Engineering and stable production of recombinant IgE for cancer immunotherapy and AllergoOncology

We developed a versatile, novel, time- and resource-effective tool for production of functional IgE at high yields that has wide application potential in basic research and clinical evaluations in allergy, cancer immunotherapy and AllergoOncology. (GMP), Tumor Associated Antigen (TAA), Enzyme-linked immunosorbent assay (ELISA), High Pressure Liquid Chromatography (HPLC), Ricinus Cummunis Agglutinin I (RCAI), Aleuria Aurantia lectin (AAL), Concavalin A (ConA).


Engineering and stable production of recombinant IgE for cancer immunotherapy and AllergoOncology
To the Editor: AllergoOncology, the emerging discipline of cancer immunology aiming to exploit features of allergy-related immunity to treat tumors (1)(2)(3), has catalyzed the development of tumor-specific IgE monoclonal antibodies as powerful alternatives to commonly-used therapeutic IgGs (4)(5)(6). IgE typically associated with the pathogenesis of allergic responses and known for Fc-mediated protective effects in parasitic infection clearance, presents exciting opportunities to unleash previously- We aimed to develop a stable expression application to generate recombinant IgE, exemplified with an antibody recognizing the melanoma-associated antigen chondroitin sulphate proteoglycan 4 (CSPG4). Our strategy incorporates seamless cloning, selection and fast antibody production at high yields (Fig. 1a). To prevent promoter silencing, we developed a novel dual-plasmid system containing Ubiquitous Chromatin Opening Elements (UCOE) sequences, located upstream of the transgene promoter (7). We isolated the coding sequences of anti-CSPG4 IgE heavy and light chains from a previously-described (pVITRO1-CSPG4 IgE/k) vector (4) (Fig. 1b) and cloned these into two UCOE-vectors (UCOE-CSPG4-HC(ε), UCOE-CSPG4-LC(κ)) ( Fig. 1c), employing Polymerase Incomplete Primer Extension (PIPE) cloning. UCOE enables higher transfection efficiency and higher proportions of medium-and highexpressing transfectomas than pVITRO1 (see Figure E1 in the Online Repository).
Vectors were linearized before transfections to allow correct integration into the host genome, and transgene-expressing cells were selected. The choice of Expi293F TM as hosts was based on human-like glycosylation profiles, ability to grow in suspension, high-density and serum-free conditions, characteristics crucial for expediting production, scaling up and adaptability to good manufacturing practice (GMP) conditions. We adapted Expi293F TM cells from suspension to adherent growth M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
conditions and vice-versa. Adherent cells were transfected and seeded in selection medium to promote host genome integration of exogenous DNA. Resistant cells were cloned by limiting dilution. We designed a cell-based flow cytometric method to detect functional IgE recognizing natively-expressed antigens to screen antibodysecreting clones (see Figure E1 in the Online Repository). High antibody-expressing clones were amplified and re-adapted to grow in high-density suspension cultures for antibody harvesting.
After selecting the highest-expressing clone, we optimized culture conditions to maximize IgE production and minimize time and resources. We observed a slow decrease of specific daily antibody productivity, consistent with cell growth rate and consumption of culture medium nutrients. This productivity decrease was due to nutrient depletion in the medium rather than cell density (see Figure E2 in the Online Repository). To maximize yields, we tested different seeding Expi-CSPG4 IgE cell concentrations in fresh medium, measuring secreted antibody daily for 5 days. As expected, higher starting cell concentration yielded faster and higher antibody production, with cells seeded at 11x10 6 cells/mL generating 2mg/day (Fig. 1d).
Using high cell concentrations (5x10 6 and 11x10 6 cells/mL), that place cells under stress, we analysed production consistency over time by passaging every two days at 5x10 6 (5M/mL 2D), or every day at 5x10 6 (5M/mL 1D) or 11x10 6 cells/mL (11M/mL 1D), replacing media at every passage. After 4 days all conditions yielded consistent antibody production (Fig. 1e). The 11M/mL 1D and 5M/mL 2D conditions yielded similar production per passage. However, 11M/mL 1D resulted in the highest production per day (see Figure E3 in the Online Repository), suggesting this is optimal for reducing resources and time.
IgE production reached yields of up to 87mg/L/day (83±4 mg/L/day mean±SD), with ability to repeat the process with the same cells at least 3 times without losing production efficiency. Yields in 4 days, in small shaking flask cultures, were approximately 33-fold higher than the most optimal 14-day stable IgE production recorded in shaking flask conditions, and 13-fold higher than 14-day IgE production reported using bioreactors (4). Optimized high-density conditions allowed maximized yields and substantially-reduced medium volumes, achieving 2mg/25mL/day, with similar yields scaling down to 15mL and up to 300mL cultures.
Different culture conditions can affect antibody quality, structural and functional properties, including post-translational modifications such as glycosylation (8). This is highly pertinent for IgE, based on larger size and higher glycosylation levels compared with IgG. We performed structural, glycosylation and functional analyses comparing affinity chromatography-purified antibodies from high-and low-density cultures alongside IgE produced with the previous pVITRO method (4).
Size-exclusion High Pressure Liquid Chromatography (HPLC), showed very similar antibody main peak profiles (Fig. 2a). Lectin blot and LC-MS glycosylation analyses, revealed no significant differences among IgE produced with our method (Fig. 2b, 2c), particularly with regards to oligomannose structures, whose removal is reported to abrogate anaphylaxis (9). LC-MS glycosylation analyses showed a reduction in MAN5 oligomannose structure in new preparations compared to pVITRO IgE.
Antibodies from all conditions showed comparable binding characteristics to target antigen on A375 melanoma cells and to rat basophilic leukemia RBL-SX38 cells expressing human FcεRI (Fig. 2d). All preparations triggered significant and comparable levels of mast cell degranulation when cross-linked by polyclonal anti-IgE or by target antigen on CSPG4 high melanoma cells (Fig. 2e). The hapten-specific NIP-IgE cross-linked by polyclonal anti-IgE, but not by CSPG4 high cells, triggered significant degranulation. These suggest that different preparations and density conditions preserve receptor recognition and antibody potency. Importantly, reduced MAN5 in the new IgE is insufficient to impact IgE binding or functionality.
We therefore present a novel process for serum-free production of IgE with comparable structural and functional characteristics to previous pVITRO method, but at higher yield and in less time than any documented stable platforms and with less resources than transient systems (see Table E1

CSPG4 IgE expression vectors
The pVITRO1-CSPG4-IgE/k vector (Fig 1b) was previously developed in our group (2). The dual UCOE-vector system (UCOE-CSPG4-HC(ɛ) and UCOE-CSPG4-LC(κ)) (Fig 1c) was developed by Polymerase Incomplete Primer Extension (PIPE) cloning as described before (2). Briefly, the coding sequence of CSPG4 IgE heavy and light chains were isolated from pVITRO1-CSPG4-IgE/k vector and UCOE® Mu-H vector (Merck Millipore) was linearized in three portions by using the primers in Table E2 and the PCR conditions reported in Table E3 in the Online Repository. The PCR products were then digested for 2 hours at 37°C with DpnI (NEB) to eliminate template DNA, and two separate mixes were prepared using the three vector portions and one heavy or light chain. The two mixes were incubated overnight at room temperature and then used to transform chemically competent One Shot TM Top10 bacteria (Invitrogen) according to the manufacturer's instruction.

Development of stable IgE expressing Expi293F TM cells
The Expi293F TM cell line (Gibco) was cultured according to the manufacturer's instructions in Expi293 TM Expression Medium (Gibco). One day before transfection, Different clones were frozen and the best clone (Expi-CSPG4 IgE) was used for the optimization of antibody production.

Flow cytometry-based assay for the detection and quantification of antigenspecific IgEs in cell supernatants
This method is based on the detection, in cell supernatants, of IgE antibodies that specifically bind to CSPG4 high tumor (A375 melanoma) cells. A375 cells were detached using PBS supplemented with 5mM EDTA and resuspended in PBS Laboratories) was incubated at 30µg/mL in FACS buffer for 30 minutes at 4 °C followed by one wash as above. Samples were resuspended in 100µL of FACS buffer and analysed using a FACS Canto II (BD Biosciences).

Size exclusion chromatography
Purified antibodies were analysed by size exclusion chromatography as previously described (3). Briefly, gel filtration was performed on a Gilson HPLC system using a Superdex TM 200 10/300 GL column (GE Healthcare), suitable for purifying proteins between 10-300 kDa, at a flow rate of 0.75 mL/min in PBS (pH 7.0, 0.2µm filtered).

Lectin blot
Purified IgE samples (150ng) were reduced with 50 mM DTT and boiled at 95 °C for

Procainamide labeling and cleanup
Released glycans were labelled with procainamide using a procainamide labelling kit (LT-KPROC-24, Ludger Ltd) in a similar manner to that previously reported (5). In short, 150 µL of 30% glacial acetic acid in DMSO was added to a vial of procainamide followed by 150 µL of water. This solution was transferred to a vial of sodium cyanoborohydride to make the final labelling reagent. 20 µL of labeling reagent was added to each sample. The samples were then incubated for 1 hour at 65°C. A HILIC method was performed to clean up the samples and remove free dye using an LC-PROC-96 plate (Ludger Ltd) on a vacuum manifold. Samples were eluted from the cleanup plate in 300 µL of water.

LC-MS analysis of procainamide labeled N-glycans
Samples were analysed by HILIC-LC on an Ultimate 3000 UHPLC using a BEH-

Evaluation of antibody-specific productivity
Expi-CSPG4-IgE cells were cultured in different conditions and IgE secretion and cell viability were monitored daily. Antibody-specific productivity q mAb (pg cell -1 day -1 ) was calculated according to the following equation (8): with m mAb being secreted IgE, N and N 0 being the final and the initial viable cell values, respectively, and t being the days in culture.

Statistical analysis
Error bars represent the standard deviation (SD) or the standard error of the mean (SEM