Selective depletion of myelin-reactive T cells with the anti-OX-40 antibody ameliorates autoimmune encephalomyelitis.

The OX-40 protein was selectively upregulated on encephalitogenic myelin basic protein (MBP)-specific T cells at the site of inflammation during the onset of experimental autoimmune encephalomyelitis (EAE). An OX-40 immunotoxin was used to target and eliminate MBP-specific T cells within the central nervous system without affecting peripheral T cells. When injected in vivo, the OX-40 immunotoxin bound exclusively to myelin-reactive T cells isolated from the CNS, which resulted in amelioration of EAE. Expression of the human OX-40 antigen was also found in peripheral blood of patients with acute graft-versus-host disease and the synovia of patients with rheumatoid arthritis during active disease. The unique expression of the OX-40 molecule may provide a novel therapeutic strategy for eliminating autoreactive CD4+T cells that does not require prior knowledge of the pathogenic autoantigen.

Cytotoxicity of a cell-reactive immunotoxin containing ricin A chain is potentiated by an anti-immunotoxin containing ricin B chain.

In vitro killing of the human Daudi cell line by either univalent [F(ab')] or divalent (IgG) forms of rabbit anti-human Ig (RAHIg) coupled to ricin A chain can be specifically potentiated by a "piggyback" treatment with ricin B chain coupled to goat anti-rabbit Ig (GARIg). When cells are treated with univalent immunotoxin (IT) [F(ab') RAHIg-A] and then cultured, IT can be detected on the cell surface for at least 5 h, since GARIg-B can still enhance killing at this time. These results provide a strategy for in vivo use of A chain- and B chain-containing IT.

Molecular dissection of the murine antibody response to streptococcal group A carbohydrate.

Monoclonal antibodies (mAb) to streptococcal group A carbohydrate (GAC) are encoded by a minimum of two VH, four JH, four V kappa, three J kappa, one V lambda, and one J lambda gene segments. The IdX, IdI-1, and Id5 idiotypic determinants are expressed by anti-GAC mAb and are found on free kappa chains. Each pattern of these determinants is encoded by a distinct V kappa gene segment, apparently without the requirement for a particular J kappa, VH, or JH gene segment, or somatic mutation. In contrast, the binding site-associated idiotypic determinant IdI-3a does not correlate with any single V or J gene segment.

Redesigning nature's poisons to create anti-tumor reagents.

Immunotoxins are conjugates of cell-reactive antibodies and toxins or their subunits. In this report, the chemistry, biology, pharmacokinetics, and anti-tumor effects of first generation immunotoxins; the preparation of improved second generation immunotoxins that display greater anti-tumor efficacy; and the role of genetic engineering in creating third-generation immunotoxins are discussed.

In vivo therapy of the BCL1 tumor: effect of immunotoxin valency and deglycosylation of the ricin A chain.

The in vivo therapeutic effects of Fab' and IgG anti-delta (anti-IgD)-ricin A chain immunotoxins were compared in mice bearing the surface IgD-positive BCL1 leukemia. The immunotoxins were prepared with either native or deglycosylated ricin A chain. Immunotoxin therapy was assessed both by the number of cells bearing the BCL1 immunoglobulin idiotype which remained in the spleen 24-48 h after injection with immunotoxin and by adoptive transfer of these spleen cells into normal mice. Immunotoxins prepared with either Fab'-anti-delta or IgG-anti-delta and native A chain induced a dose-dependent reduction in the number of idiotype-positive BCL1 cells present in the spleens of the tumor-bearing mice. The maximal therapeutic response was achieved with 250 micrograms of immunotoxin (containing 80-100 micrograms A chain) per mouse for both immunotoxins, resulting in tumor reduction of approximately 90%. This represents an elimination of 3-4 X 10(8) tumor cells. Reduction in the number of tumor cells was not observed with control reagents including antibody alone, antibody mixed with A chain, or an immunotoxin of irrelevant specificity. A Fab' immunotoxin prepared with deglycosylated ricin A chain was approximately 5-fold more effective as an antitumor reagent than the same immunotoxin prepared with native A chain; thus, optimal therapy was achieved after injection of 50 micrograms of immunotoxin (containing 15-20 micrograms A chain). Since the immunotoxins prepared with deglycosylated A chain were only 2-3-fold more toxic to the mice than those prepared with native A chain, the former resulted in a 2-3-fold increase in the therapeutic index.

Pharmacokinetics of tumor-reactive immunotoxins in tumor-bearing mice: effect of antibody valency and deglycosylation of the ricin A chain on clearance and tumor localization.

The blood clearance and tissue distribution of immunotoxins composed of either intact or Fab' fragments of tumor-reactive, monoclonal rat antimurine immunoglobulin D (anti-delta) or nonreactive, normal rat immunoglobulin G and ricin A chain (native or chemically deglycosylated) were determined in BCL1 tumor-bearing mice. In the presence of accessible target cells, neither the valency of the antibody nor deglycosylation of the A chain affect the initial rate of clearance (alpha phase) of immunotoxins from the blood; however, both factors affect the levels of tumor-specific and nonspecific localization of the immunotoxins. Thus, the use of immunotoxins with deglycosylated A chain greatly reduced the levels of nonspecific uptake by the liver and concomitantly increased tumor-specific localization. Immunotoxins prepared with Fab' fragments of anti-delta antibody and deglycosylated A chain were twice as effective at localizing to splenic tumor than immunotoxins prepared with intact immunoglobulin G and deglycosylated A chain. Approximately 90% of tumor-specific localization of anti-delta immunotoxins occurred within 1 h after injection and less than 5% of the immunotoxin remained in the circulation 4 h after injection. In addition, the antigen-binding capacity of the remaining circulating immunotoxins decreased in a linear manner over the first 10 h to approximately 20% of the initial binding activity. Thus, by 10 h after injection, only 2-3% of injected immunotoxin remained in the circulation and this material expressed little antigen reactivity. Analysis of the in vivo stability of circulating 125I-labeled immunotoxins in tumor-bearing mice demonstrated that Fab' immunotoxins are more stable than IgG immunotoxins prepared with N-succinimidyl-3-(2-pyridylthio)propionate. In BCL1-bearing mice, significant splitting of the anti-delta immunotoxins did not occur until tumor localization was virtually complete.

Characterization of a plasma membrane-associated plasminogen activator on thymocytes.

Plasma membranes isolated from normal thymocytes of hamster and rats were found to exhibit neutral protease activity toward 125I-labeled casein. The plasma membrane-associated proteases were completely inhibited by the serine protease inhibitors, diisopropyl fluorophosphate, phenylmethylsulfonyl fluoride and p-nitrophenyl-p-guanidinobenzoate, partially inhibited by soybean trypsin inhibitor and antipain, but were only weakly inhibited by L-1-tosylamino-2-phenylethyl chloromethyl ketone. The plasma membrane-associated proteases were also completely inhibited by ZnCl2 (75--100 mu M), but they were not affected by several other divalent cations. The plasma membrane fraction contained a plasminogen activator activity which was specifically localized in this fraction. The plasma membrane-associated plasminogen activator activity was inhibited by all of the inhibitors which inhibited plasma membrane-associated proteases except L-1-tosylamido-2-phenylethyl chloromethyl ketone. Labeling of plasma membrane-associated serine esterases with [3H] diisopropyl fluorophosphate followed by separation of the proteins by sodium dodecyl sulfate polyacrylamide gel electrophoresis revealed that this fraction contained a single major 3H-labeled protein of Mr 105 000. Both the plasminogen activator and the Mr 105 000 esterase were shown to be glycoproteins by affinity chromatography on lentil lectin-Sepharose. These results indicate that the plasminogen activator of thymocytes is a glycosylated serine protease with an active site-containing subunit of Mr 105 000 which is specifically localized in the plasma membrane.

C-type natriuretic peptide mediates the hypothalamic actions of the natriuretic peptides to inhibit luteinizing hormone secretion.

Both A- and C-type natriuretic peptides (ANP and CNP, respectively) significantly reduce LH secretion when injected into the third cerebral ventricle of conscious rats. To establish which natriuretic peptide receptor subtype transduces these inhibitory messages, we have employed novel cytotoxin cell targeting techniques to selectively destroy cells in the hypothalamus that respond to ANP or CNP. Rats pretreated with ANP conjugated to the toxic A-chain of the plant cytotoxin ricin failed 1 week later to respond to central injection of ANP with the normal inhibition of LH secretion. These rats did, however, respond with significant inhibition of LH secretion to central injection of CNP. In fact, the LH inhibition observed after CNP injection was significantly greater than that expressed after similar injection of CNP in rats pretreated with unconjugated ricin A-chain (toxin control). Those control rats displayed significant reduction of LH levels in response to ANP injection as well. Plasma LH levels were not significantly affected by central administration of either ANP or CNP in rats pretreated with ricin A-chain conjugated to CNP. These results further demonstrate the power of this novel technology and provide positive evidence supporting our hypothesis that ANP exerts its LH-inhibiting effect by displacing endogenous CNP from clearance receptors within the brain. This endogenous CNP, then, like exogenously applied CNP, activates the guanyl cyclase-B receptors on cells, which are part of the network controlling the release of LHRH.

A nonoxytocinergic prolactin releasing factor and a nondopaminergic prolactin inhibiting factor in bovine neurointermediate lobe extracts: in vitro and in vivo studies.

Several peptidergic PRL-releasing factors (PRFs) have been described; however, none have been proven to be of primary physiological importance in the control of hormone release. Similarly dopamine withdrawal alone cannot completely explain the profiles of PRL secretion observed under a variety of conditions. We describe here the isolation in semipurified form of both a PRF and a PRL-inhibiting factor (PIF) from bovine neurointermediate lobe (NIL) extracts. Acid extracts of bovine NILs stimulated, in a dose-related manner, PRL release from cultured anterior pituitary cells, even after immunoabsorption of endogenous oxytocin from the extract. PIF and PRF activities were semipurified from NIL extracts by Sephadex chromatography and detected by in vitro and in vitro bioassays. The PRF material could be separated from oxytocin by gel sieving and was active in the presence of dopamine in vitro unlike synthetic oxytocin and in cell preparations in which the oxytocin-responsive lactotrophs had been removed by selective cytotoxin cell targeting using an oxytocin-ricin A chain cytotoxic conjugate. The PRF material stimulated PRL secretion in a dose-dependent fashion in conscious male rats after iv injection. The PIF material comigrated on sizing gel chromatography with immunoreactive oxytocin and was active in vitro during dopamine blockade with domperidone and in vivo in the presence of endogenous dopaminergic tone. These data suggest that novel factors present in the NIL might exert physiologically relevant control over lactotroph function and add to the growing literature on the presence of a PRF in the NIL.

Opposing neuroendocrine actions of the natriuretic peptides: C-type and A-type natriuretic peptides do not interact with the same hypothalamic cells controlling prolactin secretion.

The two major members of the family of natriuretic peptides (NPs) in brain, A-type natriuretic peptide (ANP) and C-type natriuretic peptide (CNP) exert opposing actions on the neuroendocrine regulation of prolactin (PRL) secretion. We have targeted for compromise and destruction cells within the diencephalon which bear receptors for ANP (NRP-A receptors), CNP (NRP-B receptors), or both peptides (NPR-C receptors) using novel cytotoxin cell targeting methodology in order to determine if the neuroendocrine effects of these two peptides are exerted on similar cell systems. In animals pretreated with ANP conjugated to the cytotoxic A chain of ricin, central administration of a dose of ANP which is known to inhibit PRL secretion did not alter PRL levels in plasma; however, subsequent administration of CNP elicited the stimulation of PRL secretion. In rats pretreated with CNP-ricin A chain conjugate, a treatment we hypothesize targets for destruction CNP responsive cells, ANP injection did inhibit PRL secretion, while the stimulatory effect of CNP was absent. These results suggest that the neuroendocrine effects of these two natriuretic peptides on PRL secretion are expressed on different cellular elements of the hypothalamo-pituitary axis. Furthermore, they reveal that neither peptide acts directly on the tuberoinfundibular dopamine system since pretreatment with either cytotoxin conjugate failed to alter basal PRL levels. Thus ANP and CNP do not appear to express opposing actions on the same cell systems, suggesting the recruitment of each peptide individually by differing, unique stimuli for PRL release.

Influence of the immunoglobulin heavy chain locus on expression of the VK1GAC light chain.

The VK1GAC light chain represents the dominant V kappa structure employed in the antibody response of A/J mice to streptococcal group A carbohydrate ( GAC ). Two anti-idiotypic antisera, anti- Id5 and anti- Id20 , with specificity for the VK1GAC light chain were used to examine anti- GAC antibody responses in a series of inbred mouse strains that differ at the heavy chain constant region ( IgCH ) allotype locus. Both idiotypes were expressed in normal and immune sera from mice of most IgCH allotypes, except IgCHb (C57BL/6J) and IgCHf (CE/J). C57BL/6J mice expressed Id5 , but not Id20 , whereas CE/J mice did not express either idiotype. Testing of recombinant inbred strains between BALB/c and C57BL/6 indicated that the pattern of idiotype expression did not correlate with IgCH allotype. The C X B recombinants expressed all three idiotype patterns that were observed in the panel of inbred strains. Testing of allotype congenic mice between BALB/c and C57BL/6 showed that CB.20 and BC.8 mice were Id20 -, whereas BAB-14 mice were Id20 +, indicating that both VH and background (V kappa or regulatory) loci must be derived from BALB/c to obtain Id20 expression. The difference in the frequency of idiotype expression observed between BALB/c and BAB-14 mice indicates that the IgCH locus may exert a quantitative influence on the expression of this light chain. To examine the Id20 -, Id5 + antibodies of C57BL/6 mice, anti- GAC hybridomas were prepared. Of 16 C57BL/6-derived anti- GAC monoclonal antibodies, six were reactive with anti- Id5 and not with anti- Id20 . Isoelectric focusing of the purified kappa light chains from three of these antibodies revealed two distinct spectrotypes that co-migrated with the two known VK1GAC spectrotypes observed with A/J anti- GAC light chains. Idiotypic analysis of in vitro recombinants between the heavy and light chains of A/J and C57BL/6 monoclonal antibodies demonstrated that the C57BL/6 light chains were idiotypically similar to A/J light chains when they were free in solution or paired with A/J heavy chains. These results demonstrate that C57BL/6 mice can express a light chain that is very similar, if not identical, to the VK1GAC light chain, although the light chain is expressed in lower frequency and is paired with a distinct VH structure, which can mask expression of one of the VK1GAC idiotypes. These effects on V kappa expression map to at least three genetic loci: VH, CH, and an unlinked locus.

Kirsten murine sarcoma virus transformed cell lines and a spontaneously transformed rat cell-line produce transforming factors.

We have examined culture fluids from a variety of Kirsten murine sarcoma virus (KiMSV) transformed rat and mouse cells for the presence of factors which induce normal Rat-1 cells to assume the transformed phenotype. All KiMSV transformants produced transforming factor (TF). Revertants of KiMSV transformed rat or mouse failed to release TF as did normal rat or mouse cells. Cells transformed by a temperature sensitive mutant of KiMSV produced TF at the permissive temperature but not at the nonpermissive temperature. Further, cells from a spontaneous transformant of Rat-1 cells also produced TF. TF is a small polypeptide which competes for the epidermal growth factor receptor. Its effect upon normal cells is reversible and requires de novo RNA and protein synthesis. Cells treated with TF lose the actin fibers observed in normal fibroblasts, assume a transformed cell morphology, become anchorage independent for growth, grow in low concentrations of serum, grow to a high cell density, and have an increased rate of hexose uptake.

Analysis of anti-streptococcal group A carbohydrate idiotope levels in sera: correlation of magnitude of expression with idiotope position and VK haplotype.

We have employed monoclonal anti-idiotopes to map the corresponding idiotopes on the variable domain of a prototype antibody specific for streptococcal group A carbohydrate (GAC). Idiotope variability, as assessed by direct and competitive binding assays, previously was shown to correlate with idiotope position; determinants farther from the binding site were less variable. We now describe a relationship between idiotope position and idiotope concentration in normal and GAC-immune sera from mice of several inbred and recombinant inbred strains. Employing sera as inhibitors in a competitive radioimmunoassay, we demonstrate that idiotopes farther from the binding site (more proximal) tend to be present at higher levels in GAC-hyperimmune sera. Only the most proximal idiotope was detected in normal serum, and this idiotope was present in normal sera from all 12 strains tested. Finally, significant interstrain differences in patterns of idiotope expression were observed, and some of these differences appear to be correlated with allelic differences at immunoglobulin structural gene loci.

Species differences in the expression of caseinolytic proteinases and plasminogen activators by bone marrow cells.

Intact viable bone marrow cells from hamsters (Mesocricetus aruatus), rat (Rattus norvegicus), guinea pig (Cavia porcella) and rabbit (Oryctolagus cuniculus) were analyzed for caseinolytic proteinases and plasminogen activator activity. Species specific quantitative and qualitative differences in caseinolytic activity were detected. Quantitative and qualitative differences in plasminogen activator expression were observed. Cellular fractionation experiments revealed a heterogenous distribution of plasminogen activator activity among subpopulations of cells. The plasminogen activator activity associated with bone marrow cells behaved as ectoenzymes. These results indicated that different species may regulate bone marrow cell proteinase activity in a species-specific manner, compatible with their unique regulatory requirements.

A variant of exotoxin A that forms potent and specific chemically conjugated immunotoxins.

To introduce a free sulfhydryl into Pseudomonas aeruginosa exotoxin A (ETA), methionine 161 in domain I of the toxin was changed to cysteine by site-directed mutagenesis. The free sulfhydryl provides a convenient site for covalent attachment of ETA to other proteins in the production of chimeric toxins. The mutation was then introduced into a variant of ETA that is impaired in receptor binding, termed ETA-60EF61, that has the dipeptide Glu-Phe inserted between residues 60 and 61. The resulting double mutant, ETA-60EF61 Cys161, was conjugated to three different monoclonal antibodies via a thioether linkage, and the immunotoxins were tested for cytotoxicity with cells in culture. Each immunotoxin was extremely potent against cells that expressed surface determinants for the monoclonal antibodies but had little cytotoxicity for cells that did not bind the antibodies. For comparison, we also conjugated ricin A chain to each of the three monoclonal antibodies and found that the resulting immunotoxins were at least two-orders of magnitude less potent than the corresponding immunotoxins made with ETA-60EF61Cys161. This study demonstrates that ETA-60EF61Cys161 makes potent and specific immunotoxins and may potentially be useful in selectively eliminating subpopulations of cells in vitro and in vivo.

Monoclonal antibodies to streptococcal group A carbohydrate. II. The VK1GAC light chain is preferentially associated with serum IgG3.

A kappa-light chain variable region (V kappa) dominantly employed in the serum antibody response of A/J mice to streptococcal group A carbohydrate (GAC) has been termed VK1GAC. Examination of in vitro recombinants between the isolated heavy and light chains of VK1GAC+ and VK1GAC-anti-GAC hybridomas and non-GAC-binding myeloma proteins indicated that two antisera (anti-Id5 and anti-Id20) recognized the VK1GAC light chain when it was free in solution or paired with several heterologous heavy chains. Screening of a panel of A/J anti-GAC monoclonal antibodies with these antisera showed almost complete concordance between Id5 and Id20 expression and the presence of VK1GAC light chain as detected by its unique isoelectric focusing spectrotype. These antisera were used to examine serum expression of the VK1GAC light chain in normal and hyperimmune serum of A/J mice. Normal A/J serum contained from 20 to 100 micrograms Id5/ml serum, whereas only 1 to 10 micrograms Id20/ml serum was detected. The levels of both VK1GAC idiotypes increased dramatically 10- to 20-fold after hyperimmunization of mice with group A vaccine. When serum IgG from normal and immune mice was fractionated into the IgG subclasses (IgG1, IgG2a, and IgG3), it was found that the VK1GAC light chain does not pair randomly with heavy chains of the IgG subclasses, but rather is associated preferentially with heavy chains of the IgG3 subclass whether or not it is associated with antibodies to GAC. These results suggest that the heavy chain pairing exhibited by this VK product may not be random.

The effect of antibody valency and lysosomotropic amines on the synergy between ricin A chain- and ricin B chain-containing immunotoxins.

The cytotoxicity of A chain immunotoxins containing IgG or Fab fragments specific for the surface immunoglobulin of the Daudi cell line was assessed in the presence of B chain immunotoxins (IgG or Fab) or lysosomotropic amines, or both. The concentration required for 50% inhibition of protein synthesis (IC50) in Daudi cells was 1.3 X 10(-8) M for IgG-A and 5 X 10(-8) M for Fab-A. The toxicity of both A chain immunotoxins was enhanced twofold by ammonium chloride. In the presence of A chain immunotoxins and ammonium chloride, a maximum of 99 and 90% reduction of clonal precursors was obtained with IgG and Fab-A chain immunotoxins respectively. Immunotoxins containing ricin B chain and IgG or Fab fragments specific for the antibody portion of A chain immunotoxins were used as secondary "piggyback" immunotoxins to treat cells that were pretreated with A chain immunotoxins. Both B chain immunotoxins were nontoxic at 1 X 10(-6) M. When added to target cells pretreated with specific A chain immunotoxins, the IC50 of the A chain immunotoxins was decreased up to 16-fold in the absence of ammonium chloride. In contrast to the results obtained with A chain immunotoxins alone, ammonium chloride significantly increased the toxicity of the complete piggyback system, resulting in the killing of 99.999% or five logs of target cells in the clonal assay. This decreased the IC50 of A chain immunotoxins up to 116-fold when compared with A chain immunotoxin alone. This enhanced toxicity was independent of the valency of either immunotoxin.

Pyodermatitis-pyostomatitis vegetans: a specific marker for inflammatory bowel disease.

In pyodermatitis-pyostomatitis vegetans annular pustular cutaneous lesions may precede, accompany, or follow the usually extensive vegetating oral disease. Sometimes only cutaneous or only oral lesions occur and previously have been described as separate entities. Clinical, histopathologic, and immunopathologic evidence clearly indicates this is one disease and suggests that it is distinct from pemphigus vegetans. The association between pyodermatitis-pyostomatitis vegetans and inflammatory bowel disease, most commonly ulcerative colitis, has been amply confirmed. Pyodermatitis-pyostomatitis vegetans should be considered a marker for inflammatory bowel disease.