Identification of the Antigenic Epitopes in Staphylococcal Enterotoxins A and E and Design of a Superantigen for Human Cancer Therapy

Abstract

Monoclonal antibodies have a potential for cancer therapy that may be further improved by linking them to effector molecules such as superantigens. Tumor targeting of a superantigen leads to a powerful T cell attack against the tumour tissue. Encouraging results have been observed preclinically and in patients using the superantigen staphylococcal enterotoxin A, SEA. To further improve the concept, we have reduced the reactivity to antibodies against superantigens, which is found in all individuals. Using epitope mapping, antibody binding sites in SEA and SEE were found around their MHC class II binding sites. These epitopes were removed genetically and a large number of synthetic superantigens were produced in an iterative engineering procedure. Properties such as decreased binding to anti-SEA as well as higher selectivity to induce killing of tumour cells compared to MHC class II expressing cells, were sequentially improved. The lysine residues 79, 81, 83 and 84 are all part of major antigenic epitopes, Gln204, Lys74, Asp75 and Asn78 are important for optimal killing of tumour cells while Asp45 affects binding to MHC class II. The production properties were optimised by further engineering and a novel synthetic superantigen, SEA/E-120, was designed. It is recognised by approximately 15% of human anti-SEA antibodies and have more potent tumour cell killing properties than SEA. SEA/E-120 is likely to have a low toxicity due to its reduced capacity to mediate killing of MHC class II expressing cells. It is produced as a Fab fusion protein at approximately 35 mg/l in Escherichia coli.

Introduction

Antibody based cancer therapies have been investigated for two decades. A high number of different antibodies have been studied in clinical trials, but due to several challenges only a few antibodies have become important tools in the fight against cancer.¹ Except for problems with immunogenicity of non-human antibodies, the main obstacles have been associated with poor tumour uptake, difficulties to find antigens that are selectively expressed on a large number of tumour cells and finally a low cytotoxic potency of antibodies. Several different ways to potentiate antibody therapies have been investigated such as radio-labelling or fusion to cytotoxic molecules.² We have developed a slightly alternative approach by fusing the superantigen staphylococcal enterotoxin A, SEA, to tumour reactive antibody fragments.³ The conceptual idea behind this fusion protein is to stimulate and direct potent effector cells of the immune system against the tumour.

Bacterial superantigens, activate a large fraction of the T-cell population by their binding to the T cell receptor, TCR Vα⁴ or Vβ chain.5., 6. Despite a relatively low sequence similarity, this family of proteins appears to share the same three-dimensional structure.⁷ SEA belongs to the group of the most extensively characterised superantigens.8., 9. Upon presentation on MHC class II, SEA stimulate T cells to divide and secrete cytokines.10., 11. Interestingly, only a couple of superantigen molecules need to be bound to a cell to generate a potent T cell stimulation.¹² Certain superantigens, like SEA, have two distinct MHC class II binding sites13., 14. that enable them to cross-link MHC class II molecules on antigen presenting cells and this lead to secretion of inflammatory cytokines.¹⁵

Fusion proteins between antibody moieties and SEA trigger cytotoxic T cells to kill tumour cells in the absence of MHC class II molecules.3., 16. In addition, activated T cells produce tumouricidal and pro-inflammatory cytokines, counteracting the problems of tumor heterogeneity and macromolecular uptake.17., 18. In murine tumour models, the anti-tumour effects of Fab-SEA fusion proteins are clearly dependent on targeting of superantigen activity to the tumour site.3., 19. In contrast, dose-limiting systemic immune activation is largely due to MHC class II-dependent targeting of superantigen activity to the spleen and other lymphoid tissues.20., 21. In preclinical models, the FabSEA fusion proteins have shown potent effects against colorectal cancer, lymphoma, melanoma and non-small cell lung cancer.3., 19., 22., 23., 24.

A fusion protein consisting of the C242Fab moiety and wild-type SEA has been investigated in clinical trials of colorectal and pancreatic cancer.²⁵ Even though encouraging results were obtained, this protein showed toxic effects and caused fever and hypotension. The most likely reason for this is that the product accumulated in MHC class II containing organs, instead of only in the tumour tissue. Most residues in SEA that interact with MHC class II have been identified13., 14., 26. and this has given insight into the molecular mechanisms causing toxicity. Based on these findings a novel fusion protein, 5T4FabSEA_D227A, with 100–1000-fold lower affinity for MHC class II¹⁹ has been designed for later clinical studies. Despite the reduced binding to MHC class II and decreased toxicity, this product still has very powerful T cell activating properties. However, there is also another complication due to that all individuals have preformed antibodies against SEA which makes dosing complex. The 5T4FabSEA_D227A fusion protein is currently in clinical trials and, although less pronounced compared to the wild-type construct, toxicity and anti-SEA antibodies may be limiting factors for the therapy.

Here we have mapped the major antigenic sites in SEA and we present a novel generation of synthetic superantigens with reduced seroreactivity and toxicity. The approach was based on iterative engineering using experimental data and modelling. Taking advantage of this novel generation of superantigens, the clinical development of superantigen therapy of cancer may be substantially facilitated.