Abstract
Non-alcoholic steatohepatitis (NASH) is a potential serious disease, which almost has no available medicine for effective treatment today. Efruxifermin is a bivalent Fc-FGF21 candidate drug developed by Akero Therapeutics that has shown promising results in preclinical and clinical trials for NASH and may be approved in future. However, it is produced by Escherichia coli (E. coli) expressing system, which has no glycosylation modifications and is hard to purify for inclusion body. Suspension mammalian cell expression systems, human embryonic kidney 293 (HEK293), and Chinese hamster ovary (CHO) are good choice for protein expression of biopharmaceutical use. In this report, the objective was to produce Fc-FGF21 by mammalian cell expression systems, which enabled necessary glycosylation modifications to occur on the Fc-FGF21 protein and was relatively easy for future large-scale production. We observed that the target protein Fc-FGF21 could be easily degraded in CHO system, such as CHOK1SV or CHOZN, and it was hard to purify, whereas it was more stable in the HEK293 expressing system. Then, similarity between Fc-FGF21 from E. coli and Fc-FGF21 from HEK293 was focused by in vitro and in vivo studies, and we observed no significant difference between the proteins expressed from the two different expressing systems, indicating that a biosimilar of Efruxifermin has been developed successfully. Proteomics analysis from in vivo study samples further identified some potential biomarkers or FGF21 related signaling pathways. Taken together, this study demonstrates a good example of how to develop and validate a biosimilar for therapeutic purposes. In future, more efforts, such as mutation to FGF21 or linking FGF21 with effective antibody to form dual targets, could be done to obtain more effective FGF21 analogs.
Keywords: Efruxifermin, Fc-FGF21, Biosimilar, Expression system, Similarity
Introduction
Non-alcoholic fatty liver disease (NAFLD) is a prevalent disease associated with obesity, dyslipidemia, and type 2 diabetes mellitus (T2D). Non-alcoholic steatohepatitis (NASH) is a serious form of NAFLD and is a potential serious disease, which almost has no available medicine treatment today [1, 2]. Recently, NAFLD and NASH have the trend to be replaced by new names, metabolic dysfunction-associated steatotic liver disease (MASLD) and metabolic dysfunction-associated steatohepatitis (MASH) [2–4]. In this study, NAFLD is equal to MASLD and NASH is equal to MASH. Steatosis, inflammation, fibrosis, and ballooning are typical characteristics of NAFLD and NASH [5]. The global prevalence of NAFLD is about 25% in adults and lean individuals can also develop NAFLD [6]. Fibroblast growth factor 21 (FGF21) is an endogenous protein, which is approximately 100 pg/mL in healthy humans but is increased in NASH patients to about 1000 pg/mL [7]. This protein belongs to FGF19 subfamily, and the liver is its main secreting tissue [8]. There are two types of FGF21 receptor: fibroblast growth factor receptor (FGFR) and β-Klotho (KLB) [9].
To date, several FGF21 analogs have been designed and validated in preclinical and clinical trials. Efruxifermin (AKR-001 or AMG 876) is a fusion protein in which Fc fragment of antibody is linked to FGF21 with amino acid mutations, developed by Akero Therapeutics (South San Francisco, USA) [10]. Pegozafermin is a glycoPEGylated FGF21 analog, which is in clinical trials for NASH or severe hypertriglyceridemia (SHTG) treatment, developed by 89bio, Inc. (Herzliya, Israel) [11]. There are other FGF21 drug candidates, such as Pegbelfermin (BMS-986036, Bristol-Myers Squibb) [12], LY2405319 (Eli Lilly) [13], PF-05231023 (Pfizer Inc.) [14], and BOS-580 (Novartis Pharmaceuticals) [15].
Bacterial expression system, such as Escherichia coli (E. coli), is a conventional way to express recombinant proteins. However, one of its disadvantages is its inability to produce complex proteins with post-translational modifications (such as glycosylation) [16]. On the other hand, mammalian expression systems (such as Chinese hamster ovary (CHO) and human embryonic kidney 293 (HEK293)) are more often used by biopharmaceutical industry for therapeutic protein production due to the capability of including necessary post-translational modifications during protein production [17–20], which are similar to those produced in humans.
In this study, Fc-FGF21 (the same amino acid sequence as Efruxifermin) is expressed by four different expression systems. Physiochemical characterization and functional assays were carried out using different analytical tools and models. In addition, proteomic analysis was also performed with a goal to identify related biomarkers and pathways.
Materials and Methods
Reagents
Fc-FGF21 from E. coli, Fc-FGF21 from CHO, and Fc-FGF21 from HEK293 were expressed by Shanghai Junshi Biosciences Co., Ltd. (Shanghai, China).
3D NAC-Organ Model
NASH 3D NAC-Organ model was provided by Puheng Technology (Suzhou, China).
Animals
NASH mice with Dyets AMLN feed were purchased from Charles River, and all mouse studies were conducted at Shanghai Junshi Biosciences Co., Ltd. (Shanghai, China) and approved by the Animal Care and Use Committee. Mice were maintained in rooms with a 12-h light/dark cycle, temperature between 20 and 26℃ and humidity between 30 and 70%. Mice had free access to food and water.
Protein Expression and Purification
The DNA encoding the Fc-FGF21 (FGF21 181 amino acids) was synthesized by GENEWIZ (Suzhou, China), cloned into PET28a vector (69864–3, Merck) for Escherichia coli (69450, Merck), pXC17.4 vector (Lonza) for CHO-K1 (Lonza) and CHO-ZN (Merck), HX1 vector (Shanghai Junshi Biosciences Co., Ltd.) for HEK293 (Thermo Fisher) with different restriction sites by restriction enzymes. Transformation and transfection were proceeded according to protocols from Shanghai Junshi Biosciences Co., Ltd.
For E. coli, the inclusion bodies of recombinant proteins were purified and then refolded. Soluble proteins were further purified by protein A chromatography (rProtein A, MabSelect SuRe, GE Healthcare) by elution buffer (citrate buffer solution, pH 3.8–3.9), similar with supernatants from CHO or HEK293. Finally, sample buffer was changed to DPBS (Gibco) by ultrafiltration.
Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis (SDS-PAGE)
SDS-PAGE analysis of the samples was carried out under reducing and non-reducing conditions in 4–12% pre-cast polyacrylamide gels (180-9217H, Tanon). For reducing conditions, samples had final concentration 1 mg/mL containing 2μL NuPAGE™ Sample Reducing Agent (10X, NP0009, Invitrogen) and 5μL NuPAGE™ LDS Sample Buffer (4X, NP0007, Invitrogen) with final volume of 20 μL. For non-reducing conditions, 20 μg of each sample was first diluted with ultrapure water to 15 μL, then mixed with 5μL NuPAGE™ LDS Sample Buffer (4X). Gels were stained with 0.125% Coomassie blue R250. It was scanned and analyzed by using the Image Lab software (Bio-Rad, Version 5.0).
Size-Exclusion Chromatography (SEC)-HPLC Analysis
SEC-HPLC was conducted using U3000 HPLC instrument (Thermo-Fisher Scientific) with TSKgel G3000SWxl column (7.8 × 300 mm, 5 μm, TOSOH). Mobile phase was phosphate buffer (50 mM PB, 300 mM NaCl, pH 7.0).
Imaged Capillary Isoelectric Focusing (icIEF)
icIEF was conducted on the Maurice instrument from ProteinSimple (Bio-techne, Minneapolis, USA). Data was acquired and evaluated using the associated software Compass for iCE (v3.0.0). The separation occurred in icIEF cartridges contained fluorocarbon-coated silica capillary with an effective capillary length of 5 cm and 100 μm inner diameter, and the entire capillary was monitored in real-time using a CCD camera imaging with both native FL and/or absorbance at 280 nm. Filled the two electrolyte tanks of the cartridge with 2 mL anolyte (100 mM phosphoric acid in 0.1% methyl cellulose) and 2 mL catholyte (80 mM NaOH in 0.1% methyl cellulose), respectively, and closed the two tanks using the included covers.
Ten microliters of protein, dissolved in ultrapure water (2 mg/mL), was mixed with 35 μL of 1% methyl cellulose, 50 μL 4 M urea, 4 μL pharmalyte 3–10, 0.5 μL each of pI markers 5.12 and 9.22, resulting in a total volume of 100 μL. Samples, vortexed and centrifuged to be pipetted into a 96-well plate, were placed on the autosampler sample tray and maintained at 10 °C during the run. Sample load was done for 55 s. Subsequently, the injection was prefocused at 1500 V for 1 min and focused at 3000 V for 7 min, respectively.
N-Glycan Analysis
Samples were adjusted to 2.0 mg/ml. Glycan chains were released and labeled by Glycoworks RapiFluor-MS (186008841, 186008091, Waters Corporation) according to instructions of supplier. Analysis was conducted using ACQuity H class (Waters Corporation) with BEH Glycan Amide column (2.1 × 150 mm, 1.7 μm, Waters Corporation). Mobile phases were 100 mM ammonium formate (pH 4.5) and acetonitrile.
Differential Scanning Calorimetry (DSC)
DSC was run by MicroCal VP-DSC (Malvern Panalytical) for analysis of thermostability. Samples were diluted to 0.5 mg/ml, thermal scanning from 20 to 110 ℃; then, thermal transition temperature (Tm) was calculated.
Mass Spectrometry (MS) Analysis
MS analysis was carried on a Vanquish Flex UHPLC system (Thermo-Fisher Scientific) coupled to a Q Exactive Plus mass spectrometer (Thermo-Fisher Scientific). A reversed phase (RP) column (Waters™ BioResolve™ RP mAb polyphenyl column, 450 Å, 2.7 μm, 2.1 × 150 mm, Waters Corporation) was used. The deconvolution analysis to the raw mass spectra for molecular weight (MW) determination was performed with the BioPharma Finder 3.2 software (Thermo-Fisher Scientific).
Approximately 50 μg of protein was dissolved in 100 μL of water. The PNGaseF enzyme (P0704L, New England BioLabs) was added to the sample and it was incubated at 37 °C for 4 h. Reduced the sample by addition of 2 μl of 0.5 M DTT, incubated for 30 min at 37 °C.
The initial elution mobile phase contained 80% solvent A (0.10% formic acid + 0.02% trifluoroacetic acid in water) and 20% solvent B (0.08% formic acid + 0.02% trifluoroacetic acid in acetonitrile). After equilibrating with initial mobile phase for 2 min, the linear gradient to solvent B climbed to 32% in 2 min, then to 37% in 8 min, and maintained for 3 min at 90% at a constant flow rate of 0.4 mL/min. For MS scanning, the mode was set as positive, spray voltage was 3.8 kV and sheath gas was 35 Arb. The scan range was 500 to 3,500 m/z.
Enzyme-Linked Immunosorbent Assay (ELISA)
Human FGFR1(10616-H08H) and mouse FGFR1 (50186-M08H) were purchased from SinoBiological (China). Human KLB (RPH756Hu01) and mouse KLB (RPH756Mu01) were purchased from Cloud-clone (China). Other reagents were obtained from Shanghai Junshi Biosciences Co., Ltd. (China).
Ninety-six-well polystyrene microtiter plates were coated by different receptors, FGFR1 or KLB. After incubation, washing, blocking, and washing, samples are added. After washing, primary antibody and secondary antibody with horse radish peroxidase (AF2539 and HAF109, R&D Systems) were added. Microplate reader was used to read numbers under 450 nm.
Surface Plasmon Resonance (SPR)
SPR study was performed on Fortebio Octet RED384(Sartorius). Protein A biosensor (18–5010, Sartorius) was used to capture samples (5 ug/ml), height 1 nm. Gradient diluted β-klotho (KLB, RPH756Hu01) was added to bind from 100 nM, 25 nM, 12.5 nM to 6.25 nM. Kinetic constants were calculated by the Data Analysis HT 12.0 software.
ELK Luciferase Assay
HEK293-KLB/ELK1-LBD/GAL4-LUC cells were plated at 2.2*105 cells per mL, 90 μL/well. After incubation at 37℃, 5% CO2 for 18 h, gradient-diluted FGF21 was added, 10 μL/well; then, 50 μL supernatant was discarded and 50 μL ONE-Glo (E6120, Promega) was added. M5 (molecular devices) was used to read luminescence.
3D NAC-Organ Study
Primary human hepatocytes (PHH), liver sinusoidal endothelial cells (LSEC), hepatic stellate cells (HSC), and Kupffer cells (KC) were purchased from LV BioTech (China), 1500 PHH:692 LSEC:462 HSC:346 KC were cocultured to form 3D liver model for 48 h which was then induced to 3D NASH model for 10 days by Puheng Technology (China). Drugs were added at day 4. Supernatants were collected at day 6 for IL-6 test (Zhuochai biology, China). Liver microspheres were collected for total triglyceride (Applygen, China), hematoxylin eosin (Servicebio, China), and sirius red (HEAD Biotechnology, China) staining at day 12.
Animal Studies
All mouse studies were conducted at Shanghai Junshi Biosciences Co., Ltd. (China) and were approved by Animal Care and Use Committee.
Four groups of mice were caged for the studies, normal group with chewing feed, NASH vehicle group with AMLN feed, and Fc-FGF21E and Fc-FGF21H with AMLN feed. Drugs were administrated by intraperitoneal administration once a week for 8 weeks. Body weight was examined every week. TC, TG, AST, ALT, HDL-c, LDL-c, and glucose were tested every 2 weeks. At endpoint, animals were sacrificed and liver weight was recorded. Liver section was dyed by hematoxylin eosin and sirius red.
Proteomics Analysis
Label-free proteomics study was carried out at XJTLU Center for Pharmaceutical Analysis (CPA) according to the previously published protocols [21, 22].
Statistical Analysis
Prism 8.3 was used to draw figures. All data was means ± standard deviation (S.D.). Values were compared using Student’s paired t-test. p < 0.05 was considered statistically significant.
Results
Expression of Fc-FGF21 by 4 Different Expression Systems
Fc-FGF21 was expressed with E. coli, HEK293, CHOK1SV, and CHOZN system, respectively. When Fc-FGF21 was expressed by CHOK1SV or CHOZN system, it degraded quickly during cell culture, likely due to protease digestion. SDS-PAGE result showed that the main peak (non-reduced) was about 60 kDa, which was about 2/3 of its theoretical molecular weight (92 kDa). In E. coli and HEK293 expressing systems, however, the observed molecular weight was consistent with the theoretical value of 92 kDa. The observed molecular weight of Fc-FGF21 from HEK293 (Fc-FGF21H) was roughly 3 kDa more than that from E. coli (Fc-FGF21E), most likely due to its post-translational modifications (PTMs) (Fig. 1).
Fig. 1.
SDS-PAGE showed Fc-FGF21 from 4 different expression systems, CHOK1SV, CHOZN, E. coli, and HEK293. 1A Fc-FGF21 from CHOK1SV cell line under reduced and non-reduced conditions. Lane 1: reduced condition, Lane 2: molecular weight markers, Lane 3: non-reduced condition; 1B Fc-FGF21 from CHOZN cell line under reduced and non-reduced conditions. Lane 1: reduced condition, Lane 2: molecular weight markers, Lane 3: non-reduced condition; 1C Fc-FGF21 from E. coli and HEK293, respectively, under non-reduced and reduced conditions. Lane 1: molecular weight markers, Lane 2: from E. coli under non-reduced condition, Lane 3: from HEK293 under non-reduced condition, Lane 4: from E. coli under reduced condition, Lane 5: from HEK293 under reduced condition
Identification of Clipping Site
In order to find out which site was the clipping site, mass spectrometry (MS) was applied to identify clipping sites. Fc-FGF21 expressed by CHOK1SV and CHOZN were analyzed by LC–MS/MS. Data showed that the target protein was cleaved at position 18 from N-terminus (glutamine) and 19 from N-terminus (arginine) in both CHOK1SV and CHOZN systems (Fig. 2). The mass spectrometry result also showed that Fc-FGF21 from E. coli had no glycosylation, whereas Fc-FGF21 from HEK293 had glycosylation, which was mainly biantennary N-glycan. After reduction by DTT, their molecular weights were 46 kDa and 47 kDa, respectively.
Fig. 2.
Identification of clipping sites and confirmation by Mass spectrometry under reduced condition. 2A and 2B showed that Fc-FGF21 from CHOK1SV or CHOZN, respectively, had several clipping sites; 2C showed that Fc-FGF21 from E. coli had no N-glycan; 2D showed that Fc-FGF21 from HEK293 had a N-glycan (with an increase in molecular weight of 3 kDa under non-reduced condition)
Site Mutation to Reduce Degradation
Both glutamine (at position 18 from N-terminus) and arginine (at position 19 from N-terminus) were replaced by asparagine and lysine, respectively, and expressed by CHO system. SDS-PAGE result showed that this replacement had no significant effect on reducing degradation (Fig. 3).
Fig. 3.
SDS-PAGE showed the degraded Fc-FGF21 from CHO expression system even after amino acid mutation. 3A reduced condition; 3B non-reduced condition. Lane 1: molecular weight markers, Lane 2: Fc-FGF21 with one amino acid mutation at 18, Lane 3: Fc-FGF21 with one amino acid mutation at 19
Physiochemical Characterization of Fc-FGF21 from E. coli (Fc-FGF21E) and Fc-FGF21 from HEK293 (Fc-FGF21H)
SDS-PAGE
SDS-PAGE showed both proteins were pure, more than 90.0% (Fig. 1).
SEC-HPLC
SEC-HPLC showed both proteins were pure, monomer was more than 95.0% (Fig. 4A).
Fig. 4.
Physiochemical analyses of Fc-FGF21 from E. coli (Fc-FGF21E) and from HEK293 (Fc-FGF21H), respectively. 4A SEC-HPLC; 4B icIEF; 4C N-glycan analysis with HPLC; 4D differential scanning calorimetry
icIEF
icIEF shows their pI was 6.5, consistent with theoretical value 6.1, although their charge distribution seemed to be different (Fig. 4B).
N-Glycan
N-Glycan showed there was no N-glycan in Fc-FGF21E, but N-glycan in Fc-FGF21H was typical (Fig. 4C). Most glycosylation was G0F (one kind of N-glycan) modification.
Differential Scanning Calorimetry (DSC)
Fc-FGF21E had Tm1 65.1 ℃ and Tm2 79.87 ℃, Fc-FGF21H had Tm1 69.72 ℃, and Tm2 80.39 ℃ (Fig. 4D). Fc-FGF21H seemed to be slightly stable than Fc-FGF21E.
FGFR1 and KLB Binding ELISA
Binding activity of Fc-FGF21E and Fc-FGF21H with FGFR1 from human and mouse and KLB from human and mouse was tested by ELISA. EC50 of Fc-FGF21E and Fc-FGF21H were 3.5 nM and 4.3 nM in human FGFR1 ELISA assay; the ratio was 81%. Because the top platform was not reached in mouse FGFR1 ELISA assay, mouse KLB ELISA assay, and human KLB ELISA assay, EC50 was not fit, but the trends were almost same (Fig. 5).
Fig. 5.
ELISA of FGF21 binding with FGF receptors. 5A Human FGFR1 binding with Fc-FGF21; 5B mouse FGFR1 binding with Fc-FGF21; 5C human KLB binding with Fc-FGF21; 5D mouse KLB binding with Fc-FGF21
Surface Plasmon Resonance (SPR) Study
SPR study showed Fc-FGF21E and Fc-FGF21H had similar Kd, 10.1 nM and 9.6 nM (Fig. 6A and B). This result was consistent with reference [23].
Fig. 6.
SPR study and luciferase reporter assay of FGF21. 6A SPR of Fc-FGF21E; 6B SPR of Fc-FGF21H; 6C luciferase reporter assay of Fc-FGF21E and Fc-FGF21H
Cell-Based Assay (Luciferase Reporter Assay)
ELK luciferase assay showed Fc-FGF21E and Fc-FGF21H had similar EC50, 0.09 nM and 0.08 nM, ratio 112% (Fig. 6C). This result also showed it was a good assay to differentiate different FGF21.
Efficacy Studies in NASH 3D NAC-Organ Model
In order to evaluate functions of Fc-FGF21 further, 3D NAC-Organ liver model was established and induced to NASH model. After drug treatment for 10 days, several parameters were tested (Fig. 7 and Table 1).
Fig. 7.
Efficacy study in 3D NAC-Organ model. 7A Results from IL-6, triglyceride (TG), and sirius red staining, respectively; 7B results from hematoxylin eosin staining; 7C photos of sirius red staining
Table 1.
Average parameters in NASH 3D NAC-Organ model study
| Lean ± SD | NASH vehicle ± SD | Fc-FGF21E ± SD | Fc-FGF21H ± SD | |
|---|---|---|---|---|
| IL-6 (pg) | 6.1 ± 0.4 | 8.3 ± 1.2 | 5.9 ± 0.9 | 5.3 ± 0.2 |
| TG(nM) | 5.8 ± 0.1 | 7.6 ± 0.2 | 7.5 ± 0.2 | 7.6 ± 0.1 |
| Sirius red area ratio | 0.04 ± 0.02 | 0.1 ± 0.01 | 0.04 ± 0.01 | 0.05 ± 0.01 |
IL-6 level was elevated in NASH group compared to lean group from 6.1 to 8.3 pg, which meant inflammation was induced in NASH model. After treatment, IL-6 level was downregulated in both Fc-FGF21 group with similar effect, about 5.3 pg in Fc-FGF21H, 5.9 pg in Fc-FGF21E. The level of IL-6 in Fc-FGF21 group could be reduced to normal level, similar with lean group.
TG level was elevated in NASH group compared to lean group from 5.8 nM to 7.6 nM. Both groups of Fc-FGF21 cannot reduce level of TG, which meant TG level was not a sensitive parameter for evaluating FGF21’s function in this model.
Results of HE-staining showed NASH group had obvious features of steatosis compared to lean group. Both groups of Fc-FGF21 could not reverse level of steatosis.
Results of sirius red staining showed NASH group had more level of fibrosis compared to lean group, 0.04 area percentage versus 0.1 area percentage. Both groups of Fc-FGF21 could reverse level of fibrosis and show no difference, 0.05 area percentage and 0.04 area percentage.
Efficacy Studies in NASH Mice
Both Fc-FGF21E and Fc-FGF21H could reduce weight of liver, body weight in NASH mouse model. The average liver weight of Fc-FGF21E and Fc-FGF21H was 1796 mg and 1288 mg, respectively, at endpoint, similar with normal group 1420 mg, whereas liver weight of NASH vehicle was 2429 mg. The average body weight of Fc-FGF21E and Fc-FGF21H was 30 g and 29 g, respectively, at endpoint, similar with normal group 31 g, whereas body weight of NASH vehicle was 40 g.
Total cholesterol (TC), alanine transaminase (ALT), aspartate aminotransferase (AST), high density lipoprotein cholesterol (HDL-c) and low-density lipoprotein cholesterol (LDL-c) were reduced at endpoint. Total triglyceride (TG) and glucose (Glu) were stable and in normal level; see Table 2 and Fig. 8.
Table 2.
Average biochemical parameters of endpoint in animal study
| Normal group ± SD | NASH vehicle ± SD | Fc-FGF21E ± SD | Fc-FGF21H ± SD | |
|---|---|---|---|---|
| Liver weight (mg) | 1420 ± 86 | 2429 ± 536 | 1796 ± 210 | 1288 ± 205 |
| Body weight (g) | 31 ± 2 | 40 ± 3 | 30 ± 2 | 29 ± 2 |
| ALT (U/L) | 36.0 ± 7.6 | 126.3 ± 61.8 | 40.2 ± 14.4 | 25.5 ± 5.6 |
| AST (U/L) | 59.5 ± 15.2 | 136.7 ± 44.2 | 124.3 ± 64.9 | 75.3 ± 17.6 |
| TC (mmol/L) | 2.1 ± 0.3 | 5.6 ± 1.1 | 3.4 ± 0.5 | 3.3 ± 0.3 |
| LDL-c (mmol/L) | 0.3 ± 0 | 1.1 ± 0.3 | 0.5 ± 0.1 | 0.6 ± 0.1 |
| HDL-c (mmol/L) | 1.7 ± 0.2 | 3.9 ± 0.6 | 2.4 ± 0.2 | 2.4 ± 0.2 |
| Glucose (mmol/L) | 7.3 ± 1.3 | 7.0 ± 1.2 | 9.0 ± 2.2 | 7.4 ± 1.2 |
| TG (mmol/L) | 0.9 ± 0.2 | 0.7 ± 0.1 | 0.6 ± 0.1 | 0.7 ± 0.1 |
Fig. 8.
Efficacy study in NASH mouse model. 8A Liver weight at endpoint; 8B body weight at endpoint; 8C curves of body weight; 8D TC curves; 8E ALT curves; 8F AST curves; 8G HDL-c curves; 8H LDL-c curves; 8I TG curves; 8 J Glu curves
Proteomics Analysis
In proteomics study, 231 proteins were identified to be differentially expressed when the dosing group was compared with the NASH vehicle group; the number of proteins that had a significant expression change was 438 when the NASH vehicle group was compared with the normal group. Ingenuity pathway analysis (IPA) showed LXR/RXR activation, DHCR24 signaling pathway, IL-12 signaling, and production in macrophages are the related pathways (Fig. 9).
Fig. 9.
The results from the proteomic studies. 9A Heat map of the data from the proteomic study of NASH vehicle group versus normal group; 9B heat map of the data from the proteomic study of dosing group versus NASH vehicle group
Discussion
As the first approved drug for the management of metabolic dysfunction-associated steatohepatitis in 2024, Resmetirom has shown positive results for patients. But its effect is limited. In the phase 3 MAESTRO-NASH trial [3], treatment with Resmetirom led to a fibrosis improvement by at least one stage with no worsening of the NAFLD activity score in 24.2% of the patients in the 80-mg group and 25.9% of those in the 100-mg Resmetirom group vs 14.2% in the placebo group (p < 0.001 for both comparisons vs placebo). Response rate is lower than Efruxifermin, which 54% in the 28-mg group, 77% in the 50-mg group and 36% in the 70-mg group [1]. So, more effective drugs for MASH are needed. Efruxifermin, which has been administrated for once per week or once per two weeks in clinical trials for its longer half-life period, could be a potentially more effective and easy-to-use drug for MASH patients.
FGF21 is not only a candidate therapeutic drug for MASH treatment, it also can be a candidate for body weight loss. Today, the most popular medicines for body weight loss are Semaglutide, Dulaglutide, and Liraglutide. In STEP8 clinical trial, the mean weight change from baseline was − 15.8% with Semaglutide (once-weekly subcutaneous, 2.4 mg) vs − 6.4% with liraglutide (once-daily subcutaneous, 3.0 mg) [24]. Dulaglutide dosages > 1 mg/day significantly decreased body weight (weighted mean differences: − 1.94 kg) [25]. Once-weekly glucagon-like peptide 1 and glucagon receptor dual agonist Mazdutide can reduce body weight up to − 7.1% [26]. Although Efruxifermin can reduce body weight by 3.3 kg in non-alcoholic steatohepatitis patients only [1], Fc-FGF21E and Fc-FGF21H could reduce body weight 23.7% and 13.8% in NASH mouse, respectively, which was consistent with Efruxifermin in cynomolgus monkey [9]. This means Fc-FGF21 could potentially be a good candidate for body weight loss.
Four different expression systems had been attempted to express Fc-FGF21. It was hard to get intact Fc-FGF21 from CHO system, whereas it was easy to express intact Fc-FGF21 from E. coli or HEK293. Although degradation site was similar in different CHO cells, and amino acid mutation had been explored, the effect was almost nothing. HEK293 system was better than E. coli, because it did not need to crush inclusion body, denature, refold, and it had post-translational modifications which were similar to proteins expressed by human beings themselves, mainly G0F, which was consistent with paper report [27]. It should have some kind of enzymes in CHO which could degrade FGF21, whereas it does not exist in HEK293. Fibroblast activation protein inhibitor and esterase inhibitor had been also attempted to inhibit activity of enzymes, but it had no effect too.
SDS-PAGE, SEC-HPLC, icIEF, N-glycan, DSC, binding ELISA, SPR, cell-based assay, 3D organoid, and NASH mouse model were analyzed and results showed Fc-FGF21 from E. coli was similar to Fc-FGF21 from HEK293. This meant post-translational modification does not have impact on bioactivity of FGF21. No matter prokaryotic system or eukaryotic system, all of them can be used to produce FGF21. Although Fc-FGF21 could not reverse TG or steatosis in 3D NASH model, it could reverse fibrosis and IL-6 significantly. 3D NASH model could be a good choice for drug screening in future.
Proteomics results showed several pathways were related to the treatment of FGF21. More validation studies need to be done in the future to confirm these findings. Since there is no good NASH cellular model available, it will be hard to validate our findings in animal studies directly. The best way is to establish a cell-based model just like work from other schoolmates of our laboratory, such as comprehensive analysis of the lysine succinylome in fish oil-treated prostate cancer cells [28].
Although CHO expression system and HEK293 expression system are suspension systems and easy for production, CHO expression system is more attractive than HEK293 expression system in industry. Food and Drug Administration has approved more than 100 antibodies [29], which most of them were expressed by CHO expression system. So, the best way to express Fc-FGF21 is CHO expression system, and more work to reduce degradation during cell culture should be done to fix degradation question in CHO system.
Conclusions
Fc-FGF21 has been expressed and validated successfully. This study is also a typical way to develop a biosimilar. If Efruxifermin is approved by medicine administration agent in future, it is easy to develop its biosimilar according to this study. Some other characterization or cell-based assay could be done to validate their similarity further.
Acknowledgements
This study was supported by Shanghai Junshi Biosciences Co., Ltd. and Xi’an Jiaotong-Liverpool University (XJTLU). We thank Dr. Chao He, Aidi Gao, Yifan Jiang, and the other members of Professor Mu Wang’s research group. We thank the scientific staffs of XJTLU Center for Pharmaceutical Analysis (CPA) for Mass Spectrometry and the staffs of Wuhan Heyan Biotech for cell lines. We also thank Jinwei Zhou, Minlu Fan, Longjian Xue, Minze Lu, Xueting Li, Honghong Liu, and Ying Wang from Shanghai Junshi Biosciences Co., Ltd. for their technical assistance.
Author Contribution
Xujia Wang and Qin Meng: conceptualization, investigation, methodology, and writing original draft. Aijuan Jia, Yuehua Zhou, Dandan Song, Shaokang Ma, and Wei Li: methodology and sample preparation. Zhuobing Zhang, Christopher Goldring, Hui Feng, and Mu Wang: conceptualization, supervision, review, and editing. All authors approved the final version and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding
The research was funded by an internal research fund from Shanghai Junshi Biosciences Co., Ltd.
Data Availability
Materials described in this manuscript including all relevant raw data will be freely available from the corresponding author for non-commercial use upon request.
Declarations
Ethical Approval
Animal study protocols (ASP) were approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Junshi Biosciences Co., Ltd. (Shanghai, China), and all in vivo experiments were conducted according to IACUC guidelines.
Consent to Participate
Not applicable.
Consent for Publication
Not applicable.
Conflict of Interest
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Xujia Wang and Qin Meng contributed equally to this work.
References
- 1.Harrison, S. A., et al. (2021). Efruxifermin in non-alcoholic steatohepatitis: a randomized, double-blind, placebo-controlled, phase 2a trial. Nature Medicine, 27(7), 1262–1271. [DOI] [PubMed]
- 2.Kokkorakis, M., et al. (2024). Resmetirom, the first approved drug for the management of metabolic dysfunction-associated steatohepatitis: Trials, opportunities, and challenges. Metabolism, 154, 155835. [DOI] [PubMed]
- 3.Borozan, S., et al. (2024). Metabolic dysfunction-associated steatohepatitis and cardiovascular disease prevention: Is resmetirom useful? Current Vascular Pharmacology. [DOI] [PubMed]
- 4.Karim, G., & Bansal, M. B. (2023). Resmetirom: An orally administered, smallmolecule, liver-directed, beta-selective THR agonist for the treatment of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis. touchREVIEWS in Endocrinology, 19(1), 60–70. [DOI] [PMC free article] [PubMed]
- 5.Smeuninx, B., Boslem, E., & Febbraio, M. A. (2020). Current and future treatments in the fight against non-alcoholic fatty liver disease. Cancers (Basel),12(7), 1714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bhandari, P., et al. (2022). Nonalcoholic fatty liver disease: Could it be the next medical tsunami? Cureus, 14(4), e23806. [DOI] [PMC free article] [PubMed]
- 7.Kaufman, A., et al. (2020). AKR-001, an Fc-FGF21 analog, showed sustained pharmacodynamic effects on insulin sensitivity and lipid metabolism in Type 2 Diabetes Patients. Cell Reports Medicine, 1(4), 100057. [DOI] [PMC free article] [PubMed]
- 8.Flippo, K. H., & Potthoff, M. J. (2021). Metabolic messengers: FGF21. Nature Metabolism, 3(3), 309–317. [DOI] [PMC free article] [PubMed]
- 9.Stanislaus, S., et al. (2017). A Novel Fc-FGF21 with improved resistance to proteolysis, increased affinity toward beta-Klotho, and enhanced efficacy in mice and cynomolgus monkeys. Endocrinology. 158(5), 1314–1327. [DOI] [PubMed]
- 10.Harrison, S. A., et al. (2023). A randomized, double-blind, placebo-controlled phase IIa trial of efruxifermin for patients with compensated NASH cirrhosis. JHEP Reports, 5(1), 100563. [DOI] [PMC free article] [PubMed]
- 11.Rosenstock, M., et al. (2023). The Novel GlycoPEGylated FGF21 analog pegozafermin activates human FGF receptors and improves metabolic and liver outcomes in diabetic monkeys and healthy human volunteers. Journal of Pharmacology and Experimental Therapeutics. 387(2), 204–213. [DOI] [PubMed]
- 12.Sanyal, A., et al. (2019). Pegbelfermin (BMS-986036), a PEGylated fibroblast growth factor 21 analogue, in patients with non-alcoholic steatohepatitis: a randomised, double-blind, placebo-controlled, phase 2a trial. Lancet, 392(10165), 2705–2717. [DOI] [PubMed]
- 13.Le, C. T., et al. (2018). LY2405319, an analog of fibroblast growth factor 21 ameliorates alpha-smooth muscle actin production through inhibition of the succinate-G-protein couple receptor 91 (GPR91) pathway in mice. PLoS One, 13(2), e0192146. [DOI] [PMC free article] [PubMed]
- 14.Talukdar, S., et al. (2016). A Long-Acting FGF21 molecule, PF-05231023, decreases body weight and improves lipid profile in non-human primates and type 2 diabetic subjects. Cell Metabolism, 23(3), 427–40. [DOI] [PubMed]
- 15.Raptis, D. D., Mantzoros, C. S., & Polyzos, S. A. (2023). Fibroblast growth factor-21 as a potential therapeutic target of nonalcoholic fatty liver disease. Therapeutics and Clinical Risk Management, 19, 77–96. [DOI] [PMC free article] [PubMed]
- 16.Dumont, J., et al. (2015). Human cell lines for biopharmaceutical manufacturing: history, status, and future perspectives. Critical Reviews in Biotechnology, 36(6), 1110–1122. [DOI] [PMC free article] [PubMed]
- 17.Zhang, M., et al. (2024). Advances in serum-free media for CHO cells: From traditional serum substitutes to microbial-derived substances. Biotechnology Journal, 19(6), e2400251. [DOI] [PubMed]
- 18.Zeh, N., et al. (2024). The new frontier in CHO cell line development: From random to targeted transgene integration technologies. Biotechnology Advances, 75, 108402. [DOI] [PubMed]
- 19.Liu, S., et al. (2024). Systematic comparison of rAAV vectors manufactured using large-scale suspension cultures of Sf9 and HEK293 cells. Molecular Therapy, 32(1), 74–83. [DOI] [PMC free article] [PubMed]
- 20.Tan, E., et al. (2021). HEK293 cell line as a platform to produce recombinant proteins and viral vectors. Frontiers in Bioengineering and Biotechnology, 9, 796991. [DOI] [PMC free article] [PubMed]
- 21.Fitzpatrick, D. P., et al. (2007). Searching for potential biomarkers of cisplatin resistance in human ovarian cancer using a label-free LC/MS-based protein quantification method. Proteomics: Clinical Applications, 1(3), 246–63. [DOI] [PubMed]
- 22.Wang, M., et al. (2008). Label-free mass spectrometry-based protein quantification technologies in proteomic analysis. Briefings in Functional Genomics and Proteomics, 7(5), 329–39. [DOI] [PubMed]
- 23.Micanovic, R., et al. (2009). Different roles of N- and C- termini in the functional activity of FGF21. Journal of Cellular Physiology, 219(2), 227–34. [DOI] [PubMed]
- 24.Rubino, D. M., et al. (2022). Effect of weekly subcutaneous semaglutide vs daily liraglutide on body weight in adults with overweight or obesity without diabetes: The STEP 8 randomized clinical trial. JAMA, 327(2), 138–150. [DOI] [PMC free article] [PubMed]
- 25.Li, Y., et al. (2023). The effect of subcutaneous dulaglutide on weight loss in patients with Type 2 diabetes mellitus: Systematic review and meta-analysis of randomized controlled trials. European Journal of Clinical Investigation, e14125. [DOI] [PubMed]
- 26.Zhang, B., et al. (2023). Efficacy and safety of mazdutide in chinese patients with type 2 diabetes: A randomized, double-blind, placebo-controlled phase 2 Trial. Diabetes Care. [DOI] [PMC free article] [PubMed]
- 27.Goetze, A. M., et al. (2011). High-mannose glycans on the Fc region of therapeutic IgG antibodies increase serum clearance in humans. Glycobiology. 21(7), 949–59. [DOI] [PubMed]
- 28.Jiang, Y., et al. (2023). Comprehensive analysis of the lysine succinylome in fish oil-treated prostate cancer cells. Life Science Alliance,6(11), e202302131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kaplon, H., et al. (2023). Antibodies to watch in 2023. MAbs, 15(1), 2153410. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Materials described in this manuscript including all relevant raw data will be freely available from the corresponding author for non-commercial use upon request.