Search

Glycosylation inhibiting factor (GIF) and macrophage migration inhibitory factor (MIF) share an identical structure gene. Here we unravel two steps of posttranslational modifications in GIF/MIF molecules in human suppressor T (Ts) cell hybridomas. Peptide mapping and MS analysis of the affinity-purified GIF from the Ts cells revealed that one modification is cysteinylation at Cys-60, and the other is phosphorylation at Ser-91. Cysteinylated GIF, but not the wild-type GIF/MIF, possessed immunosuppressive effects on the in vitro IgE antibody response and had high affinity for GIF receptors on the T helper hybridoma cells. In vitro treatment of wild-type recombinant human GIF/MIF with cystine resulted in preferential cysteinylation of Cys-60 in the molecules. The cysteinylated recombinant human GIF and the Ts hybridoma-derived cysteinylated GIF were comparable both in the affinity for the receptors and in the immunosuppressive activity. Polyclonal antibodies specific for a stretch of the amino acid sequence in alpha2-helix of GIF bound bioactive cysteinylated GIF but failed to bind wild-type GIF/MIF. These results strongly suggest that cysteinylation of Cys-60 and consequent conformational changes in the GIF/MIF molecules are responsible for the generation of GIF bioactivity.

An ELISA system for the human glycosylation inhibiting factor (GIF) was established using polyclonal antibodies against highly purified 13 kDa recombinant human GIF, and the concentration of GIF in the sera of healthy donors and patients with various diseases was determined. GIF was detected in the sera of most healthy individuals and its concentration tended to increase with age. It was also found that the serum GIF levels markedly increased in some patients with rheumatoid arthritis or malignant tumors. Analysis of serum samples by SDS-PAGE and immunoblotting revealed a 55 kDa protein that has both the GIF antigenic determinant and the TCR alpha chain determinant. A 13 kDa GIF was not detected in the sera. In view of our previous findings on antigen-specific GIF from murine suppressor T cell hybridomas indicating that the 55 kDa GIF is a post-translationally formed conjugate of a TCR alpha chain with 13 kDa GIF, we suspect that the 55 kDa GIF detected in human sera is a human homologue of the murine 55 kDa GIF.

Receptors for bioactive glycosylation-inhibiting factor (GIF) were demonstrated using a bioactive mutant of recombinant human (rh) GIF, which is comparable to the suppressor T (Ts) cell-derived bioactive GIF in its affinity for the receptors on helper T (Th) hybridoma cells. Both naive T and B cells in normal mouse spleen lacked GIF receptors. However, presentation of specific antigen to naive T cells resulted in the expression of the receptors on activated T cells. Furthermore, activation of small resting B cells with F(ab')2 fragments of anti-mouse IgM plus IL-4, lipopolysaccharide (LPS) plus IL-4 or LPS plus dextran sulfate induced the expression of the receptors within 48 h of B cell stimulation. It was also found that NK T cells freshly isolated from mouse spleen, but not conventional NK cells, expressed receptors for GIF. CD4(+) and CD4(-) subpopulations of NK T cells showed a similar binding capability. Mature dendritic cells derived from bone marrow did not bear the receptors. The dissociation constant (Kd) of the interaction between the bioactive rhGIF mutant and the high-affinity receptors was 10-100 pM, whereas inactive wild-type rhGIF failed to bind to the receptors. A bioactive derivative of rhGIF suppressed both IgG1 and IgE synthesis by purified B cells activated by LPS and IL-4, indicating that the binding of bioactive GIF to its receptors on activated B cells results in suppression of their differentiation.

We have shown previously that glycosylation-inhibiting factor (GIF) in culture supernatants of suppressor T cell (Ts) hybridomas had bioactivity, while the same cells contained a substantial quantity of inactive GIF in cytosol. Mass-spectrometric analysis of GIF in the culture supernatant and cytosol of a Ts hybridoma provided direct evidence that GIF protein was posttranslationally modified in the Ts cells, and that the GIF bioactivity is associated with the posttranslationally modified species. Assuming that conformational changes induced by the posttranslational modifications are responsible for generation of bioactivity, we constructed cysteine mutants of human rGIF (rhGIF) in which cysteine at position 57, 60, or 81 was replaced with Ala, and the mutants were expressed in Escherichia coli. Replacement of Cys57 or Cys60 with Ala resulted in generation of bioactivity, while replacement of Cys81 with Ala failed to do so. It was also found that replacement of Cys57 with Ala and carboxymethylation of a sulfhydryl group in Cys60 synergistically increased the GIF bioactivity of the GIF derivatives. A mutated GIF protein, in which Cys57 and Asn106 in the rhGIF were replaced with Ala and Ser, respectively, had immunosuppressive effects on the IgE and IgG1 Ab responses of BDF1 mice to DNP-OVA, while wild-type rhGIF did not. Evidence was obtained that the mutated GIF suppressed Ag priming of Th cells for the Ab responses and proliferative response.

High-affinity binding was demonstrated between suppressor-T-cell-derived bioactive glycosylation-inhibiting factor (GIF) and helper T hybridomas and natural killer cell line cells. Inactive GIF present in cytosol of suppressor T cells and Escherichia coli-derived recombinant human GIF (rhGIF) failed to bind to these cells. However, affinity of rhGIF for the target cells was generated by replacement of Cys-57 in the sequence with Ala or of Asn-106 with Ser or binding of 5-thio-2-nitrobenzoic acid to Cys-60 in the molecule. Such mutations and the chemical modification of rhGIF synergistically increased the affinity of GIF molecules for the target cells. The results indicated that receptors on the target cells recognize conformational structures of bioactive GIF. Equilibrium dissociation constant (Kd) of the specific binding between bioactive rGIF derivatives and high-affinity receptors was 10-100 pM. Receptors for bioactive GIF derivatives were detected on Th1 and Th2 T helper clones and natural killer NK1.1(+) cells in normal spleen but not on naive T or B cells. Neither the inactive rGIF nor bioactive rGIF derivatives bound to macrophage and monocyte lines or induced macrophages for tumor necrosis factor alpha production.

Escherichia coli-derived recombinant human glycosylation inhibiting factor (rhGIF) contains three cysteine residues (Cys-57, -60, and -81). All SH groups in the cysteine residues are free, and the GIF molecule had no biologic activity. Carboxymethylation of the SH group of Cys-60 in the molecule resulted in the generation of bioactivity, although the activity of the carboxymethylated GIF was 10- to 20-fold less than that of suppressor T cell (Ts)-derived GIF. However, treatment of the inactive rhGIF with ethylmercurithiosalicylate or 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB) resulted in the generation of derivatives whose bioactivity was comparable to that of the Ts-derived bioactive GIF. The activity of these derivatives was lost by treatment with DTT. Isolation and chemical analysis of the DTNB-treated GIF derivative revealed that binding the 5-thio-2-nitrobenzoic acid group with Cys-60 was responsible for the generation of the highly bioactive derivative. Inactive cytosolic GIF from mammalian cells could also be converted to bioactive derivative by treatment with the SH reagent, while Ts-derived bioactive GIF was inactivated by DTT. These results, together with an x-ray crystal structure of GIF molecules, strongly suggest that the generation of bioactivity of GIF in Ts cells is due to posttranslational modifications that result in conformational changes in the molecule.

Upon stimulation with anti-CD3, suppressor T-cell (Ts) hybridomas and homologous transfectants of T-cell receptor a (TCRalpha) cDNA in the T-cell hybridoma formed a 55-kDa TCRalpha chain derivative that bound both the monoclonal anti-TCRalpha chain and polyclonal antibodies against glycosylation inhibiting factor (GIF). The peptide is a subunit of antigen-specific suppressor T-cell factor (TsF), and is considered to be a posttranslationally-formed conjugate of TCRalpha chain with GIF peptide. The TCRalpha derivative is synthesized by the transfectant after stimulation with anti-CD3, and not derived from TCR present on the cell surface. Stimulation of the stable homologous transfectants with anti-CD3 induced translocation of the 13-kDa GIF peptide into endoplasmic reticulum (ER). When a helper Ts hybridoma or a stable transfectant of the same TCRalpha cDNA in a helper cell-derived TCRalpha- clone was stimulated with anti-CD3, translocation of GIF peptide was not detected, and these cells failed to secrete a TCRalpha derivative. However, further transfection of a chimeric cDNA encoding a procalcitonin-GIF fusion protein into the helper cell-derived stable transfectant of TCRalpha cDNA resulted in translocation of the GIF protein and formation of bioactive 55-kDa GIF. The results indicated that translocation of GIF peptide through ER is unique for Ts cells, and that this process is essential for the formation/secretion of the soluble form derivative of TCRalpha chain by T cells.

Glycosylation-inhibiting factor (GIF) is a cytokine that is involved in the regulation of IgE synthesis. The crystal structure of recombinant human GIF was determined by the multiple isomorphous replacement method. The structure was refined to an R factor of 0.168 at 1.9 angstrom resolution. The overall structure is seen to consist of three interconnected subunits forming a barrel with three 6-stranded beta-sheets on the inside and six alpha-helices on the outside. There is a 5-angstrom-diameter "hole" through the middle of the barrel. The barrel structure of GIF in part resembles other "trefoil" cytokines such as interleukin 1 and fibroblast growth factor. Each subunit has a new class of alpha + beta sandwich structure consisting of two beta-alpha-beta motifs. These beta-alpha-beta motifs are related by a pseudo-twofold axis and resemble both interleukin 8 and the peptide binding domain of major histocompatibility complex protein, although the topology of the polypeptide chain is quite different.

We isolated a unique TCR alpha-chain derived from the OVA-specific Ts hybridoma, 231F1. The TR alpha-chain consists of V alpha 11.3, a unique J alpha and a complete sequence of C alpha region. Transfection of the TCR-alpha cDNA into a TCR-alpha-, TCR-beta+ T cell line, 175.2, resulted in the expression of TCR-alpha beta, and the transfectant contained a 35-kDa peptide having the TCR- alpha-specific antigenic determinant. However, the stable transfectant failed to release a peptide with the TCR-alpha determinant upon stimulation with anti-CD3. In contrast, overexpression of the cDNA in the 231 F1 cells markedly increased the formation of the 55-kDa peptide, which reacted with both anti-glycosylation-inhibiting factor (GIF) and the mAb H28-710. Definitive evidence for the relationship between the 55-kDa peptide and the TCR alpha-chain was obtained by transfection of the cDNA of the TCR alpha-chain with histidine tag into the 231F1 cells. The 55 kDa GIF peptide formed by stable transfectants of the TCR-alpha-tag cDNA bound to Ni+-nitrilotriacetic acid-agarose. Upon stimulation with anti-CD3, a stable transfectant of the TCR-alpha cDNA formed OVA-specific GIF which contained the 55-kDa GIF peptide, and bound not only to anti-TCR-alpha column but also to anti-TCR-beta column. The results indicate that the OVA-specific GIF consists of the TCR-alpha+ 55-kDa GIF and another peptide with TCR-beta determinant. It was found that the association of the TCR-beta+ peptide with the 55-kDa GIF is required for binding of the factor to OVA, but not essential for the formation and release of the latter peptide.

Stimulation of OVA-specific suppressor T cell (Ts) hybridoma and bee venom phospholipase A2 (PLA2)-specific Ts hybridoma with Ag-pulsed APC or by cross-linking of CD3 resulted in the formation of Ag-specific glycosylation-inhibiting factor (GIF). Affinity-purified Ag-specific GIF preparations, obtained by using Ag-coupled Sepharose or anti-TCR alpha-chain-coupled Affi-Gel, contained a 55-kDa peptide, which bound both polyclonal anti-GIF Abs and anti-TCR-alpha mAb in immunoblotting. The same hybridomas constitutively secrete 13-kDa bioactive GIF peptide that has no affinity for homologous Ag, but neither the Ag-specific GIF activity nor 55-kDa GIF peptide was detectable in culture supernatants of unstimulated cells. Northern blot analysis of mRNA from the anti-CD3-stimulated hybridoma with 32P-labeled GIF cDNA revealed only 0.6 kb mRNA, which encodes the 13-kDa nonspecific GIF. No mRNA of the 55-kDa GIF was detectable. A representative OVA-specific Th hybridoma, DO 11.10 cells contain the 0.6 kb GIF mRNA and constitutively secrete inactive GIF peptide. However, the Th hybridoma failed to secrete the 55-kDa peptide or any peptide with the TCR-alpha determinant upon stimulation with anti-CD3. It appears that the formation of the 55-kDa peptide with the TCR-alpha determinant is unique for a subset of T cells including Ts cells that form bioactive GIF.

Murine suppressor T-cell hybridoma cells (231F1) secrete not only bioactive glycosylation-inhibiting factor (GIF) but also an inactive peptide comparable to bioactive GIF peptide in its molecular size and reactivity with anti-GIF; the amino acid sequence of the inactive peptide is identical to that of the bioactive homologue. The inactive GIF peptide in culture supernatant of both the 231F1 cells and a stable transfectant of human GIF cDNA in the murine suppressor T hybridoma selectively bound to Affi-Gel 10, whereas bioactive GIF peptides from the same sources failed to bind to the gel. The inactive cytosolic human GIF from the stable transfectant and Escherichia coli-derived recombinant human GIF also had affinity for Affi-Gel 10. Both the bioactive murine GIF peptide from the suppressor T hybridoma and bioactive recombinant human GIF from the stable transfectant bound to the anti-I-J monoclonal antibody H6 coupled to Affi-Gel. However, bioactive hGIF produced by a stable transfectant of human GIF cDNA in BMT10 cells failed to be retained in H6-coupled Affi-Gel. These results indicate that the I-J specificity is determined by the cell source of the GIF peptide and that the I-J determinant recognized by monoclonal antibody H6 does not represent a part of the primary amino acid sequence of GIF. It appears that the epitope is generated by a posttranslational modification of the peptide.

Secretion of bioactive glycosylation-inhibiting factor (GIF) appears to be restricted to suppressor T (Ts) cells, although various human and murine cell line cells secrete the 13-kDa peptide that reacts with anti-GIF. Nucleotide sequences of GIF cDNA from the Ts and non-Ts cells are identical, indicating that bioactive GIF and inactive GIF have an identical amino acid sequence. A stable transfectant of human GIF (hGIF) cDNA in BMT10 cells secretes inactive GIF peptide, whereas transfection of a chimeric cDNA encoding a fusion protein consisting of the N-terminal region of procalcitonin precursor and hGIF into the same cells results in secretion of bioactive GIF. Evidence was obtained that the fusion protein goes into the endoplasmic reticulum and is cleaved for the secretion of mature GIF peptide, whereas the inactive 13-kDa peptide synthesized by the former transfectant does not go through the endoplasmic reticulum. However, a stable transfectant of hGIF cDNA in mouse Ts hybridoma contains inactive GIF in the cytosol and secretes bioactive hGIF without participation of the endoplasmic reticulum-Golgi system. Heterogeneity of the 13-kDa hGIF from the transfectant was detected in two-dimensional electrophoresis. The results suggested that Ts cells have a machinery that converts a portion of inactive cytosolic GIF peptide to bioactive GIF during secretion.

We obtained human T-cell hybridomas that are specific for bee venom phospholipase A2 (PLA2) and constitutively secrete glycosylation inhibiting factor (GIF). Upon crosslinking of CD3, the hybridoma produced GIF having affinity for PLA2. When affinity-purified PLA2-binding GIF was used as an immunogen, monoclonal antibodies specific for the antigen-binding GIF were obtained. Monoclonal antibody 110BH3 bound the antigen-binding GIF but failed to bind the 13-kDa nonspecific GIF, as determined by both bioassay and ELISA. In contrast, 388F1, a monoclonal antibody against nonspecific GIF, gave ELISA signals with both the nonspecific GIF and the antigen-binding GIF. Gel filtration of affinity-purified antigen-binding GIF revealed the presence of a 72- to 80-kDa protein which gave ELISA signals with both 110BH3 and 388F1 and contained GIF bioactivity. Upon reduction and alkylation, the antigen-binding GIF dissociated into a 62- to 64-kDa protein which gave positive ELISA with antibody 110BH3 but no signal with antibody 388F1, and a 15-kDa protein, which gave ELISA signal with the 388F1 but not with 110BH3. Immunoblotting of a PLA2-binding GIF preparation revealed that under reducing conditions, the antigen-binding GIF dissociated a 13-kDa peptide which reacted with polyclonal antibodies against recombinant GIF. The results indicate that the 13-kDa nonspecific GIF is a subunit of antigen-binding GIF. The PLA2-binding GIF has affinity for an epitope, representing amino acid residues 19-28 in PLA2 which appears to be an external structure in the antigen.

By using probes based on partial amino acid sequence of glycosylation-inhibiting factor (GIF) from a mouse T-cell hybridoma, a full-length cDNA encoding mouse GIF was isolated. A cDNA clone encoding human GIF was isolated from cDNA libraries of a GIF-producing human T-cell hybridoma by using mouse GIF cDNA as a probe. The cDNAs encode a putative 12.5-kDa peptide of 115 amino acids. Northern blot analysis demonstrated a single, 0.6-kb transcript. Polyclonal rabbit antibodies against the Escherichia coli-derived recombinant 13-kDa peptide bound hybridoma-derived GIF. Although the peptide did not contain a signal peptide sequence, transfection of the cDNA into COS-1 cells resulted in secretion of 13-kDa peptide, but the peptide had substantially less bioactivity than the hybridoma-derived GIF. However, expression of a chimeric cDNA encoding a fusion protein consisting of the N-terminal pro region of calcitonin precursor and human GIF and cotransfection with furin cDNA to allow intracellular cleavage of the fusion protein resulted in secretion of 13-kDa peptide that was comparable to hybridoma-derived GIF in its bioactivity. Both the 13-kDa peptide and GIF bioactivity in the transfected COS-1 supernatant bound to a monoclonal antibody against hybridoma-derived human GIF. These results indicate that the 13-kDa peptide represents recombinant GIF, but posttranslational modification of the peptide is important for generation of the bioactivity. The GIF cDNA had high homology with the cDNA encoding macrophage migration inhibitory factor. However, the recombinant GIF failed to inhibit migration of human monocytes, and recombinant human macrophage migration inhibitory factor did not have GIF bioactivity.

From the spleen cells of BALB/c mice primed with bee venom phospholipase A2 (PLA2), we established seven T cell hybridomas which constitutively secreted glycosylation inhibiting factor (GIF), expressed both CD3 and TCR alpha beta, and responded to antigen-pulsed antigen presenting cells (APC) for the formation of IgE-binding factor. Upon stimulation with antigen-pulsed APC, four of the seven hybridomas produced GIF having affinity for native PLA2. The antigen-binding GIF could suppress the anti-hapten antibody response of BALB/c mice to dinitrophenyl (DNP)-PLA2 conjugates in a carrier-specific manner and bound to immunosorbents coupled with either the mAb 14-12 or anti-TCR alpha chain, H28-710. Analysis of the epitope specificity of the TCR on the GIF-producing T hybridomas indicated that all of the hybridomas which could produce antigen-binding GIF upon antigenic stimulation recognized the synthetic peptide representing amino acid residues 19-34 in PLA2 molecules in the context of the product of the I-Ad subregion and the antigen-binding GIF formed by the cells had affinity for the peptide. The 3-D structure of bee venom PLA2 indicates that the sequence of amino acid 14-24 forms a loop in the PLA2 molecule and represents an external structure of the antigen, while peptide 25-37 forms an alpha helix. Evidence was obtained which suggests that the sequence of 25-34 contains amino acid residues interacting with Ia molecules, while peptide 19-24 contains residues involved in the interaction of p19-34-Ia complexes with TCR on the hybridomas. It was also found that not only the synthetic peptide 19-34, but also the peptides 13-28 and 19-30 inhibited the binding of antigen-binding GIF to PLA2-coupled Sepharose, while peptide 25-40 failed to do so. The results collectively indicate that the antigen-binding GIF and TCR on the cell source of the factor interact with a common epitope which is exposed on the surface of a nominal antigen.

The human T cell hybridoma AC5 has been shown to produce an IgE-binding factor (IgEBF) upon stimulation with T cell mitogens, or anti-CD3 antibody. In this study, the line was established, propagated long term in a newly available serum-free, protein-free medium, and the factor it produced was analyzed. Serologic analysis, utilizing an ELISA assay and biotin-labeled human Ig of various isotypes, revealed that the IgEBF thus obtained was highly specific for the Ig epsilon-chain and was released primarily within the first 24 h after mitogen stimulation. Using biotin-labeled IgE as the detecting reagent, Western blot analysis of this factor demonstrated that the molecule was a single chain moiety of m.w. 64,000, and could be purified to apparent homogeneity by either DEAE or affinity (IgE column) chromatography. Detection of the IgEBF by ELISA and Western blot correlated well with activity in the previously employed rosette inhibition assay. Finally, purified IgEBF was found to suppress in vitro the production of IgE in the monoclonal human myeloma line U266, but not IgA or IgM-producing myelomas, providing evidence for the direct and specific regulatory action of this molecule on IgE-producing cells.

The ovalbumin (OVA)-specific T cell hybridoma 71B1, which constitutively secretes glycosylation inhibiting factor (GIF) and is specific for the immunogenic epitope represented by amino acids 323-339 in the OVA molecules, failed to form GIF having affinity for nominal antigen upon stimulation with OVA-pulsed antigen-presenting cells (APC). However, the GIF produced by the antigen-stimulated 71B1 cells bound to the mAb 14-12, which is specific for the antigen-binding chain of effector type suppressor T cell factor (TseF), and to mAb specific for TCR. The GIF constitutively released from unstimulated 71B1 cells failed to bind to any of these antibodies. Gel filtration of GIF preparations showed that the 14-12+ GIF from the antigen-stimulated 71B1 cells are composed of 80-100 and 25-35 kDa species, while the GIF from unstimulated cells was 12-15 kDa. Reduction and alkylation treatment of the GIF from the antigen-stimulated cells resulted in the disappearance of the 80-100 and 25-35 kDa GIF, which was accompanied by the formation of the 12-15 kDa GIF. Thus, the GIF from the antigen-stimulated 71B1 cells was similar to the previously described OVA-binding GIF from the 231F1 cells with respect to their antigenic structures and molecular size, and both factors appear to be composed of the 14-12+ polypeptide chain and 12-15 kDa non-specific GIF. However, the GIF from the antigen-stimulated 71B1 cells lacked affinity for the native OVA or synthetic peptide 323-339, and failed to suppress the in vivo antibody response to dinitrophenyl (DNP)-OVA. In contrast, the OVA-binding GIF has affinity for native OVA and the peptide 307-317, to which the cell source of the factor is specific, and suppressed the in vivo anti-hapten antibody response to DNP-OVA. The results suggest that formation of antigen-specific TsF is confined to T cells with certain epitope specificities. It was also found that the OVA-binding GIF failed to suppress the in vivo anti-hapten antibody response to DNP-conjugates of urea-denatured OVA (UD-OVA), which does not bind OVA-binding GIF. However, APC pulsed with UD-OVA appear to express the epitope 307-317 for which the OVA-binding GIF has affinity. The results collectively suggest that the affinity of GIF for an immunizing antigen, rather than processed antigen, is required for immunosuppression.

Human T cell hybridomas, which constitutively secrete glycosylation inhibiting factor (GIF), were constructed from PBL of an allergic individual who was sensitive to honey bee venom. PBMC of the patient were stimulated with either denatured or cyanogen bromide-treated bee venom phospholipase A2 (PLA2), and Ag-activated cells were propagated by IL-2 in the presence of human recombinant lipocortin I. T cells obtained in the cultures were fused with a HAT-sensitive mutant of the human lymphoblastoid cell line CEM. Approximately one-third of hybridoma clones constitutively secreted GIF. The GIF-producing hybridomas were CD3+ and bore TCR-alpha beta. GIF formed by unstimulated hybridomas lacked affinity for bee venom PLA2. Upon cross-linking of CD3, however, a majority of the GIF-producing hybridomas formed IgE-binding factors and GIF, the latter of which had affinity for bee venom PLA2. Both nonspecific GIF and Ag-binding GIF from the hybridomas bound to an immunosorbent coupled with the anti-lipomodulin mAb 141-B9. Using an affinity-purified GIF as an immunogen, we established mouse B cell hybridomas that secreted monoclonal anti-human GIF. In order to characterize human nonspecific GIF, one of the GIF-producing hybridomas was adapted to a serum-free medium, and culture supernatant was fractionated by DEAE-Sepharose column chromatography and by gel filtration. The majority of nonspecific GIF in the culture supernatant was recovered from DEAE-Sepharose by elution of the column with 10 mM Tris-HCl buffer, pH 8.0, containing 50 mM NaCl. Affinity-purification of GIF in the DEAE Sepharose fraction by using anti-GIF-coupled Affigel, and analysis of the purified GIF by SDS-PAGE revealed that human GIF is a single polypeptide chain of 14 to 15 kDa. Gel filtration of both crude and affinity-purified GIF preparations confirmed the molecular size of the cytokine.

The glycosylation-inhibiting factor (GIF) was isolated from serum-free culture supernatants of the murine T-cell hybridoma, 231F1 cells, by using an immunosorbent coupled with the monoclonal anti-lipomodulin antibody. The isolated lymphokine is a 14-kDa protein with a pI of 5.5, as determined by SDS/PAGE and two-dimensional gel electrophoresis. Fractionation of a mixture of radiolabeled GIF with culture supernatant of the 231F1 cells on ion-exchange and reverse-phase columns and by gel filtration demonstrated homogeneity of the 14-kDa GIF and confirmed that the bioactivity of GIF and the antigenic determinant recognized by the monoclonal anti-GIF antibody are associated with the 14-kDa protein. The 125I-labeled 14-kDa protein binds to the murine T-cell hybridoma 12H5 cells, which have been used for bioassay of GIF, and the murine B-cell line A20.3 cells, but the binding of the protein to resting murine splenic lymphocytes was barely detectable. Under the same experimental conditions, binding of the 125I-labeled recombinant human lipocortin I to the 12H5 cells was not detectable. In contrast, the 125I-labeled lipocortin, but not the 14-kDa GIF, bound to phosphatidylserine vesicles. The results indicate that GIF does not belong to the anexin family.

The glycosylation inhibiting factor (GIF) was detected in EGTA extracts of the OVA-specific Ts cell hybridoma, 231F1 cells and 71B4 cells, which constitutively secrete GIF. The lymphokine in both culture supernatants and EGTA extracts failed to bind to OVA-Sepharose. Association of GIF with the plasma membrane was confirmed by surface labeling of the 231F1 cells with 125I. The major species of GIF in the extract was 14.4-kDa peptide as determined by SDS-PAGE, and was identical to that detected in culture supernatants. Pretreatment of the cells with monoclonal anti-GIF switched the cells from the formation of unglycosylated IgE-BF to the formation of glycosylated IgE-BF, indicating that the membrane-associated GIF is involved in the determination of the nature of IgE-binding factor during their biosynthesis. When the hybridoma was stimulated with OVA-pulsed APC, EGTA extracts of the cells contained GIF having affinity for OVA. The binding of the OVA-binding GIF in the EGTA extracts to OVA-Sepharose was inhibited by a synthetic peptide, which corresponds to amino acid residues 307-317 in the OVA molecule and represents the epitope recognized by TCR on the cells. The OVA-binding GIF in the extracts bound to the monoclonal anti-TCR-alpha chain, H-28-710 and the mAb 14-12, which is specific for the Ag-binding chain of effector type suppressor factor, and suppressed the in vivo antibody response of BDF1 mice to DNP-OVA in a carrier-specific manner. Evidence was obtained that indicated that the Ag-binding chain was associated with nonspecific GIF chain on the cell surface of the Ag-stimulated cells.

Peripheral blood mononuclear cells of patients allergic to honey bee venom were stimulated with denatured bee venom phospholipase A2, and the antigen-activated T cells were propagated for 4 days by human IL-2 in the presence or absence of recombinant human lipocortin I. Upon antigenic stimulation with the denatured phospholipase A2 and autologous monocytes or by cross-linking of CD3 by anti-CD3 antibody, the activated T cells, which had propagated by IL-2 alone, formed N-glycosylated IgE-binding factors and glycosylation enhancing factor (GEF), while those propagated in the presence of lipocortin formed unglycosylated IgE-binding factors and glycosylation inhibiting factor (GIF). The GEF and GIF formed by the antigen- or anti-CD3-stimulated T cells had affinity for bee venom phospholipase A2 and could be purified by using anti-lipomodulin Sepharose. In the mouse lymphocyte system, the major cell source of GIF is antigen-specific suppressor T cells, and the antigen-binding GIF from the cells suppressed the in vivo antibody response in an antigen (carrier)-specific manner. In view of the findings in the mouse system, the present results may provide an immunological maneuver to generate allergen-specific suppressor T cells, and to obtain allergen-specific suppressor factor from T cell populations in the peripheral blood of allergic patients.

Rat monoclonal antibody against mouse glycosylation inhibiting factors was obtained, and radiolabeled glycosylation inhibiting factors from the mouse T-cell hybridoma, 231F1 cells, were purified by using the monoclonal antibody and antilipomodulin antibody. Analysis of the affinity-purified lymphokine by PAGE demonstrated two proteins of 14.4 kDa and 41 kDa. Both proteins migrated similarly under the reduced and unreduced conditions, indicating that each of the two species consist of a single polypeptide chain. Biologic activity of the lymphokine was recovered by extraction of the proteins from gel slices followed by renaturation. Evidence was obtained that suggested the 14.4-kDa peptide was a degradation product of the 41-kDa molecules. The 14.4-kDa peptide was also recovered by extraction of the surface-labeled 231F1 cells with Ca2+ chelator, indicating that the glycosylation inhibiting factor is associated with the cell surface.

The mouse T cell hybridoma 12H5 cells constitutively form glycosylation-enhancing factor (GEF) and produce both IgE-potentiating factor and ovalbumin (OVA)-binding GEF upon antigenic stimulation with OVA-pulsed macrophages. Culture of the 12H5 cells either with nonspecific glycosylation inhibiting factor (GIF) or with a phospholipase A2 (PLA2) inhibitor, ONO-RS-082, stopped the formation of GEF and induced the same cells to form GIF. Induction of the GIF formation by a PLA2 inhibitor was observed even when the 12H5 cells had been treated with mitomycin C, indicating that the switching from the GEF formation to the GIF formation was not due to selective proliferation of a GIF-producing subclone. The OVA-binding GIF produced by the PLA2-inhibitor-treated, antigen-stimulated 12H5 cells binds to homologous antigen (ovalbumin), and shares both antigenic determinant recognized by the monoclonal antibody 14-30 and the lipomodulin-determinant with antigen-specific suppressor inducer factor (TsiF). The present experiments also showed that a typical helper T cell clone, D10, G4.1 cells, constitutively formed GEF and that preculture of the T cell clone with IL-2 and the PLA2 inhibitor switched the cells from the formation of GEF to the formation of GIF. Upon stimulation with antigen-pulsed macrophages, the inhibitor-treated D10.G4.1 cells formed GIF having affinity for conalbumin. The results indicated that the same T cells have the capacity to form either GIF or GEF under different conditions, and suggested that the GIF-producing suppressor T cells may be a phenotype of a subset of helper T cells. Switching of the same cells from the GEF formation to the GIF formation by the PLA2 inhibitor and the ability of the inhibitor to enhance GIF formation suggested that PLA2-inhibitory activity or GIF activity of TsiF is involved in the suppressor T cell cascade.

Stimulation of normal rat splenic T cells with pertussigen (lymphocytosis-promoting factor, LPF, from Bordetella pertussis) resulted in the release of a soluble factor that enhanced the glycosylation of IgE-binding factors during their biosynthesis. The soluble factor was detected by the ability of a culture filtrate of LPF-stimulated spleen cells to switch a T cell hybridoma, 23A4, from the formation of unglycosylated IgE-binding factor to the formation of glycosylated IgE-binding factor. The glycosylation-enhancing factor (GEF) had affinity for D-galactose, and the binding of the factor to hybridoma cells via a cell surface galactose was essential for modulation of IgE-binding factors. The GEF was inactivated by irreversible inhibitors of serine proteases such as phenylmethylsulfonyl fluoride, diisopropylfluorophosphate, and p-nitrophenyl ethylpentylphosphonate but was not affected by nonphosphorylating analogues of the organophosphorus compounds. Benzamidine, a competitive and reversible inhibitor of trypsin, also inhibited the glycosylation of IgE-binding factors by GEF. The factor could be purified by absorption to p-aminobenzamidine agarose followed by elution with benzamidine. The capacity of GEF to enhance the glycosylation of IgE-binding factors was inhibited by synthetic substrates of trypsin but not by substrates of chymotrypsin, indicating that GEF is a trypsin-like enzyme. Indeed, trypsin, plasmin, and kallikrein enhanced the glycosylation of IgE-binding factors during their biosynthesis. An inhibitor of trypsin-like enzyme(s), N-alpha-p-tosyl-L-lysine chloromethylketone (TLCK), inhibited trypsin and plasmin but not kallikrein, and TLCK failed to inhibit the GEF-mediated enhancement of glycosylation. It was also found that bradykinin, the biologically active product of cleavage of kininogen by kallikrein, enhanced the glycosylation of IgE-binding factors. The results indicate that GEF is a kallikrein-like enzyme.

Peripheral blood mononuclear cells of human immunodeficiency virus infected patients were cultured in the presence or absence of 10 micrograms/ml human IgE, and culture filtrates were fractionated on IgE-Sepharose. IgE-binding factors were assessed in the acid eluate fraction from IgE-Sepharose by both rosette inhibition assay and radioimmunoassay. Among 22 cases studied, mononuclear cells from 9 patients formed IgE-binding factors in the absence of IgE, and those of 13 cases formed the factors upon incubation with IgE. The major cell source of IgE-binding factors was T cells. In 2 cases of human immunodeficiency virus infection, Fc epsilon receptors were detected on both Leu-2+ and Leu-3+ T cells.

BDF1 mice were immunized with a protein antigen, such as ovalbumin (OA) or keyhole limpet hemocyanin (KLH), absorbed to aluminum hydroxide gel, and their spleen cells were stimulated by homologous antigen for the formation of glycosylation-enhancing factor (GEF). It was found that GEF obtained from OA-primed spleen cells had affinity for OA, whereas those derived from KLH-primed spleen cells had affinity for KLH. Nonspecific GEF, which was obtained by stimulation of normal spleen cells with pertussis toxin, failed to bind OA or KLH. Both antigen-specific GEF and nonspecific GEF are inactivated by phenylmethylsulfonyl fluoride, but not by N-alpha-p-tosyl-L-lysyl-chloromethyl ketone. Both factors can be partially purified by binding to p-aminobenzamidine agarose and elution with benzamidine. These findings suggest that not only non-specific GEF but also antigen-specific GEF are serine protease(s). The antigen-specific GEF consisted of two m.w. species, of 65 to 85 kilodaltons (Kd) and 40 to 55 Kd, whereas nonspecific GEF consisted of 50 to 70 Kd and 20 to 30 Kd molecules. The OA-specific GEF augmented the in vitro secondary indirect PFC response of DNP-OA-primed cells to the homologous antigen, but failed to affect the PFC response of DNP-KLH-primed cells to DNP-KLH. Similarly, KLH-specific GEF enhanced the response of DNP-KLH-primed cells but not the response of DNP-OA-primed cells. However, OA-specific GEF failed to replace the requirement for antigen-primed helper T cells. Antigen-specific GEF bound to alloantibodies reactive to the products of the I region of the major histocompatibility complex. The results collectively suggest that antigen-specific GEF is identical to antigen-specific augmenting factors described by other investigators.

BDF1 [(C57BL/6 X DBA/2)F1] mice were primed with alum-absorbed ovalbumin, and their spleen cells were cultured with ovalbumin to activate antigen-primed T cells. The activated T cells were then propagated in interleukin 2-containing medium in the presence or absence of affinity-purified glycosylation-inhibiting factor (GIF). Upon incubation with ovalbumin-pulsed macrophages, T cells propagated in the absence of GIF produced IgE-potentiating factor and glycosylation-enhancing factor. In contrast, T cells propagated in the presence of GIF produced IgE-suppressive factor and GIF. The ovalbumin-primed T cells propagated in the presence of GIF constitutively produced the 13-kDa GIF that lacked affinity for ovalbumin. However, stimulation of the same T cells with ovalbumin-pulsed macrophages resulted in the production of 80- and 35-kDa GIF that had affinity for ovalbumin. Both the antigen-specific GIF and nonspecific GIF from the activated BDF1 T cells had I-Jb determinants. Since the ovalbumin-specific GIF is derived from Lyt-2+, I-J+ ovalbumin-specific suppressor T cells and suppresses the anti-hapten antibody response to dinitrophenyl-coupled ovalbumin, the results strongly suggest that the presence of GIF during the propagation of antigen-primed T cells facilitates the generation of antigen-specific suppressor T cells.

Incubation of rat-mouse T cell hybridoma cells, 23B6, with rat immunoglobulin E (IgE) results in the formation of the 15,000-dalton IgE-suppressive factor and the 30,000-dalton IgE-binding factor, which has neither potentiating activity nor suppressive activity on the IgE response. Another T cell hybridoma, 23A4 cells, produces the 30,000-dalton "inactive" IgE-binding factor upon incubation with IgE. Both the 15,000-dalton IgE-suppressive factor and the 30,000-dalton IgE-binding factor lacked affinity for lentil lectin but bound to peanut agglutinin. When the 23B6 cells were incubated with IgE in the presence of lysolecithin, the majority of the 15,000-dalton IgE-binding factor formed by the cells gained affinity for lentil lectin, and this factor selectively potentiated the IgE response. The glycosylation-enhancing factor, which was formed by stimulation of normal spleen cells with lymphocytosis-promoting factor (LPF or pertussigen), also switched 23B6 cells from the formation of IgE-suppressive factor to the formation of IgE-potentiating factor. It was also found that the 30,000-dalton "inactive" IgE-binding factor, formed by both 23B6 and 23A4 cells, gained the ability to potentiate the IgE response, when the cells were cultured with IgE in the presence of glycosylation-enhancing factor. The results indicate that IgE-potentiating factor and IgE-suppressive factor share common precursors, and that biologic activities of IgE-binding factors are decided by their carbohydrate moieties. Incubation of the two hybridoma cells with lysolecithin or glycosylation-enhancing factor results in an increase in the proportion of FC epsilon R+ cells, suggesting that the assembly of N-linked oligosaccharide to precursor molecules is intrinsic for the expression of FC epsilon R.

Cells of the T cell hybridoma 23A4 produce IgE-binding factors lacking N-linked oligosaccharides (unglycosylated form) when they are incubated with IgE alone. In the presence of glycosylation-enhancing factor (GEF) or bradykinin, however, the same cells produce IgE-binding factors with N-linked oligosaccharides (glycosylated form). Switching the cells from the formation of unglycosylated IgE-binding factors to the formation of glycosylated factors was accompanied by the release of both glycosylation-inhibiting factor (GIF) in its phosphorylated form, i.e., phosphorylated lipomodulin, and arachidonate from the cells. Analysis of the biochemical processes for the release of GIF from 23A4 cells showed that affinity-purified GEF or bradykinin induced transient phospholipid methylation and diacylglycerol (DAG) formation, and enhanced 45Ca uptake into the cells. Inhibitors of methyltransferases, i.e., 3-deaza-adenosine plus L-homocysteine thiolactone, inhibited not only phospholipid methylation but also DAG formation and GIF release. Exogenously added 1-oleoyl-2-acetyl glycerol, i.e., a DAG that is permeable to the plasma membrane, induced the release of GIF from the cells. It was also found that 12-O-tetradecanoyl-phorbol 13-acetate (TPA) switched 23A4 cells and normal lymphocytes to the selective formation of N-glycosylated IgE-binding factor, and induced the release of GIF from the cells. 32PO4-labeled lipomodulin was detected in the extract of 23A4 cells 3 to 5 min after the addition of GEF, bradykinin, or TPA. These results indicate that GEF and bradykinin induced the activation of methyltransferases and phospholipase C for the formation of DAG, which in turn activated Ca2+-activated, phospholipid-dependent protein kinase (protein kinase C) for the phosphorylation of lipomodulin. Because lipomodulin loses phospholipase inhibitory activity after phosphorylation, increased phospholipase A2 activity would be expressed by this process.

The T cell hybridomas 231F1 and 12H5 constitutively secrete glycosylation-inhibiting factor (GIF) and glycosylation-enhancing factor (GEF), respectively, which lack affinity for OVA-coupled Sepharose. When the 231F1 and 12H5 cells were stimulated by OVA-pulsed syngeneic macrophages, however, GIF and GEF produced by the cells had affinity for OVA. Both the OVA-binding GIF from the 231F1 cells and OVA-binding GEF from the 12H5 cells bound to a mAb against TCR-alpha beta and a mAb against TCR-alpha, suggesting a serologic relationship between TCR and OVA-binding factors. However, the OVA-binding GIF and GEF bound to mAb 14-12 and 14-30, respectively. Because these mAb do not bind TcR alpha beta-chains, it appears that the Ag-binding factors are different from TCR itself. The OVA-binding factors from both 12H5 cells and 231F1 cells do not bind to urea-denatured OVA. The binding of the factors to OVA Sepharose was inhibited by a peptide corresponding to residues 307-317 (P307-317) in the native OVA, but not by the peptide corresponding to residues 323-339 (P323-339). Furthermore, the OVA-binding factors bound to P306-319-coupled Sepharose but not to P323-339-coupled Sepharose, and were recovered by elution of the former Sepharose at acid pH. The binding of OVA to anti-OVA antibodies was not inhibited by either peptide. Inasmuch as the 231F1 cells and 12H5 cells can be stimulated by P307-317 in the context of a MHC product, it appears that the Ag-binding factors and TCR-alpha beta on the cell sources of the factors may recognize the same epitope in the OVA molecules. The results also showed that Ag-binding factors and antibodies recognize distinct epitopes in the Ag molecules.

Since our previous experiments suggested that glycosylation-inhibiting factor (GIF) is a phosphorylated derivative of a phospholipase inhibitory protein, we determined whether other well-known phospholipase inhibitors may have similar biological activities. The results showed that phospholipase A2 (PLA2) inhibitors, such as recombinant human lipocortin I and ONO-RS-082, could switch T cell hybridoma 12H5 cells from the formation of glycosylated IgE-binding factors (IgE-BF) to the formation of unglycosylated IgE-BF, whereas neomycin, a phospholipase C inhibitor, failed to affect the nature of IgE-BF formed by the cells. The minimum concentrations of lipocortin I and ONO-RS-082 required for switching the 12H5 cells to the formation of unglycosylated IgE-BF were comparable to or less than IC50 of the inhibitors for PLA2. The ability of partially purified GIF to switch the 12H5 cells to the formation of unglycosylated IgE-BF was markedly enhanced by treatment of the preparation with alkaline phosphatase. It was also found that lipocortin I and ONO-RS-082, but not neomycin, facilitated the generation of GIF-producing T cells. When spleen cells of ovalbumin (OVA)-primed BDF1 mice were stimulated with homologous antigen and the activated T cells were propagated by recombinant IL-2 in the presence of GIF, lipocortin I, or ONO-RS-082, T cells obtained in the cultures constitutively produced their own GIF. Antigenic stimulation of the T cells induced the formation of unglycosylated IgE-BF and GIF with an affinity for OVA.(ABSTRACT TRUNCATED AT 250 WORDS)

Stimulation of normal rat splenic T cells with pertussigen (lymphocytosis-promoting factor from Bordetella pertussis) resulted in the release of a soluble factor that enhanced the assembly of N-linked oligosaccharides to IgE-binding factors during their biosynthesis. The glycosylation-enhancing factor (GEF) is a kallikrein-like enzyme and is purified by absorption to p-aminobenzamidine-Agarose followed by elution with benzamidine. Incubation of normal mouse mast cells with affinity-purified GEF or bradykinin, a product of cleavage of kininogen by kallikrein, resulted in the release of histamine and arachidonate from the cells. Passive sensitization of mast cells with mouse IgE antibody, followed by pretreatment of the cells with a suboptimal concentration of GEF, resulted in an enhancement of antigen-induced histamine release. It was found that GEF and bradykinin induced the same biochemical events in mast cells as those induced by bridging of IgE receptors. Both GEF and bradykinin induced phospholipid methylation and an increase in intracellular cyclic AMP (cAMP). Incorporation of 3H-methyl groups into phospholipids and intracellular cAMP levels both reached a maximum 30 sec after challenge with GEF or bradykinin, and then declined to base-line levels within 2 to 3 min. These biochemical events were followed by 45Ca influx and histamine release; 45Ca uptake reached a plateau value at 2 min, and histamine release reached a maximum at 5 to 8 min. The initial rise in cAMP induced by GEF (or bradykinin) was not inhibited by indomethacin, indicating that the activation of adenylate cyclase is not the result of prostaglandin synthesis. In both IgE-mediated and GEF-induced histamine release, inhibitors of methyltransferases, such as 3-deaza adenosine and L-homocysteine thiolactone, inhibited not only phospholipid methylation but also the cAMP rise and subsequent Ca2+ uptake and histamine release. The results indicate that GEF induces activation of methyltransferases and that phospholipid methylation is involved in the cAMP rise, Ca2+ uptake, and histamine release. The induction of the same biochemical events in the same sequence by bridging of IgE receptors and by GEF (bradykinin) supports the hypothesis that receptor bridging induces the activation of serine protease(s) and cleavage products of this enzyme in turn activate methyltransferases in mast cells.

Peripheral blood mononuclear cells of 10 out of 26 patients with human immunodeficiency virus type 1 (HIV-1) released IgE-binding factors, as determined by two independent assays. The formation of the factors by the mononuclear cells was enhanced by incubation of the cells with homologous IgE. In the presence of IgE, peripheral blood mononuclear cells from 15 of the 26 patients formed a detectable amount of IgE-binding factors, whereas those of normal individuals of allergic rhinitis patients failed to do so. The major source of IgE-binding factors was the T cells of the HIV-1-infected patients. The CD8+ T cells from a HIV-1-infected patient formed IgE-binding factors upon incubation with IgE, and type II receptors for Fc epsilon were detected on both CD4+ T cells and CD8+ T cells in one of five patients studied. It was also found that culture supernatants of mononuclear cells from HIV-1-infected patients released soluble factors that induce normal human T cells to form IgE-binding factors. The results suggest that lymphocytes of some HIV-1-infected patients are activated to produce lymphokines regulating formation of IgE-binding factors.

Upon antigenic stimulation with OVA-pulsed syngeneic macrophages, the mouse T cell hybridoma 231F1 produced glycosylation inhibiting factor (GIF) having affinity for OVA and IgE-suppressive factors, whereas another T cell hybridoma, 12H5, cells produced OVA-binding glycosylation enhancing factor (GEF) and IgE-potentiating factor. The OVA-binding GIF from the 231F1 cells is an Ag-specific Ts cell factor, whereas OVA-binding GEF from the 12H5 cells is an Ag-specific augmenting factor. Both hybridomas express CD3 complex and functional TCR-alpha beta. Cross-linking of TCR-alpha beta or CD3 molecules on the hybridomas by anti-TCR-alpha beta mAb or anti-CD3 mAb and protein A resulted in the formation of the same factors as those obtained by the stimulation of the cells with OVA-pulsed syngeneic macrophages. It was also found that both the 231F1 cells and 12H5 cells formed IgE-binding factors upon incubation with H-2d and H-2b APC, respectively, with a synthetic peptide corresponding to residues 307-317 in the OVA molecules (P307-317). Six other synthetic peptides, including those containing the major immunogenic epitope, i.e., P323-339, failed to stimulate the hybridomas in the presence of APC. Indeed, all of the 10 T cell hybridoma clones, which could produce either OVA-binding GIF or OVA-binding GEF, responded to P307-317 and APC for the formation of IgE-binding factors. In contrast, GIF/GEF derived from six other hybridoma clones, whose TCR recognized P323-339 in the context of a MHC product, failed to bind to OVA-coupled Sepharose. The results indicate the correlation between the fine specificity of TCR and the affinity of GIF/GEF to the nominal Ag. The amino acid sequence of P307-317 suggested that TCR on the cell sources of Ag-binding factors are specific for an external structure of the Ag molecules.

Glycosylation inhibiting factor (GIF) was purified from culture filtrates of a T cell hybridoma, 23A4, by affinity chromatography on anti-lipomodulin Sepharose. The factor exhibited phospholipase inhibitory activity upon dephosphorylation. Immunization of BDF1 mice with aluminum hydroxide gel (alum)-absorbed dinitrophenyl derivatives of ovalbumin (DNP-OA) resulted in persistent IgE and IgG antibody formation. However, repeated injections of the affinity-purified GIF into the DNP-OA-primed mice beginning on the day of priming prevented the primary anti-hapten antibody responses of both the IgE and the IgG1 isotypes. Treatment with GIF also diminished on-going IgE antibody formation in the DNP-OA-primed mice. The treatment changed the nature of IgE-binding factors formed by BDF1 spleen cells. Incubation of spleen cells from OA + alum-primed mice with OA resulted in the formation of IgE-potentiating factor, whereas spleen cells of OA-primed, GIF-treated mice formed IgE-suppressive factor upon antigenic stimulation. It was also found that Lyt-2+ T cells in the OA-primed, GIF-treated mouse spleen cells released GIF, which had affinity for OA and bore I-Jb determinant(s). Transfer of a Lyt-1+ cell-depleted fraction of the OA-primed, GIF-treated mouse spleen cells into naive syngeneic animals resulted in suppression of the primary anti-DNP IgE antibody response of the recipients to alum-absorbed DNP-OA, but failed to affect the anti-DNP antibody response to DNP-keyhole limpet hemocyanin. The results indicate that GIF treatment during the primary response to OA facilitated the generation of antigen-specific suppressor T cells.

cDNA clones encoding rodent IgE-binding factors (IgE-BF) were isolated from cDNA libraries of a rat-mouse T hybridoma that secretes IgE-suppressive factor (IgE-SF) upon incubation with rat IgE. COS7 cells transfected with two of the cDNAs expressed IgE-BF, which selectively potentiate an in vitro IgE response. IgE-BF expressed in COS7 cells are glycoproteins of approximately equal to 60 and approximately equal to 11 kDa. DNA sequence analysis of an IgE-BF cDNA revealed a 556-amino acid (62 kDa) protein coding region. The results suggest that IgE-potentiating and IgE-suppressive factors share common precursor polypeptides and that the 11-kDa IgE-BF is derived from a 60-kDa precursor.

The IgE synthesis is regulated by T cell derived IgE-binding factors in an isotype-specific manner. The IgE-potentiating and IgE-suppressive factors may share common structural genes; therefore, a common polypeptide chain and their biologic activities are determined by posttranslational glycosylation processes. Under the physiological conditions, the carbohydrate moieties in the IgE-binding factors formed by a subset of T cells are determined by the ratio between two T cell factors, i.e., glycosylation-enhancing factors and glycosylation-inhibiting factors (GIF), in their environment. GIF is an immunosuppressive lymphokine. Repeated injections of this lymphokine into antigen-primed mice facilitated the generation of antigen-specific suppressor T cells and suppressed both IgE and IgG antibody responses. This effect of GIF was reproduced in vitro. Activation of antigen-specific T cells by antigen-pulsed macrophages, followed by propagation of the antigen-activated T cells by interleukin 2 in the presence of GIF resulted in the generation of suppressor T cells which produced antigen-specific GIF upon antigenic stimulation. Some of the T cell hybridomas constructed from the antigen-specific suppressor T cells formed antigen-specific GIF upon antigenic stimulation. The antigen-specific GIF formed by the T cell hybridoma share several common properties with antigen-specific suppressive factor and suppressed the antibody response of syngeneic mice in a carrier-specific manner. The results obtained with mouse lymphocytes suggest a maneuver to suppress the IgE antibody formation to an allergen in atopic patients.

We previously established an ovalbumin (OA)-specific T cell clone from spleen cells of BDF1 mice, which had been treated by i.v. injections of OA, and constructed antigen-specific T cell hybridomas from the T cell clone. One of the hybridomas constitutively released glycosylation-inhibiting factor (GIF) which lacked affinity for OA, and was called non-specific GIF. Incubation of the same hybridoma cells with OA-pulsed syngeneic macrophages or OA-pulsed B lymphoblastoid cells of BALB/c origin resulted in the formation of GIF molecules that had affinity for OA but not for bovine serum albumin or keyhole limpet hemocyanin. Both the OA-specific GIF and nonspecific GIF bound to monoclonal anti-lipocortin and possessed I-Jb determinants. The OA-specific GIF consisted of two species of molecules, of m.w. 80,000 and 30,000 to 40,000, respectively, whereas the nonspecific GIF from unstimulated cells had an m.w. of 15,000. Intravenous injections of OA-specific GIF or nonspecific GIF into BDF1 mice suppressed both the IgE and IgG1 anti-hapten antibody responses of the animals to dinitrophenyl derivatives of OA (DNP-OA), but OA-specific GIF was much more effective than nonspecific GIF in suppressing the antibody responses. When the same preparations of GIF were injected into DNP-KHL-primed mice, OA-specific GIF and nonspecific GIF were comparable in suppressing the anti-DNP antibody response. In contrast to the 40,000 m.w. species of OA-specific GIF, the 80,000 m.w. OA-specific GIF had carrier-specific suppressive effects. The similarities of antigen-specific GIF to antigen-specific TsF suggest that the phospholipase-inhibiting activity of the molecules may be involved in the immunosuppressive effects of some antigen-specific TsF.

Attempts were made to generate Ag-specific suppressor T cells from Ag-primed spleen cells by using glycosylation inhibiting factor (GIF). BDF1 mice were primed with alum-absorbed OVA and their spleen cells were stimulated with OVA. Ag-activated T cells were then propagated in IL-2-containing conditioned medium. Incubation of the T cells with OVA-pulsed syngeneic macrophages resulted in the formation of IgE-potentiating factor and glycosylation-enhancing factor that has affinity for OVA, i.e., OVA-specific glycosylation-enhancing factor. However, if the same Ag-activated splenic T cells were propagated in the IL-2-containing medium in the presence of GIF T cells obtained in the cultures formed IgE-suppressive factors and OVA-specific GIF on antigenic stimulation. Thus we constructed T cell hybridomas from the Ag-activated T cells propagated by IL-2 in the presence of GIF. A representative hybridoma, 71B4, formed OVA-specific GIF on incubation with OVA-pulsed macrophages of BDF1 mice or C57B1/6 mice. However, if the same hybridoma cells were incubated with OVA alone or with OVA-pulsed macrophages of H-2k or H-2d strains, they produced GIF that had no affinity for OVA. The OVA-specific GIF bound to OVA-Sepharose but did not bind to BSA-Sepharose or KLH Sepharose. Intravenous injections of the OVA-specific GIF from the hybridoma suppressed the IgE and IgG1 anti-DNP antibody response of BDF1 mice to DNP-OVA, but failed to suppress the anti-hapten antibody responses of the strain to DNP-keyhole limpet hemocyanin, indicating that the factors suppressed the antibody response in a carrier-specific manner. However, the same OVA-specific GIF failed to suppress the anti-hapten antibody response of DBA/1 mice to DNP-OVA, suggesting that the immunosuppressive effects of the factors is MHC restricted.

Major research on the isotype-specific regulation of antibody responses was carried out in the IgE antibody response, and extended to the other isotypes. Because of dissociation between the IgE and IgG antibody responses, it was anticipated that each isotype may have distinct antigen-specific helper and suppressor T cells. However, current studies explain the dissociation by collaboration between antigen-specific (helper and suppressor) T cells and antigen-nonspecific, but isotype-specific T cells, which form soluble factors having the affinity for the isotype. In the IgE response, soluble factors having affinity for IgE, i.e., IgE-binding factors, are consisted of IgE-potentiating factors and IgE-suppressive factors. The main difference between the two IgE-binding factors appears to be carbohydrate moieties in the molecules. Accumulated evidence indicates that the same T cells have capacities to form both IgE-potentiating factors and IgE-suppressive factors, and suggests that the nature and biologic activities of the factors are decided in the process of the glycosylation of the same precursor molecules. Mechanisms for the regulation of the glycosylation of the factors will also be discussed in this review.

A biologic cell product that in an experimental system completely and and specifically suppresses the IgE response! This is one of the intriguing findings that give rise to optimism that the growing understanding of mechanisms that regulate normal IgE responses will, in the not-too-distant future, provide rational and effective therapy for the prevention and management of atopic diseases.

BDF1 mice were immunized with alum-absorbed OVA and T cell hybridomas were constructed from their splenic T cells. Many of the hybridomas constitutively produced glycosylation enhancing factor (GEF), which could switch the T cell hybridoma 23A4 cells from the formation of IgE-suppressive factors to the formation of IgE-potentiating factors. When one of the hybridoma clones, 12H5, was incubated with OVA-pulsed syngeneic or semi-syngeneic (H-2b) macrophages, the hybridoma produced GEF that have affinity for OVA, but not for either keyhole limpet hemocyanin or BSA. However, the same hybridoma constitutively produced nonspecific GEF, that lacked affinity for OVA. Upon incubation with OVA-pulsed macrophages, the same hybridoma produced both IgE-potentiating factors and IgG-potentiating factors which selectively enhance the IgE response and IgG response, respectively. Both Ag-specific GEF and nonspecific GEF from the hybridoma bind to p-aminobenzamidine-agarose, and are recovered by elution with benzamidine. It was also found that both OVA-specific GEF and nonspecific GEF from the hybridoma induced the release of arachidonic acid from phospholipids of mouse fibrosarcoma cell line, HSDM1C1 cells. GEF formed by the 12H5 hybridoma bound to alloantibodies reactive to the product(s) of the I-Ab subregion of major histocompatibility complex. The Ag-specific GEF consisted of two Mr species, of 70 to 90 kDa and 50 to 60 kDa, whereas nonspecific GEF consisted of 50 to 60 kDa and 25 to 30 kDa molecules. Reduction and alkylation treatment of the OVA-specific GEF resulted in the formation of nonspecific GEF, suggesting that Ag-specific GEF is composed of Ag-binding polypeptide chain and nonspecific GEF.

Glycosylation enhancing factor (GEF) from rat T cells is a kallikrein-like enzyme and enhances the assembly of N-linked oligosaccharides to IgE binding factors during their biosynthesis, whereas another T cell factor, i.e., glycosylation inhibiting factor (GIF), is a fragment of phosphorylated lipomodulin (i.e., phospholipase inhibitor), which when dephosphorylated inhibits phospholipase and the glycosylation process. The two T cell factors compete with each other when they are added to normal mesenteric lymph node cells during the formation of IgE binding factors. The addition of GEF to T cell hybridoma 23A4 cell switches the cells from the formation of unglycosylated IgE binding factor to the formation of N-glycosylated IgE binding factor. However, GEF neither inactivated GIF nor inhibited the formation of GIF by the T cell hybridoma. Stimulation of the T cell hybridoma with either affinity-purified GEF or bradykinin resulted in the release of GIF from the cells. GIF released by GEF stimulation had a m.w. of approximately 15,000 and bound to monoclonal antibody against lipomodulin. GEF and bradykinin also induced normal mesenteric lymph node cells to release GIF. Incorporation of 14C-arachidonic acid into 23A4 cells, followed by stimulation of the cells with GEF, resulted in the release of 14C-arachidonate. The results suggest that lipomodulin, a phospholipase inhibitory protein, is present in lymphocytes, and indicate that GEF and bradykinin induce the activation of phospholipase by stimulating cells to release lipomodulin.

The IgE-potentiating and IgE-suppressive factors share a common structural gene and therefore a common polypeptide chain, and their biologic activities are decided by a post-translational glycosylation process. Under physiological conditions, this process is controlled by two T cell factors, i.e., the glycosylation-enhancing factor (GEF) and glycosylation-inhibiting factor (GIF). GIF is a fragment of phosphorylated lipocortin and has immunosuppressive effects. Repeated injections of this lymphokine into antigen-primed mice switched their T cells from the formation of IgE-potentiating factor to the formation of IgE-suppressive factor and facilitated the generation of antigen-specific suppressor T cells, which form antigen-specific GIF upon antigenic stimulation. The antigen-specific GIF suppressed the antibody response in a carrier-specific manner and has properties similar to antigen-specific suppressor T cell factors.