Optimizing recombinant antibody fragment production: A comparison of artificial intelligence and statistical modeling.

Maximizing the recombinant protein yield necessitates optimizing the production medium. This can be done using a variety of methods, including the conventional "one-factor-at-a-time" approach and more recent statistical and mathematical methods such as artificial neural network (ANN), genetic algorithm, etc. Every approach has advantages and disadvantages of its own, yet even when a technique has flaws, it is nevertheless used to get the best results. Here, one categorical variable and four numerical parameters, including post-induction time, inducer concentration, post-induction temperature, and pre-induction cell density, were optimized using the 232 experimental assays of the central composite design. The direct and indirect effects of factors on the yield of anti-epithelial cell adhesion molecule extracellular domain fragment antibody were examined using statistical methods. The analysis of variance results indicate that the response surface methodology (RSM) model is effective in predicting the amount of produced single-chain fragment variable (p-value = 0.0001 and R2 = 0.905). For ANN modeling, the evaluation using normalized root mean square error (NRMSE) and R2 values shows a good fit (R2 = 0.942) and accurate predictions (NRMSE = 0.145). The analysis of error parameters and R2 of a dataset, which contained 30 data points randomly selected from the complete dataset, showed that the ANN model had a higher R2 value (0.968) compared to the RSM model (0.932). Furthermore, the ANN model demonstrated stronger predictive ability with a lower NRMSE (0.048 vs. 0.064). Induction at the cell density of 0.7 and an isopropyl ß-D-1-thiogalactopyranoside concentration of 0.6 mM for 32 h at 30°C in BW25113 was the ideal culture condition leading to the protein yield of 259.51 mg/L. Under the optimum conditions, the output values predicted by the ANN model (259.83 mg/L) were more in line with the experimental data (259.51 mg/L) than the RSM (276.13 mg/L) expected value. This outcome demonstrated that the ANN model outperforms the RSM in terms of prediction accuracy.

Genome-scale metabolic model-based engineering of Escherichia coli enhances recombinant single-chain antibody fragment production.

PURPOSE: Escherichia coli is an attractive and cost-effective cell factory for producing recombinant proteins such as single-chain variable fragments (scFvs). AntiEpEX-scFv is a small antibody fragment that has received considerable attention for its ability to target the epithelial cell adhesion molecule (EpCAM), a cancer-associated biomarker of solid tumors. Due to its metabolic burden, scFv recombinant expression causes a remarkable decrease in the maximum specific growth rate of the scFv-producing strain. In the present study, a genome-scale metabolic model (GEM)-guided engineering strategy is proposed to identify gene targets for improved antiEpEX-scFv production in E. coli. METHODS: In this study, a genome-scale metabolic model of E. coli (iJO1366) and a metabolic modeling tool (FVSEOF) were employed to find appropriate genes to be amplified in order to improve the strain for incresed production of antiEpEX-scFv. To validate the model predictions, one target gene was overexpressed in the parent strain Escherichia coli BW25113 (DE3). RESULTS: For improving scFv production, we applied the FVSEOF method to identify a number of potential genetic engineering targets. These targets were found to be localized in the glucose uptake system and pentose phosphate pathway. From the predicted targets, the glk gene encoding glucokinase was chosen to be overexpressed in the parent strain Escherichia coli BW25113 (DE3). By overexpressing glk, the growth capacity of the recombinant E. coli strain was recovered. Moreover, the engineered strain with glk overexpression successfully led to increased scFv production. CONCLUSION: The genome-scale metabolic modeling can be considered for the improvement of the production of other recombinant proteins.

Machine learning modeling for solubility prediction of recombinant antibody fragment in four different E. coli strains.

The solubility of proteins is usually a necessity for their functioning. Recently an emergence of machine learning approaches as trained alternatives to statistical models has been evidenced for empirical modeling and optimization. Here, soluble production of anti-EpCAM extracellular domain (EpEx) single chain variable fragment (scFv) antibody was modeled and optimized as a function of four literature based numerical factors (post-induction temperature, post-induction time, cell density of induction time, and inducer concentration) and one categorical variable using artificial neural network (ANN) and response surface methodology (RSM). Models were established by the CCD experimental data derived from 232 separate experiments. The concentration of soluble scFv reached 112.4 mg/L at the optimum condition and strain (induction at cell density 0.6 with 0.4 mM IPTG for 24 h at 23 °C in Origami). The predicted value obtained by ANN for the response (106.1 mg/L) was closer to the experimental result than that obtained by RSM (97.9 mg/L), which again confirmed a higher accuracy of ANN model. To the author's knowledge this is the first report on comparison of ANN and RSM in statistical optimization of fermentation conditions of E.coli for the soluble production of recombinant scFv.

RSM-based Model to Predict Optimum Fermentation Conditions for Soluble Expression of the Antibody Fragment Derived from 4D5MOC-B Humanized Mab in SHuffle™ T7 E. coli.

Overexpression of the EpCAM in epithelial-derived neoplasms makes this receptor a promising target in antibody-based therapy. Due to the lack of N-glycosylation, Escherichia coli (E. coli) seems to be the most appropriate choice for the expression of antibody fragments. However, developing a robust and cost-effective process that produces consistent therapeutic proteins from inclusion bodies is a major challenge. Undoubtedly, it can be circumvented by the soluble expression of these proteins. Utilization of numerous genetically modified hosts and optimization of cultivation conditions are two effective approaches widely used to overcome the insolubility problem. Due to the cytoplasmic expression of DsbC and the ability to the correct formation of disulfide bonds, the Shuffle™ T7 strain can be a suitable host for the soluble production of recombinant proteins. Here, Box-Behnken design (BBD)- Response surface methodology (RSM) modeling was employed to develop optimized culture conditions for 4D5MOC-B scFv fragment production in SHuffle™ T7 strain while solubility and production level were considered as responses. Although both responses were significantly influenced by post-induction temperature, cell density at induction time, and IPTG concentration, the temperature had the largest effect. The maximum experimental soluble protein obtained by adding 1 mM of IPTG into the M9 medium when the cell density reached 0.7 at 23 ᵒC was 693.56 µg/mL which was in good correlation with the predicted value of 720.742 µg/mL. Predictable total expression value was also experimentally verified. This strategy can be scaled-up for the production of large amounts of scFvs from SHuffle™ T7 E. coli to facilitate their potential applications as therapeutic and diagnostic agents.

Statistical optimization of culture conditions for expression of recombinant humanized anti-EpCAM single-chain antibody using response surface methodology.

BACKGROUND AND PURPOSE: The epithelial cell adhesion molecule (EpCAM), is one of the first cancer- associated markers discovered. Its overexpression in cancer stem cells, epithelial tumors, and circulating tumor cells makes this molecule interesting for targeted cancer therapy. So, in recent years scFv fragments have been developed for EpCAM targeting. EXPERIMENTAL APPROACH: In this study, an scFv against EpCAM extracellular domain (EpEX) derived from 4D5MOC-B humanized mAb was expressed in Escherichia coli k12 strain, and in order to obtain the optimum culture conditions in chemically defined minimal medium, response surface methodology (RSM) was employed. According to the RSM-CCD method, a total of 30 experiments were designed to investigate the effects of various parameters including isopropyl-b-D-thiogalactopyranoside (IPTG) concentration, cell density before induction, post-induction time, and post-induction temperature on anti EpEX-scFv expression level. FINDINGS/RESULTS: At the optimum conditions (induction at cell density 0.8 with 0.8 mM IPTG for 24 h at 37 °C), the recombinant anti EpEX-scFv was produced at a titer of 197.33 µg/mL that was significantly consistent with the prediction of the model. CONCLUSION AND IMPLICATION: The optimized-culture conditions obtained here for efficient production of anti EpEX-scFv in shake flask cultivation on a chemically defined minimal medium could be applied to large- scale fermentation for the anti EpEX-scFv production.

Solubility assessment of single-chain antibody fragment against epithelial cell adhesion molecule extracellular domain in four Escherichia coli strains.

BACKGROUND: Overexpression of the EpCAM (epithelial cell adhesion molecule) in malignancies makes it an attractive target for passive immunotherapy in a wide range of carcinomas. In comparison with full-length antibodies, due to the small size, the scFvs (single-chain variable fragments) are more suitable for recombinant expression in E. coli (Escherichia coli). However, the proteins expressed in large amounts in E. coli tend to form inclusion bodies that need to be refolded which may result in poor recovery of bioactive proteins. Various engineered strains were shown to be able to alleviate the insolubility problem. Here, we studied the impact of four E. coli strains on the soluble level of anti-EpEX-scFv (anti-EpCAM extracellular domain-scFv) protein. RESULTS: Although results showed that the amount of soluble anti-EpEX-scFv obtained in BL21TM (DE3) (114.22 ± 3.47 mg/L) was significantly higher to those produced in the same condition in E. coli RosettaTM (DE3) (71.39 ± 0.31 mg/L), and OrigamiTM T7 (58.99 ± 0.44 mg/L) strains, it was not significantly different from that produced by E. coli SHuffleTM T7 (108.87 ± 2.71 mg/L). Furthermore, the highest volumetric productivity of protein reached 318.29 ± 26.38 mg/L in BL21TM (DE3). CONCLUSIONS: Although BL21TM (DE3) can be a suitable strain for high-level production of anti-EpEX-scFv protein, due to higher solubility yield (about 55%), E. coli SHuffleTM T7 seems to be better candidate for soluble production of scfv compared to BL21TM (DE3) (solubility yield of about 30%).