Difference between revisions of "Team:FAU Erlangen/Description"

Since the application of bispecific antibodies is a relatively new approach, current protocols for production are not yet standardized. In most cases mammalian cells are used for the expression of these proteins. However, this expression system is associated with very high costs and proves to be very time-consuming. The use of the bacterial host Escherichia coli poses an alternative for the expression of recombinant protein. It is well published and provides many advantages over the expression in mammalian cells, such as a far cheaper cultivation due to the higher proliferation rate and faster protein production. Despite these benefits, producing proteins of mammalian origin in E. coli has been shown to be difficult since post-translational modifications are often vital to protein folding and function and cannot be replicated in these bacterial systems. Especially for complex proteins like antibodies, misfolding is often detrimental for their solubility and performance.
Devising a reliable and well applicable procedure to successfully express a functional bispecific antibody in bacterial cells is crucial in fully accessing their therapeutic potential by facilitating cost- and time-effective research. However, published protocols vary immensely and offer no conclusive approach to producing recombinant, bispecific antibodies in bacteria.

For our experiments we chose to create a bispecific antibody against CD3 on T-lymphocytes and GPA33. GPA33 can be found on 95% of all colorectal cancer cells, thereby allowing them to be specifically targeted. By also binding CD3, the bispecific antibody can bring cytotoxic T-lymphocytes (CTLs) into direct contact with the cancer cells, which activates the CTLs to release their cytotoxic granules and kill the malignant cell. Our protein is designed as a Bispecific T-cell Engager (BiTE) and consists of two distinct single-chain variable Fragments (scFv) that are connected via a flexible linker.
We are going to use a pET27b plasmid with a T7 promoter and lacZ control in the E. coli strain Arctic Express. This strain is engineered to tolerate a high level of protein expression and shows improved protein folding capacity. It is well published that periplasmic expression of recombinant protein offers an oxidizing environment that is beneficial for correct protein folding, as well as a reduced number of proteases that may degrade our protein. For this reason, we plan to use a PelB leader sequence to transport our constructs to the periplasm of E. coli for synthesis.

In summary, we pursue three approaches:

In our first approach, we plan on expressing a complete BiTE molecule that will serve as reference for our other production strategies.

In our second approach, we express the two scFvs separately as fusion proteins with SpyTag or SpyCatcher, respectively. This system relies on a split domain of the FbaB protein from Streptococcus pyogenes to form a covalent, isopeptide bond upon co-culturing. This allows us to avoid the possible problem of producing a large, complex protein that could result in the formation of insoluble aggregates. Furthermore, the SpyTag/SpyCatcher system is fashioned after a modular concept. If this approach proves to be successful, it will be possible to express several effector subunits with varying affinities or functions, i.e. different antigen specificities or enzymatic activity, and connect them. This will allow for a flexible modular system that might be adapted with regards to current requirements in a fast and easy manner.

In our last approach, we intend to produce CD3- and GPA33-specific antibody fragments, so called Fabs, and link them together via disulfide bonds. This chemical conjugation is achieved by using a bis-dibromomaleimide cross-linker, providing a versatile method to produce bispecific antibodies.

To test the binding activity of our products, we will perform killing assays with blood derived CTLs. We will further test their stability at different temperatures and pH conditions.

By programming a neural network, it will also be possible to predict potential allergenicity of our bispecific antibodies. This will also prove beneficial as a prognostic tool to produce other recombinant proteins and their possible clinical application.

Altogether, the assays will provide a comparative evaluation of the properties of our different products and permit the assessment of their respective value for clinical or scientific application.

We are a team that is tackling problems of the emerging field of molecular immunology - no team in Erlangen has done this before. Here in Erlangen, we have an optimal professional network, since the newly established Masters-Program has been opened to sustain the ongoing research in this field, from which we can benefit greatly. Furthermore, our Team has an interdisciplinary approach as we have sub-teams working on chemical ligation of Fab-Fragments and a deep-learning approach to predicting immune reactions concerning different allergens. Our interdisciplinary approach is comprehensive as it covers the cure as well as the side effects by predicting these with our neural network software.