of in situ cyanogen bromide cleavage and OPA. Briefly, plural sequences should appear starting from internal methionine residues after in situ cyanogen bromide cleavage of the protein on the PVDF membrane. However, once a cycle number where proline residue appears is identified in the first run, OPA is added at that cycle in the second run to

block all the N-termini except proline. That would enable us to have only one internal sequence starting from proline, and, although the location of the sequence in the whole peptide cannot be mapped (Hewick, R.M., Hunkapillar, M.W., Hood, L.E. and Dreyer, W.J., /. Biol. Chem. 1981

256:7990; Gross, J., Brauer, A., Bringhurst, R., Borbett, C and Margolies, M.N., Proc. Natl Acad. Sci. USA 1985. 82:5627), this would give us enough information to make comparison with known proteins and to generate

oligonucleotide probes. We obtained a single pure sequence Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala 1 5 10 15

Lys Lys (SEQ ID NO:2)

(Al^APKKEEKKKKAAAKK), indicating that εRMP has only one internal methionine residue. This sequence is distinct from all of the known sequences on record including serine proteases such as granzymes. The most homologous protein in amino acid levels identified by using program TfastA of the Wisconsin package (Devereux, J., Haeberli, P. and Smithies, O.,

Nucleic Acids Research 1984. 12:387) was human HMG-14, one of the non- histone DNA-building proteins, which showed 70.6% identity in 17 amino- acid overlap out of 18 residues (Landman, D., McBride, O.W., Soares, N.,

Crippa, M.P., Srikantha, T., and Bustin, M., /. Biol Chem. 1989. 264:3421).-

The dissociation of CD23-modulating activity form IgE-suppressive activity of crude SFA and the purification of each protein molecule from serum-free culture supernatants of the T cell hybridoma line MBI-1.15 was an unexpected result of efforts to provide isolated, purified SFA. It should be noted that this hybridoma was shown to not produce any of the other regulatory proteins involved in the IgE system. Results of the purification of crude SFA show that 1) purified CD23-modulating activity decreases avidity between IgE and CD23-bearing B cells, and is thus designated ε receptor modulating protein (εRMP; mw 17 kDa); and 2) purified IgE-suppressive activity (SFA; mw 30 kDa) suppresses IgE and IgGl responses but is distinct from IFNy. It should be noted that the CD23-ιnodulating activity is unlikely to reflect non-specific effects of εRMP since: 1) many other T cell hybridomas generated in the fusion population from which εRMP-secreting

MBI-1.15 was selected were devoid of such activity, and 2) data shown in Figure 7 indicate that εRMP activity requires the co-participation of CD4⁺

T cells. These proteins were purified from 3 liters of serum-free MB 1-1.15 culture fluid by using both rosette assay for functional expression of CD23 and LPS-rTL-4-stimulated IgE production system in vitro. Approximately 0.5 μg of εRMP and 2.0 μg of SFA were recovered. Total εRMP activity increased after RP-HPLC, suggesting that MBI-1.15 culture supernatants contain inhibitor(s) of εRMP activity or counteracting activity that might be co-produced with εRMP from MBI-1.15 cells and removed by purification. On the other hand, total yield of SFA activity decreased after purification procedures. A possible explanation for this decreases is that TFA and ACN worked on SFA to change its conformation. Alternative purification schemes to retain its full activity need to be established.

The purity of εRMP was indicated by a single band detected on SDS- PAGE. The elution of εRMP activity from the polyacrylamide gel clearly confirmed the activity only in the 17 kDa area where the major band migrates, and also ruled out any non-specific effect of contaminating small molecules, because the dye front did not show any activity. The migration pattern did not change under reducing conditions, indicating that the protein is a single-chain peptide without any inter-chain or intra-chain disulfide bonds. Interestingly, εRMP did not suppress the amount of CD23 molecules expressed on the cell surface, but functionally suppressed CD23 by decreasing avidity of binding to IgE. With regard to avidity of Fc receptors to their ligands, a similar phenomenon has been reported with IgG Fc receptors. Stimulation of B cells with 1L-4 resulted in less binding of IgG to IgG Fc receptors, but probing with anti-receptor antibodies revealed the same number of total receptors on the cell surface (Laszlo, G. and Dickler, H.B. (1988) /. Immunol. 141,3416-3422). Alteration of receptor-ligand binding without changing the number of receptors has also been reported in IgA Fc receptors (Sandor, M., Waldschmidt, T.J., Williams, K.R. and Lynch, R.G. (1990) /. Immunol. 144,4562-4569). The molecular mechanisms and physiological significance of the activity of εRMP, and possible effects on other species of Fc receptors are now being studied. In addition to modulation of CD23 molecules themselves, We are examining processes such as receptor-receptor association (Rao, M., Van Dusen, R. and Conrad, D.H. (1989) 7th Int. Congress Immunol. 19 (Abstract)), and spatially associated molecules such as immunoglobulins (Lee, W.T. and Conrad, D.H. (1985) /. Immunol. 134,518-524) and MHC (Rao, M., Van Dusen, R. and Conrad, D.H. (1989) 7th Int. Congress Immunol. 19 (Abstract); Lee, W.T. and Conrad, D.H. (1985) /. Immunol. 134,518-524; Bonnefoy, J.-Y., Guillot, O., Spits, H., Blanchard, D. and Ishizaka, K. (1988) J.Exp. Med. 167,57-63). The antibody synthesis-suppressing protein, SFA was purified by reversed-phase HPLC followed by molecular sieving HPLC. Studies with the purified SFA protein yielded two new pieces of information. First, in terms of structure, the mw of reduced SFA estimated by SDS-PAGE was 3 kDa higher (33 kDa) than that of non-reduced SFA (30 kDa). This might indicate the presence of intra-chain disulfide bonds. Second, it is clear from the data in Table IV, that purified SFA suppresses IgGl in addition to IgE synthesis. In previous work we have analyzed total IgG synthesis in vivo and not observed suppression by SFA (total IgG would not be appreciably changed by suppression of IgGl above). The present results, in which subtypes of IgG (e.g. Gl, G2a, G2b, G3) were examined, in an in vitro system, extended the activity of SFA to the IgGl subclass.

Previous studies with partially-purified SFA showed that it is produced by CD4⁺ T cells (Marcelletti, J.F. and Katz, D.H. (1984) /. Immunol. 133,2837-2845; Marcelletti, J.F. and Katz, D.H. (1989) CelLImmunol 12,314- 320 ;Marcelletti, J.F. and Katz, D.H. (1984)/. Immunol. 133,2821-2830.). In this respect, the TH1 cell population (Mosmann, T.R. and Coffman, R.L. (1987) Immunology Today 8,223-226) appears similar to the SFA-producing lymphocyte population, in that both 1) are CD4-positive, 2) suppress IgE and IgGl synthesis, and 3) are characterized as irradiation-sensitive suppressor T cells for IgE immune responses (Tung, A.S., Chiorazzi, N. and Katz, D.H. (1978) /. Immunol. 120,2050-2057; Marcelletti, J.F. and Katz, D.H. (1989) CelLImmunol. 12,314-320). The TH1 population also produces IFNy which functions as a suppressive molecule for IgE synthesis (Marcelletti, J.F. and Katz, D.H. (1984)/. Immunol. 133,2821-2830; Mosmann, T.R. and Coffman, R.L. (1987) Immunology Today 8,223-226). However, the studies presented herein clearly distinguish SFA from IFNy because: 1) anti-IFNy monoclonal antibody does not block the activity of SFA, 2) the mw of SFA (30 kDa) is clearly different from that of IFNy, 3) SFA does not enhance IgG2a secretion, whereas IFNy does (Snapper, CM. and Paul, W.E. (1987) Science 236,944-947; Haveli, E.A. and Spitanly, G.L. (1983) Virology 129,508-513; Snapper, CM., Peschel, C and Paul, W.E. (1988) /. Immunol. 140,2121- 2127; Coffman, R.L., Seymour, B.W.P., Lebman, D.A., Hiraki, D.D., Christiansen, J.A., Shrader, B., Cherwinski, H.M., Savelkoul, H.F.J., Finkelman, F.D., Bond, M.W. and Mosmann, T.R. (1988) ImmunolRev. 102,5-15), and 4) purified SFA unlike IFNy, does not suppress expression of CD23.

The possible existence of a suppressive mechanism of IgE responses distinct from IFNy has been speculated by Finkelman, el al. (Finkelman, F.D., Katona, I.M., Mosmann, T.R. and Coffman, R.L. (1988) /. Immunol. 140,1022-1030.) who studied murine immune responses to Brucella abortus (BA). BA, which stimulates IFNy secretion, suppressed IgGl and IgE, while enhancing IgG2a, secretion. However, treatment of these mice with anti- IFNy did not reverse the suppressive effect of BA on IgE secretion. These observations, and ours, indicate the existence of suppressive mechanisms for IgE synthesis other than IFNy (Coffman, R.L., Seymour, B.W.P., Lebman, D.A., Hiraki, D.D., Christiansen, J.A., Shrader, B., Cherwinski, H.M., Savelkoul, H.F.J., Finkelman, F.D., Bond, M.W. and Mosmann, T.R. (1988) ImmunolRev. 102,5-15; Finkelman, F.D., Katona, I.M., Mosmann, T.R. and Coffman, R.L. (1988); /. Immunol. 140,1022-1030; Swain, S.L. and Dutton, R.W. (1987) ImmunolRev. 99,263-280). Approximately 20 ng/ml of IFNy (100 U/ml) and purified SFA (10 U/ml) were added to culture system of IL- 4-induced IgE synthesis by LPS-activated whole spleen cells with or without anti-IFNy monoclonal antibody. Results are depicted in Figure 5 with the geometric means and standard deviations of triplicate cultures. One such molecule, distinct from IFNy, is IFNα (Pene, J., Rousset, F., Briere, F., Chretien, I., Bonnefoy, J-Y., Spits, H., Tokota, T., Arai, R, Arai, K., Banchereau, J. and deVries, J.E. (1988) Pwc.NallAcad.Sci. USA 85,6880- 6885; Finkelman, F.D., Holmes, J.M., Katona, I., Urban, Jr., J.F., Beckman, M.P., Park, L.S., Schooley, K.A., Coffman, R.L., Mosmann, T.R. and Paul, W.E. (1990) Ann.Rev. Immunol. 8,303-318). IFNα is also a potent enchancer of IgG2a synthesis. The absence of such activity of SFA rules out the possible identity of SFA with IFNα. Taken collectively, SFA is a lymphokine involved in regulating IgE synthesis and which appears to be a novel molecule, although definitive proof of this must await sequence data which is expected in the near future.

With regard to εRMP, all the data clearly demonstrate that εRMP is a novel serine protease that controls functional expression of CD23 through

CD4⁺ T cells. The precise manner in which this occurs is still unknown.

However, the possible mechanisms of functional modulation of CD23 include:

1) εRMP stimulates CD4⁺ T cells to make them functionally modulate CD23 via a factor-mediated or a contact-dependent interaction with CD23 on B cells; 2) εRMP works as a "cofactor" of a second factor produced by CD4⁺

T cells during the culture period. In this regard, Ishizaka, el al have reported that kallikrein (one of the serine proteases)-like glycosylation enhancing factor (GEF) stimulates CD4⁺ T cells to produce IgE-potentiating factor (a rodent retrovirus-related IgE-binding factors; Moore, K.W., Jardieu, P., Mietz, J.A., Trounstine, M.L., Kuff, E.L., Ishizaka, K., and Martens, C.L., /.

Immunol. 1986. 136:4283.) which suppresses FcεRII-rosette formation

(Ishizaka, K., /. Immunol. 1985. 135:i). The characteristics of the molecule and the pathway of action resemble those of εRMP. It is clear, however, that

IFNy is not involved since the cell surface expression of CD23, which is markedly inhibited by IFNy, is not diminished by εRMP (Matsushita, S.,

Marcelletti, J.F., Katz, L.R. and Katz D.H., Proc. Natl Acad. ScL USA, 88:4718, 1991). Soluble 17-kDa εRMP Induces IgE Synthesis bv B Cells in the Presence of IL-4. The effects of εRMPs on in vitro IgE synthesis by murine B cells are shown in Table VII. Whereas neither IL-4 (group 3) nor εRMPs (group 4) alone had any effect, the two factors combined (without LPS) clearly induced IgE synthesis (groups 6 and 7 vs group 3, p = 0.002 and 0.03, respectively). The fact that εRMPs alone does not induce IgE synthesis (group 4) also rules out possible cross-reaction of εRMPs with anti-IgE antibody in the ELISA system. The IL-4 costimulating activity of εRMPs in inducing IgE synthesis was observed reproducibly in many experiments, although the magnitude of IgE production never exceeded 20 ng/ml. Unfortunately, 10 μg/ml of εRMPs was the highest concentration that could be tested, because of the limited amounts of purified εRMPs that can be obtained (three liters of culture supernatants of MBI-1.15 cells yield only 0.5 μg of εRMPs; (Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23- modulating and IgE-suppressive proteins. Proc. Natl. Acad. Sci. USA. 88:4718).

TABLE VII

SOLUBLE εRMP (εRMPs) INDUCES IgE SYNTHESIS BY B CELLS

IN THE PRESENCE OF IL-4a

Group LPS IL-4

+

3 4

5

6 10 13± 3

7 + 5± 2

+ 0.1 2± 1

0 97 ± 13 154 ± 19

133 ± 13 12 + + 0.1 111 ± 12

^a Murine splenic B cells were cocultured with various combinations of

LPS (20 μg/ml), IL-4 (1,000 U/ml), and εRMPs at the indicated concentrations for 7 days. Total IgE concentration i the culture supernatants was determined using ELISA. Results are expressed as geometric mean ± standard deviation of triplicate determinations from triplicate cultures. Statistically significant differences existed between groups 3 and 6 (p = 0.002), groups 3 and 7 (p — 0.03, groups

9 and 10 (p = 0.012), groups 9 and 11 (p = 0.027). +, added to culture; -, not added to culture.

Interestingly, addition of εRMPs to B cell cultures costimulated with

IL-4 plus LPS resulted in a significant further increase in quantities of IgE

synthesized (Table VII, group 9 vs group 10, p = 0.012; vs group 11, p — 0.027). Since saturating amounts of LPS were employed in these cultures,

this finding suggests that εRMPs acts through a signalling mechanism distinct from that provided by LPS.

MBI-1.15 T Cell Membranes Help B Cells Synthesize IgE in Response

to IL-4. Since T_H cell membranes have been shown to costimulate with IL-4 the synthesis of IgE by B cells (Hodgkin, P.D., Yamashita, L.C, Coffman, R.L., & Kehry, M.R. 1990. Separation of events mediating B cell prolifera¬ tion and Ig production by using T cell membranes from MBI-1.15 T

hybridoma cells for such activity (Table VIII, Exp. 1).

TABLE VIII MBI-1.15 T CELL MEMBRANES HELP B CELLS SYNTHESIZE IgE

IN RESPONSE TO IL-4^a

Exp. 2

+ MBI-1.15 33 ± 6

2 + MBI-1.15 (cycloheximide added)^b <1

3 + DMSO-treated MBI-1.15^C 32 ± 13

4 + PMSF-treated MBI-1.15^C 2 ± 1 5 + MBI-1.15 (culture sonicated on day 0)^d <1

^a Purified B cells were cocultured with/without 1L-4 and/or MBI-1.15 cell membranes (1 x 10⁷ cell equivalent/ml), in 10% CO,/atmosphere for 7 days at 37°C IgE concentrations in culture supernatant fluids were determined by ELISA. Results are expressed as geometric mean

± standard deviation of triplicate determinations of triplicate cultures. Differences were found to be statistically significant between groups

4 and 3 (p<0.001) in Exp. 1, and between groups 1 and 2 (p = 0.001), and groups 3 and 4 (p = 0.019) in Exp. 2. +, added to culture; -, not

added to culture.

^b Cycloheximide (100 μg/ml) was co-cultured with B cells, 1L-4, and

MBI-1.15 membrane for 7 days.

^c MBI-1.15 membranes were either mock-incubated with dimethylsulfox-

ide (DMSO) or incubated with ImM PMSF in DMSO for 30 min at

37°C, and washed twice with PBS prior to incorporating into cultures.

^d B cells, IL-4, and MBI-1.15 membranes were sonicated on day 0, and

IgE concentrations in the supernatants were determined by ELISA.

Whereas MBI-1.15 T cell membranes alone had no effect (group 3), they exerted strong IL-4 costimulating activity for IgE synthesis (group 4 vs 3, p < 0.001); in striking contrast, membranes prepared from the M12 B cell line had no activity (group 5). Effects of PMSF or cycloheximide on the activity of MB1-1.15 membranes were studied in Table VIII, Exp. 2. The IL- 4 costimulating activity for IgE biosynthesis was significantly inhibited by either PMSF treatment of the membranes or incorporating cycloheximide into the culture (group 1 vs 2; p = 0.001, group 3 vs 4; p = 0.019). Moreover, cultures of B cells sonicated at the beginning of the culture did not contain detectable levels of IgE (group 5). These data indicate that de novo IgE biosynthesis is induced by these stimulants in vitro, and that the activity of MBI-1.15 cell membranes is dependent on intact serine protease activity on the membranes.

IgE Costimulating Activities of Both Soluble εRMP and MBI-1.15 T

Cell Membranes Are Specifically Blocked by Anti-εRMP mAb. The effects on the anti-εRMP mAb established as described in Materials and Methods were next tested on these IgE costimulating activities.

TABLE IX

IgE CO-STIMULATING ACTIVITY OF BOTH SOLUBLE εRMP AND

MBI-1.15 T CELL MEMBRANES ARE SPECIFICALLY BLOCKED

Group LPS IL-4 εRMPs or Anti-εRMPmAb IgE(ng/ml)

MBI-1.15^n,eι-^b (2E5B)

Exp.l^a

1 - + εRMPs - 10 ± 4

2 - + εRMPs + 1 ± 0

3 - + εRMPs control 11 ± 3

Exp.2^b 1 + MBI-1.15^merab - 40 ± 10

2 + MBI-1.15^memb + 8 ± 3

None 154 ± 40

+ + None 145 ± 17

B cells were cocultured with 1L-4 (1,000 U/ml), εRMPs (3

μg/ml), with or without 2E5B anti-εRMP mAb (1:20 as 2E5B cell-culture supernatant). For the control of 2E5B, culture supernatant fluid of B cell hybridoma parent cell line FO was added at the same final dilution. Statistically significant

difference was found between groups 1 and 2 in Exp. 1 (p =

5 0.015). ^b Plasma membranes prepared from MB1-1.15 T cells were

cocultured at a final concentration of 1 x 10⁷ cell equivalent/ml.

Statistically significant difference was found between groups 1 and 2 in Exp. 2 (p — 0.007), but not between groups 3 and 4 (p

10 > 0.3).

In Table IX, Exp. 1, anti-εRMP mAb (2E5B) was found to significantly inhibit IL-4 + εRMPs-induced IgE synthesis (group 2 vs 1, p = 0.015). The control culture fluid of B cell hybridoma parent FO cells had no

15. effect on IgE synthesis (group 3 vs 1, p > 0.3). Thus, 2E5B mAb not only binds to εRMP, but also neutralizes its biological activity in this system. Moreover, as shown in Table IX, Exp. 2, the costimulating activity of MBI- 1.15 T cell membranes was likewise significantly inhibited by the addition of 2E5B anti-εRMP mAb (groups 2 vs 1, p = 0.007), the specificity of which is

0 supported by the absence of any effect of this mAb on the induction of IgE

synthesis by IL-4 plus LPS (group 4 vs 3, p > 0.3).

Anti-εRMP MAb 2E5B Immunoprecipitates Several Protein Species, of Which A 36-kDa Protein Is the Major Epitope-Bearing Protein. MBI-1.15 cells were externally labeled with ¹²⁵I and the lysates were immunoprecipitated with 2E5B mAb-coupled Sepharose. As shown in Figure 8, ("A" tabulation), 80-, 50-, 40-, 38-, 36-, and 17-kDa bands were immuno¬

precipitated in non-reducing conditions (lane 1). The mobilities of the 80-

kDa and 38-kDa bands seem to be slightly higher in reducing conditions, (lane 2). To identify protein species bearing an epitope recognized by the 2E5B mAb, Western immunoblot analysis was conducted. As shown in

Figure 8 ("B" tabulation), a 36-kDa band showed the highest binding of 2E5B mAb both in reducing and non-reducing conditions, and three other faint bands (28 kDa, 25 kDa, and 17 kDa) could be observed, only one of which

(17-kDa band) was clearly shown in immunoprecipitation. In both Figure 8, "A" and "B" tabulations, use of control antibody prepared from ascites of FO

cells failed to illuminate any bands.

Induction of Germ-Line IgE Heavy Cham Transcripts by εRMP- Bearing Membranes Is Blocked by Anti-εRMP MAb. Recent studies by

other have suggested that, in contrast to IgE biosynthesis (Le., appearance of

mature mRNA for Cε), immature (shorter) germ-line transcripts which cannot be translated into heavy chain protein, can be induced by the stimulus of T cell membranes alone. IL-4 alone also induces the germ-line IgE heavy chain transcripts, whereas LPS alone does not (Gauchat, F.-F., Lebman, D.A. Coffman, R.L. Gascan, H., & de Vries, J.E. 1990. Structure and expression of gennline ε transcripts in human B cells induced by interleukin 4 to switch

to IgE production. /. Exp. Med. 172:463). To reproduce this observation in our system and to specify the effect of anti-εRMP mAb on B cells at molecular and clonal levels, we stimulated the IgM-bearing mouse B cell line,

1-29 (Stavnezer, J. G., Radcliffe, Y-C Lin, L.Berggren, R. Sitia, and E.

Severinson, 1988, Immunoglobulin heavy-chain switching may be directed by prior induction transcripts from constant-region genes, Proc. Nail Acad. Sci.

USA 85:7704), with MBI-1.15 membranes or IL-4 in the presence or absence of anti-εRMP mAb or anti-IL-4 mAb, and analyzed expression of germ-line

transcripts. A genomic DNA fragment which contains a sequence (lε) specific to the germ-line IgE heavy chain transcripts was used as a probe

(Rothman, P., Y-Y. Chen, S. Lutzker, S.C Li, V. Stewart, R. Coffman, and

F.W. Alt, Structure and expression of germ line immunoglobulin heavy-chain ε transcripts: interleukin-4 plus lipopolysaccharide-directed switching to C..,

1990). Since 1) Iε probe showed cross hybridization with ribosomal RNA; and 2) the levels of Iε expression were relatively low, poly (A⁺) RNA was used for analysis. Unfortunately, it was not possible to study soluble εRMP in this system because the quantities required were considerably greater than

could be practically purified.

As shown in Figure 9, Iε was induced in 1-29 cells as 1.9-kb poly

(A⁺) RNA species after coculture with either IL-4 (lane 2) or MB1-1.15 cell

membranes (lane 5). Anti-εRMP mAb blocked Iε induction by MBI-1.15 cell membranes (lane 7), but not that induced by IL-4 (lane 4). Conversely, anti-

IL-4 mAb blocked Iε induction by IL-4 (lane 3), but not that by MB1-1.15 cell membranes (lane 6). This specific inhibition by anti-εRMP vs anti-IL-4 mAbs clearly indicates that induction of Iε by MBI-1.15 cell membranes is not due to IL-4, and εRMP on T cell membranes is directly involved in inducing lε through a pathway distinct from IL-4.

DISCUSSION We demonstrated that 1) both εRMPs and plasma membranes from εRMPs-producing MBI-1.15 T cells can induce IgE synthesis by purified

B cells in the presence of IL-4; 2) the IgE-costimulating activity of both

preparations are specifically blocked by anti-εRMP mAb; 3) εRMP apparently exists as both a 17-kDa soluble form as previously described (Matsushita, S.,

Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and IgE- suppressive proteins. Proc. Natl Acad. Sci. USA. 88:4718), and a 36-kDa

membrane form (hereafter denoted εRMPm); and 4) MBI-1.15 T cell membranes alone can induce germ-line IgE heavy chain RNA transcripts in 1-29 B cells, and this induction is also specifically blocked by anti εRMP mAb. These data strongly suggest that εRMPm molecules expressed on MBI-1.15 T cells are involved in inducing Ig class switching of B cells to the IgE isotype.

The activities of both soluble and membrane-forms of εRMP on IgE production are interesting to compare to that of LPS. In only one respect are there similarities, namely the fact that εRMPm, like LPS, induces de novo biosynthesis of IgE when cultured B cells are costimulated with IL-4 (Table VIII, Exp. 1). The quantities of IgE produced, however, are around 10-fold lower (10-50 ng ml) with εRMP than with LPS as costimulant, and more closely approximate the levels observed in antigen-stimulated murine

(Marcelletti, J.F., & Katz, D.H. 1991. Elicitation of antigen-induced primary and secondary murine IgE antibody responses in vitro. Cell, Immunol,

135:471) and human (Matsushita, S., Muto, M., Suemura, M., Saito, Y., &

Sasazuki, T. 1987. HLA-lϊnked nonresponsiveness to cryptomeria japonica pollen antigen. I. Nonresponsiveness is mediated by antigen-specific suppressor T cell. /. Immunol. 138:109) IgE synthesis in vitro (up to 40 ng ml). This may reflect that 1) a smaller population of B cells, presumably those possessing εRMP receptors, respond to εRMP than to LPS, 2) the signaling intensity of LPS is significantly higher than εRMP or 3) both differences in responding cell numbers and signalling intensities contribute

to the response magnitudes. In any case, these differences in IgE production levels suggest the existence of distinct mechanisms for IgE induction by εRMP and LPS, respectively, Further evidence that εRMP and LPS act by distinct mechanisms on B cells is shown by the observations that 1) εRMPs

significantly enhances the costimulating activity of LPS for IL-4-induced IgE synthesis (Table VII); 2) anti εRMP mAb 2E5B blocks the costimulating activity of εRMP, but not that of LPS, in inducing IgE synthesis (Table IX); and 3) εRMPm, but not LPS (Jabara, H.H., Schneider, L.C, Shapira, S.K., Alfieri, C, Moody, C.T., Kieff, E., Geha, R.S., & Vercelli, D. 1990. Induction of germ-line and mature Cε transcripts in human B cells stimulated

with rIL-4 and EBV. /. Immunol. 145:3468, Gauchat, J.-F., Lebman, D.A. Coffman, R.L. Gascan, H., & de Vries, J.E. 1990. Structure and expression of gennline ε transcripts in human B cells induced by interleukin 4 to switch to IgE production. /. Exp. Med. 172:463), is singularly capable of inducing transcription of Iε in B cells (Fig. 9).

We have also shown that εRMPs is a novel serine protease

(Matsushita, S., & Katz, D.H., 1991. The murine ε receptor modulating protein: A novel serine protease which modulates CD23 binding of IgE. Cell.

Immunol. 137:252). Enhancing effects of other serine proteases on IgE biosynthesis have been previously reported. Ishizaka, et al, (Ishizaka, K. 1985 Twenty years with IgE. /. Immunol. 135 :i.) demonstrated a kallikrein-like

factor, designated glycosylation-enhancing factor, which induces CD4⁺ T cells to produce IgE-potentiating factor which in turn, facilitates the differentiation of IgE-B memory cells. Verwaerde, et al, (Verwaerde, C, Auriault, C, Neyrinck, J.L., & Capron, A. 1988. Properties of serine proteases of schistosoma mansoni schistosomula involved in the regulation of IgE synthesis. Scand. J. Immunol. 27:17.) reported that serine proteases derived

from Schistosoma mansoni schistosomula enhance IgE synthesis of DNA-OA- immunized rats. There is also direct evidence that serine proteases provide mitogenic signals to B cells (Simon, M.M., Hoschuetzky, H., Fruth, U.,

Simon, H-G., & Kramer, M.D. 1986. Purification and characterization of a T cell specific serine proteinase (TSP-1) from cloned cytolytic T lymphocytes. EMBO J. 5:3267.). The IgE-enhancing activity of εRMP described herein appears to directly correlate with its previously shown proteolytic activity (Matsushita,

S., & Katz, D.H., 1991. The murine ε receptor modulating protein: A novel

serine protease which modulates CD23 binding of IgE. Cell Immunol 137:252.), since the activity of εRMPm is abrogated by treatment of MBI-1.15

cell membranes with PMSF (Table VIII). In this regard, some serine proteases which have similar substrate specificity as εRMP, such as kallikrein or α-chymotrypsin (Matsushita, S., & Katz, D.H., 1991. The murine ε receptor modulating protein: A novel serine protease which modulates CD23 binding of IgE. Cell Immunol 137:252.), significantly enhanced IgGl and slightly enhanced IgE synthesis in cultures costimulated with LPS and IL-4.

These observations may indicate that any serine protease expressed on T_ιr

cells (Scholz, W., Mentlein, R., Heymaήn, E., Feller, AC, Ulmer, A.J. and

Flad, H.D. 1985. Interleukin 2 production by human T lymphocytes identified by antibodies to dipeptidyl peptidase IV. CelLImmunol 93:199.)

with εRMP-like substrate specificity may provide signals from T_π membranes to B cells. There is a strong indication of the physiological significance of εRMP in view of the finding that anti-εRMP mAb specifically blocks the IL-4 costimulating activity of εRMPm in the induction of IgE synthesis (Table IX). Although kallikrein is expressed on some leukocytes and is reportedly activated in ongoing allergic reactions in the skin (Atkins, P.C, Miragliotta,

G., Talbot, S.F., Zweiman, B., & Kaplan, A.P. 1987. Activation of plasma hageman factor and kallikrein in ongoing allergic reactions in the skin. /.

Immunol.. 139:2744.), a membrane form has not yet been described.

In general, although it is now well established that proteolytic

activities can be detected on the surface of some cell types, the physiological roles of these membrane-bound proteases remain poorly understood. During

embryogenesis, tissue repair and tumor metastasis, cell invasion is considered to require the turnover and the destruction of tissue matrices through the action of specific cell-surface proteases (O'Donnel-Tormey, J. and Quigley,

J.P. 1983. Detection and partial characterization of a chymostatin-sensitive

endopeptidase in transformed fibroblasts. Proc. Nail. Acad. ScL USA 80:344.). Several surface-associated proteases have already been character¬ ized on immunocompetent cells: dipeptidyl aminopeptidase IV has been shown to be expressed predominantly on CD4⁺ T cells (Scholz, W., Mentlein, R., Heymann, E., Feller, AC, Ulmer, AJ. and Flad, H.D. 1985. Interleukin 2 production by human T lymphocytes identified by antibodies to dipeptidyl peptidase IV. CelLImmunol 93:199.), and specific inhibition of this enzyme decreased Ig synthesis in vitro by pokeweed mitogen-activated mononuclear cells (Schoen, E., Jahn, S., Kiessig, S.T., Demuth, H.U., Neubert, K., Barth, A., Von Baehr, R., & Ansorge, S. 1987. The role of dipeptidyl peptidase IV in human T lymphocyte activation. Inhibitors and antibodies against dipeptidyl peptidase IV suppress lymphocyte proliferation and immunoglobu¬ lin synthesis in vitro. Eur. J. Immunol. 17:1821.), suggesting its involvement in Ig synthesis. In terms of T cell subpopulations, our previous studies demonstrated that εRMPs is secreted from CD4⁺, but not from CD8⁺, T cells

(Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and

IgE-suppressive proteins. Proc. Nail. Acad. ScL USA. 88:4718.); this is consistent with the finding by others that only activated CD4⁺ T cell membranes exert IL-4-costimulating activity in inducing IgE synthesis

(Hodgkin, P.D., Yamashita, L.C., Coffman, R.L., & Kehry, M.R. 1990.

Separation of events mediating B cell proliferation and Ig production by using T cell membranes and lymphokines. /. Immunol. 145:2025.).

Expression analyses of εRMP at protein, mRNA, and functional levels using

T_H0-, T_H1-, and TY_H2-cell clones are underway.

Results obtained from immunoprecipitation and Western immunoblot studies indicate that a 36-kDa protein is the major form of εRMPm. Larger protein species, which were shown only by immunoprecipitation from MBI-1.15 cell membranes (80-kDa, 50-kDa, 40- kDa, and 38-kDa proteins), may be non-covalently associated with εRMP.

Conversely, smaller species, especially the 17-kDa protein which was shown both in immunoprecipitation and Western analysis, may be derived from proteolytic cleavage of the 36-kDa εRMPm. Since our purification scheme for εRMPs was based on the biological activity of the εRMPs molecule

(Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and

IgE-suppressive proteins. Proc. Natl Acad. ScL USA. 88:4718.), we did not use serine protease inhibitors, such as PMSF, in the purification. This might have allowed εRMP itself or non-specific proteases to digest the native 36- kDa εRMP molecule during purification, thus resulting in the appearance of

the 17-kDa εRMP. Referring to Figure 8, data depicting the results of immunoprecipitation and Western blot analyses with 2E5B anti-εRMP mAb are depicted graphically and numerically. (A) Radioiodinated MBI-1.15 cell- surface proteins were immunoprecipitated with 2E5B anti-εRMP mAb-bound

Sepharose and analyzed on 13.5 %SDS-P AGE. Lane 1, non-reducing; Lane 2, reducing conditions. (B) MB1-1.15 cell lysates separated on 13.5% SDS-

PAGE in non-reducing (lane 1) or reducing (lane 2) conditions, transferred onto nitrocellulose membrane, and incubated with radioiodinated 2E5B anti- εRMP mAb. The 17-kDa band of Figure 8 might correspond to the 17-kDa

εRMPs (Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991.

Purification of murine suppressive factor of allergy into distinct CD23- modulating and IgE-suppressive proteins. Proc. Natl Acad. Sci. USA. 88:4718.), but the physiological significance of εRMPs is still unclear. In this regard, Leclercq, et al. (Leclercq, L., Bismuth, B., & Theze, J. 1984. Antigen- specific helper T-cell clone supernatant is sufficient to induce both polyclonal

proliferation and differentiation of small resting B lymphocytes. Proc. Natl. Acad. Sci. USA. 81:6491.) have reported that antigen-specific helper T cell- clone supernatant is sufficient to induce both polyclonal proliferation and differentiation of small resting B cells. This observation might be attributable

to lymphokines secreted from T_H cells, such as εRMPs. In light of the observation by others that contact-dependent T-B cell interactions are necessary for IL-4-induced IgE synthesis (Parronchi, P., Tiri, A., Macchia, D., De Carli, M., Biswas, P., Simonelli, C, Maggi, E., Del Prete, G., Ricci, M., and Romagnani, S. 1990. Noncognate contact-dependent B cell activation can promote IL-4-dependent in vitro human IgE synthesis. /. Immunol. 144:2102.), it is also possible that εRMPs functions only in the micro-

environment of contact-dependent T-B cell interactions. In this regard, the activity of purified εRMPs is well maintained in PBS but not in FCS, suggesting the existence of a natural inhibitor(s) of εRMP in serum. If this is the case, εRMP may only be effectively active in the context of direct cell-

cell interactions or factor-mediated interactions in the immediate micro- environment.

MBI-1.15 cells were established by fusion of AKR (H-2k)-derived thymoma cells (BW5147) and DBA-2 (H-2^d)-derived T cell blasts (Matsushita, S., Marcelletti, J.F., Katz, L.R., & Katz, D.H. 1991. Purification of murine suppressive factor of allergy into distinct CD23-modulating and IgE-suppressive proteins. Pioc. Natl Acad. ScL USA. 88:4718.). Since MB1- 1.15 membranes induced IgE synthesis by C57BL/6 (H-2^b) B cells, this indicates that the interactions between isolated T cell membranes and B cells

in this system are not MHC-restricted, which corroborates the previous finding of others (Parronchi, P., Tiri, A, Macchia, D., De Carli, M., Biswas, P., Simonelli, C, Maggi, E., Del Prete, G., Ricci, M., and Romagnani, S. 1990. Noncognate contact-dependent B cell activation can promote IL-4- dependent in vitro human IgE synthesis. /. Immunol. 144:2102, Hodgkin,

P.D., Yamashita, L.C., Coffman, R.L., & Kehry, M.R. 1990. Separation of events mediating B cell proliferation and Ig production by using T cell

membranes and lymphokines. /. Immunol. 145:2025.). Although interaction

between T cell-derived εRMPm and B cells is not genetically restricted, and εRMP can function polyclonally on B cells in vitro, such interactions in vivo,

especially in secondary responses, may occur immediately after T cell activation as a consequence of antigen-specific, cognate and MHC-restricted

T-B cell interactions.

The finding that T cell-derived εRMPm shares the same property as the 17 kDa εRMPs in costimulating (with 1L-4) the induction of IgE synthesis, see Table VIII, suggests that εRMPm expressed on T cells could

directly transduce class-switch signals to B cells. This distinguishes its function from that of an adhesion molecule, such as CD4, between T and B

cells.

Interestingly, there appears to be some difference in isotype selectivity between εRMPm and εRMPs; preliminary studies indicate that, in costimulation with IL-4, εRMPm is a more effective inducer of IgE than εRMPs but has no effect of IgGl, whereas εRMPs induces IgGl synthesis.

Explanations for these apparently distinct class-switch pressures between εRMPm and εRMPs require further study.

Germ-line IgE heavy chain transcripts have been described as early-stage transcripts (Rothman, P., Lutzker, S., Cook, W., Coffman, R., Alt, F.W. 1988. Mitogen plus interleukin 4 induction of Cε transcripts in B lymphoid cells. /. Exp. Med. 168:2385) which appear prior to the transcrip¬ tion of mature message for the constant region of heavy chain IgE (Cε). It

has an exon designated Iε, which locates at 5' upstream of the switch recombination site for Cε, and mRNA for the germ-line transcripts consist of Iε and four exons that code for Cε. Because the Iε exon contains

termination codons throughout the reading frames, it does not encode any protein (Gerondakis, S. 1990. Structure and expression of murine germ-line immunoglobulin ε heavy chain transcripts induced by interleukin 4. Proc.

Natl Acad. ScL USA. 87:1581). However, the appearance of the germ-line transcripts may be an important prerequisite for class switching to Cε, since accessibility of the recombinase may increase (Stavnezer-Nordgren, J., &

Sirlin, S. 1986. Specificity of immunoglobulin heavy chain switch correlates with activity of germline heavy chain genes prior to switching. EMBO J. 5:95). Our observation that anti-εRMP mAb blocks the induction by MBI- 1.15 T cell membranes of both IgE synthesis and germ-line IgE heavy chain

transcription in B cells, attests to the likely significance of εRMP in physiological class switching IgE. It is thus highly conceivable that εRMP expressed on T cell membranes interacts with B cells and induces germ-line

IgE heavy chain transcripts. IL-4, either produced internally by T_H cells or added exogenously, should act on εRMP-activated B cell populations and induce mature Cε message, resulting in IgE secretion by B cells. Significantly, both IL-4 and εRMP induce Iε transcription, but neither molecule by itself can induce class switching to IgE synthesis. This clearly suggests that both εRMP and IL-4 each have another important

molecular effect (which may or may not be identical) on B cells, that is essential for class-switching, but only when both molecules are present

simultaneously is the complete mechanism of class-switching manifested. The eventual cloning of the εRMP molecule should facilitate understanding of its structure, the biological pathways to its function, and its possible natural receptors on B cells. Ultimately, such information should shed light on the role of CD23 and perhaps explain why this molecule is so relatively widely

displayed on lymphocytes.

Purified εRMP and purified IgE-suppressive protein (SFA) are useful in scientific and industrial research in testing for various immunologi¬ cal substances and testing cells for immunological response. These proteins may be used, for example, to modulate the immunological response to

various allergens to aid in characterizing the responsive mechanism of the

cells and in studying the allergens.

These new compositions of matter, Le. the purified protein, εRMP, characterized in that it is a serine protease, an ε receptor modulating protein that modulates the low affinity Fc receptor for IgE on lymphocytes CD23, has a molecular weight of approximately 17 kDa, decreases the binding avidity of IgE of CD23-bearing B cells and does not diminish the cell surface expression of CD23 and the purified IgE-suppressive protein characterized in having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy,

are considered to be useful in immunological therapy to modulate or prevent the inflammation and other immunogenic response to various allergens.

These new compositions of matter may also be used in research to treat laboratory mammals to evaluate the effect of other materials on the

immunological system of the mammal or to modulate the mammal's immunological response to enable the mammal to tolerate testing without undue inflammation. In therapeutic applications and in the treatment of mammals, either of the compositions, in a physiologically acceptable carrier, of which there are many, e.g. buffered saline, is injected into the mammal,

human or otherwise, in an amount between 5 and 500 μgrams/kg of body weight, preferably in the range of from 50 to 250 μg/kg of body weight. Periodic injections, approximately daily, are expected to be required to maintain adequate control of the immune response of the mammal or

individual being treated.

The invention, in one form, is embodied in an immunologically active liquid consisting essentially of purified soluble 17-Kda protein, εRMP, wherein substantially the sole immunological activity arises from said εRMP that is useful as a diagnostic reagent or as a therapeutic composition. The protein εRMP is characterized in that it diminishes the avidity of binding

between IgE and CD23 low affinity Fc receptor for IgE without decreasing the quantitative expression of the CD23, is inactivated by phenylmethyl- sulfonyl fluoride (PMSF), decomposes N,α-benzyloxycarbonyl-L-lysine thiobenzyl ester (BLT) and N-succinyl-Ala-Ala-Pro-Phe- -nitroanilide, does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells and comprises the partial internal amino acid sequence:

Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala

1 5 10 15

Lys Lys (SEQ ID NO:2:).

The invention is also embodied in a method of modifying the immune response of a mammal. In accordance with this method, the mammal is injected intravenously, intramuscularly or subdermally with an

effective amount between 5 and 500 μgrams/kg of body weight of the mammal an immunologically active liquid, substantially the sole immunologi- cally active species in said liquid consisting essentially of soluble, purified εRMP that is substantially free of other components of SFA, said εRMP being characterized as stated in the next preceding paragraph.

The invention is also embodied in a membranous composition having immunological activity, the immunological activity arises substantially

solely from purified εRMP on or forming the membrane. The εRMP is separated from other components of SFA and purified, and characterized as stated above. This membranous composition is useful in separating

immunological species in diagnostic methods and in immunological research. In an alternative aspect, the invention is embodied in a

composition of matter consisting essentially of a purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses

IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy.

The invention is also embodied in a method of modifying the

immune response of a mammal comprising injecting into the mammal an effective amount between 5 and 500 μgrams/kg of body weight of the mammal of purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy.

The method of modifying the immune response of a mammal comprising injecting into the mammal substantially contemporaneously an effective amount between 5 and 500 μgrams/kg of body weight of the

mammal of IL-4 and an effective amount between 5 and 500 μgrams/kg of body weight of the mammal soluble, purified εRMP that is substantially free of other components of SFA ^s-id εRMP being characterized as stated above is another embodiment of the invention.

The invention is embodied, in addition, in a method of modifying the immune response of a mammal comprising injecting into the mammal an effective amount between 5 and 500 μgrams/kg of body weight of the mammal of a liquid consisting essentially of εRMP and 1L-4 and, as well, in a method of modifying the immune response of a mammal comprising injecting into the mammal an effective amount between 5 and 500 μgrams/kg of body weight of the mammal a liquid consisting essentially of purified IgE-

suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy in combination with said effective amount of IL-4.

A composition of matter comprising soluble εRMP, said composition being substantially free of all other proteins and peptides is also within the present invention, as well as a composition of matter comprising membranous εRMP, said composition being substantially free of all other proteins and peptides.

In another embodiment, the invention comprises a liquid composition of matter consisting essentially of purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy in combination with IL-4 and, also, a liquid composition of matter consisting essentially of εRMP and IL-4. Viewed differently, the invention may be embodied in the cell line MBI-1.15 deposited in the American Type Culture Collection identified by the access number ATCC .

Other embodiments of the invention and applications thereof will become apparent from the above teachings to those skilled in the art. Industrial Application

This invention is useful in the development of immunological methods and compositions, in immunological research, and in immunological therapy on humans and other animals.

Claims

WHAT IS CLAIMED IS:

1. An immunologically active liquid consisting essentially of purified soluble 17-Kda protein, εRMP, wherein substantially the sole immunological activity arises from said εRMP, said εRMP being character- ized in that said protein:

(a) diminishes the avidity of binding between IgE and CD23 low affinity Fc receptor for IgE without decreasing the quantitative expression of the CD23;

(b) is inactivated by phenylmethylsulfonyl fluoride (PMSF); (c) decomposes N,α-benzyloxycarbonyl-L-lysine thiobenzyl ester

(BLT) and N-succinyl-Ala-AIa-Pro-Phe- >-nitroanilide;

(d) does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells; and

(e) comprises the partial internal amino acid sequence: Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala

1 5 10 15

Lys Lys (SEQ ID NO:2:).

2. A method of modifying the immune response of a mammal comprising injecting into the mammal an effective amount between 5 and 500 μgrams/kg of body weight of the mammal an immunologically active liquid, substantially the sole immunologically active species in said liquid consisting essentially of soluble, purified εRMP that is substantially free of other components of SFA, said εRMP being characterized in that said protein: (a) diminishes the avidity of binding between IgE and CD23 low affinity Fc receptor for IgE without decreasing the quantitative expression of the CD23; (b) is inactivated by phenylmethylsulfonyl fluoride (PMSF); (c) decomposes N,α-benzylo_^~ carbonyl-L-Iysine thiobenzyl ester (BLT) and N- succinyl-AIa-Ala-Pro-Phe- 7-nitroanilide; (d) does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells; and (e) comprises the partial internal amino acid sequence: Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala

1 5 10 15

Lys Lys (SEQ ID NO:2:).

3. A membranous composition consisting essential of membrane and purified εRMP, said membrane being separated from other components of SFA and purified, and characterized by an internal sequence Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala 1 5 10 15

Lys Lys (SEQ ID NO:2:) and being a serine protease, an ε receptor modulating protein that modulates the low affinity Fc receptor for IgE on lymphocytes CD23, has a molecular weight of approximately 17 kDa, decreases the binding avidity of IgE of CD23-bearing B cells and does not diminish the cell surface expression of CD23.

4. Purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy.

5. A method of modifying the immune response of a mammal comprising injecting into the mammal an effective amount between 5 and 500 μgrams/kg of body weight of the mammal of purified IgE- suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy.

6. A method of modifying the immune response of a mammal comprising injecting into the mammal substantially contemporane¬ ously an effective amount between 5 and 500 μgrams/kg of body weight of the mammal of IL-4 and an effective amount between 5 and 500 μgrams/kg of body weight of the mammal soluble, purified εRMP that is substantially free of other components of SFA, said εRMP being characterized in that said protein:

(a) diminishes the avidity of binding between IgE and CD23 low affinity Fc receptor for IgE without decreasing the quantitative expression of the CD23;

(b) is inactivated by phenylmethylsulfonyl fluoride (PMSF);

(c) decomposes N,α-benzyloxycarbonyl-L-lysine thiobenzyl ester (BLT) and N-succinyl-Ala-Ala-Pro-Phe-/?-nitroanilide;

(d) does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells; and

(e) comprises the partial internal amino acid sequence: Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala 1 5 10 15

Lys Lys (SEQ ID NO:2:).

7. A method of modifying the immune response of a mammal comprising injecting into the mammal an effective amount between

5 and 500 μgrams/kg of body weight of the mammal of a liquid consisting essentially of εRMP and IL-4.

8. A method of modifying the immune response of a mammal comprising injecting into the mammal an effective amount between

5 and 500 μgrams/kg of body weight of the mammal a liquid consisting essentially of purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy in combination with said effective amount of IL-4.

9. A composition of matter comprising soluble εRMP, said composition being substantially free of all other proteins and peptides.

10. A composition of matter comprising membranous εRMP, said composition being substantially free of all other proteins and peptides.

11. A liquid composition of matter consisting essentially of purified IgE-suppressive protein having a molecular weight of approxi¬ mately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS- activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy in combination with IL-4.

12. A liquid composition of matter consisting essentially of εRMP and IL-4.

13. The cell line MBI-1.15 deposited in the American Type Culture Collection identified by the access number ATCC .

14. In the manufacture of an immunologically active medicament for altering the immunological characteristics of a mammal, the use of purified soluble 17-Kda protein, εRMP characterized in that it:

(a) diminishes the avidity of binding between IgE and CD23 low affinity Fc receptor for IgE without decreasing the quantitative expression of the CD23;

(b) is inactivated by phenylmethylsulfonyl fluoride (PMSF);

(c) decomposes N,α-ben__yloxycarbonyl-L-lysine thiobenzyl ester (BLT) and N-succinyl-Ala-Ala-Pro-Phe-^-nitroanilide;

(d) does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells; and

(e) comprises the partial internal amino acid sequence: Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala 1 5 10 15

Lys Lys (SEQ ID NO:2:).

15. Purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy.

16. In the manufacture of a medicament for modifying the immune response of a mammal the use of IL-4 and purified εRMP that is substantially free of other components of SFA, said εRMP being characterized in that said protein:

(a) diminishes the avidity of binding between IgE and CD23 low affinity Fc receptor for IgE without decreasing the quantitative expression of the CD23;

(b) is inactivated by phenylmethylsulfonyl fluoride (PMSF);

(c) decomposes N,α-benzyloxycarbonyl-L-lysine thiobenzyl ester (BLT) and A/-succinyl-AIa-Ala-Pro-Phe-/?-nitroanilide;

(d) does not work directly on B cells but requires CD4⁺ T cells to decrease functional expression of CD23 on B cells; and

(e) comprises the partial internal amino acid sequence: Ala Lys Pro Ala Pro Lys Lys Glu Glu Lys Lys Lys Lys Ala Ala Ala 1 5 10 15

Lys Lys (SEQ ID NO:2:).

17. In the manufacture of a medicament for modifying the immune response of a mammal the use of εRMP and IL-4.

18. In the manufacture of a medicament for modifying the immune response of a mammal the use of purified IgE-suppressive protein having a molecular weight of approximately 30 kDa that suppresses IL-4-induced IgE and IgGl synthesis by LPS-activated spleen cells, but has no effect on other antibody isotypes and is distinct from IFNy in combination with IL-4.