Diphtheria toxin-based bivalent human IL-2 fusion toxin with improved efficacy for targeting human CD25(+) cells.

Regulatory T cells (Treg) constitute a major inhibitory cell population which suppresses immune responses. Thus, Treg have proven to be key players in the induction of transplantation tolerance, protection from autoimmune disease and prevention of the development of effective anti-tumor immune reactions. Treg express high levels of the high affinity interleukin-2 receptor (IL-2R) consisting of IL-2Rα (CD25) together with IL-2Rß (CD122) and the common γ-chain (CD132). An effective reagent capable of depleting Treg in vivo would facilitate better cancer treatment and allow mechanistic studies of the role of Treg in transplantation tolerance and the development of autoimmune disease. In this study, we have developed a novel bivalent human IL-2 fusion toxin along with an Ontak®-like monovalent human IL-2 fusion toxin and compared the functional ability of these reagents in vitro. Here we show that genetically linking two human IL-2 domains in tandem, thereby generating a bivalent fusion toxin, results in significantly improved capacity in targeting human CD25(+) cells in vitro. Binding analysis by flow cytometry showed that the bivalent human IL-2 fusion toxin has notably increased affinity for human CD25(+) cells. In vitro functional analysis demonstrated that the bivalent isoform has an increased potency of approximately 2 logs in inhibiting cellular proliferation and protein synthesis in human CD25(+) cells compared to the monovalent human IL-2 fusion toxin. Additionally, we performed two inhibition assays in order to verify that the fusion toxins target the cells specifically through binding of the human IL-2 domain of the fusion toxin to the human IL-2 receptor on the cell surface. These results demonstrated that 1) both monovalent and bivalent human IL-2 fusion toxins are capable of blocking the binding of biotinylated human IL-2 to human CD25 by flow cytometry; and 2) human IL-2 blocked the fusion toxins from inhibiting protein synthesis and cellular proliferation in vitro, thus confirming that the human IL-2 fusion toxins target the cells specifically through binding to the human IL-2 receptor. We believe that the bivalent human IL-2 fusion toxin will be a more potent, and therefore, more optimal agent than the current clinically-used monovalent fusion toxin (denileukin diftitox, Ontak®) for in vivo depletion of Treg.

Development of a diphtheria toxin-based recombinant porcine IL-2 fusion toxin for depleting porcine CD25+ cells.

Regulatory T cells (Tregs) have been widely recognized as crucial players in controlling immune responses. Because their major role is to ensure that the immune system is not over reactive, Tregs have been the focus of multiple research studies including those investigating transplantation tolerance, autoimmunity and cancer treatment. On their surface Tregs constitutively express CD25, a high affinity receptor for the cytokine interleukin-2 (IL-2). The reagents constructed in this study were generated by genetically linking porcine IL-2 to the truncated diphtheria toxin (DT390). This reagent functions by first binding to the cell surface via the porcine IL-2/porcine CD25 interaction then the DT390 domain facilitates internalization followed by inhibition of protein synthesis resulting in cell death. Four versions of the porcine IL-2 fusion toxin were designed in an interest to find the most effective isoform: 1) monovalent glycosylated porcine IL-2 fusion toxin (Gly); 2) monovalent non-N-glycosylated porcine IL-2 fusion toxin (NonGly); 3) bivalent glycosylated porcine IL-2 fusion toxin (Bi-Gly); 4) bivalent non-N-glycosylated porcine IL-2 fusion toxin (Bi-NonGly). Using a porcine CD25(+) B cell lymphoma cell line (LCL13271) in vitro analysis of the fusion toxins' ability to inhibit protein synthesis demonstrated that the Bi-NonGly fusion toxin is the most efficient reagent. These in vitro results are consistent with binding affinity as the Bi-NonGly fusion toxin binds strongest to CD25 on the same LCL13271 cells. The Bi-Gly fusion toxin significantly prolonged the survival (p=0.028) of tumor-bearing NOD/SCID IL-2 receptor γ(-/-) (NSG) mice injected with LCL13271 cells compared with untreated controls. This recombinant protein has great potential to function as a useful tool for in vivo depletion of porcine CD25(+) cells for studying immune regulation.

Molecular basis of cross-species reactivities of human versus porcine CTLA-4.

The binding motif of human CTLA-4 is well known to be MYPPPY and for porcine CTLA-4 the binding motif is LYPPPY. Is this single amino acid difference of methionine (M) versus leucine (L) critical for the CTLA-4 binding? Recently, we have reported that the recombinant soluble porcine CTLA-4 was incapable of binding to human CD80. In this study we mutated L to M in the binding motif of the soluble porcine CTLA-4 and mutated M to L in the binding motif of the soluble human CTLA-4. We then analyzed how these mutations affected the binding affinity of the mutants to both porcine and human CD80(+) cells. The soluble porcine CTLA-4-L97M mutant partially lost its binding affinity to porcine CD80 compared to the wild-type and conferred very weak binding ability to human CD80. These results indicate that the L in the binding motif of porcine CTLA-4 is important for determining its binding ability to porcine CD80. Wild-type soluble human CTLA-4 binds to both human and porcine CD80 with comparable affinity, however, the soluble human CTLA-4-M97L mutant almost lost its binding ability to human CD80 and increased its binding ability to porcine CD80. These results indicate that M in the human CTLA-4 binding motif is extremely critical for its binding to human CD80. Those data suggest that the human CTLA-4 based recombinant protein drugs such as human CTLA-4-Ig can be used and/or tested in a porcine model. Conversely, the use of porcine CTLA-4 based recombinant protein drugs such as porcine CTLA-4-Ig is restricted to swine models. The difference in binding specificity of CTLA-4 observed in this study may be useful for studies such as pig to nonhuman primate xeno-transplantation. Porcine CTLA-4- and human CTLA-4-M97L mutant-based recombinant protein drugs can be used to specifically block the direct presentation by donor antigen presenting cells in pig to nonhuman primate xeno-transplantation. Human CTLA-4-M97L mutant-based recombinant protein drugs will be more ideal as it is without immunogenicity to human being.

A truncated diphtheria toxin based recombinant porcine CTLA-4 fusion toxin.

Targeted cell therapies are possible through the generation of recombinant fusion proteins that combine a toxin, such as diphtheria toxin (DT), with an antibody or other molecule that confers specificity. Upon binding of the fusion protein to the cell of interest, the diphtheria toxin is internalized which results in protein synthesis inhibition and subsequent cell death. We have recently expressed and purified the recombinant soluble porcine CTLA-4 both with and without N-glycosylation in yeast Pichia pastoris for in vivo use in our preclinical swine model. The glycosylated and non-N-glycosylated versions of this recombinant protein each bind to a porcine CD80 expressing B-cell lymphoma line (LCL13271) with equal affinity (K(D)=13 nM). In this study we have linked each of the glycosylated and non-N-glycosylated soluble porcine CTLA-4 proteins to the truncated diphtheria toxin DT390 through genetic engineering yielding three versions of the porcine CTLA-4 fusion toxins: 1) monovalent glycosylated soluble porcine CTLA-4 fusion toxin; 2) monovalent non-N-glycosylated soluble porcine CTLA-4 fusion toxin and 3) bivalent non-N-glycosylated soluble porcine CTLA-4 fusion toxin. Protein synthesis inhibition analysis demonstrated that while all three fusion toxins are capable of inhibiting protein synthesis in vitro, the non-N-glycosylated porcine CTLA-4 isoforms function most efficiently. Binding analysis using flow cytometry of the porcine CTLA-4 fusion toxins to LCL13271 cells also demonstrated that the non-N-glycosylated porcine CTLA-4 isoforms bind to these cells with higher affinity compared to the glycosylated fusion toxin. The monovalent non-N-glycosylated porcine CTLA-4 fusion toxin was tested in vivo. NSG (NOD/SCID IL-2 receptor γ(-)/(-)) mice were injected with porcine CD80(+) LCL13271 tumor cells. All animals succumbed to tumors and those treated with the monovalent non-N-glycosylated porcine CTLA-4 fusion toxin survived longer based on a symptomatic scoring system compared to the untreated controls. This recombinant protein may therefore provide a novel approach for in vivo depletion of porcine antigen presenting cells (APCs) for studies investigating the induction of transplantation tolerance, autoimmune disease and cancer treatment.

Expression and characterization of recombinant soluble porcine CD3 ectodomain molecules: mapping the epitope of an anti-porcine CD3 monoclonal antibody 898H2-6-15.

The porcine CD3 specific monoclonal antibody 898H2-6-15 has been used in allo- and xeno-transplantation studies as a porcine CD3 marker and as an effective T cell depletion reagent when conjugated to the diphtheria toxin mutant, CRM9. A recombinant anti-porcine CD3 immuntoxin was recently developed using single-chain variable fragments (scFv) derived from 898H2-6-15. In this study, using published sequence data, we have expressed the porcine CD3 ectodomain molecules in E. coli through inclusion body isolation and in vitro refolding approach. The expressed and refolded porcine CD3 ectodomain molecules include CD3ε, CD3γ, CD3δ, CD3εγ heterodimer, CD3εδ heterodimer, CD3εγ single-chain fusion protein and CD3εδ single-chain fusion protein. These refolded porcine CD3 ectodomain molecules were purified with a strong anion exchange resin Poros 50HQ. ELISA analysis demonstrated that only the porcine CD3εγ ectodomain single-chain fusion protein can bind to the porcine CD3 specific monoclonal antibody 898H2-6-15. The availability of this porcine CD3εγ ectodomain single-chain fusion protein will allow screening for affinity matured variants of scFv derived from 898H2-6-15 to improve the recombinant anti-porcine CD3 immunotoxin. Porcine CD3εγ ectodomain single-chain fusion protein will also be a very useful reagent to study the soluble phase interaction between porcine CD3εγ and porcine CD3 antibodies such as 898H2-6-15.