High-Yield Biosynthesis Process and Characterization of Brolucizumab

Abstract

Brolucizumab was the first single-chain fragment variable (scFv) antibody approved by the FDA for the treatment of age-related macular degeneration (AMD). However, the manufacturing process of brolucizumab remains rarely reported. This study aimed to explore a bioprocess for the production of brolucizumab, where it is expressed as inclusion bodies (IBs) in Escherichia coli (E. coli) BL21 (DE3) cells. In this work, IBs were initially obtained via high cell density fermentation (HCDF) at a high expression level of 30 g/L, followed by denaturation, refolding, and purification to obtain brolucizumab. The refolding parameters were systematically optimized to ensure a high yield of brolucizumab, with 413 mg of the target protein from a 1-L fermentation broth, and purity exceeding 98%. In addition, the amino acid sequence coverage and disulfide bond pairing of the protein were further verified. The results confirmed that brolucizumab has excellent structural integrity, high purity, and notable biological activity. The biosynthetic process holds significant potential for therapeutic applications of brolucizumab and provides valuable insights for the further development of additional scFv bioprocesses.

 

MissingLinkText

Keywords

age-related macular degeneration - brolucizumab - inclusion bodies refolding - recombinant expression

Introduction

Age-related macular degeneration (AMD) involves abnormal retinal/choroidal neovascularization that disrupts retinal structure through fluid leakage, causing macular damage.[1] [2] AMD is the leading cause of severe vision loss in adults over 65 years of age. However, with the global aging trends, its prevalence is increasing, creating an urgent need for clinical treatment of the disease.[3] [4] Vascular endothelial growth factor (VEGF), a primary driver of AMD pathophysiology, can be targeted by anti-VEGF therapies.[5] [6] [7] [8] Intravitreal anti-VEGF injections have now become the standard treatment to preserve vision and improve patient outcomes globally.[9]

In 2019, brolucizumab, as an anti-VEGF drug, was approved by the Food and Drug Administration (FDA) for the treatment of AMD. It inhibits the VEGF pathway by blocking ligand–receptor interactions and impeding the growth of diseased neovascularization.[10] [11] Small molecular weight, higher molar doses, extensive tissue penetration, and good local efficacy are features of brolucizumab.[12] [13] Additionally, it has no Fc region in its molecular structure and therefore has lower immunogenicity.[7] It consists of 252 amino acids, with a theoretical molecular weight of 26 kDa, has no glycosylation sites, and contains two intra-chain disulfide bonds.[13] It can be produced by a prokaryotic expression system. However, there are few reports illustrating its specific manufacturing process. Given the high cost of the original drug (1,850 USD for 6 mg per vial at the recommended therapeutic dosage),[14] there is an urgent need to develop an efficient brolucizumab production strategy to meet the increasing clinical demand for this drug.[3]

Using Escherichia coli (E. coli) to produce brolucizumab in the form of inclusion bodies (IBs) may represent an optimal solution. First, E. coli is a preferred host for the production of recombinant, non-glycosylated proteins, owing to its well-characterized genetic background, high expression levels, cost-effectiveness, and ease of manufacture.[15] [16] [17] Second, the expression through IBs provides numerous benefits, including resistance to proteolytic degradation and simplification of the separation and purification processes.

Encouragingly, this study established a biosynthesis process to produce brolucizumab using E. coli BL21 (DE3) cells. However, significant challenges persist in process development, particularly in the refolding of the target protein containing two disulfide bond pairs. This is because the refolding process may be prone to misfolding and aggregation, reducing the yields of the protein in a biologically active form.[18] [19] In this work, the refolding step is optimized. The physicochemical characterizations of brolucizumab, including disulfide bond mapping, were assessed to verify the structural integrity. Furthermore, a biological activity assay was conducted to confirm the functionality of the drug. This study lays a solid foundation for the advancement of biosimilars of brolucizumab and provides valuable insights for the development of other single-chain fragment variable (scFv) bioprocesses.

 

MissingLinkText

Material and Methods

Material and Reagents

The materials used in this study were yeast extract (Angel, Hubei, China), dithiothreitol (DTT) (Solarbio, Beijing, China), polyethylene glycol-800 (Thermo Fisher, Massachusetts, United States), human umbilical vein endothelial cells (HUVEC) (PromoCell, Baden-Württemberg, Germany), VEGF (ACRO, Shanghai, China), serum-free medium (Thermo Fisher, Massachusetts, United States), alamar blue (Yeasen, Shanghai, China), and adhesion factor (Thermo Fisher, Massachusetts, United States). Other analytical-grade chemicals (e.g., kanamycin, guanidine hydrochloride [GuHCl], Urea, Arg, glutathione [GSH], glutathione disulfide [GSSG], etc.) were sourced from Sinopharm Chemical Reagent (Shanghai, China), while the plasmid, strains, and ranibizumab were obtained from storage in our laboratory.

Luria-Bertani (LB) medium, consisting of peptone (10 g/L), yeast extract (5 g/L), and NaCl (10 g/L), was used for seed culture.[20] The high cell density fermentation (HCDF) medium included glucose (30 g/L), yeast extract (7.5 g/L), KH₂PO₄ (8 g/L), trisodium citrate (7 g/L), NaCl (5 g/L), (NH₄)₂SO₄ (5 g/L), MgSO₄ (0.5 g/L), and CaCl₂ (0.25 g/L). The feeding medium included 80% glucose and 2% MgSO₄.

 

MissingLinkText

Construction of Recombinant Expression Vector and Expression Verification

The target gene was designed based on the protein sequence of brolucizumab (MEIVMTQSPSTLSASVGDRVIITCQASEIIHSWLAWYQQKPGKAPKLLIYLASTLASGVPSRFSGSGSGAEFTLTISSLQPDDFATYYCQNVYLASTNGANFGQGTKLTVLGGGGGSGGGGSGGGGSGGGGSEVQLVESGGGLVQPGGSLRLSCTASGFSLTDYYYMTWVRQAPGKGLEWVGFIDPDDDPYYATWAKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCAGGDHNSGWGLDIWGQGTLVTVSS) and subsequently cloned into the pET28a plasmid to construct the expression vector pET28a-Bro. Following electroporation into competent E. coli BL21 (DE3) cells, transformants were selected on medium plates supplemented with kanamycin. Kanamycin selects E. coli transformants carrying the kanamycin-resistant plasmid by inhibiting the growth of non-transformed bacteria lacking the resistance gene.

Several transformants were then inoculated into LB shaking flasks containing 25 μg/mL kanamycin. After cultivation at 37°C and 220 rpm for 4 hours, isopropyl-β-D-1-thiogalactopyranoside (IPTG) was added to induce the expression of the target protein. After 12 hours of incubation, cells from the fermentation broth were collected by centrifugation (8000 × g, 15 minutes, 4°C) and then resuspended in 25 mmol/L Tris-HCl buffer (pH 8.0). Cells were disrupted in an ice bath using an ultrasonic cell disruptor (SCIENTZ, Ningbo, China). Cell debris was removed by centrifugation at 10,000 × g. The supernatant and the resuspended pellet were collected for the next 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis.

 

MissingLinkText

High Cell Density Fermentation

The seed flasks containing kanamycin were inoculated (220 rpm, 37°C, 6 hours) with cells from frozen glycerol stock at a concentration of 1% (v/v) until the cell density measured as optical density at 600 nm (OD₆₀₀) reached approximately 1.5 to 2.0. Subsequently, the culture was inoculated at a ratio of 10% (v/v) into 3 L of pre-sterilized fermentation medium within a 5-L fermenter (Bxbio, China). The initial fermentation temperature was 37°C, with an airflow rate of 3 L/min, a pH level controlled around 7.0, and an agitation speed set to 400 rpm.

During the fermentation process, the dissolved oxygen (DO) level in the fermenter was sustained between 30 and 40% by adjusting the agitation speed (ranging from 400 to 1,000 rpm), airflow (between 3 and 6 L/min), and container pressure, thus ensuring an adequate oxygen supply for the growth of E. coli cells. The pH was adjusted using a 2 mol/L sulfuric acid solution and a 50% (v/v) ammonia solution, while an anti-foaming agent was introduced to mitigate the formation of foam. When the pH and DO levels increased significantly (indicating depletion of the carbon sources of the medium), the feeding medium was added to the medium. After approximately 12 hours, the OD₆₀₀ of the culture reached about 60. At this juncture, 0.3 mmol/L IPTG was added to induce the expression of the target protein. The culture was maintained for an additional 11 to 13 hours, after which centrifugation (8000 × g, 4°C, 15 minutes) was performed to harvest the cells post-fermentation.

 

MissingLinkText

Denaturation

The effects of different concentrations of denaturants, pH levels, and DTT on the dissolution of IBs were investigated by a series of single-factor experiments and an orthogonal experiment. DTT was used as a reducing agent to reduce and cleave disulfide bonds in proteins.

Denaturation solutions were prepared by adding 2, 4, or 6 mol/L GuHCl, or 2, 4, 6, and 8 mol/L urea to a basic buffer containing 100 mg/mL IBs and 25 mmol/L Tris-HCl. The mixtures were stirred at 25°C for 1 hour, followed by centrifugation (10,000 × g, 15°C, 20 minutes). The supernatants were collected to assess the clarity of the solutions.[21] Subsequently, the pH levels of the buffer (7.0, 7.5, 8.0, 8.5, 9.0, 9.5, and 10.0) and DTT concentrations (0, 10, 20, 30, 40, 50, and 100 mmol/L) were screened. The supernatant was analyzed by reversed-phase high-performance liquid chromatography (RP-HPLC) to assess the areas of the denaturation peaks.

Based on the outcomes of the single-factor experiments, a three-level, four-factor orthogonal experiment (L9 3[4]) was designed to assess the interaction effects among the factors. As a result, nine experimental groups were established in accordance with the design generated by the software (SPSS Statistics 26), as detailed in [Table 1].

Table 1 The design of the denaturation orthogonal experiment

	Denaturing factors				Results
Experiments No.	pH	DTT concentration	Tris-HCl concentration	Reaction time	Peak areaa
1	8.0	15 mmol/L	25 mmol/L	1 hour	248051
2	8.0	20 mmol/L	50 mmol/L	2 hours	282558
3	8.0	25 mmol/L	75 mmol/L	3 hours	276272
4	8.5	15 mmol/L	50 mmol/L	3 hours	264827
5	8.5	20 mmol/L	75 mmol/L	1 hour	293351
6	8.5	25 mmol/L	25 mmol/L	2 hours	286460
7	9.0	15 mmol/L	75 mmol/L	2 hours	284112
8	9.0	20 mmol/L	25 mmol/L	3 hours	275958
9	9.0	25 mmol/L	50 mmol/L	1 hour	282793
Mean of level 1	268960	265663	270156	274732
Mean of level 2	281546	283956	276726	284377
Mean of level 3	280954	281842	284578	272352
Range	12586	18293	14422	12025
Validation of optimal conditions
Experiments No.	pH	DTT concentration	Tris-HCl concentration	Reaction time	Peak areaa
1	8.5	20 mmol/L	75 mmol/L	2 hours	297158
2	8.5	20 mmol/L	75 mmol/L	2 hours	289674
3	8.5	20 mmol/L	75 mmol/L	2 hours	294506

Abbreviation: DDT, dithiothreitol.

Note: ^aDetermined by a reversed-phase high-performance liquid chromatography method. The chromatographic peak areas were integrated using Waters Empower 3 Software (Waters, Massachusetts, United States).

 

MissingLinkText

Refolding

In vitro refolding, as a significant rate-limiting step, affects the overall throughput of downstream processing and the production of scFv generated in E. coli.[22] In our study, a dilution refolding method was used to gradually decrease the concentration of the denaturant, thereby promoting the transition of the target protein from its fully unfolded denatured state to its native folded structure, while ensuring the correct formation of disulfide bonds.[23]

The denaturation solution was diluted 50-fold into the refolding buffer containing 2 mol/L urea, 0.1 mol/L arginine (Arg), 1 mmol/L reduced GSH, 1 mmol/L GSSG, and 50 mmol/L Tris-HCl (pH 8.5), at a flow rate of 1 mL/min. The reaction was gently stirred at 4°C for 4 hours. After centrifugation (12,000 × g, 4°C, 20 minutes), the supernatant was collected for the 15% SDS-PAGE and size exclusion chromatography–high performance liquid chromatography (SEC-HPLC) analysis for the detection of aggregates and the calculation of refolding yield.

Subsequently, the effects of various factors on the refolding efficiency were assessed by a series of single-factor studies, including buffer pH levels (7.0, 8.0, 9.0, 10.0, and 11.0), additives (5% sucrose; 5% D-sorbitol; 5% polyethylene glycol-800, abbreviated as PEG-800; 5% glycerol; 5% Triton X-100; and 5% mannitol), Arg concentrations (0, 0.1, 0.2, 0.3, 0.4, and 0.5 mol/L), urea concentrations (0, 0.5, 1.0, 1.5, 2.0, and 2.5 mol/L), and refolding times (0, 2, 4, 6, 8, 10, and 24 hours).[24] [25] [26] [27] On this basis, the key factors were selected for further optimization through a three-level, four-factor orthogonal experiment (L9 3⁴), as detailed in [Table 2]. Ultimately, the original and optimized refolding conditions at 4°C were compared.

Table 2 The design of the refolding orthogonal experiment

	Refolding factors				Results
Experiments No.	pH	Arg concentration	Urea concentration	Tris-HCl concentration	Peak areaa
1	9.0	0.2 mol/L	1.5 mol/L	25 mmol/L	39783
2	9.0	0.3 mol/L	2.0 mol/L	50 mmol/L	44822
3	9.0	0.4 mol/L	2.5 mol/L	75 mmol/L	48767
4	9.5	0.2 mol/L	2.0 mol/L	75 mmol/L	47482
5	9.5	0.3 mol/L	2.5 mol/L	25 mmol/L	53617
6	9.5	0.4 mol/L	1.5 mol/L	50 mmol/L	41121
7	10.0	0.2 mol/L	2.5 mol/L	50 mmol/L	56943
8	10.0	0.3 mol/L	1.5 mol/L	75 mmol/L	53904
9	10.0	0.4 mol/L	2.0 mol/L	25 mmol/L	51519
The mean of level 1	44457	48069	44936	48306
The mean of level 2	47407	50781	47941	47629
The mean of level 3	54122	47136	53109	50051
Range	9665	3645	8173	2422
Validation of optimal conditions
Experiments No.	pH	Arg concentration	Urea concentration	Tris-HCl concentration	Peak areaa
1	10.0	0.3 mol/L	2.5 mol/L	75 mmol/L	72909
2	10.0	0.3 mol/L	2.5 mol/L	75 mmol/L	69882
3	10.0	0.3 mol/L	2.5 mol/L	75 mmol/L	67209

Note: ^aDetermined by a SEC-HPLC method. The chromatographic peak areas were integrated using Waters Empower 3 Software (Waters, Massachusetts, United States). SEC-HPLC, size exclusion chromatography–high performance liquid chromatography.

 

MissingLinkText

Purification of Brolucizumab

An anion exchange chromatography method was selected for the initial purification, considering the isoelectric point (pI) of brolucizumab being 4.8 and its other physicochemical properties. Prior to this, a hollow fiber tube membrane module (Asahi-KASEI, Japan) with a molecular cut-off of 10 kDa was employed to desalinate the refolding solution through ultrafiltration. Ultrafiltration was performed to exchange the initial solution with 25 mmol/L Tris-HCl buffer (pH 8.0) while maintaining a constant conductivity of 5 mS/cm for protein stabilization. Following filtration through a 0.2-μm filter (Merck Millipore, Germany), the sample was obtained for the purification step.

In this study, five 5 mL Cytiva anion exchange chromatography resins, namely, Q FF, Q XL, 30 Q, Capto Q, and Q HP (Cytiva, Washington, United States), were optimized. The columns were pre-equilibrated with five column volumes (CVs) of a buffer consisting of 25 mmol/L Tris-HCl at pH 8.0. Brolucizumab was eluted and harvested using an elution buffer of 25 mmol/L Tris-HCl with 1 mol/L NaCl at pH 8.0. 15% SDS-PAGE, SEC-HPLC, and RP-HPLC were used for analysis.

After determining the optimal resin, the pH values (7.5, 8.0, 8.5, and 9.0) of the flow buffer and the sample were further optimized. In addition, the collected eluted fractions were used for a second step of purification to improve the purity of the target protein. As shown in [Table 3], four purification methods were evaluated. The purity and yield were compared. Both SP HP and SP FF chromatography columns were pre-equilibrated with five CVs of 25 mmol/L NaOAc buffer at pH 3.5. Brolucizumab was subsequently eluted using a buffer consisting of 25 mmol/L NaOAc and 1 mol/L NaCl buffer at pH 3.5. For phenyl HP hydrophobic interaction chromatography (HIC), the column was equilibrated with five CVs of 25 mmol/L Tris-HCl and 0.6 mol/L NaCl at pH 8.0, followed by elution of brolucizumab with 25 mmol/L of Tris-HCl at pH 8.0. In contrast, the Adhere multimodal chromatography column underwent equilibration using five CVs of 25 mmol/L Tris-HCl containing 0.2 mol/L NaCl at pH 8.0, after which brolucizumab was eluted with 100 mmol/L citric acid buffer at pH 3.0.

Table 3 Chromatographic methods screening for the second step of separation and purification

Column	Sample pH	Sample conductivity (mS/cm)	Mobile phase buffer A	Mobile phase buffer B
SP HP	3.5	5.1	25 mmol/L NaOAc, pH 3.5	25 mmol/L NaOAc, 1 mol/L NaCl, pH 3.5
SP FF	3.5	5.1	25 mmol/L NaOAc, pH 3.5	25 mmol/L NaOAc, 1 mol/L NaCl, pH 3.5
Phenyl HP	8.0	50.2	25 mmol/L Tris-HCl, 0.6 mol/L NaCl, pH 8.0	25 mmol/L Tris-HCl, pH 8.0
Adhere	8.0	18.0	25 mmol/L Tris-HCl, 0.2 mol/L NaCl, pH 8.0	0.1 mol/L citric acid, pH 3.0

15% SDS-PAGE

SDS-PAGE analysis of brolucizumab was performed using precast 15% polyacrylamide gels and multicolor protein markers (Epizyme, Shanghai, China). Gels were assembled in a Bio-Rad electrophoresis chamber filled with running buffer. Protein samples (80 μL) were combined with 20 μL Laemmli buffer and denatured at 100°C for 5 minutes. Electrophoresis commenced at 70 V for 15 minutes, followed by an increase to 150 V until the tracking dye reached the gel bottom. Subsequently, the gels were stained with Coomassie Blue (100°C, 5 minutes), destained for 25 minutes, and imaged with a gel imaging system (Genetimes, Shanghai, China) for analytical evaluation.

 

MissingLinkText

Size Exclusion Chromatography–High Performance Liquid Chromatography (SEC-HPLC) Analysis

The refolding supernatant of brolucizumab was examined via SEC-HPLC employing a Welch Xtimate SEC-200 column (Shanghai, China), with the mobile phase consisting of 90% 50 mmol/L phosphate buffer and 10% acetonitrile (ACN) at a detection wavelength of 280 nm. The column temperature was 25°C. The flow rate was 0.5 mL/min with the same gradient elution for 35 minutes.

 

MissingLinkText

Reversed-Phase High-Performance Liquid Chromatography (RP-HPLC) Analysis

Thermo MabPac™ RP column (Massachusetts, United States) was utilized for RP-HPLC analysis, with mobile phase A comprising 0.1% (v/v) trifluoroacetic acid (TFA) in water and mobile phase B consisting of 0.1% (v/v) TFA in ACN. Gradient conditions were applied from 20 to 90% of mobile phase B in 35 minutes at a flow rate of 0.6 mL/min. The column temperature was set at 60°C. The detection wavelength was 280 nm.

 

MissingLinkText

Liquid Chromatography–Mass Spectrometry (LC-MS) Analysis

The primary structure of brolucizumab was confirmed by the liquid chromatography–mass spectrometry (LC-MS) with a YMC Triart-C4 column (Kyoto, Japan), an ACQUITY UPLC system, and an Xevo G2-XS QTOF mass spectrometer (Waters, Massachusetts, United States). The raw data were processed with MassLynx V4.1 software (Waters, Massachusetts, United States). The flow rate was 0.3 mL/min. The total run time was 10 minutes. The detection wavelength was 280 nm.

 

MissingLinkText

 

MissingLinkText

Amino Acid Sequence Determination and Disulfide Bond Mapping

Brolucizumab was denatured by the addition of 6 mol/L GuHCl and 0.1 mol/L DTT. The mixture was incubated at 37°C for 60 minutes. Subsequently, alkylating agent iodoacetamide (IAM) was added at a concentration of 0.2 mol/L, and the reaction was carried out for 45 minutes. The purpose of IAM was to alkylate free cysteine residues, preventing the reformation of disulfide bonds after the reduction with DTT, and ensuring stable peptide fragments for accurate LC-MS analysis. After the reaction, the solution was exchanged for 25 mmol/L Tris-HCl (pH 7.5). An enzyme solution was added to cleave the peptide chain at specific sites, and the reaction was maintained at 37°C. The mixture was centrifuged (12,000 × g, 4°C, 5 minutes). The supernatant was collected for LC-MS analysis. No reducing agents should be added when confirming disulfide bonds.

 

MissingLinkText

Cell Proliferation Inhibition Assay

The in vitro activity of brolucizumab was assessed by determining the HUVEC proliferation inhibition.[28] Brolucizumab and the positive control drug, ranibizumab, were diluted at different concentrations in a detection solution consisting of 100 mL of serum-free medium (SFM) and 5 mL of adhesion factor (Gibco, California, United States). The resulting samples were combined with 120 ng/mL of VEGF₁₆₅ in a 1:1 ratio (v:v) in a 96-well plate and incubated at 25°C for 30 minutes. The purpose of VEGF₁₆₅ was to stimulate HUVEC proliferation, creating a model to assess the inhibitory effects of brolucizumab and ranibizumab on VEGF-induced cell growth.

HUVEC were dispersed in the detection solution to reach a concentration of 2 × 10⁵ cells/mL. Subsequently, 50 μL of the cell suspension was added to each well of a black 96-well plate, followed by the addition of 50 μL of the VEGF₁₆₅–sample mixture. The plate was incubated at 37°C in 5% CO₂ for 90 hours. Then 25 μL of Alamar Blue was added. The plate was incubated for another 6 hours. Fluorescence intensity was measured using an excitation wavelength of 530 nm and an emission wavelength of 590 nm.

 

MissingLinkText

Statistical Analysis

Data represent mean ± standard deviation (SD) of triplicate injections. Statistical analysis was conducted using commercial software (GraphPad Prism, version 9.5, California, United States) with Duncan's multiple range test (SPSS Inc., Chicago, IL, United States) being used to compare groups with a p-value of <0.05 indicating statistical significance. In addition, chromatographic peak areas were integrated using Waters Empower 3 Software (Waters, Massachusetts, United States).

 

MissingLinkText

 

MissingLinkText

Results and Discussion

Successful Establishment of Recombinant Plasmids Expressing Brolucizumab in E. coli BL21 (DE3) Strain

The target gene was inserted into the plasmid. The schematic of the pET28a-Bro expression vector is presented in [Fig. 1A]. The E. coli BL21 (DE3) strain, which contains T7 RNA polymerase, is capable of efficiently expressing foreign genes, ensuring the reliability and repeatability of the experiments.[29] Since the plasmid harbors an antibiotic resistance gene, the successfully transformed strains were able to grow on plates containing kanamycin,[30] which was consistent with our experimental results.

Fig. 1 Construction of plasmids and verification of recombinant expression. (A) Schematic representation of the expression vector pET28a-Bro. (B) Reduced SDS-PAGE analysis of changes in the supernatant and precipitate of the fermentation lysate before and after induction. The arrow indicated the expression band of Brolucizumab. M, marker (Epizyme, wj103). SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis.

After fermentation at the shake flask level, the supernatant and precipitate from the cell lysate were collected for analysis. As illustrated in [Fig. 1B], a new distinct band, with a molecular weight ranging between 25 and 27 kDa, appeared when compared to the non-induced cells. This band's molecular weight was close to the theoretical molecular weight of brolucizumab (26,313 Da). Therefore, it can be concluded that the strains were able to effectively express brolucizumab as IBs in precipitation within the E. coli BL21 (DE3) cells following induction.

 

MissingLinkText

High Cell Density Fermentation in a 5-L Fermenter

The HCDF method has the advantages of favoring pH control during fermentation, reducing foam production, increasing DO, and enabling nutrient feeding, when compared to a shaking flask fermentation, thus contributing to a higher biomass yield.[31] Furthermore, during subsequent process development, we were faced with a high demand for brolucizumab that could not be met with the biomass produced by shaking flask fermentation. Consequently, the HCDF method for the brolucizumab strain in a 5-L fermenter became essential.

Similar to the results in [Fig. 1B], HCDF expression analysis revealed that the expression of the target protein in the supernatant remained undetectable with the induction conditions employing 0.3 mmol/L IPTG at 37°C for 12 hours. This finding aligns with previous reports indicating that the use of protease-deficient strains, such as E. coli BL21 (DE3), along with higher induction temperatures and inducer concentrations, reduces the degradation of target proteins by proteases and promotes the formation of IBs.[32] Additionally, the presence of four unpaired cysteine residues in the amino acid sequence contributes to the formation of IB.[33] At the end of the fermentation process, we obtained 180 g of wet cell mass per liter of fermentation broth and 30 g of IBs following high-pressure homogenization.

 

MissingLinkText

Denaturation of IBs

As shown in [Supplementary Fig. S1] (available in the online version) the clarification of dissolved IBs with different concentrations of denaturants showed that both 6 mol/L GuHCl and 8 mol/L urea were effective in denaturing the protein, with 6 mol/L GuHCl yielding superior results. GuHCl was more effective than urea in altering the hydrophobic and hydrophilic interactions among the surface residues, thereby enhancing the protein's solubility.[34] Furthermore, there is evidence that high concentrations of denaturants induce peptide unfolding by disrupting various chemical bonds within and between protein molecules in IBs through ionic interactions,[35] which is consistent with our observations.

In this work, the peak area of the denaturation of brolucizumab was highest at pH 9.0, significantly different from the results at other pH levels ([Fig. 2A], all p < 0.05). This may be attributed to changes in the three-dimensional structure of proteins prompted by changes in environmental pH, which subsequently affect their solubility.[36] In contrast, alkaline conditions tend to promote the solubility of a wide range of proteins.[29] In addition, the peak area of the denaturation of brolucizumab was highest when the concentration of DTT in the solution reached 20 mmol/L, suggesting that the concentration of DTT in the solution needed to exceed 20 mmol/L to fully disrupt its disulfide bonds. In combination with the results of the orthogonal experiment ([Table 1]), our data showed that the optimal denaturation conditions were: a solution containing 6 mol/L GuHCl, 20 mmol/L DTT, and 75 mmol/L Tris-HCl, with a pH of 8.5, and a reaction time of 2 hours.

Fig. 2 Denaturation results under different (A) pH values and (B) DTT concentrations. Data were expressed as mean ± SD, which was calculated from the area of the denatured chromatographic peak of brolucizumab with three repeats. Different letters indicated statistically significant differences between the means (p < 0.05). DTT, dithiothreitol; SD, standard deviation.

 

MissingLinkText

Refolding of IBs

A systematic refolding strategy was developed to enhance brolucizumab's refolding accuracy and yield. Initial optimization focused on buffer pH and additives, followed by refining Arg/urea concentrations and reaction time, with orthogonal experiments validating critical factors.

The proteins could rapidly form aggregates with exposed hydrophobic amino acids under neutral pH conditions, while increasing the pH value could reduce aggregation rates but enhance disulfide bond formation and minimize misfolding.[37] Our data showed that the buffer pH values significantly influence the refolding efficiency, with an optimal pH of 10.0 for refolding ([Fig. 3A]).

Fig. 3 Refolding results under various conditions. The peak area changes were observed at different (A) pH levels, (B) additives, (C) Arg concentrations, (D) urea concentrations, and (E) reaction times. (F) SDS-PAGE comparison was made between refolding under initial conditions and after optimization by orthogonal testing. The initial condition was 2 mol/L urea, 0.1 mol/L Arg, 50 mmol/L Tris-HCl, 1 mmol/L reduced GSH, 1 mmol/L GSSG, at pH 8.5, with a reaction duration of 4 hours at 4°C. The optimal condition was 2.5 mol/L urea, 0.3 mol/L Arg, 75 mmol/L Tris-HCl, 1 mmol/L GSH, 1 mmol/L GSSG, at pH 10.0, with a reaction duration of 2 hours at 4°C. The arrow indicated the expression band of Brolucizumab. Data were expressed as mean ± SD, which were calculated from the area of the reductive chromatographic peak of brolucizumab with three repeats. Different letters indicated statistically significant differences between the means (p < 0.05). Arg, arginine; GSH, glutathione; GSSG, glutathione disulfide; SD, standard deviation; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis.

Additives affect the solubility and stability of unfolded proteins, folding intermediates, and fully folded proteins.[38] However, this study revealed that refolding was more effective in the absence of additional additives ([Fig. 3B]). This may be attributed to a mismatch between these additives and the physicochemical properties of brolucizumab.[39]

Arg, in conjunction with low concentrations of urea, can effectively disrupt hydrophobic interactions within protein cores, thereby enhancing solubility and reducing aggregate formation.[40] As illustrated in [Fig. 3C], [D], the study demonstrated that the yield of brolucizumab increased as the concentrations of Arg and urea increased, with the maximum effect being seen at 0.5 mol/L Arg and 2.5 mol/L urea. However, when Arg is excessive, it will enhance conductivity and complicate the downstream purification. When urea is employed at a high concentration, it will risk the re-denaturation. Therefore, strict monitoring is required, and it is critical to find the optimal concentration balance between the two in the following study.

The refolding duration was investigated. Our data suggested that the peak area of brolucizumab reached a maximum at 2  hours ([Fig. 3E]), whereas it significantly reduced with prolonged refolding time (24  hours). The decrease in refolding efficiency may be attributed to an increase in aggregate precipitation. This observation aligned with prior studies demonstrating that extended refolding time correlated with reduced solution clarity, increased aggregate formation, and diminished protein activity and recovery rates.[41]

We further analyzed the interactions of the factors through orthogonal experiments ([Table 2]) and established the following order of influence on refolding: pH > urea concentration > Arg concentration > Tris-HCl concentration. As a result, the optimal conditions for refolding were identified as follows: 2.5 mol/L urea, 0.3 mol/L Arg, 75 mmol/L Tris-HCl, 1 mmol/L GSH, 1 mmol/L GSSG, at pH 10.0, with a reaction duration of 2 hours at 4°C. In addition, the optimized conditions significantly enhanced the clarity of the post-refolding solution compared to pre-optimized methods. SDS-PAGE results indicated a marked reduction in aggregate bands and a substantial darkening of the target protein band ([Fig. 3F]). Furthermore, RP-HPLC analysis demonstrated an increase in refolding yield from 9.5 to 28% ([Supplementary Fig. S2], available in the online version), greatly reducing the loss of refolded proteins, and confirming a highly efficient optimized refolding method.

 

MissingLinkText

Purification Process of Brolucizumab

After refolding, anion exchange chromatography was preferred to purify the solution based on its isoelectric point, and it was evaluated in terms of yield and purity. Our data showed that Q HP resin showed the best performance under identical elution conditions ([Supplementary Table S1], available in the online version). HP resin's high performance may be attributed to several factors, including its optimized pore size and particle size, rapid mass transfer kinetics, stable substrate, and high chemical compatibility. These properties make it excellent for biomolecular purification. In addition, the effect of pH on protein purification is multifaceted, encompassing the charge state, stability, and solubility of the protein.[42] In this present work, pH 8.5 was the most appropriate when Q HP resin was used, resulting in a target protein purity of 85.8% and a yield of 83.3% ([Supplementary Table S1], available in the online version).

The second-step purification experiments demonstrated that HIC provided the best purification effect compared to others ([Supplementary Table S2], available in the online version). This may be attributed to the addition of sodium sulfate (Na₂SO₄), which increased the conductivity of the solution and fully exposed the hydrophobic groups of brolucizumab, facilitating efficient binding to the Phenyl HP chromatography column. As a result, the final product had a purity exceeding 98% and a yield of 85.8%, demonstrated by the SEC-HPLC and RP-HPLC analysis ([Fig. 4]). Given the above, we have established an efficient downstream purification process that takes advantage of the physicochemical characteristics of brolucizumab, as shown in [Fig. 5], which provides a summary of the developed technical process, where 413 mg of the final brolucizumab product could be harvested from 1 L of fermentation medium.

Fig. 4 Purification of brolucizumab. (A) SDS-PAGE results of upstream and downstream processes. The reduced state denoted the purified protein solution supplemented with reducing agents. (B) SEC-HPLC analysis of brolucizumab. (C) RP-HPLC analysis of brolucizumab. RP-HPLC, reversed-phase high-performance liquid chromatography; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; SEC-HPLC, size exclusion chromatography–high performance liquid chromatography.

Fig. 5 Brief process flow for brolucizumab production. IBs, inclusion bodies; IPTG, isopropyl-β-D-1-thiogalactopyranoside.

 

MissingLinkText

Structural Confirmation

The deconvoluted mass result of brolucizumab ([Fig. 6A]) closely matched the theoretical value (26,313 Da) calculated using the amino acid sequence and the Compute pI/Mw tool. We then assessed the amino acid sequence of its primary structure through peptide mapping. Brolucizumab was trypsinized into characteristic peptides and then subjected to spectrometric analysis to assess the molecular weight and fragment ion spectra of the individual peptides. The experimental masses were systematically adapted to the theoretical sequence through peptide mass matching. As shown in [Fig. 6B], the collective coverage of the peptides achieved 100% of the theoretical amino acid sequence. These findings indicated that brolucizumab produced by this bioprocess possessed the correct primary structure and did not experience oxidation or deamidation. In addition, the intramolecular disulfide bonds remained intact in the absence of a reducing agent. The two pairs of disulfide bonds (Cys24-Cys89 and Cys154-Cys228) within the molecule were correctly formed, as illustrated in [Supplementary Fig. S3], available in the online version.

Fig. 6 Brolucizumab primary structure confirmation. (A) Mass spectrometry analysis of brolucizumab. (B) Sequence coverage map of brolucizumab analyzed by PepFinder 2.0 (Thermo Fisher Scientific, Massachusetts, United States).

 

MissingLinkText

In vitro Activity

The HUVEC proliferation inhibition assay indicated that both brolucizumab and ranibizumab inhibited VEGF-induced HUVEC proliferation in a dose-dependent manner ([Fig. 7]), with the average half-maximal inhibitory concentration (IC₅₀) values for brolucizumab and ranibizumab being 0.30 ± 0.07 and 0.89 ± 0.09 nmol/L (n = 3), respectively, which were consistent with the activities reported in the previous report.[43] These data also suggested that brolucizumab was more effective than ranibizumab in inhibiting HUVEC proliferation.

Fig. 7 Brolucizumab dose-dependently inhibited the proliferation of HUVEC cells induced by VEGF. Data represent mean ± SD of triplicate. SD, standard deviation; VEGF, vascular endothelial growth factor.

 

MissingLinkText

 

MissingLinkText

Conclusion

In this study, the biosynthesis process for the production of brolucizumab was explored for the first time, and the process flow was simplified. Initially, the expression of brolucizumab in the E. coli strain was successfully constructed. Subsequently, the denaturation process was refined to ensure complete unfolding of the polypeptide chains, followed by refolding to obtain correct protein folding with appropriate disulfide bond pairing. A streamlined purification protocol was then developed based on the characteristics of brolucizumab, yielding 413 mg of target product per liter of fermentation broth. Furthermore, we confirmed the theoretical molecular weight, the amino acid sequence, and the disulfide bond locations of brolucizumab, and a HUVEC proliferation inhibition assay was established to evaluate its activity. The bioprocess described herein, particularly concerning scFv refolding, can be adapted for the production of other multi-disulfide bond fusion proteins. This research established a foundational framework for the preparation of brolucizumab biosimilars and encouraged further research and development.

 

MissingLinkText

 

MissingLinkText

Conflict of Interest

None declared.

Supporting Information

Effects of denaturant type and concentration on inclusion bodies solubilization ([Supplementary Fig. S1] [available in the online version]); RP-HPLC analysis of refolding yield before and after optimization, HPLC spectrum of brolucizumab ([Supplementary Fig. S2] [available in the online version]); purification process of brolucizumab and optimization, screening of chromatographic resins and mobile phase pH for primary ion-exchange chromatography ([Supplementary Table S1] [available in the online version]); as well as screening chromatography for the secondary purfication experiments ([Supplementary Table S2] [available in the online version]); brolucizumab undergoing tryptic digestion under non-reducing conditions to preserve disulfide bonds, with resultant fragments confirmed by LC-MS/MS analysis ([Supplementary Fig. S3] [available in the online version]), are included in “[Supplementary Material], available in the online version” section of this article's webpage.

• References

• 1 Van Cleemput L, Peeters F, Jacob J. Brolucizumab for neovascular age-related macular degeneration (BEL Study). Clin Ophthalmol 2023; 17: 1077-1085
• 2 Kazemi MS, Shoari A, Salehibakhsh N. et al. Anti-angiogenic biomolecules in neovascular age-related macular degeneration; therapeutics and drug delivery systems. Int J Pharm 2024; 659: 124258
• 3 Tan CS, Ngo WK, Chay IW, Ting DS, Sadda SR. Neovascular age-related macular degeneration (nAMD): a review of emerging treatment options. Clin Ophthalmol 2022; 16: 917-933
• 4 Ghanchi F, Bourne R, Downes SM. et al. An update on long-acting therapies in chronic sight-threatening eye diseases of the posterior segment: AMD, DMO, RVO, uveitis and glaucoma. Eye (Lond) 2022; 36 (06) 1154-1167
• 5 Apte RS, Chen DS, Ferrara N. VEGF in signaling and disease: beyond discovery and development. Cell 2019; 176 (06) 1248-1264
• 6 Bressler SB. Introduction: understanding the role of angiogenesis and antiangiogenic agents in age-related macular degeneration. Ophthalmology 2009; 116 (10) S1-S7
• 7 Nguyen QD, Das A, Do DV. et al. Brolucizumab: evolution through preclinical and clinical studies and the implications for the management of neovascular age-related macular degeneration. Ophthalmology 2020; 127 (07) 963-976
• 8 Jiang D, Xu T, Zhong L. et al. Research progress of VEGFR small molecule inhibitors in ocular neovascular diseases. Eur J Med Chem 2023; 257: 115535
• 9 Sharma A, Kumar N, Parachuri N. et al. Understanding retinal vasculitis associated with brolucizumab: complex pathophysiology or occam's razor?. Ocul Immunol Inflamm 2022; 30 (06) 1508-1510
• 10 Enríquez AB, Baumal CR, Crane AM. et al. Early experience with brolucizumab treatment of neovascular age-related macular degeneration. JAMA Ophthalmol 2021; 139 (04) 441-448
• 11 Markham A. Brolucizumab: first approval. Drugs 2019; 79 (18) 1997-2000
• 12 Ferro Desideri L, Traverso CE, Nicolò M. Brolucizumab: a novel anti-VEGF humanized single-chain antibody fragment for treating w-AMD. Expert Opin Biol Ther 2021; 21 (05) 553-561
• 13 Tadayoni R, Sararols L, Weissgerber G, Verma R, Clemens A, Holz FG. Brolucizumab: a newly developed anti-VEGF molecule for the treatment of neovascular age-related macular degeneration. Ophthalmologica 2021; 244 (02) 93-101
• 14 Siddiqui ZA, Dhumal T, Patel J, LeMasters T, Almony A, Kamal KM. Cost impact of different treatment regimens of brolucizumab in neovascular age-related macular degeneration: a budget impact analysis. J Manag Care Spec Pharm 2022; 28 (12) 1350-1364
• 15 Sandomenico A, Sivaccumar JP, Ruvo M. Evolution of Escherichia coli expression system in producing antibody recombinant fragments. Int J Mol Sci 2020; 21 (17) 6324
• 16 Rosano GL, Morales ES, Ceccarelli EA. New tools for recombinant protein production in Escherichia coli: a 5-year update. Protein Sci 2019; 28 (08) 1412-1422
• 17 Rosano GL, Ceccarelli EA. Recombinant protein expression in Escherichia coli: advances and challenges. Front Microbiol 2014; 5: 172
• 18 Yamaguchi H, Miyazaki M. Refolding techniques for recovering biologically active recombinant proteins from inclusion bodies. Biomolecules 2014; 4 (01) 235-251
• 19 Sarker A, Rathore AS, Gupta RD. Evaluation of scFv protein recovery from E. coli by in vitro refolding and mild solubilization process. Microb Cell Fact 2019; 18 (01) 5
• 20 Khanchezar S, Hashemi-Najafabadi S, Shojaosadati SA, Babaeipour V. High cell density culture of recombinant E. coli in the miniaturized bubble columns. Bioprocess Biosyst Eng 2021; 44 (10) 2075-2085
• 21 Burgess RR. Refolding solubilized inclusion body proteins. Methods Enzymol 2009; 463: 259-282
• 22 Tungekar AA, Fulewar P, Kumthekar R, Bhambure R. Understanding in-vivo refolding of antibody fragments (Fab): biosimilar ranibizumab a case study. Process Biochem 2024; 146: 484-497
• 23 Farokhi-Fard A, Bayat E, Beig Parikhani A. et al. Bacterial production and biophysical characterization of a hard-to-fold scFv against myeloid leukemia cell surface marker, IL-1RAP. Mol Biol Rep 2023; 50 (02) 1191-1202
• 24 Vagenende V, Yap MGS, Trout BL. Mechanisms of protein stabilization and prevention of protein aggregation by glycerol. Biochemistry 2009; 48 (46) 11084-11096
• 25 Kachhawaha K, Singh S, Joshi K, Nain P, Singh SK. Bioprocessing of recombinant proteins from Escherichia coli inclusion bodies: insights from structure-function relationship for novel applications. Prep Biochem Biotechnol 2023; 53 (07) 728-752
• 26 Manissorn J, Tonsomboon K, Wangkanont K, Thongnuek P. Effects of chemical additives in refolding buffer on recombinant human BMP-2 dimerization and the bioactivity on SaOS-2 osteoblasts. ACS Omega 2023; 8 (02) 2065-2076
• 27 Mousavi SB, Davarpanah SJ. Solvent extraction of recombinant interferon alpha-2b from inclusion bodies and efficient refolding at high protein concentrations. Protein Expr Purif 2022; 197: 106110
• 28 Yu L, Liang XH, Ferrara N. Comparing protein VEGF inhibitors: in vitro biological studies. Biochem Biophys Res Commun 2011; 408 (02) 276-281
• 29 Wong RS, Liew MWO, Ong EBB. Production of recombinant human epidermal growth factor in Escherichia coli: strategic upstream and downstream considerations for high protein yield. Process Biochem 2024; 146: 81-96
• 30 Zhao Y, Wang ZS, Wang Q. et al. Efficient transformation and genome editing in a nondomesticated, biocontrol strain, Bacillus subtilis GLB191. Phytopathol Res 2024; 6 (01) 69
• 31 Zhang Y, Wang Y, Lu J. et al. High-yield and cost-effective biosynthesis process for producing antimicrobial peptide AA139. Protein Expr Purif 2024; 219: 106475
• 32 Gutiérrez-González M, Farías C, Tello S. et al. Optimization of culture conditions for the expression of three different insoluble proteins in Escherichia coli . Sci Rep 2019; 9 (01) 16850
• 33 Moghadam M, Ganji A, Varasteh A, Falak R, Sankian M. Refolding process of cysteine-rich proteins: chitinase as a model. Rep Biochem Mol Biol 2015; 4 (01) 19-24
• 34 Basak P, Kundu N, Pattanayak R, Bhattacharyya M. Denaturation properties and folding transition states of leghemoglobin and other heme proteins. Biochemistry (Mosc) 2015; 80 (04) 463-472
• 35 Paladino A, Vitagliano L, Graziano G. The action of chemical denaturants: from globular to intrinsically disordered proteins. Biology (Basel) 2023; 12 (05) 754
• 36 Bhat N, Roy T, Sengupta S. et al. Degradation and molecular docking of Curli to scout aggregation complexion. Bioresour Technol Rep 2024; 26: 101830
• 37 Andlinger DJ, Röscheisen P, Hengst C, Kulozik U. Influence of pH, temperature and protease inhibitors on kinetics and mechanism of thermally induced aggregation of potato proteins. Foods 2021; 10 (04) 796
• 38 Bauer KC, Suhm S, Wöll AK, Hubbuch J. Impact of additives on the formation of protein aggregates and viscosity in concentrated protein solutions. Int J Pharm 2017; 516 (1-2): 82-90
• 39 Leibly DJ, Nguyen TN, Kao LT, Hewitt SN, Barrett LK, Van Voorhis WC. Stabilizing additives added during cell lysis aid in the solubilization of recombinant proteins. PLoS One 2012; 7 (12) e52482
• 40 Chen J, Liu Y, Li X. et al. Cooperative effects of urea and L-arginine on protein refolding. Protein Expr Purif 2009; 66 (01) 82-90
• 41 Wang Y, van Oosterwijk N, Ali AM. et al. A systematic protein refolding screen method using the DGR approach reveals that time and secondary TSA are essential variables. Sci Rep 2017; 7 (01) 9355
• 42 Ryan BJ, Kinsella GK, Henehan GT. Protein extraction and purification by differential solubilization. In: Loughran ST, Milne JJ. eds. Protein Chromatography. Methods in Molecular Biology. New York, NY: Humana; 2023. ;2699: 349-368
• 43 Lowe J, Araujo J, Yang J. et al. Ranibizumab inhibits multiple forms of biologically active vascular endothelial growth factor in vitro and in vivo . Exp Eye Res 2007; 85 (04) 425-430

Publication History

Received: 18 February 2025

Accepted: 24 June 2025

Article published online:
18 July 2025