Automated Online-Sampling Multidimensional Liquid Chromatography with Feedback-Control Capability as a Framework for Real-Time Monitoring of mAb Critical Quality Attributes in Multiple Bioreactors.

Real-time monitoring of biopharmaceutical reactors is becoming increasingly important as the processes become more complex. During the continuous manufacturing of monoclonal antibodies (mAbs), the desired mAb product is continually created and collected over a 30 day process, where there can be changes in quality over that time. Liquid chromatography (LC) is the workhorse instrumentation capable of measuring mAb concentration as well as quality attributes such as aggregation, charge variants, oxidation, etc. However, traditional offline sampling is too infrequent to fully characterize bioprocesses, and the typical time from sample generation to data analysis and reporting can take weeks. To circumvent these limitations, an automated online sampling multidimensional workflow was developed to enable streamlined measurements of mAb concentration, aggregation, and charge variants. This analytical framework also facilitates automated data export for real-time analysis of up to six bioreactors, including feedback-controlling capability using readily available LC technology. This workflow increases the data points per bioreactor, improving the understanding of each experiment while also reducing the data turnaround time from weeks to hours. Examples of effective real-time analyses of mAb critical quality attributes are illustrated, showing substantial throughput improvements and accurate results while minimizing labor and manual intervention.

Cell Free Remodeling of Glycosylation of Antibodies.

The N-glycosylation profile of a monoclonal antibody (mAb) is a critical quality attribute in relation to its therapeutic application. The control of this profile during biomanufacture is difficult because of the multiple parameters that affect the glycosylation metabolism within the cell and the environment in which the cell is grown. One of the approaches that can be used to produce a preferred glycan profile or a single glycoform is through chemoenzymatic remodeling during the isolation of a mAb. Here we describe protocols that can be utilized to produce preferred glycoforms that include galactosylated, agalactosylated, or sialylated glycoforms following isolation of a mAb. Methods for analysis and assignment of structures of the samples following glycoengineering are also described. Chemoenzymatic modeling of mAb glycans has the potential for scale-up and to be introduced into biomanufacturing of mAbs with higher specific therapeutic activities.

HDX-MS and MD Simulations Provide Evidence for Stabilization of the IgG1-FcγRIa (CD64a) Immune Complex Through Intermolecular Glycoprotein Bonds.

Previous reports present different models for the stabilization of the Fc-FcγRI immune complex. Although accord exists on the importance of L235 in IgG1 and some hydrophobic contacts for complex stabilization, discord exists regarding the existence of stabilizing glycoprotein contacts between glycans of IgG1 and a conserved FG-loop (171MGKHRY176) of FcγRIa. Complexes formed from the FcγRIa receptor and IgG1s containing biantennary glycans with N-acetylglucosamine, galactose, and α2,6-N-acetylneuraminic terminations were measured by hydrogen-deuterium exchange mass spectrometry (HDX-MS), classified for dissimilarity with Welch's ANOVA and Games-Howell post hoc procedures, and modeled with molecular dynamics (MD) simulations. For each glycoform of the IgG1-FcγRIa complex peptic peptides of Fab, Fc and FcγRIa report distinct H/D exchange rates. MD simulations corroborate the differences in the peptide deuterium content through calculation of the percent of time that transient glycan-peptide bonds exist. These results indicate that stability of IgG1-FcγRIa complexes correlate with the presence of intermolecular glycoprotein interactions between the IgG1 glycans and the 173KHR175 motif within the FG-loop of FcγRIa. The results also indicate that intramolecular glycan-protein bonds stabilize the Fc region in isolated and complexed IgG1. Moreover, HDX-MS data evince that the Fab domain has glycan-protein binding contacts within the IgG1-FcγRI complex.

Strategic feeding of NS0 and CHO cell cultures to control glycan profiles and immunogenic epitopes of monoclonal antibodies.

The control of glycosylation profiles is essential to the consistent manufacture of therapeutic monoclonal antibodies that may be produced from a variety of cell lines including CHO and NS0. Of particular concern is the potential for generating non-human epitopes such as N-glycolylneuraminic acid (Neu5Gc) and Galα1-3 Gal that may be immunogenic. We have looked at the effects of a commonly used media supplements of manganese, galactose and uridine (MGU) on Mab production from CHO and NS0 cells in enhancing galactosylation and sialylation as well as the generation of these non-human glycan epitopes. In the absence of the MGU supplement, the humanized IgG1 antibody (Hu1D10) produced from NS0 cells showed a low level of mono- and di-sialylated structures (SI:0.09) of which 75 % of sialic acid was Neu5Gc. The chimeric human-llama Mab (EG2-hFc) produced from CHO cells showed an equally low level of sialylation (SI: 0.12) but the Neu5Gc content of sialic acid was negligible (<3%). Combinations of the MGU supplements added to the production cultures resulted in a substantial increase in the galactosylation of both Mabs (up to GI:0.78 in Hu1D10 and 0.81 in EG2-hFc). However, the effects on sialylation differed between the two Mabs. We observed a slight increase in sialylation of the EG2-hFc Mab by a combination of MG but it appeared that one of the components (uridine) was inhibitory to sialylation. On the other hand, MG or MGU increased sialylation of Hu1D10 substantially (SI:0.72) with an increase that could be attributed predominantly to the formation of Neu5Ac rather than Neu5Gc. The increased level of galactosylation observed with MG or MGU was attributed to an activation of the galactosyl transferase enzymes through enhanced intracellular levels of UDP-Gal and the availability of Mn2+ as an enzymic co-factor. However, this effect not only increased the desirable beta 1-4 Gal linkage to GlcNAc but unfortunately in NS0 cells increased the formation of Galα1-3 Gal which was shown to increase x3 in the presence of combinations of the MGU supplements. Supplementation of media with fetal bovine serum (FBS) increased the availability of free Neu5Ac which resulted in a significant increase in the sialylation of Hu1D10 from NS0 cells. This also resulted in a significant decrease in the proportion of Neu5Gc in the measured sialic acid from the Mab.

Solid-Phase Enzymatic Remodeling Produces High Yields of Single Glycoform Antibodies.

Antibodies are synthesized in mammalian cell culture as heterogeneous mixtures of glycoforms. Production of single glycoforms remains a challenge despite their value as therapeutics. The authors report a method of sequential enzymatic-based changes to antibodies while immobilized on an affinity column. Various antibodies (monoclonal and polyclonal) are isolated on Protein A or G columns and their glycans modified by sequential addition of enzymes for a desired transformation. Galactosylated antibodies (>90% yield) are produced by a one stage reaction process with sialidase to remove any sialic acid residues and addition of galactose with galactosyltransferase and UDP-Gal. Sialylated antibodies (>90%) are produced by a 2 stage conversion involving α(2,3) sialidase and galactosyltransferase followed by treatment with α(2,6) sialyltransferase in the presence of CMP-NANA. By this method, >90% of a disialylated human-llama antibody (EG2-hFc) and equimolar quantities of monosialylated and disialylated forms of human antibodies (αIL8-hFc and human polyclonal) are produced. Such high levels of sialylation are very difficult to obtain by typical cell culture methods. This method of transformation while the antibody is held on a solid-phase column is superior to previous methods because it allows a series of enzymatic steps without the need for intermediate purification. This is an efficient and rapid method to generate therapeutic antibodies with predefined glycosylation profiles. This should also assist in investigating the structure-function relationship of antibody glycans to find the desired glycosylation profile for high functional activity. With further optimization the method can be used to modify antibodies in large-scale manufacturing.

A systematic study of glycopeptide esterification for the semi-quantitative determination of sialylation in antibodies.

RATIONALE: In the expression of recombinant proteins, an important parameter to control or influence is their level of sialylation. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometric (MS) methods tend to either underestimate (positive mode) or overestimate (negative mode) the content of sialylated vs. neutral glycans in glycoproteins. Esterification methods have been developed for free sialylated glycans and sialylated Asn-glycans, allowing these acidic groups to ionize with the same efficiency as neutral sugars. METHODS: Here we describe a method which modifies glycopeptides by esterification. This simple procedure is applied to glycopeptides isolated from tryptic digests of monoclonal antibodies (mAbs), some highly sialylated. To better understand the effect of esterification on the peptide backbone, synthetic EEQYNSTYR was esterified and studied by tandem mass spectrometry (MS/MS). Acetamidation of EEQYNSTYR was also studied as some mAb samples had been overalkylated prior to tryptic digestion. RESULTS: As a general trend, ethyl-esterification or lactonization is observed for each sialic acid on glycoforms of EEQYNSTYR (the N-glycosylated tryptic peptide of IgG Fc), depending on the branching position of the sialic acid (α2,3 or α2,6). Esterification also affects the carboxyl groups in the peptide, including the C-terminal COOH. CONCLUSIONS: For antibody analysis, MALDI-MS ion abundances give a better semi-quantitative estimate of sialylation levels for esterified than for unreacted glycopeptides. The method is simple to use and helps to differentiate the branching patterns of sialic acids in antibodies.

Low glucose depletes glycan precursors, reduces site occupancy and galactosylation of a monoclonal antibody in CHO cell culture.

Controlled feeding of glucose has been employed previously to enhance the productivity of recombinant glycoproteins but there is a concern that low concentrations of glucose could limit the synthesis of precursors of glycosylation. Here we investigate the effect of glucose depletion on the metabolism, productivity and glycosylation of a chimeric human-llama monoclonal antibody secreted by CHO cells. The cells were inoculated into media containing varying concentrations of glucose. Glucose depletion occurred in cultures with an initial glucose ≤5.5 mM and seeded at low density (2.5 × 10(5) cells/mL) or at high cell inoculum (≥2.5 × 10(6) cells/mL) at higher glucose concentration (up to 25 mM). Glucose-depleted cultures produced non-glycosylated Mabs (up to 51%), lower galactosylation index (GI <0.43) and decreased sialylation (by 85%) as measured by mass spectrometry and HPLC. At low glucose a reduced intracellular pool of nucleotides (0.03-0.23 fmoles/cell) was measured as well as a low adenylate energy charge (<0.57). Low glucose also reduced GDP-sugars (by 77%) and UDP-hexosamines (by 90%). The data indicate that under glucose deprivation, low levels of intracellular nucleotides and nucleotide sugars reduced the availability of the immediate precursors of glycosylation. These results are important when applied to the design of fed-batch cultures.

Mathematical model of the rate-limiting steps for retrovirus-mediated gene transfer into mammalian cells.

A quantitative understanding of the process of retrovirus-mediated gene transfer into mammalian cells should assist the design and optimization of transduction protocols. We present a mathematical model of the process that incorporates the essential rate-limiting transduction steps including diffusion, convection and decay of viral vectors, their binding at the cell surface and entry into the cell cytoplasm, reverse transcription of uncoated RNA to form DNA intermediates, transport of the latter through the cytosol to the cell nucleus and, finally, nuclear import and integration of the delivered DNA into the target cell genome. Cell and virus population balances are used to account for the kinetics of multiple vector infections which influence the transduction efficiency and govern the integrated copy number. The mathematical model is validated using gibbon ape leukemia virus envelope pseudotyped retroviral vectors and K562 target cells. Viral intermediate complexes derived from the internalized retroviral vectors are found to remain stable inside the K562 cells and the cytoplasmic trafficking time is consistent with the time scale for retrovirus uncoating, reverse transcription and transport to the cell nucleus. The model predictions of transduction efficiency and integrated copy number agree well with experimental data for both static (i.e., standard gravity) and centrifugation-based gene transfer protocols. The formulation of the model can also be applied to transduction protocols involving lenti- or foamy-viruses and so should prove to be useful for the optimization of several types of gene transfer processes.