Skip to main content
NCBI home page
ACS Pharmacology & Translational Science logoLink to ACS Pharmacology & Translational Science
. 2025 Jan 23;8(2):394–408. doi: 10.1021/acsptsci.4c00460

Enzymatic Desialylation Enables Reliable Charge Variant Characterization of Highly Glycosylated and Sialylated Fc Fusion Proteins

Xiaona Wen †,*, Anita P Liu †, Jing Song †, Chuan Leng †, Jingzhou Wang ‡, Briana Russo §, Geetha Thiagarajan †, Hongxia Wang †, Ximeng Y Dow †, Xiaoqing Hua †, Xiaoping Ao †, Sarita Mittal †, Lynn Gennaro †, Rico Gunawan †
PMCID: PMC11833721  PMID: 39974635

Abstract

graphic file with name pt4c00460_0008.jpg

Fusion proteins constitute a class of engineered therapeutics and have emerged as promising candidates for disease treatment. However, the structural complexity and heterogeneity of fusion proteins make their characterization extremely challenging, and thus, an innovative and comprehensive analytical toolbox is needed. Here, for the first time, we demonstrate a novel and robust workflow to evaluate charge variants for a highly glycosylated fusion protein with heavy sialylation using imaged capillary isoelectric focusing (icIEF). In the development of the icIEF method, key factors that were systematically investigated include the desialylation level, the stability of the desialylated molecule, incubation time and temperature of desialylation, protein concentrations, urea and l-arginine effects on the tertiary structure, and instrumental comparability. Multivariate and correlation analyses were subsequently applied to confirm the impacts of the parameters evaluated. Furthermore, a microfluidic chip-based icIEF system coupled with ultraviolet detection and mass spectrometry (icIEF-UV/MS) was utilized to identify critical post-translational modifications and ameliorate the understanding of charge variants. Our study demonstrates that this workflow enables a mechanistic understanding of charge variants for heavily sialylated therapeutics.

Keywords: fusion protein, glycosylation, sialylation, charge variant characterization, imaged capillary isoelectric focusing, icIEF-UV/MS

Fusion proteins are genetically engineered molecules consisting of a fragment crystallizable (Fc) domain of an antibody with another biologically active protein domain or peptide. The unique structure–function properties often provide these Fc fusion proteins with significant therapeutic potentials.1−3 To date, ∼13 fusion proteins have been approved by the U.S. Food and Drug Administration (FDA) and European Medicines Agency (EMA) and 37 molecules are currently in clinical trials.4 Fusion proteins have emerged as promising therapeutics mainly due to their distinct physicochemical and biological properties. Typically, the bioactive moieties have a relatively short half-life due to proteolytic metabolism through receptor-mediated cellular uptake5−7 and renal clearance, which is known to have significant impacts on products with a molecular mass smaller than the glomerular filtration threshold (∼60 kDa).8−10 Coupling these bioactive moieties with the Fc domain can extend circulation half-live via its interaction with the neonatal Fc receptor (FcRn) in a pH-mediated recycle process.11−13 Further, the added Fc domain can increase the hydrodynamic diameter of molecules to decrease the kidney filtration rate and prolong the circulation time.14 Therefore, the exposure of target tissue to the pharmacologically active moiety is enhanced, leading to greater therapeutic potential.15,16

Next, the Fc domain also significantly improves the stability and solubility of relatively hydrophobic ligands or receptors.2,17 The coupling of the moiety to the Fc region is usually achieved through mammalian cell-based production processes. Chinese hamster ovary (CHO) cell lines are one of the most commonly used host cell lines for the commercial expression of Fc fusion proteins to ensure the correct conformation and post-translational modifications (PTMs).18−20 Glycosylation is one of the most important PTMs that plays important roles in the immunogenicity and clinical efficacy of therapeutic proteins.21−23 Similar to monoclonal antibodies (mAbs), Fc fusion proteins contain conserved N-linked glycosylation at asparagine 297 (N297) in the heavy chain of CH2 domain-bearing complex biantennary glycans with low sialic acid content.24−26 These glycans have often been employed to increase affinity for the Fc γ receptor (FcγR) to consequently enhance Fc effector functions.27−29 These Fc glycans can also support the folding and stabilization of the Fc domain by maintaining a closed Fc conformation.30−32 Compared to mAbs, Fc fusion proteins often carry multiple N-glycosylation and additional O-linked glycosylation sites in the nonimmunoglobulin G (IgG) domain.31,33 These oligosaccharide structures include a wide variety of complex bi-, tri-, and tetra-antennary structures terminating with sialic acid and N-acetylglucosamine (GlcNAc) residues, which have far-reaching effects on the in vivo clearance of molecules.34−37 Hence, comprehensive and site-specific analysis of glycans is essential for the quality control strategy of Fc fusion proteins.

Similar to other biological modalities, Fc fusion proteins are also susceptible to many other PTMs that could result in size, charge, and other product variants that can compromise the product efficacy and safety.38,39 For example, modifications of Fc fusion proteins can result in changes to surface-exposed charged residues or acid dissociation constant, which lead to overall changes in the surface charge distribution.40−42 These modifications are generally referred to as charge variants, including acidic species with lower apparent isoelectric point (pI) values and basic species with higher pI values. The most common modifications resulting in charge variants are sialylation, deamidation, C-terminal lysine variations, and N-terminal cyclization.4 In general, fusion proteins with a high degree of glycosylation and complex sialylation patterns are extremely difficult to analyze charge variants at an intact level.4 Despite the challenges in analysis caused by the inherent heterogeneity of fusion proteins, the regulatory authorities expect a thorough characterization of charge variants throughout all developmental and commercial phases of a therapeutic protein to ensure consistent product quality.43

Imaged capillary isoelectric focusing (icIEF) is an established platform release method for charge variant analysis.44−46 Compared to conventionally used ion-exchange chromatography (IEX) methods, icIEF separates species based on the pI of various species within a provided pH gradient and is not affected by stationary phase interaction.45 The relatively better separation efficiency, shorter separation time, and higher sample throughput render icIEF an attractive alternative compared to traditional IEX.47 Recent advances in icIEF technology development have enabled direct coupling to mass spectrometry (MS), allowing for high-resolution separation, rapid pI measurement, and ultraviolet (UV)-based charge variant quantitation and mass characterization in one workflow,48 and improved instrumental operations.49 However, complex and highly heterogeneous therapeutic proteins are still particularly difficult to analyze via icIEF.

In this work, we developed a workflow by employing icIEF and intrinsic fluorescence in conjunction with multivariate and correlation analysis and icIEF-UV/MS to address the challenges of the charge variant characterization of structurally heterogeneous fusion proteins. Such a workflow consisted of four steps. First, our workflow started with desialylation of protein to simplify its complex structure. Key parameters in desialylation that need thorough examination include the desialylation level, the stability of the desialylated molecule, and incubation time and temperature of desialylation. Second, our workflow evaluated the impact of protein concentrations, urea, l-arginine, and instrumental comparability on the overall icIEF performance. Third, our workflow applied multivariate and correlation analyses as complementary tools to confirm the effects of a variety of experimental parameters evaluated. Lastly, our workflow employed icIEF-UV/MS to identify critical PTMs and decipher each charge variant.

For the first time, we showed the correlation between the level of desialylation and the variabilities caused by changes in the storage conditions of desialylated materials, incubation time and temperature of desialylation, and protein concentration. Our results demonstrated the criticality of thoroughly understanding the desialylation step robustness during method development. We also used intrinsic fluorescence to understand the effects of desialylation and the concentrations of urea and l-arginine, which are commonly used components in icIEF assay buffers, on the protein tertiary structure. Although there was no impact derived from desialylation and l-arginine, urea with a concentration of at least 6.5 M has significantly changed the protein tertiary structure in solution. By integrating icIEF with multivariate and correlation analysis, we highlighted the use of this complementary platform to corroborate the impacts of instrumental changes and a variety of method parameters evaluated. To ameliorate the understanding of charge variants, we applied icIEF-UV/MS to decode the molecular mass of each charge variant and determine critical PTMs that need to be closely monitored. Our study goes beyond the traditional ways of charge variant characterization and provides guidance for the in-depth characterization of highly glycosylated and sialylated fusion proteins, which ultimately will be safe and effective therapeutics received by patients.

Experimental Section

Materials

One Fc fusion protein comprising the extracellular domain (ECD) linked to an IgG1 Fc fragment was produced in-house by Merck & Co., Inc., Rahway, NJ. Acetonitrile, acetic acid, ammonium formate, formic acid, β-mercaptoethanol (BME), carboxypeptidase B, dithionite, 1,2-diamino-4,5-methylene-dioxybenzene (DMB), N-acetyl-neuraminic acid (NANA), N-glycolyl-neuraminic acid (NGNA), 2-(N-morpholino)ethanesulfonic acid (MES), sodium dodecyl sulfate (SDS), sodium phosphate monobasic monohydrate, sodium phosphate dibasic dihydrate, l-arginine monohydrochloride, urea, iminodiacetic acid, formamide, l-arginine, carboxypeptidase B (CpB), and Zeba 7K MWCO spin desalting columns were purchased from Sigma-Aldrich. pI markers at 3.59, 4.05, 4.22, and 7.05, Pharmalyte 2.5–5, Pharmalyte 5–8, 1% methyl cellulose, Maurice, and iCE3 were purchased from ProteinSimple. PNGase F was purchased from New England Biolabs. The GlycoWorks Rapi-Fluor-MS N-glycan kit was purchased from Waters. The 10 kDa internal standard was purchased from Sciex. The 40% Servalyt 2–11 was purchased from Servalyt. 51 kDa sialidase A was purchased from Agilent. Milli-Q water (18 MΩ) was obtained from a Milli-Q direct-Q 3UV system. The Acquity BEH amide column and Acquity BEH column were purchased from Waters. The GlycoSep R column was purchased from Prozyme.

Hydrophilic Interaction Ultra-Performance Liquid Chromatography (HILIC-UPLC) Rapi-Fluor Labeled N-Glycan Analysis

15 μg of intact protein was first mixed with 15 μL of Milli-Q water and 6 μL of 5% (w/v) Rapi-Gest SF solution. The mixture was heated at 90 °C for 3 min, followed by the addition of 1.2 μL of PNGase F and incubation at 50 °C for 5 min. This deglycosylated solution was then labeled with glycosylamines at room temperature (RT) for at least 5 min by adding 12 μL of Rapi-Fluor-MS solution and 358 μL of 100% acetonitrile. Afterward, the labeled glycosylamines were washed twice with 600 μL of a washing solution made of 1% formic acid and 90% acetonitrile. Samples were then collected using 30 μL of SPE elution buffer, followed by dilution using 100 μL of the GlycoWorks sample diluent. The collected samples were tested using an Acquity UPLC glycan BEH amide column (130 Å, 1.7 μm, 2.1 × 150 mm) along with two mobile phases, including 50 mM ammonium formate at pH 4.5 (mobile phase A) and 100% acetonitrile (mobile phase B). Data were measured using fluorescence detection with excitation at 265 nm and emission at 425 nm and analyzed with Waters Empower 3 FR5.

Reversed-Phase Ultra-Performance Liquid Chromatography (RP-UPLC) Sialic Acid Analysis

10 μg of intact protein was mixed with 50 μL of 4 M acetic acid and then incubated at 80 °C for 2 h. Then, 10 μL of hydrolyzed samples were labeled with 40 μL of 7 mM DMB solution at 50 °C for 3 h. The DMB solution was made by mixing 1 mg of dithionite with 319 μL of derivatization solution composed of 1.4 M acetic acid and 1 M BME. Two calibration standards for NANA and NGNA were also prepared by first reconstituting them with 7 mM DMB solution to reach a concentration of 14.9 pmol/μL and then incubating them at 50 °C for 3 h. Afterward, 950 μL of Milli-Q water was added to each sample and calibration standard. The labeled NANA and NGNA standards were serially diluted with Milli-Q water to 0.000745 pmol/μL to generate standard curves. All prepared samples and standards were tested using a GlycoSep R column (Prozyme, 3 μm, 4.6 × 150 mm) with Milli-Q water as mobile phase A and 100% acetonitrile as mobile phase B. Data were measured using fluorescence detection with excitation and emission at 373 and 448 nm, respectively, and analyzed with a Waters Empower 3 FR5.

Ultra-Performance Size Exclusion Chromatography (UP-SEC) Analysis

30 μg of protein was prepared using a mobile phase (50 mM phosphate, 450 mM arginine mono HCl, pH 7.0) and then injected onto an Acquity UPLC protein BEH SEC column (200 Å, 1.7 μm, 4.6 × 300 mm). Data were collected at a wavelength of 280 nm by using a UV detector. Data analysis was performed by using Waters Empower 3 FR5.

Intrinsic Tryptophan Fluorescence Analysis

Intrinsic tryptophan fluorescence was performed using a FluoroMax-4 fluorometer (HORIBA). The protein samples were analyzed at ∼0.3 mg/mL in their corresponding buffers. The excitation wavelength was set to 295 nm, and the emission spectra were acquired from 300 to 550 nm with a scanning speed of 120 nm/min and a slid width of 2 nm. The integration time was set to 0.1 s, and the data interval was at 1 nm with three accumulations per sample. The measurements were performed at 20 °C. The sample spectra were then processed to subtract the corresponding buffer spectra prior to plotting and comparison.

Desialylation of Protein with Different Amounts of the Sialidase A Enzyme

Desialylation of protein was prepared by first diluting to 4.8 mg/mL with 5× reaction buffer, 51 kDa sialidase A, and Milli-Q water. Every 104 μL of diluted sample which contained 1× reaction buffer and 2, 4, 8, 12, 16, 20, or 30 μL of sialidase A was incubated at 37 °C for 2 h. Buffer blanks were prepared in a similar manner by replacing the protein with Milli-Q water. Prepared samples were then tested using the icIEF.

Stability of Desialylated Protein Treated with Different Amounts of the Enzyme

Desialylated protein treated with 4 or 20 μL of 51 kDa sialidase A at 37 °C for 2 h was stored at 4 °C for 1, 2, 3, 4, 5, 6, and 24 h, along with desialylated protein treated with 16 μL of 51 kDa sialidase A at 37 °C for 2 h and stored at room temperature for 1, 2, 3, 4, 5, 6, and 24 h; they both were subjected to stability evaluation using icIEF.

One Factor at a Time (OFAT) and Design of Experiment (DoE) Evaluation on Effects of Enzyme Amounts, Desialylation Incubation Time, and Temperature

In OFAT experiments, the protein was treated with either 4 or 20 μL of the enzyme and then exposed to 37 °C for 105, 120 (set point), or 135 min prior to addition of the master mix followed by icIEF analysis using Maurice. Additionally, the protein was treated with either 4 or 20 μL of the enzyme and then exposed to 35, 37 (set point), or 39 °C for 120 min prior to addition of the master mix followed by icIEF analysis using Maurice. For DoE experiments, a full factorial and completely randomized DoE with three parameters, including enzyme amount (16 ± 2 μL), incubation time (120 ± 15 min), and temperature (37 ± 2 °C), was performed. The DoE design with six runs at set points and a total of 14 runs was tested for icIEF analysis.

Linearity of Desialylated icIEF

The linearity of the desialylated protein was prepared by first diluting to seven different levels from 25 to 175%, corresponding to protein concentrations ranging from 1.20 to 8.41 mg/mL with 5× reaction buffer, either 4 or 20 μL of 51 kDa sialidase A and Milli-Q water. Every 104 μL of the diluted sample, which contained 1× reaction buffer and 4 or 20 μL of sialidase A, was incubated at 37 °C for 2 h. Prepared samples were then tested using icIEF.

icIEF Analysis

12 μL of the desialylated sample was mixed with 188 μL of the master mix containing pI markers (4.22 and 7.05), 40% Servalyt 2–11, 1% methyl cellulose, 10 mM l-arginine, and 8 M urea. Note: pI markers at 4.05 and 7.05 were used instead for intact icIEF evaluation. The samples were loaded onto Maurice or iCE3 (ProteinSimple) with the following instrument parameters: 10 °C autosampler temperature, 280 nm detection wavelength, absorbance at 0.005 s, fluorescence at 3, 5, 10, 20, and 30 s (Maurice only), 90 s sample injection load, and two focusing periods including 1 min at 1500 V and 9 min at 3000 V. Data analysis was performed using Waters Empower 3 FR5.

Desialylation and N-Deglycosylation of the Protein

Desialylation and N-deglycosylation of the protein were performed by first diluting to 3 mg/mL with PNGase F, 51 kDa sialidase A, 10 × glycobuffer 2, and 5× reaction buffer. The diluted sample, which contained 30% (v/v) PNGase F, 15% (v/v) 51 kDa sialidase A, 1× glycobuffer, and reaction buffer, was incubated at 37 °C for 48 h. A negative control was also prepared in a similar manner by replacing protein with Milli-Q water. After 48 h of incubation, the samples were aliquoted and stored at −80 °C until analysis.

Desialylation and CpB Treatment of the Protein

Desialylation and CpB treatment of the protein were performed by first diluting to 3 mg/mL with 51 kDa sialidase A, CpB, and 5× reaction buffer. The diluted sample, which contained 5% (v/v) 51 kDa sialidase A, 0.08 mg/mL CpB, and 1× reaction buffer, was incubated at 37 °C for 2 h. A corresponding negative control was also prepared in a similar manner by replacing protein with Milli-Q water. After incubation, the samples were aliquoted and stored at −80 °C until analysis.

Nonreduced Capillary Electrophoresis with Sodium Dodecyl Sulfate (CE-SDS) Analysis

Nonreduced CE-SDS was performed by first diluting samples to 1.5 mg/mL with a dilution buffer made of 50 mM MES at pH 6.0. Every 90 μL of the diluted sample was mixed with 114 μL of the nonreducing master mix containing sample buffer (50 mM MES and 2% SDS), 10 kDa internal standard, and 250 mM iodoacetamide to obtain a final concentration of 0.66 mg/mL. Samples were incubated at 70 °C for 5 min. Following incubation, samples were placed onto a Beckman Coulter/Sciex PA800 Plus CE instrument with a PDA detector at 220 nm. Data analysis was performed using Waters Empower 3 FR5.

Reduced CE-SDS Analysis

Reduced CE-SDS was performed by first diluting samples to 1.5 mg/mL with a dilution buffer made of 50 mM MES at pH 6.0. Every 90 μL of the diluted sample was mixed with 114 μL of the reducing master mix containing sample buffer (50 mM MES and 2% SDS), 10 kDa internal standard, and BME to obtain a final concentration of 0.66 mg/mL. Samples were incubated at 70 °C for 5 min. Following incubation, samples were placed onto a Beckman Coulter/Sciex PA800 Plus CE instrument with a PDA detector at 220 nm. Data analysis was performed using Waters Empower 3 FR5.

icIEF-UV/MS Analysis

30 μL of desialylated and N-deglycosylated materials were desalted with Zeba 7K MWCO spin desalting columns and then mixed with 120 μL of a proprietary master mix solution from SCIEX containing pI markers (3.59 and 7.05), Pharmalyte 2.5–5, Pharmalyte 5–8, iminodiacetic acid, and formamide. The samples were loaded onto an Intabio ZT system coupled to a SCIEX ZenoTOF 7600 system, where the samples were separated for 1 min at 1500 V, followed by 1 min at 3000 V, and then 5.5 min at 4500 V before mobilization to the mass spectrometer at 9000 V for 12.5 min. The UV detection wavelength was 280 nm. The TOF-MS acquired m/z 1500–6000 with 25 psi curtain gas, 8 psi CAD gas, 100 °C source, 200 eV declustering potential, and 60 eV collision energy to identify charge variants. Data analysis was performed by using Expressionist Genedata (version 17.0).

Statistical Analysis

Statistical analysis was performed with a Python script (Python version 3.8.5 configured with Anaconda3 on Win32). To assess correlations and differences for peaks (% area and pI) generated under different instruments or enzyme amounts, Pearson correlation between peaks of each identical sample and 95% confidence intervals of peak differences under different conditions were calculated with the SciPy package (version 1.5.2). Calculations of the coefficient of determination between peaks of each sample under different conditions and principal component analysis for all peaks under each condition were performed with the scikit-learn package (version 0.23.2). For the DoE robustness study, a multivariate least-squares model was created using JMP (version 17.2.0) for each icIEF species with the effects of enzyme amount, incubation time, and temperature, along with all quadratics and interactions. A backward stepwise was then performed to remove any nonsignificant (p ≥ 0.05) quadratics and interactions. The main effects were retained in the model regardless of the statistical significance. The relevance of each effect was assessed with an estimate scaled to 1/2 of the range of each effect.

Results and Discussion

N-Glycan and Charge Variant Profiles of IgG1 Fc Fusion Protein

To introduce a new paradigm for characterizing structurally heterogeneous molecules, we chose a recombinant human homodimer fusion protein comprising the ECD linked to the Fc domain of IgG1 as a model protein. To illustrate the complexity of this model protein, we first evaluated the oligosaccharide profile of this Fc fusion protein by using HILIC-UPLC to separate PNGase F enzymatically released and Rapi-Fluor labeled N-glycans based on their hydrophilicity. This molecule exhibited a complex N-glycan pattern with 47 distinguished peaks identified by mass spectrometry (Figure 1a), where ∼53% of N-glycan species were mono-, di-, tri-, and tetra-sialylated (Figure 1b). Compared to the Fc fusion protein, it is well-known that monoclonal antibodies contain fewer N-glycan species, typically exhibiting ∼20 N-glycan peaks.50,51 We further quantitated the sialic acid content by RP-UPLC with fluorescence detection and observed the major species as NANA with 9.56 pmol of NANA per pmol of protein and minor species as NGNA with 0.18 pmol of NGNA per pmol of protein (Figure S1). To identify the charge variant profiles, we performed the icIEF with Maurice using an intact molecule. The intact icIEF showed a broad pI profile without a distinguishable main peak (Figure 1c). Our findings were consistent with literature4,45,52,53 that Fc fusion proteins possess intricate glycosylation and sialylation patterns, which make it challenging to characterize their charge variants at an intact level. These collective data further affirmed the suitability of the selected model protein for our evaluation.

Figure 1.

Figure 1

Open in a new tab

N-glycan and charge variant profiles of a highly glycosylated and sialylated Fc fusion protein. (a) Chromatogram of N-glycan in an Fc fusion protein by Rapi-Fluor HILIC-UPLC. (b) Relative abundance of N-glycan species. SA represents sialic acid. Data were presented as mean ± standard deviation (n = 3). (c) Electropherogram of the intact Fc fusion protein determined by icIEF with a fluorescence feature of Maurice. Two peaks at pI values of 4.22 and 7.05 are pI markers.

Evaluation of the Desialylation Level and the Stability of Desialylated IgG1 Fc Fusion Protein

Due to the presence of heavy sialylation on the molecule, we sought to simultaneously reduce charge complexity and assess the impact of sialic acid on the charge variant profile by treating the Fc fusion protein with different amounts of the sialidase A enzyme. We first evaluated whether the utilized sialidase A enzyme to remove sialic acid had a signature peak in icIEF that could potentially interfere with the integration of electropherograms of the Fc fusion protein. The buffer blank with the sialidase A enzyme was subjected to icIEF analysis using absorbance detection of Maurice at a 0.005 s exposure time. A relatively flat baseline was obtained between two pI markers, and no sialidase A enzyme peak was found (Figure S2a). It is known that fluorescence detection has better sensitivity compared with ultraviolet (UV) absorbance detection. We then exposed the buffer blank to fluorescence detection using five different exposure times, including 3, 5, 10, 20, and 30 s. Although the pI marker of 7.05 started reaching saturation at 20 s exposure time, a peak at pI of 5.4 was observed, and its intensity increased with increasing exposure time (Figure S2b). These results confirmed that the sialidase A enzyme had a signature peak at a pI of 5.4, and thus, we did not include this peak in the integration of electropherograms of the molecule. Further, we observed that the relative % area of the main summing of main 1 and 2 and basic peaks increased with sialidase A enzyme amounts, whereas the acidic peak showed a decreasing trend (Figure 2a). All peaks started reaching a plateau around 12 μL of the sialidase A enzyme. It is noteworthy that among three acidic peaks, acidic 1 decreased significantly as sialidase A enzyme amounts increased (Figure 2b). To confirm these observations, the sialic acid content was performed on the desialylated Fc fusion protein treated with 2, 4, 10, 16, and 30 μL of the sialidase A enzyme. Interestingly, we noticed that the major species NANA attached to the molecule was fully removed using as little as 4 μL of the sialidase A enzyme (Figure S3a), whereas the minor species NGNA attached to the molecule decreased at different rates as sialidase A enzyme amounts increased (Figure S3b). These results were consistent with literature where the sialidase A enzyme cleaved NANA residues more preferentially than NGNA residues.54 These sialic acid content results also implied the roles of NGNA and NANA in charge variant profiles from icIEF.

Figure 2.

Figure 2

Open in a new tab

Desialylation of a highly glycosylated and sialylated Fc fusion protein and stability of a desialylated molecule as a function of desialylation level and storage temperature. (a) Relative % area of acidic, basic, main, main 1, and main 2 peaks as a function of sialidase A enzyme amount determined by icIEF with absorbance feature of Maurice (n = 2). (b) icIEF electropherograms of the Fc fusion protein treated with various sialidase A enzyme amounts (2, 4, 8, 12, 16, and 20 μL). Acidic, main 2, main 1, and basic variants are color-coded in red, blue, green, and purple, respectively. The sum of main 1 (%) and main 2 (%) results in main (%). (c) UP-SEC chromatograms of the desialylated Fc fusion protein indicating no significant changes in HMW and LMW compared to the untreated Fc fusion protein. Chromatograms were normalized to the monomer peak. (d) Relative % area of (i) main, (ii) acidic, and (iii) basic peaks of icIEF when the Fc fusion protein was treated with 4, 16, or 20 μL of the sialidase A enzyme and stored at 4 °C or RT, respectively, up to 24 h as a function of storage time (n = 2).

Sialic acid is important in protein stability,55,56 and thus, we evaluated the impact of desialylation on protein attributes. UP-SEC coupled with UV detection is a widely used methodology for analyzing the purity of protein samples. UP-SEC provides critical information on protein aggregation and chemical or proteolytic fragment levels of the sample, informing on sample quality and stability.4 Sample components are separated by size according to their ability to migrate through a porous chromatography column. During the separation process, large molecules elute from the column first, and smaller molecules elute later. We utilized UP-SEC to first examine the aggregation and fragment levels of the desialylated Fc fusion protein treated with various sialidase A enzyme amounts at 37 °C for 2 h. The results demonstrated minimal changes in high (HMW) and low (LMW) molecular weights compared to the untreated Fc fusion protein treated with various amounts of the sialidase A enzyme up to 20 μL (Figure 2c and Table S1). UP-SEC chromatograms also showed two peaks at 13.4 and 14.2 min increasing with the amount of the sialidase A enzyme (Figure S4a). We confirmed these two peaks contributed by the sialidase A enzyme by comparing chromatograms of buffer blanks with 0, 4, and 20 μL of the sialidase A enzyme (Figure S4b) and thus did not include these two peaks in the UP-SEC integration.

Additionally, we evaluated the stability of the desialylated Fc fusion protein up to 24 h as a function of the level of desialylation and storage temperature evaluated by icIEF prior to addition of the master mix followed by icIEF analysis using Maurice. We treated intact molecules with 4, 16, or 20 μL of the sialidase A enzyme and then stored them at 4 °C or room temperature (RT) as a function of storage time. We found that the relative % area of the main peak did not change significantly up to 24 h when storing the desialylated protein with 20 μL of the sialidase A enzyme at 4 °C, whereas major changes were noticed when the protein was treated with 4 μL of the sialidase A enzyme and stored at 4 °C (Figure 2d-i). Our results implied that the Fc fusion protein with a high degree of desialylation was more stable, driven by the kinetics of the desialylation process. To assess the effect of storage temperature on the stability of the desialylated protein, we exposed the desialylated molecule to RT using 16 μL of the sialidase A enzyme and observed an increasing trend in the relative % area of the main peak by comparing it with that subjected to 20 μL of the sialidase A enzyme and stored at 4 °C. Our previous results (Figure 2a) demonstrated that the Fc fusion protein treated with 16 or 20 μL of the sialidase A enzyme had a similar relative % area of the main peak. As expected, the highly desialylated protein was more stable when stored at 4 °C. By comparing the slope of the relative % area of the main peak using linear fitting (Figure 2d-i), we also concluded that the stability of the desialylated molecule was more affected by the level of desialylation than that of storage temperature. Overall, the relative % areas of main 1 and 2 (Figure S5), acidic (Figure 2d-ii), and basic (Figure 2d-iii) peaks were well aligned with our discussion above on the effect of sialidase A enzyme amount and storage temperature.

Impacts of Incubation Time, Temperature, and Protein Concentration on Desialylation

To evaluate the impact of incubation time and temperature on desialylation, the Fc fusion protein was treated with either 4 or 20 μL of the sialidase A enzyme and then incubated at 37 °C for 105, 120 (set point), or 135 min, followed by icIEF analysis using Maurice. The relative % area of main, acidic, basic, and main 1 and 2 peaks showed minor changes when varying the incubation time (Figures 3a and S6a). To understand how these changes correlated with the degree of desialylation, we calculated the absolute difference of relative % area for each peak by subtracting the relative % area using 120 min incubation time from that with 105 or 135 min and compared the obtained differences between 4 and 20 μL of the sialidase A enzyme. The results demonstrated that the impact of incubation time was intensified when 4 μL of the sialidase A enzyme was used and could be mitigated by increasing the amount of the sialidase A enzyme (Figure S6b). Further, we examined the effect of incubation temperature by subjecting the Fc fusion protein with either 4 or 20 μL of the sialidase A enzyme and then incubating at 35, 37 (set point), or 39 °C for 120 min, followed by icIEF analysis using Maurice (Figures 3b and S7a). Similarly, the various incubation temperatures did not have a significant impact on the relative % area for each peak, and this impact was minimized with increased sialidase A enzyme amount (Figure S7b). All these data indicated that a high level of desialylation of the Fc fusion protein prior to icIEF analysis could reduce reportable variabilities caused by deviation of incubation time and temperature.

Figure 3.

Figure 3

Open in a new tab

Desialylation of a highly glycosylated and sialylated Fc fusion protein as a function of incubation time and temperature. (a) Relative % area of main, acidic, and basic peaks as a function of incubation time where the Fc fusion protein was treated with either 4 or 20 μL of the sialidase A enzyme and then exposed to 37 °C for 105, 120 (set point), or 135 min prior to addition of the master mix followed by icIEF analysis using Maurice (n = 2). (b) Relative % area of main, acidic, and basic peaks as a function of incubation temperature where the Fc fusion protein was treated with either 4 or 20 μL of the sialidase A enzyme and then exposed to 35, 37 (set point), or 39 °C for 120 min prior to addition of the master mix followed by icIEF analysis using Maurice (n = 2). (c) Design of experiment results for relative % area of main, acidic, and basic peaks. The dashed gray lines represent the acceptable response range.

To further evaluate the effect of desialylation parameters, a full factorial and completely randomized DoE with three parameters including sialidase A enzyme amount (16 ± 2 μL), incubation time (120 ± 15 min), and temperature (37 ± 2 °C) was performed. The DoE design with six runs at set points defined by sialidase A enzyme amount as 16 μL, incubation time as 120 min, and incubation temperature as 37 °C is summarized in Table S2. The acceptable response range for the relative % area of main and main 1 and 2 peaks was defined by data that did not differ by more than 10% from the average data obtained at the set point, whereas for the relative % area for acidic and basic peaks, the acceptable response range was less than 30% of that obtained at the set point. The acceptable response range for each peak is summarized in Table S3. The parameter estimates were also calculated to understand the magnitude of changes caused by main effects and interactions on relative % area. The results of parameter estimates demonstrated that the relative % area of some peaks caused by certain parameters was statistically significant (Tables S4–S8). However, the relative % area of all peaks fell within its corresponding acceptable response range (Figures 3c and S8), which was also confirmed by the prediction profiler of each peak, which is a graphical assessment of the statistical and practical impacts of the parameters (Figure S9). These results further indicated that the impact of sialidase A enzyme amount (16 ± 2 μL), incubation time (120 ± 15 min), and temperature (37 ± 2 °C) were not practically significant. The DoE results were well aligned with our above conclusions that the highly desialylated Fc fusion protein can mitigate the data variabilities caused by the changes in incubation time and temperature.

We also assessed the linearity over seven different levels from 25 to 175%, corresponding to a final protein concentration ranging from 0.07 to 0.50 mg/mL, a function of desialylation level by using the Fc fusion protein treated either with 4 or 20 μL of the sialidase A enzyme. The results indicated that each peak area increased linearly with increasing protein concentrations (Figure S10), and at least 80% recovery was achieved at each level regardless of the desialylation level (Figure S11). However, regression analysis showed that the Fc fusion protein treated with 20 μL of the sialidase A enzyme had a better coefficient of determination (R2) compared to that with 4 μL of the sialidase A enzyme (Tables S9 and S10). These linearity data jointly suggested the advantages of using the Fc fusion protein with a higher degree of desialylation for icIEF analysis.

Multivariate and Correlation Analysis of icIEF

To closely examine the changes in % area and pI of icIEF upon exposure to different amounts of the sialidase A enzyme, we prepared 80 samples that were first subjected to different conditions, such as deliberative variations in method parameters and a variety of forced degradation conditions (e.g., heat and light stress) followed by treatment with either 4 or 20 μL of the sialidase A enzyme for side-by-side evaluation. We calculated the difference in % area and pI obtained by subtracting the % area and pI of samples treated with 4 μL of the sialidase A enzyme from those with 20 μL of the sialidase A enzyme. The box plot indicated that with increasing sialidase A enzyme amounts, the % area of acidic peaks decreased significantly by 9.4%, basic peaks slightly increased by 1.6%, and main peaks increased significantly by 7.8%, which was contributed by a 12.5% increase in main peak 1 and a 4.7% decrease in main peak 2 (Figure 4a). Further, no changes in pI were observed for all peaks upon exposure to 4 or 20 μL of the sialidase A enzyme (Figure 4b). The lower and upper 95% confidence intervals (CIs) for % area and pI differences are summarized in Tables S11 and S12. To identify major patterns in % area and pI that distinguish the samples treated with 4 μL of the sialidase A enzyme from those with 20 μL of the sialidase A enzyme, principal component analysis (PCA), which is a dimensionality reduction and unsupervised clustering and classification tool,57−62 was applied to the data sets. Each dot in the PC scatter plot represents a sample with dimension reduction. The first and second principal components, PC1 and PC2, presented in a two-dimensional PC scatter plot showed a clear clustering in % area between two groups of samples with a variance level of 71.7% for PC1 and 27.6% for PC2 (Figure 4c). However, no clustering was found in the PC scatter plot when comparing the pI of samples treated with 4 or 20 μL of the sialidase A enzyme (Figure 4d). We further explored the Pearson correlation to find patterns in % area and pI obtained from samples treated with 4 or 20 μL of the sialidase A enzyme and used the coefficient of determination to determine how well fittings match data. For % area, we obtained a Pearson correlation of 0.91, confirming the existence of a relationship between the two groups evaluated (4 vs 20 μL of the sialidase A enzyme) and the coefficient of determination of 0.80, indicating a good model fitting (Figure S12a). For pI, we found a Pearson correlation and coefficient of determination of 1.00 (Figure S12b). These PCA and R2 statistical analyses were well aligned with our previous observations on the effects of sialidase A enzyme amounts on the % area and pI of icIEF.

Figure 4.

Figure 4

Open in a new tab

Statistical analysis of % area and pI in icIEF for Fc fusion proteins exposed to different amounts of the sialidase A enzyme. The box plots of (a) % area and (b) pI differences obtained by subtracting the % area and pI of samples treated with 4 μL of the sialidase A enzyme from those with 20 μL of the sialidase A enzyme (n = 80). The PC scatter plots compare (c) % area and (d) pI of samples treated with 4 μL of the sialidase A enzyme to those treated with 20 μL of the sialidase A enzyme (n = 80).

Impacts of Urea, l-Arginine, and Desialylation Level on the Tertiary Structure of the IgG1 Fc Fusion Protein

In addition to the evaluation of sialidase A enzyme amounts in the % area and pI of icIEF, we examined the attributes of the master mix on protein properties. Briefly, the master mix consists of ampholytes, pI markers, and additives that are mixed with the sample of interest to enable charge variant separation in the icIEF assay. Here, the Fc fusion protein was treated with 4 or 20 μL of the sialidase A enzyme and then added with and without a master mix. The master mix only was also prepared as a control. UP-SEC results indicated that samples with the master mix significantly degraded compared to those without the master mix (Figure S13), which could be contributed likely by urea (0.6 M) in the master mix that is known to denature proteins. We further investigated the effects of urea and l-arginine, which are two main components of the master mix on the tertiary structure of the Fc fusion protein as a function of desialylation level. We first desialylated the Fc fusion protein with 20 μL of the sialidase A enzyme and then mixed it with urea at different concentrations up to 8.7 M. By evaluating the intensity and wavelength of peak maximum (λ Max), the intrinsic tryptophan fluorescence measurements demonstrated that the tertiary structure of the desialylated Fc fusion protein was significantly changed upon exposure to urea with a concentration of at least 6.5 M (Figure 5a). We also used intact Fc fusion protein combined with different urea concentrations. A similar trend was obtained, and at least 6.5 M urea was needed to cause changes in the tertiary structure (Figure 5b), indicating that desialylation did not have a significant impact on the protein tertiary structure. Afterward, we studied the impact of l-arginine concentrations on the protein tertiary structure using either desialylated or intact Fc fusion protein. Comparable intensity and wavelength of λ Max for tryptophan fluorescence were noted after exposure to l-arginine up to 605 mM for both desialylated (Figure 5c) and intact (Figure 5d) proteins, suggesting comparable tertiary structures for all samples tested regardless of desialylation applied. These intrinsic tryptophan fluorescence analyses confirmed our observation from UP-SEC regarding the role of urea in protein properties.

Figure 5.

Figure 5

Open in a new tab

Effects of urea and l-arginine on the tertiary structure of the Fc fusion protein as a function of desialylation level. The (i) spectra and (ii) plots of intensity and wavelength of peak maximum (λ Max) from the intrinsic tryptophan fluorescence measurements by exposing to different concentrations of urea (up to 8.7 M) using either (a) desialylated or (b) intact Fc fusion proteins to understand the impact of urea and desialylation on protein tertiary structure (n = 3). Similarly, the (i) spectra and (ii) plots of intensity and wavelength of peak maximum (λ Max) from the intrinsic tryptophan fluorescence measurements by exposing to different concentrations of l-arginine (up to 605 mM) using either (c) desialylated or (d) intact Fc fusion proteins to understand the impact of l-arginine and desialylation on protein tertiary structure (n = 3). Data were presented as mean ± standard deviation.

Impacts of icIEF Instrumentation on Desialylated IgG1 Fc Fusion Protein

Along with a wide interest of using icIEF as one of the primary tools by biopharmaceutical industries for determining charge heterogeneity and identity of therapeutic proteins, the evolution of icIEF detection technology continues to improve over the years to simplify the operations and reduce instrumentation failures.49 Many biopharmaceutical industries are currently undergoing the icIEF instrument transition from iCE3 to Maurice. Thus, it is important to demonstrate the comparability between iCE3 and Maurice in support of process development and regulatory submissions of biotherapeutic molecules. Here, we prepared 135 samples using the Fc fusion protein exposed to different experimental conditions and then tested them using iCE3 and Maurice side-by-side. Similarly, we calculated the difference in % area and pI obtained by subtracting the % area and pI of samples obtained by using iCE3 from those of Maurice. We observed a minimal decrease (0.4%) in the % area of acidic peaks and a significant decrease (5.2%) in main peaks, leading to a significant increase (5.6%) in basic peaks when using Maurice (Figure S14a). Minimal changes (<0.1) in isoelectric point (pI) were observed for all peaks obtained using iCE3 and Maurice (Figure S14b). The lower and upper 95% CIs for the % area and pI differences of samples obtained from iCE3 and Maurice are summarized in Tables S13 and S14. The PC scatter plots demonstrated a clear clustering in both % area and pI between the two groups of samples measured using iCE3 and Maurice (Figure S14c,d). PC1 and PC2 accounted for 53.2% and 45.1% of the total variance, respectively, when comparing the % area of samples, whereas a PC1 of 84.9% and a PC2 of 8.6% were found when comparing the pI of samples obtained using two instruments. We also performed the Pearson correlation and coefficient of determination of these data sets. A Pearson correlation of 0.99 and a coefficient of determination of 0.96 were obtained for the % area of samples measured using iCE3 and Maurice (Figure S14e). For pI, we observed a Pearson correlation of 0.98 and a coefficient of determination of 0.80 (Figure S14f). All of these data jointly indicated that instrumental transition from iCE3 to Maurice resulted in significant changes in % area, especially for basic and main peaks, and minimal changes in the pI of samples. Therefore, it is recommended to conduct a comprehensive bridging study empowered by statistical analysis before transitioning to a new instrument for testing.

Advanced Characterization of Charge Variants Using icIEF-UV/MS

Lastly, to understand and characterize charge variants in our model Fc fusion protein, we used icIEF-UV/MS, which is an online microfluidic chip-based system to obtain rapid charge variant separation along with molecular mass identification of charge variant peaks.48 Specifically, in the integrated icIEF-UV/MS workflow, samples were first loaded and focused in a microfluidic chip separation channel and then mobilized electrophoretically into the mass spectrometer for subsequent analysis. Since broad pI profiles from intact icIEF analysis and highly complex glycosylation from N-glycan analysis were observed in the model Fc fusion protein, the molecule was enzymatically treated with both PNGase F and sialidase A for N-glycan and sialic acid removal, respectively, at 37 °C for 48 h to reduce heterogeneity and ameliorate the understanding of charge variants. Prior to the icIEF-UV/MS analysis, we first evaluated whether enzymatic treatment could induce aggregation and fragmentation. UP-SEC showed minimal changes in the % area of HMW and LMW upon sialic acid removal for 48 h compared to T0, where samples were frozen immediately after treatment (Table S15). In contrast, a slight increase (∼0.3%) in the % area of LMW was observed following N-glycan or N-glycan coupled with sialic acid removal for 48 h. To further evaluate LMW species, reduced and nonreduced CE-SDS were performed. CE-SDS is commonly utilized for fragmentation analysis due to its high resolution, accurate quantitation, and high degree of automation.63 Reduced CE-SDS is typically employed to discriminate aggregates formed by disulfide bonds from those held together by other covalent bonds, while nonreduced CE-SDS involves the use of alkylating agents to block free thiols forming disulfide bonds.4 Our data demonstrated that minor forms in CE-SDS are likely product-related impurities, as indicated by their relative size mobility and susceptibility to glycanases and sialidase (Figure S15). Collectively, the UP-SEC and CE-SDS data demonstrated that the enzymatic treatment did not induce significant aggregation and/or fragmentation that could potentially interfere with the icIEF-UV/MS analysis. Afterward, for the icIEF-UV/MS analysis, due to the relative position of the charge variant peaks with respect to the electrospray ionization tip, the MS base peak electropherogram (BPE) showed a profile that was a mirror image of the icIEF-UV data, where basic peaks were positioned to the left of the main peak and acidic peaks are positioned to the right of the main peak (Figure 6). The charge variant profiles observed from icIEF-UV/MS revealed two acidic peaks (“Acidic 1” and “Acidic 2”), a main peak (“Main”), and one basic peak (“Basic 1”).

Figure 6.

Figure 6

Open in a new tab

Characterization of the N-deglycosylated and desialylated Fc fusion protein by icIEF-UV/MS. Representative (a) icIEF-UV profile and (b) mass spectrometry base peak electropherogram of the N-deglycosylated and desialylated Fc fusion protein.

The deconvoluted mass spectra of each charge variant peak in the N-deglycosylated and desialylated Fc fusion protein indicated that the predominant charge variant species observed were product-related species and not a result of impurities (Figure 7 and Table 1). For the main peak, the deconvoluted mass spectra revealed an O-glycan profile. The observed masses were close to theoretical masses and indicated that the distribution of the O-glycans ranged from three to eight O-glycans, with six of the O-glycans being the most abundant species. The deconvoluted mass spectra of Acidic 1 and Acidic 2 peaks indicated that deamidation was present, since a stepwise increase in mass was observed between the mass of the intended product in the main peak and the mass of the most abundant proteoform within each acidic peak. Deamidation is a post-translational modification (PTM) that results from the conversion of asparagine to aspartic acid, leading to a decrease in pI and a mass increase of 0.98 Da. In the Acidic 1 peak, another distribution with mass shifts of approximately +2352 Da was observed. The size of this mass shift suggested that it was likely attributed to a glycan, which could either be the result of an O-glycan, as the sample was N-deglycosylated but the O-glycans remained, or an O-glycan that was not removed, which can occur if the glycosylation site was located in a region difficult for PNGase F to access. The Acidic 2 peak profiles showed slight (+ 0.98 Da) mass shifts compared to masses observed in the Acidic 1 peak and were likely composed of a combination of deamidation and +2352 Da mass. In the deconvoluted mass spectra of the Basic 1 variant, a mass shift of ∼128 Da was observed, likely corresponding to the presence of one C-terminal lysine. C-terminal lysine is a common PTM that imparts a positive charge, resulting in an increase in pI and a mass increase of 128.1 Da. To further confirm the basic variant identity, the molecule was desialylated with and without carboxypeptidase B (CpB), an enzyme that can selectively cleave the peptide bond at the C-terminal lysine. The desialylated and the desialylated coupled with CpB-treated samples were analyzed by icIEF (Figure S16 and Table S16). The results demonstrated a decrease in basic variants from 9.7 to 6.5% upon additional treatment with CpB, confirming the presence of C-terminal lysine in basic variants. Altogether, the icIEF-UV/MS data provided a better understanding of change variants in our model Fc fusion protein.

Figure 7.

Figure 7

Open in a new tab

Deconvoluted mass spectra of the desialylated and N-deglycosylated Fc fusion protein. DN, DSO, K, DA, and NO Glycan represent de-N-glycosylated, desialylated O-glycan, c-terminal lysine, deamination, and N- or O-glycan, respectively.

Table 1. Charge Variant Analysis of the Desialylated and N-Deglycosylated Fc Fusion Protein.

charge variant peak charge variant identity predicted Δintact mass (Da) observed Δintact mass (Da)
basic 1 +1× C-terminal lysine +128 +128
main      
acidic 1 +deamidation +1 +1
+unknown glycan +2352 +2355
acidic 2 +deamidation +1 +2
+deamidated unknown glycan +2353 +2354
Open in a new tab

Conclusions

In summary, this work introduces a state-of-the-art approach to examining charge variants in highly heterogeneous fusion proteins. Here, we employed icIEF to examine the effects of desialylation on method performance. Additionally, intrinsic fluorescence was used to assess the influences of desialylation and concentrations of urea and l-arginine on the protein tertiary structure. Our findings of this study had several implications. First, a higher level of enzymatic desialylation enabled mitigation of data variabilities induced by changes in storage conditions of desialylated materials, incubation time and temperature of desialylation, and protein concentration. Second, the integration of icIEF with multivariate and correlation analysis demonstrated a powerful platform to confirm the impacts of instrumental changes and different parameters evaluated during method development. Third, urea with a concentration of at least 6.5 M considerably altered the tertiary structure of the fusion protein, whereas no impacts from desialylation and l-arginine were observed. Furthermore, we pioneered the application of icIEF-UV/MS to rapidly obtain the molecular mass of each charge variant and determine important post-translational modifications that need to be closely monitored for product quality. Ultimately, we anticipate that our findings will provide valuable information for scientists involved in the characterization of complex modalities, thus benefiting the discovery and development of future therapeutics.

Acknowledgments

The authors would like to thank Jason Cheung, Douglas Richardson, Wei Xu, and Eric Routhier from Merck & Co., Inc., Rahway, New Jersey, United States, for scientific reviews.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00460.

  • Sialic acid profile of the Fc fusion protein; icIEF electropherogram of buffer blank using absorbance and fluorescence detection of Maurice; sialic acid content of the Fc fusion protein after desialylation; relative % area of HMW, monomer and LMW of UP-SEC of the untreated and desialylated Fc fusion protein; UP-SEC chromatograms of the desialylated Fc fusion protein and buffer blank; stability of the desialylated Fc fusion protein up to 24 h as a function of level of desialylation and storage temperature evaluated by icIEF; desialylation of the Fc fusion protein as a function of incubation time; desialylation of the Fc fusion protein as a function of incubation temperature; design of experiment for icIEF to evaluate the effect of sialidase A enzyme amount, desialylation incubation time, and temperature, acceptable response range for relative % area of main, main 1 and 2, and acidic and basic peaks; parameter estimates for relative % area of the main peak by sialidase A enzyme amount, desialylation incubation time, and temperature; parameter estimates for relative % area of main 1 peak by sialidase A enzyme amount, desialylation incubation time, and temperature; parameter estimates for the relative % area of main 2 peak by sialidase A enzyme amount, desialylation incubation time, and temperature; parameter estimates for the relative % area of the acidic peak by sialidase A enzyme amount, desialylation incubation time, and temperature; parameter estimates for the relative % area of the basic peak by sialidase A enzyme amount, desialylation incubation time, and temperature; design of experiment results for the relative % area of main 1 and 2 peaks; prediction profilers for the relative % area of main, main 1 and 2, and acidic and basic peaks; linearity of desialylated icIEF as a function of desialylation level; percent recovery of acidic, basic main, main 1, and main 2 peaks at different linearity levels; linearity regression analysis when the Fc fusion protein was treated with 4 μL of the sialidase A enzyme; linearity regression analysis when the Fc fusion protein was treated with 20 μL of the sialidase A enzyme; lower and upper 95% confidence intervals for % area difference by comparing samples treated with different amounts of the sialidase A enzyme; lower and upper 95% confidence intervals for pI difference by comparing samples treated with different amounts of the sialidase A enzyme; evaluation of Pearson correlation and coefficient of determination in the % area and pI of samples obtained by treating with either 4 or 20 μL of the sialidase A enzyme; UP-SEC chromatograms of the desialylated Fc fusion protein in the presence or absence of the master mix; statistical analysis of the % area and pI of Fc fusion proteins in icIEF measured using iCE3 and Maurice; lower and upper 95% confidence intervals for % area difference by comparing samples obtained using iCE3 and Maurice; lower and upper 95% confidence intervals for pI difference by comparing samples obtained using iCE3 and Maurice; relative % area of HMW, monomer, and LMW of UP-SEC of the Fc fusion protein treated with PNGase F and sialidase A to remove N-glycan and sialic acid; nonreduced and reduced CE-SDS profiles for desialylated, N-deglycosylated, and desialylated along with N-deglycosylated treated Fc fusion protein; icIEF profiles of desialylated and desialylated with CpB-treated Fc fusion protein; and icIEF results of the Fc fusion protein after desialylation and desialylation along with CpB treatment (PDF)

The authors declare no competing financial interest.

Supplementary Material

pt4c00460_si_001.pdf (796.8KB, pdf)

References

  1. Ebrahimi S. B.; Samanta D. Engineering protein-based therapeutics through structural and chemical design. Nat. Commun. 2023, 14 (1), 2411 10.1038/s41467-023-38039-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Yang C.; Gao X.; Gong R. Engineering of Fc fragments with optimized physicochemical properties implying improvement of clinical potentials for Fc-based therapeutics. Front. Immunol. 2018, 8, 1860 10.3389/fimmu.2017.01860. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Huang C. Receptor-Fc fusion therapeutics, traps, and MIMETIBODY technology. Curr. Opin. Biotechnol. 2009, 20 (6), 692–699. 10.1016/j.copbio.2009.10.010. [DOI] [PubMed] [Google Scholar]
  4. Duivelshof B. L.; Murisier A.; Camperi J.; Fekete S.; Beck A.; Guillarme D.; D’Atri V. Therapeutic Fc-fusion proteins: current analytical strategies. J. Sep. Sci. 2021, 44 (1), 35–62. 10.1002/jssc.202000765. [DOI] [PubMed] [Google Scholar]
  5. Wang L.; Wang N.; Zhang W.; Cheng X.; Yan Z.; Shao G.; Wang X.; Wang R.; Fu C. Therapeutic peptides: current applications and future directions. Sig. Transduct. Target. Ther. 2022, 7 (1), 48. 10.1038/s41392-022-00904-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Sato A. K.; Viswanathan M.; Kent R. B.; Wood C. R. Therapeutic peptides: technological advances driving peptides into development. Curr. Opin. Biotechnol. 2006, 17 (6), 638–642. 10.1016/j.copbio.2006.10.002. [DOI] [PubMed] [Google Scholar]
  7. Grevys A.; Frick R.; Mester S.; Flem-Karlsen K.; Nilsen J.; Foss S.; Sand K. M. K.; Emrich T.; Fischer J. A. A.; Greiff V.; et al. Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 2022, 25 (2), 103746 10.1016/j.isci.2022.103746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Glassman P. M.; Vladimir R. M. Pharmacokinetic and pharmacodynamic properties of drug delivery systems. J. Pharmacol. Exp. Ther. 2019, 370 (3), 570. 10.1124/jpet.119.257113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bolleddula J.; Kevin B.; Gerard B.; Anthony L.; Jennifer A. M.; Markus W.; Keyang X.; Tong-Yuan Y.; Xiaochun Z.; Hongbin Y. Absorption, distribution, metabolism, and excretion of therapeutic proteins: current industry practices and future perspectives. Drug Metab. Dispos. 2022, 50 (6), 837. 10.1124/dmd.121.000461. [DOI] [PubMed] [Google Scholar]
  10. Meibohm B.; Zhou H. Characterizing the impact of renal impairment on the clinical pharmacology of biologics. J. Clin. Pharmacol. 2012, 52 (S1), 54S–62S. 10.1177/0091270011413894. [DOI] [PubMed] [Google Scholar]
  11. Pyzik M.; Sand K. M. K.; Hubbard J. J.; Andersen J. T.; Sandlie I.; Blumberg R. S. The neonatal Fc receptor (FcRn): a misnomer?. Front. Immunol. 2019, 10, 1540 10.3389/fimmu.2019.01540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Sockolosky J. T.; Szoka F. C. The neonatal Fc receptor, FcRn, as a target for drug delivery and therapy. Adv. Drug Delivery Rev. 2015, 91, 109–124. 10.1016/j.addr.2015.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Seijsing J.; Lindborg M.; Höidén-Guthenberg I.; Bönisch H.; Guneriusson E.; Frejd F. Y.; Abrahmsén L.; Ekblad C.; Löfblom J.; Uhlén M.; Gräslund T. An engineered affibody molecule with pH-dependent binding to FcRn mediates extended circulatory half-life of a fusion protein. Proc. Natl. Acad. Sci. U.S.A. 2014, 111 (48), 17110–17115. 10.1073/pnas.1417717111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kontermann R. E. Strategies for extended serum half-life of protein therapeutics. Curr. Opin. Biotechnol. 2011, 22 (6), 868–876. 10.1016/j.copbio.2011.06.012. [DOI] [PubMed] [Google Scholar]
  15. Rosenblum D.; Joshi N.; Tao W.; Karp J. M.; Peer D. Progress and challenges towards targeted delivery of cancer therapeutics. Nat. Commun. 2018, 9 (1), 1410 10.1038/s41467-018-03705-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Mitchell M. J.; Billingsley M. M.; Haley R. M.; Wechsler M. E.; Peppas N. A.; Langer R. Engineering precision nanoparticles for drug delivery. Nat. Rev. Drug Discovery 2021, 20 (2), 101–124. 10.1038/s41573-020-0090-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Zhang J.; Carter J.; Siu S.; W O’Neill J.; H Gates A.; Delaney J.; Mehlin C. Fusion partners as a tool for the expression of difficult proteins in mammalian cells. Curr. Pharm. Biotechnol. 2010, 11 (3), 241–245. 10.2174/138920110791111898. [DOI] [PubMed] [Google Scholar]
  18. Bryan L.; Clynes M.; Meleady P. The emerging role of cellular post-translational modifications in modulating growth and productivity of recombinant Chinese hamster ovary cells. Biotechnol. Adv. 2021, 49, 107757 10.1016/j.biotechadv.2021.107757. [DOI] [PubMed] [Google Scholar]
  19. Kunert R.; Reinhart D. Advances in recombinant antibody manufacturing. Appl. Microbiol. Biotechnol. 2016, 100 (8), 3451–3461. 10.1007/s00253-016-7388-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Fliedl L.; Grillari J.; Grillari-Voglauer R. Human cell lines for the production of recombinant proteins: on the horizon. New Biotechnol. 2015, 32 (6), 673–679. 10.1016/j.nbt.2014.11.005. [DOI] [PubMed] [Google Scholar]
  21. Ma B.; Guan X.; Li Y.; Shang S.; Li J.; Tan Z. Protein glycoengineering: an approach for improving protein properties. Front. Chem. 2020, 8, 622 10.3389/fchem.2020.00622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Zhou Q.; Qiu H. The mechanistic impact of N-glycosylation on stability, pharmacokinetics, and immunogenicity of therapeutic proteins. J. Pharm. Sci. 2019, 108 (4), 1366–1377. 10.1016/j.xphs.2018.11.029. [DOI] [PubMed] [Google Scholar]
  23. Mimura Y.; Katoh T.; Saldova R.; O’Flaherty R.; Izumi T.; Mimura-Kimura Y.; Utsunomiya T.; Mizukami Y.; Yamamoto K.; Matsumoto T.; Rudd P. M. Glycosylation engineering of therapeutic IgG antibodies: challenges for the safety, functionality and efficacy. Protein Cell 2018, 9 (1), 47–62. 10.1007/s13238-017-0433-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Boune S.; Hu P.; Epstein A. L.; Khawli L. A. Principles of N-linked glycosylation variations of IgG-based therapeutics: pharmacokinetic and functional considerations. Antibodies 2020, 9 (2), 22. 10.3390/antib9020022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Vattepu R.; Sneed S. L.; Anthony R. M. Sialylation as an important regulator of antibody function. Front. Immunol. 2022, 13, 818736 10.3389/fimmu.2022.818736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kissel T.; Derksen V. F. A. M.; Bentlage A. E. H.; Koeleman C.; Hafkenscheid L.; van der Woude D.; Wuhrer M.; Vidarsson G.; Toes R. E. M. N-linked Fc glycosylation is not required for IgG-B-cell receptor function in a GC-derived B-cell line. Nat. Commun. 2024, 15 (1), 393 10.1038/s41467-023-44468-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Liu R.; Oldham R. J.; Teal E.; Beers S. A.; Cragg M. S. Fc-engineering for modulated effector functions-improving antibodies for cancer treatment. Antibodies 2020, 9 (4), 64. 10.3390/antib9040064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Cambay F.; Raymond C.; Brochu D.; Gilbert M.; Tu T. M.; Cantin C.; Lenferink A.; Grail M.; Henry O.; De Crescenzo G.; Durocher Y. Impact of IgG1 N-glycosylation on their interaction with Fc gamma receptors. Curr. Res. Immunol. 2020, 1, 23–37. 10.1016/j.crimmu.2020.06.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Saunders K. O.Conceptual approaches to modulating antibody effector functions and circulation half-life.Front. Immunol. 2019,10 10.3389/fimmu.2019.01296. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Li W.; Zhu Z.; Chen W.; Feng Y.; Dimitrov D. S. Crystallizable fragment glycoengineering for therapeutic antibodies development. Front. Immunol. 2017, 8, 1554. 10.3389/fimmu.2017.01554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Higel F.; Seidl A.; Sörgel F.; Friess W. N-glycosylation heterogeneity and the influence on structure, function and pharmacokinetics of monoclonal antibodies and Fc fusion proteins. Eur. J. Pharm. Biopharm. 2016, 100, 94–100. 10.1016/j.ejpb.2016.01.005. [DOI] [PubMed] [Google Scholar]
  32. Göritzer K.; Turupcu A.; Maresch D.; Novak J.; Altmann F.; Oostenbrink C.; Obinger C.; Strasser R. Distinct Fcα receptor N-glycans modulate the binding affinity to immunoglobulin A (IgA) antibodies. J. Biol. Chem. 2019, 294 (38), 13995–14008. 10.1074/jbc.RA119.009954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Higel F.; Sandl T.; Kao C.-Y.; Pechinger N.; Sörgel F.; Friess W.; Wolschin F.; Seidl A. N-glycans of complex glycosylated biopharmaceuticals and their impact on protein clearance. Eur. J. Pharm. Biopharm. 2019, 139, 123–131. 10.1016/j.ejpb.2019.03.018. [DOI] [PubMed] [Google Scholar]
  34. Bongers J.; Devincentis J.; Fu J.; Huang P.; Kirkley D. H.; Leister K.; Liu P.; Ludwig R.; Rumney K.; Tao L.; et al. Characterization of glycosylation sites for a recombinant IgG1 monoclonal antibody and a CTLA4-Ig fusion protein by liquid chromatography-mass spectrometry peptide mapping. J. Chromatogr. A 2011, 1218 (45), 8140–8149. 10.1016/j.chroma.2011.08.089. [DOI] [PubMed] [Google Scholar]
  35. Houel S.; Hilliard M.; Yu Y. Q.; McLoughlin N.; Martin S. M.; Rudd P. M.; Williams J. P.; Chen W. N- and O-glycosylation analysis of etanercept using liquid chromatography and quadrupole time-of-flight mass spectrometry equipped with electron-transfer dissociation functionality. Anal. Chem. 2014, 86 (1), 576–584. 10.1021/ac402726h. [DOI] [PubMed] [Google Scholar]
  36. Keck R.; Nayak N.; Lerner L.; Raju S.; Ma S.; Schreitmueller T.; Chamow S.; Moorhouse K.; Kotts C.; Jones A. Characterization of a complex glycoprotein whose variable metabolic clearance in humans is dependent on terminal N-acetylglucosamine content. Biologicals 2008, 36 (1), 49–60. 10.1016/j.biologicals.2007.05.004. [DOI] [PubMed] [Google Scholar]
  37. Liu L. Antibody glycosylation and its impact on the pharmacokinetics and pharmacodynamics of monoclonal antibodies and Fc-fusion proteins. J. Pharm. Sci. 2015, 104 (6), 1866–1884. 10.1002/jps.24444. [DOI] [PubMed] [Google Scholar]
  38. Hashii N.; Suzuki J.; Hanamatsu H.; Furukawa J.-i.; Ishii-Watabe A. In-depth site-specific O-glycosylation analysis of therapeutic Fc-fusion protein by electron-transfer/higher-energy collisional dissociation mass spectrometry. Biologicals 2019, 58, 35–43. 10.1016/j.biologicals.2019.01.005. [DOI] [PubMed] [Google Scholar]
  39. Camperi J.; Dahotre S.; Guillarme D.; Stella C. Monitoring multiple quality attributes of a complex Fc-fusion protein during cell culture production processes by mD-LC-MS peptide mapping. Talanta 2022, 246, 123519 10.1016/j.talanta.2022.123519. [DOI] [PubMed] [Google Scholar]
  40. Shi R. L.; Xiao G.; Dillon T. M.; Ricci M. S.; Bondarenko P. V. Characterization of therapeutic proteins by cation exchange chromatography-mass spectrometry and top-down analysis. mAbs 2020, 12 (1), 1739825 10.1080/19420862.2020.1739825. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Vlasak J.; Bussat M. C.; Wang S.; Wagner-Rousset E.; Schaefer M.; Klinguer-Hamour C.; Kirchmeier M.; Corvaïa N.; Ionescu R.; Beck A. Identification and characterization of asparagine deamidation in the light chain CDR1 of a humanized IgG1 antibody. Anal. Biochem. 2009, 392 (2), 145–154. 10.1016/j.ab.2009.05.043. [DOI] [PubMed] [Google Scholar]
  42. Talebi M.; Nordborg A.; Gaspar A.; Lacher N. A.; Wang Q.; He X. Z.; Haddad P. R.; Hilder E. F. Charge heterogeneity profiling of monoclonal antibodies using low ionic strength ion-exchange chromatography and well-controlled pH gradients on monolithic columns. J. Chromatogr. A 2013, 1317, 148–154. 10.1016/j.chroma.2013.08.061. [DOI] [PubMed] [Google Scholar]
  43. Berkowitz S. A.; Engen J. R.; Mazzeo J. R.; Jones G. B. Analytical tools for characterizing biopharmaceuticals and the implications for biosimilars. Nat. Rev. Drug Discovery 2012, 11 (7), 527–540. 10.1038/nrd3746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wu J.; McElroy W.; Haitjema C.; Lück C.; Heger C. Platform methods to characterize the charge heterogeneity of three common protein therapeutics by imaged capillary isoelectric focusing. Methods Mol. Biol. 2021, 2261, 93–103. 10.1007/978-1-0716-1186-9_8. [DOI] [PubMed] [Google Scholar]
  45. Peters B. H.; Bautista J.; Slaney T. R.; Guo H.; Huang R. Y.; Krause M. E.; Zeng M.; Cheng J.; Chen Z. Enzymatic removal of sialic acid enables iCIEF stability monitoring of charge variants of a highly sialylated bispecific antibody. Electrophoresis 2022, 43 (9–10), 1059–1067. 10.1002/elps.202100259. [DOI] [PubMed] [Google Scholar]
  46. Farmerie L.; Rustandi R. R.; Loughney J. W.; Dawod M. Recent advances in isoelectric focusing of proteins and peptides. J. Chromatogr. A 2021, 1651, 462274 10.1016/j.chroma.2021.462274. [DOI] [PubMed] [Google Scholar]
  47. Kahle J.; Zagst H.; Wiesner R.; Wätzig H. Comparative charge-based separation study with various capillary electrophoresis (CE) modes and cation exchange chromatography (CEX) for the analysis of monoclonal antibodies. J. Pharm. Biomed. Anal. 2019, 174, 460–470. 10.1016/j.jpba.2019.05.058. [DOI] [PubMed] [Google Scholar]
  48. He X.; ElNaggar M.; Ostrowski M. A.; Guttman A.; Gentalen E.; Sperry J. Evaluation of an icIEF-MS system for comparable charge variant analysis of biotherapeutics with rapid peak identification by mass spectrometry. Electrophoresis 2022, 43 (11), 1215–1222. 10.1002/elps.202100295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Madren S.; McElroy W.; Schultz-Kuszak K.; Boumajny B.; Shu Y.; Sautter S.; Zhao H. C.; Schadock-Hewitt A.; Chumsae C.; Ball N.; et al. Global intercompany assessment of ICIEF platform comparability for the characterization of therapeutic proteins. Electrophoresis 2022, 43 (9–10), 1050–1058. 10.1002/elps.202100348. [DOI] [PubMed] [Google Scholar]
  50. Singh S. K.; Lee K. H. Characterization of monoclonal antibody glycan heterogeneity using hydrophilic interaction liquid chromatography-mass spectrometry. Front. Bioeng. Biotechnol. 2022, 9, 805788 10.3389/fbioe.2021.805788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. De Leoz M. L. A.; Duewer D. L.; Fung A.; Liu L.; Yau H. K.; Potter O.; Staples G. O.; Furuki K.; Frenkel R.; Hu Y.; et al. NIST interlaboratory study on glycosylation analysis of monoclonal antibodies: comparison of results from diverse analytical methods. Mol. Cell. Proteomics 2020, 19 (1), 11–30. 10.1074/mcp.RA119.001677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tan Q.; Guo Q.; Fang C.; Wang C.; Li B.; Wang H.; Li J.; Guo Y. Characterization and comparison of commercially available TNF receptor 2-Fc fusion protein products. mAbs 2012, 4 (6), 761–774. 10.4161/mabs.22276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. D’Atri V.; Nováková L.; Fekete S.; Stoll D.; Lauber M.; Beck A.; Guillarme D. Orthogonal middle-up approaches for characterization of the glycan heterogeneity of etanercept by hydrophilic interaction chromatography coupled to high-resolution mass spectrometry. Anal. Chem. 2019, 91 (1), 873–880. 10.1021/acs.analchem.8b03584. [DOI] [PubMed] [Google Scholar]
  54. Minami A.; Ishibashi S.; Ikeda K.; Ishitsubo E.; Hori T.; Tokiwa H.; Taguchi R.; Ieno D.; Otsubo T.; Matsuda Y.; et al. Catalytic preference of Salmonella typhimurium LT2 sialidase for N-acetylneuraminic acid residues over N-glycolylneuraminic acid residues. FEBS Open Bio 2013, 3 (1), 231–236. 10.1016/j.fob.2013.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Krapp S.; Mimura Y.; Jefferis R.; Huber R.; Sondermann P. Structural analysis of human IgG-Fc glycoforms reveals a correlation between glycosylation and structural integrity. J. Mol. Biol. 2003, 325 (5), 979–989. 10.1016/S0022-2836(02)01250-0. [DOI] [PubMed] [Google Scholar]
  56. Yamaguchi Y.; Nishimura M.; Nagano M.; Yagi H.; Sasakawa H.; Uchida K.; Shitara K.; Kato K. Glycoform-dependent conformational alteration of the Fc region of human immunoglobulin G1 as revealed by NMR spectroscopy. Biochim. Biophys. Acta, Gen. Subj. 2006, 1760 (4), 693–700. 10.1016/j.bbagen.2005.10.002. [DOI] [PubMed] [Google Scholar]
  57. Wen X.; Ou Y. C.; Bogatcheva G.; Thomas G.; Mahadevan-Jansen A.; Singh B.; Lin E. C.; Bardhan R. Probing metabolic alterations in breast cancer in response to molecular inhibitors with Raman spectroscopy and validated with mass spectrometry. Chem. Sci. 2020, 11 (36), 9863–9874. 10.1039/D0SC02221G. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Cutshaw G.; Uthaman S.; Hassan N.; Kothadiya S.; Wen X.; Bardhan R. The emerging role of Raman spectroscopy as an omics approach for metabolic profiling and biomarker detection toward precision medicine. Chem. Rev. 2023, 123 (13), 8297–8346. 10.1021/acs.chemrev.2c00897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Synan L.; Ghazvini S.; Uthaman S.; Cutshaw G.; Lee C. Y.; Waite J.; Wen X.; Sarkar S.; Lin E.; Santillan M.; et al. First trimester prediction of preterm birth in patient plasma with machine-learning-guided Raman spectroscopy and metabolomics. ACS Appl. Mater. Interfaces 2023, 15 (32), 38185–38200. 10.1021/acsami.3c04260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Cutshaw G.; Hassan N.; Uthaman S.; Wen X.; Singh B.; Sarkar A.; Bardhan R. Monitoring metabolic changes in response to chemotherapies in cancer with Raman spectroscopy and metabolomics. Anal. Chem. 2023, 95 (35), 13172–13184. 10.1021/acs.analchem.3c02073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Mathew A. P.; Cutshaw G.; Appel O.; Funk M.; Synan L.; Waite J.; Ghazvini S.; Wen X.; Sarkar S.; Santillan M.; et al. Diagnosis of pregnancy disorder in the first-trimester patient plasma with Raman spectroscopy and protein analysis. Bioeng. Transl. Med. 2024, 9, e10691 10.1002/btm2.10691. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Cutshaw G.; Joshi N.; Wen X.; Quam E.; Bogatcheva G.; Hassan N.; Uthaman S.; Waite J.; Sarkar S.; Singh B.; et al. Metabolic response to small molecule therapy in colorectal cancer tracked with Raman spectroscopy and metabolomics. Angew. Chem., Int. Ed. 2024, 63 (43), e202410919 10.1002/anie.202410919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Li M.; Yu C.; Wang W.; Wu G.; Wang L. Interlaboratory method validation of capillary electrophoresis sodium dodecyl sulfate (CE-SDS) methodology for analysis of mAbs. Electrophoresis 2021, 42 (19), 1900–1913. 10.1002/elps.202000396. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

pt4c00460_si_001.pdf (796.8KB, pdf)

Articles from ACS Pharmacology & Translational Science are provided here courtesy of American Chemical Society

RESOURCES