Crystallographic evidence for preformed dimers of erythropoietin receptor before ligand activation.

Erythropoietin receptor (EPOR) is thought to be activated by ligand-induced homodimerization. However, structures of agonist and antagonist peptide complexes of EPOR, as well as an EPO-EPOR complex, have shown that the actual dimer configuration is critical for the biological response and signal efficiency. The crystal structure of the extracellular domain of EPOR in its unliganded form at 2.4 angstrom resolution has revealed a dimer in which the individual membrane-spanning and intracellular domains would be too far apart to permit phosphorylation by JAK2. This unliganded EPOR dimer is formed from self-association of the same key binding site residues that interact with EPO-mimetic peptide and EPO ligands. This model for a preformed dimer on the cell surface provides insights into the organization, activation, and plasticity of recognition of hematopoietic cell surface receptors.

Small peptides as potent mimetics of the protein hormone erythropoietin.

Random phage display peptide libraries and affinity selective methods were used to isolate small peptides that bind to and activate the receptor for the cytokine erythropoietin (EPO). In a panel of in vitro biological assays, the peptides act as full agonists and they can also stimulate erythropoiesis in mice. These agonists are represented by a 14- amino acid disulfide-bonded, cyclic peptide with the minimum consensus sequence YXCXXGPXTWXCXP, where X represents positions allowing occupation by several amino acids. The amino acid sequences of these peptides are not found in the primary sequence of EPO. The signaling pathways activated by these peptides appear to be identical to those induced by the natural ligand. This discovery may form the basis for the design of small molecule mimetics of EPO.

Functional mimicry of a protein hormone by a peptide agonist: the EPO receptor complex at 2.8 A.

The functional mimicry of a protein by an unrelated small molecule has been a formidable challenge. Now, however, the biological activity of a 166-residue hematopoietic growth hormone, erythropoietin (EPO), with its class 1 cytokine receptor has been mimicked by a 20-residue cyclic peptide unrelated in sequence to the natural ligand. The crystal structure at 2.8 A resolution of a complex of this agonist peptide with the extracellular domain of EPO receptor reveals that a peptide dimer induces an almost perfect twofold dimerization of the receptor. The dimer assembly differs from that of the human growth hormone (hGH) receptor complex and suggests that more than one mode of dimerization may be able to induce signal transduction and cell proliferation. The EPO receptor binding site, defined by peptide interaction, corresponds to the smaller functional epitope identified for hGH receptor. Similarly, the EPO mimetic peptide ligand can be considered as a minimal hormone, and suggests the design of nonpeptidic small molecule mimetics for EPO and other cytokines may indeed be achievable.

The structure, organization, activation and plasticity of the erythropoietin receptor.

Dimerization of the erythropoietin receptor has long been accepted as the singular step in its mechanism of activation. Recent studies have revealed a regulator process for activation that is dependent on the actual configuration of the receptor-ligand dimer assembly. This aspect of the receptor subunit assembly appears to extend to the unliganded receptor, which can dimerize on the cell surface and diminish any spontaneous background signaling in the absence of ligand. This self-recognition, as well as the multiple ligand binding capabilities of the receptor binding site, is consistent with an emerging theme of plasticity in protein-protein and ligand-receptor interactions.

Increased potency of an erythropoietin peptide mimetic through covalent dimerization.

We have synthesized a chemically defined, dimeric form of an erythropoietin mimetic peptide (EMP) that displays 100-fold increased affinity for the erythropoietin receptor (EPOR) and correspondingly elevated potency in cell-based assays and in mice. The dimeric EMP1 was synthesized using a C-terminal lysine residue as a branch point. A beta-alanine residue was coupled to the main-chain (alpha) amino group of the lysine residue in order to provide a pseudosymmetrical scaffold where both the side-chain and main-chain were of approximately equal length. Using an orthogonal protection system, independently disulphide-cylized EMP1 moieties were synthesized upon this scaffold. The proposed mechanism of increased potency of the dimer over the parental compound EMP1 is consistent with the structure of a cocrystal of EMP1 and the extracellular domain of the EPOR in which a noncovalent peptide dimer is seen spanning the cleft between two molecules of the EPOR extracellular domain.

Amino-terminal dimerization of an erythropoietin mimetic peptide results in increased erythropoietic activity.

BACKGROUND: Erythropoietin (EPO), the hormone involved in red blood cell production, activates its receptor by binding to the receptor's extracellular domain and presumably dimerizing two receptor monomers to initiate signal transduction. EPO-mimetic peptides, such as EMP1, also bind and activate the receptor by dimerization. These mimetic peptides are not as potent as EPO, however. The crystal structure of the EPO receptor (EBP) bound to EMP1 reveals the formation of a complex consisting of two peptides bound to two receptors, so we sought to improve the biological activity of EPO-mimetic peptides by constructing covalent dimers of EMP1 and other peptide mimetics linked by polyethylene glycol (PEG). RESULTS: The potency of the PEG-dimerized EPO peptide mimetics both in vitro and in vivo was improved up to 1,000-fold compared to the corresponding peptide monomers. The dimers were constructed using peptide monomers which have only one reactive amine per molecule, allowing us to conclude that the increase in potency can be attributed to a structure in which two peptides are linked through their respective amino termini to the difunctional PEG molecule. In addition, an inactive peptide was converted into a weak agonist by PEG-induced dimerization. CONCLUSIONS: The potency of previously isolated peptides that are modest agonists of the EPO receptor was dramatically increased by PEG-induced dimerization. The EPO receptor is thought to be dimerized during activation, so our results are consistent with the proposed 2:2 receptor : peptide stoichiometry. The conversion of an inactive peptide into an agonist further supports the idea that dimerization can mediate receptor activation.

Influence in vitro of IL-3/Epo fusion proteins compared with the combination of IL-3 plus Epo in enhancing the proliferation of single isolated erythroid and multipotential progenitor cells from human umbilical cord blood and adult bone marrow.

Human interleukin-3/erythropoietin (IL-3/Epo) fusion protein have been constructed, expressed, and tested for biological activity. These fusion proteins were previously shown to be active on erythroid progenitors (BFU-E) from unseparated human bone marrow. We evaluated if these fusion proteins could stimulate erythroid and multipotential progenitor cells directly at the single-cell level. Two IL-3/Epo fusion proteins containing short (SL-3E, two amino acids) and long (LL-3E, 23 amino acids) linker sequences as well as a short linker Epo/IL-3 sequence (SL-E3, three amino acids) were tested. Highly enriched CD34 or BFU-E enriched CD34 CD33- cells from human umbilical cord blood or CD34 HLA-DR+CD33- cells from normal adult bone marrow were sorted as single cells into single wells. The combination of Epo plus IL-3 synergized to enhance the proliferation of BFU-E and multipotential progenitors (CFU-GEMM) in comparison to the individual effects of these cytokines. The three fusion proteins also enhanced proliferation of BFU-E and CFU-GEMM at the single-cell level and were at least as active as the combination of Epo and IL-3, demonstrating that IL-3/Epo fusion proteins directly stimulate proliferation of BFU-E and CFU-GEMM and that biological activity of IL-3 and Epo in vitro can be maintained when these proteins are fused. The activity of the combination of Epo and IL-3 or the fusion proteins was partially neutralized by preincubation with monoclonal antibodies to either Epo or IL-3 and was neutralized by greater than 90% by the combination of both antibodies, suggesting that the Epo and IL-3 components of the fusion proteins were both involved in the enhancing activity of these proteins. Additionally, use of monoclonal antibody to the human Epo receptor completely blocked the stimulating/enhancing activity of Epo alone, Epo plus IL-3, or the fusion proteins for stimulation of colony formation by BFU-E and CFU-GEMM but not for granulocyte-macrophage progenitors (CFU-GM), suggesting that the enhancing effects of the fusion proteins are most likely mediated, at least in part, by the Epo receptor.

Interleukin-3/erythropoietin fusion proteins: in vitro effects on hematopoietic cells.

Erythropoietin (Epo) acts synergistically with interleukin-3 (IL-3) to induce proliferation and differentiation of erythroid progenitors. This synergy occurs at IL-3 concentrations that have little or no effect alone. To determine whether optimal expansion of erythroid cells results when they are targeted by a molecule with both IL-3 and Epo activities, fusion proteins were generated and analyzed. Expression vectors were constructed in which the coding regions of human IL-3 and Epo cDNAs were joined by either a short (2 to 3 amino acids) or long (23 amino acids) linker sequence and expressed in Chinese hamster ovary (CHO) cells. Analysis of equilibrium binding properties of the IL-3 and Epo moieties revealed that in all fusion proteins each retained the ability to bind receptor. When IL-3 was connected to Epo by a short linker, the binding affinity of the IL-3 moiety was lower. In vitro proliferative activity of each moiety was observed on cell lines responsive to IL-3, Epo or a combination of the two cytokines. Fusion of IL-3 to Epo through its amino terminus was found to result in partial loss of its function. All the fusion proteins were biologically active on human bone marrow. When IL-3 was located at the amino domain of the protein, induction of erythroid colonies was similar to that of a mixture of IL-3 and Epo. These results indicate that biological integrity of both IL-3 and Epo can be maintained when these cytokines are fused, but that enhancement of erythropoiesis over that observed with a mixture of the two cytokines cannot be achieved by their fusion alone. Other requirements such as the coexpression of the IL-3 and Epo receptors and the sharing of a receptor subunit are likely to be needed for an optimal cell response to the fusion growth factors.

Anti-CD3 monoclonal antibody therapy. An approach toward optimization by in vitro analysis of new anti-CD3 antibodies.

Several problems remain associated with anti-CD3 monoclonal antibody therapy, including first-dose reactions, recurrent rejection, and the host humoral response to the xenogeneic mAb. One approach toward optimization of anti-CD3 mAb therapy involves the use of anti-CD3 mAbs of defined idiotype, isotype, and epitope specificity, so that the desired T cell activation and suppression properties are obtained and neutralization by antiidiotypic antibody is avoided. The purpose of the present study was to define the contribution of mAb isotype and epitope specificity in determining T cell activation and suppression potencies and to evaluate the idiotypic relationships among ten anti-CD3 mAbs and one anti-TCR alpha beta mAb selected for this study. Epitope mapping by flow cytofluorometry indicated that one mAb, OKT3D, possesses an epitope specificity distinct from that of other anti-CD3 mAbs. Analysis of early T cell activation, proliferation, and lymphokine production (TNF-alpha, gamma-IFN, and GM-CSF) indicated that mAb isotype exerted a profound effect on activation potency (IgG2a much greater than IgG1 greater than IgG2b), whereas epitope specificity exerted a minor effect. CTL inhibition studies demonstrated an epitope effect with highest inhibition potencies observed with OKT3D IgG1 and OKT3D IgG2b mAbs. Idiotypic analysis of nine mAbs indicated that all anti-CD3 mAbs except OKT3E possess idiotypes distinct from that of OKT3. Thus, two anti-CD3 mAbs, OKT3D IgG1 and OKT3D IgG2b, possess high immune suppression potency, low activation potency, and idiotypes distinct from OKT3. In conclusion, selection of anti-CD3 mAbs of defined idiotype and appropriate T cell activation and suppression properties may provide a means for mitigating problems associated with OKT3 therapy.

OKT3 F(ab')2 fragments--retention of the immunosuppressive properties of whole antibody with marked reduction in T cell activation and lymphokine release.

Recent studies in mouse and man indicate that the first dose response to anti-CD3 mAbs likely results from in vivo T cell activation and concomitant lymphokine release. One approach toward amelioration of these effects involves the use of nonactivating digest fragment preparations of anti-CD3 mAbs. In the present study whole and F(ab')2 fragments of OKT3 were prepared and assayed for their immune-activating and -suppressing effects on human peripheral blood mononuclear cells. Immunosuppressive effects were evaluated by quantitation of TCR modulation and coating, and by inhibition of CTL activity. Whole mAb and F(ab')2 fragments both effectively coated the TCR complex. However, whole mAb was more efficient at modulating the TCR complex, suggesting that modulation is enhanced by FcR interactions. Whole and F(ab')2 fragments of OKT3 were equally efficacious in suppressing CTL activity. Immune activation was evaluated by quantitation of proliferation, activation marker expression (IL-2R and Leu-23), and lymphokine release (TNF-alpha, gamma-IFN, and GM-CSF). Rigorously purified F(ab')2 preparations demonstrated minimal T cell activation, suggesting TCR and macrophage FcR crosslinking as necessary. Whole OKT3 mAb induced expression of IL-2R and Leu-23 activation markers on the majority of CD4+ and CD8+ cells at mAb concentrations as low as 1 ng/ml, whereas F(ab')2 fragments induced detectable, but markedly reduced expression of these markers only at mAb concentrations greater than or equal to 100 ng/ml. Similarly, whole mAb induced release of TNF-alpha, gamma-IFN, and GM-CSF at low mAb concentrations, whereas F(ab')2 fragments induced detectable (though markedly reduced) levels of TNF-alpha only. However, increasing degrees of contamination with whole antibody resulted in increasing mitogenic potency of the F(ab')2 preparation, which in some cases, was actually enhanced compared with that observed with whole mAb alone. In conclusion, these studies indicate that OKT3 F(ab')2 digest fragments are markedly less potent than whole mAb in inducing T cell activation, yet they retain significant immunosuppressive effects. However, meticulous purification of F(ab')2 digest fragment preparations will likely be required to avoid T cell and macrophage activation following in vivo administration.

Nonhuman primate responses to murine and humanized OKT4A.

A nonhuman primate antimurine response (MAMA) has been observed in 17 cynomolgus renal allograft recipients of murine OKT 4A. Neither cyclosporine, nor total-lymphoid irradiation, nor donor bone marrow preparation inhibited this antixenogeneic response. In an attempt to alter the antimurine basis of the response, a humanized chimeric OKT4A (IgG4) containing the entire variable portion of the murine OKT4A and a humanized CDR grafted OKT4A mAb sharing only the Complementarity Determining Region from the murine OKT4A, were administered to 8 cynomolgus allograft recipients. MAMA was detected in each recipient. In contrast to sera from recipients of murine OKT4A, sera from recipients of humanized OKT4A displayed no reactivity to other murine mAbs. MAMA specificity did not assay constant (C) region differences between the murine and humanized mAb; however, C region homology in humans should preclude a human antimouse antibody (HAMA) to the Fc portion of a humanized mAb. Furthermore, cynomolgus recipient serum levels of the humanized OKT4A mAb were maintained (> 1 microgram/ml) for a longer period than following treatment with murine OKT4A (murine < 12 days versus between 12 and 24 days for the humanized). If the HAMA response to humanized mAb in future clinical trials, were to be predictably anti-idiotypic, then the opportunity for treatment with sequential mAbs of differing idiotypes would be retained. Moreover, these current studies also suggest that humanized construction may influence the duration of therapeutic mAb levels. Thus, anti-idiotypic reactivity may not be as consequential to the clinical administration of humanized mAb to allograft recipients.

The effects of OKT4A monoclonal antibody on cellular immunity of nonhuman primate renal allograft recipients.

Significant differences in cellular responses were found among allograft recipients treated with various OKT4A mAb protocols. Recipients of multiple infusion low-dose and 2-bolus OKT4A immunosuppressive regimens regularly showed potent donor-specific cytotoxic CD8+ and CD4+ intragraft T cells and donor-reactive PBMC in MLC tests. In contrast, PBMC isolated from recipients of high-dose OKT4A therapy generally showed very weak or no response to donor-antigens during the later posttransplant periods. Furthermore, an absence of IL2-responsive intragraft cells was found to correlate with stable graft function in these recipients. We conclude that OKT4A mAb, in high doses, can block allosensitization and induce donor-specific nonresponsiveness in vivo. An OKT4A-based therapy, therefore, may have the potential of inducing long-lasting donor-specific immunosuppression, or even tolerance.

Humanized antibodies: enhancing therapeutic utility through antibody engineering.

The promise of antibody therapeutics has been greatly expanded by the development of monoclonal antibody technology and more recently antibody humanization. By transferring the mouse antibody binding site into a human antibody gene, we can engineer a "human antibody" which retains the specificity and biological effects of the original mouse antibody but has the potential to be nonimmunogenic in humans. Additionally, antibody effector functions can be improved through manipulation of the antibody constant region genes. We have produced a humanized version of OKT3 with human IgG4 and kappa constant regions. This antibody retains all of the in vitro characteristics of murine OKT3 including induction of cytokine release and T cell activation markers. Humanized OKT3 has an affinity of 1.4 x 10(9) M-1 relative to a 1.2 x 10(9) M-1 affinity of murine OKT3. Substitution of a glutamic acid for leucine at residue 235 in the antibody constant region abrogates FcR I binding and causes a marked reduction of T cell activation. The humanized FcR mutant of OKT3 has potential to be an improved therapeutic for transplantation and may have applications in autoimmune disease treatment.

Prolonged survival of nonhuman primate renal allograft recipients treated only with anti-CD4 monoclonal antibody.

The immunosuppressive efficacy of the monoclonal antibody OKT4A reactive with human and monkey CD4 cells was evaluated in cynomolgus renal allograft recipients. Low-dose (0.1 to 0.3 mg/kg/day) intact monoclonal antibodies (10 recipients) or F(ab')2 fragments (two recipients) were administered for 12 days. High-dose OKT4A (10 mg/kg) was administered on the day of transplantation as the only suppression in five animals. Four control animals received either no therapy or a monoclonal antibody nonreactive with monkey cells (OKT3). Maximum survival of the control animals and those treated with F(ab')2 was 11 days. Mean survival in the recipients of low-dose OKT4A was 25.4 +/- 4.3 days and in the group receiving high-dose OKT4A it was 39 +/- 6.4 days. All OKT4A-treated animals showed "coating" and CD4 modulation without depletion of circulating T cells. No modulation occurred in the F(ab')2-treated recipients. Serial allograft biopsy specimens showed reduced lymphocyte infiltration that was nearly complete in recipients of high-dose OKT4A. Biopsy-derived donor-reactive cytotoxic T-cell lines were generated regularly from recipients of low-dose, but not high-dose, OKT4A during periods of stable function. All animals treated with monoclonal antibodies developed an immunoglobulin G antimurine humoral response. Thus OKT4A is a potent immunosuppressive agent administered even as a single bolus, and depletion of CD4 cells is not required to suppress rejection. Anti-CD4 monoclonal antibodies may prove useful in patients, perhaps requiring only a limited number of higher-dose injections in the peritransplant period.

Using affinity capillary electrophoresis to study the interaction of the extracellular binding domain of erythropoietin receptor with peptides.

We have shown that affinity capillary electrophoresis (ACE) can be utilized to screen peptides that bind to the extracellular binding domain of the erythropoietin receptor (EBP). The comparison of the cyclic peptides GGTYSCHFGPLTWVCKPQGG (EMP1) GGTYSCHFGPLTAVCKPQGG (EMP13), and LGRKYSCHFGPLTWVCQPAKKD (EMP37) with the linear peptides HFGPLTWV (EMP26) and FMRF as ACE buffer additives were investigated. When EMP1 and EMP37 were the buffer additives, an abrupt change in the electrophoretic mobility of EBP was observed in the electropherogram. When EMP13, EMP26, and FMRF were examined under identical ACE conditions as EMP1 and EMP37, no significant change in the electrophoretic mobility of EBP was observed. These results correlate well with previously reported IC50 competitive binding data; that is, EMP1 and EMP37 bind to EBP while EMP13 and EMP26 bind very weakly. These observations strongly infer that peptide.EBP dimerization were induced by EMP1, and EMP37 but not by EMP13, EMP26 or FMRF. This ACE method provides a rapid tool for the detection of small peptides or drugs that bind to EBP.

The energized membrane and cellular autolysis in Bacillus subtilis.

Lysis of exponential cultures of B. subtilis follows the addition of reagents that dissipate either the electrical or pH gradients of cellular membranes. Stationary-phase cells or cultures that have been inhibited in division by macromolecular-synthesis inhibitors also lyse when uncoupling agents or ionophores are added to the growth medium. Autolysis occurs after brief starvation for a carbon source. Protoplasts are unaffected by azide or other lysis-inducing agents. Electron-donating agents, such as phenazine methosulfate and ascorbate, are effective in retarding autolysis. The addition of an oxidizable carbon source to starved and lysing cultures prevents their autolysis. These results suggest that cellular lysis in B. subtilis and energized membrane are tightly coupled. The fluorescence intensity and the wavelength of maximal fluorescence of 8-anilino-1-naphthalene sulfonic acid, when added to bacterial suspensions, appear to be qualitatively related to the rate of cell lysis. Analyses show that ATP limitations are probably not involved in the elicitation of lysis by ionophores, uncoupling agents or starvation. Measurements of protonmotive forces in the lysis-prone cells suggest that a threshold force of more than 85 mV may be required to maintain cellular integrity. Lipoteichoic acids, polyelectrolytes such as dextran sulfate or phospholipids do not modify the rate of cellular lysis when added to suspensions containing azide or other reagents that eliminate transmembrane protonmotive forces. We interpret the results to suggest that the in vivo control of autolysin activity in B. subtilis is related to the energized membrane

Hydrogen ion control of autolysin-dependent functions in Bacillus subtilis.

Autolysin activity in Bacillus subtilis as reflected by cell wall turnover, was found to be maximal in slightly alkaline media. When autolysis of whole cells was measured upon the addition of azide anion, it was found that maxima were exhibited at pH 6 and pH 9. Cell walls autolyzed maximally only at pH 9. In addition, the lysis of B. subtilis by nafcillin was found to be most pronounced at pH 7, whereas the cells tended to be resistant to the lysis induced by the antibiotic at pH less than 6 and pH greater than 7.9. In contrast, the maximal rate of non-lytic killing of the bacteria by nafcillin was observed to be between pH 5 and 6. When the organisms were cultured at pH 5, a decreased growth rate, accompanied by chain formation, was observed. At pH greater than 8, growth rates were low, although long chains were not present. Because the autolysin N-acetyl-muramyl-L-alanine amidase (amidase) is minimally active at pH 5 and because the cells tend to form long chains it is suggested that the enzyme is involved in cell separation. The amidase is also primarily responsible for cell wall turnover and susceptibility to the lytic effects of nafcillin.