[Literature Review] The Impact of Amino Acid and Glucose Metabolism in Fed-Batch Cultures of CHO Cells on Antibody Production and Glycosylation

[Literature Review] The Impact of Amino Acid and Glucose Metabolism in Fed-Batch Cultures of CHO Cells on Antibody Production and Glycosylation

2023-07-28

Fed-batch cultivation of Chinese hamster ovary (CHO) cells is the most commonly used IgG production process in the biopharmaceutical industry. The consumption of amino acids and glucose, cell growth, metabolism, antibody titer, and glycosylation patterns have consistently been key issues in upstream process optimization, particularly during medium optimization. Understanding the interrelationships among these factors can provide valuable insights for achieving higher immunoglobulin G (IgG) titers and better controlling product quality related to glycosylation. In this study, two chemically defined media and different fed-batch cultivation strategies were investigated using two IgG-producing cell lines.

Experimental method

Cell lines: CHO cell lines 1030 and 4384 (DHFR-);

Culture conditions: 5% CO2, 37℃, 200 rpm;

Passage method before vaccination:

Before inoculating the fed-batch culture, seed cells at a density of 5 × 10⁵ cells/mL for a 2-day subculture or 3 × 10⁵ cells/mL for a 3-day subculture.

Fedbatch: Cells were grown in 500-mL shake flasks with an initial culture volume of 70 mL, at a temperature of 37℃, 5% CO2, and a shaking speed of 200 rpm. On day 5, the temperature was reduced from 37℃ to 33.5℃. Harvesting was performed on day 14 if cell viability dropped below 60%.

Different conditions

①A+FA8 (basic medium A, feed supplement FA, inoculum density 8×10⁵ cells/mL): Starting from Day 2, add FA once daily at a rate of 3.3% of the initial culture volume. On Days 5 and 7, adjust the glucose concentration to 8 g/L; on Days 9 and 11, adjust the glucose concentration to 10 g/L.

②The feeding strategy for A+FA4 (basic medium A, feed FA, inoculation density of 4×10⁵ cells/mL) is consistent with that for A+FA8; however, glucose does not need to be added during this process.

③ B+FB4 (basic medium B, feed FB; inoculated at a cell density of 4 × 10⁵ cells/mL): On days D2, D5, D7, D9, and D11, supplement the culture with FB (10% of the initial culture volume). For the 1030 cell line, adjust the glucose concentration to 6 g/L on day D5 and to 9 g/L on days D9 and D11. For the 4384 cell line, adjust the glucose concentration to 10 g/L on day D9; all other conditions are consistent with those for the 1030 cell line.

Detection method

1. Live Cell Density (VCD): Vi-cell XR (Beckman Coulter);

2. Glucose, glutamine, lactate, ammonium, glutamate, pH, and osmotic pressure:

Bioprofile 100plus (Nova BioMedical);

3. IgG Titer: Determined by biolayer interferometry using the Octet QK384 equipped with a protein A biosensor (ForteBio);

4. Amino acid analysis: Performed using cation-exchange chromatography, followed by post-column derivatization and fluorescence detection;

5. IgG concentration after purification: NanoDrop ND-1000 (Thermo Scientific);

6. Complete IgG mass analysis: (LC–MS) Dionex Ultimate 3000 RSLC system;

7. IgG Glycan Profiling: N-glycans in IgG were digested, labeled, and removed using the GlykoPrep1 InstantABTM Kit (Prozyme). HPLC analysis was performed on a Dionex Ultimate 3000RSLC system equipped with an Ultimate 3000 RS fluorescence detector (Dionex) and an ACQUITY UPLC BEH Glycan 1.7 mm, 2.1 × 150 mm column (Waters).

8. Total protein concentration in the supernatant: Pierce BCA Protein Assay Kit (Thermo Scientific);

Experimental results

Figure 1 shows the effects of different upstream culture processes on growth, metabolism, and IgG production in two cell lines, 1030 and 4384. The viability of A+FA4 and A+FA8 cells was slightly lower than that of B+FB4 cells. Compared to B+FB4, A+FA8 resulted in faster cell growth. Specifically, as shown in Figure 1A, when the 1030 and 4384 cell lines were cultured under B+FB4 conditions, the peak viable cell concentrations at Day 11 were approximately 5 and 15 × 10^6 cells/mL, respectively. In contrast, the A+FA8 condition led to peak viable cell densities of approximately 15 and 20 × 10^6 cells/mL for the 1030 and 4384 cell lines, respectively, by Day 9. Interestingly, compared to B+FB4, A+FA4 did not show any significant improvement in cell growth. Moreover, compared to B+FB4, both A+FA4 and A+FA8 exhibited earlier and more rapid declines in cell viability (Figure 1A). For both the 1030 and 4384 cell lines, due to the high cell density and high specific IgG productivity, the A+FA8 condition resulted in significantly higher IgG yields than other culture conditions (Figure 1B).

Figure 1

Due to differences in feeding strategies, the glucose concentration in A+FA8 exhibited greater fluctuations compared to B+FB4. Moreover, it was strikingly evident that glucose accumulated substantially during the cultivation of A+FA4 (Figure 1C). After temperature reduction, the accumulation of glutamine in A+FA4 reached 3.5 mM at harvest. Additionally, whereas the glutamine concentration in B+FB4 was approximately 0–1 mM, both cell lines showed glutamine concentrations of about 1–2 mM prior to feeding in the A+FA8 culture (Figure 1D). Throughout the entire process, the glutamate concentration in A+FA8 (approximately 3–6 mM) generally remained higher than that in B+FB4 (about 2–4 mM) before feeding (Figure 1E). It was observed that from Day 7 until harvest, the average specific consumption rates of glucose and glutamine were typically higher in A+FA8 than in B+FB4 (Figures 1D and E). In terms of byproduct formation, lactate accumulation was consistently higher in A+FA8 than in B+FB4 for both cell lines (Figure 1F). In A+FA4, lactate accumulation and its average specific production rate were particularly high, likely attributable to the elevated glucose concentration under these culture conditions (Figure 1F). From Day 5 to Day 11, the accumulation of NH4+ in A+FA8 was lower than that in B+FB4 (Figure 1G). However, in A+FA4, the accumulation of NH4+ significantly increased from 2 mM to 10 mM during the cultivation period, indicating a notably high average specific production rate of NH4+ (Figure 1G).

Figure 2 shows the differences in amino acid and glucose concentrations as well as specific consumption rates. Generally, from D5 to D13, the amino acid concentrations in the 4384 cell line were more stable in A+FA8 than in B+FB4 (Figure 2A). Specifically, in B+FB4, amino acids such as Asp, Ser, Gly, Val, Met, Ile, Leu, Lys, Arg, and all aromatic amino acids (Tyr, Phe, His, and Trp) showed a marked decline starting from D11. However, from D5 to D13, Ala accumulated in both A+FA8 and B+FB4. In B+FB4, with the exception of Glu and Gln at D5, D9, D11, and D13, and Cys, Val, and aromatic amino acids (Tyr, Phe, Trp) at D11 and D13, most amino acids exhibited higher concentrations at D5, D9, D11, and D13. Interestingly, the specific consumption rates of certain amino acids largely depended on the concentrations of the corresponding amino acids in the culture medium. Indeed, in B+FB4, from D5 to D13, the specific consumption rates of most amino acids (with the exception of Glu, Gln, and Ala) were higher. By contrast, from D5 to D13, the specific glucose consumption rate was consistently higher in A+FA8 (Figures 2B and 2C). We also observed that in B+FB4, the specific amino acid consumption rates from D9 to D11 were higher than those from D5 to D9 and from D11 to D13. Conversely, in A+FA8, the specific consumption rates of most amino acids (except Glu, Gln, Ala, Cys, Tyr, HYP, and ABU) from D9 to D11 were lower than those from D5 to D9 and from D11 to D13. Furthermore, in B+FB4, the specific glucose consumption rate during cell culture decreased after temperature changes (D5), whereas it remained stable in A+FA8.

Figure 2

As shown in Figure 3, A+FA8 and A+FA4 resulted in lower levels of relatively immature glycan combinations (Man5/Man5, Man5/A1G0F, and A1G0F/G0F), but higher levels of relatively mature glycan combinations (G0F/G1F, G1F/G1F, or G0F/G2F and G1F/G2F) compared to B+FB4. In particular, the Man5/Man5 combination from A+FA8 was below the detection limit of the analytical method used. It was also observed that, in both A+FA8 and B+FB4, the 4384 cell line produced more relatively immature glycan combinations than the 1030 cell line. Under the same culture conditions, the 4384 cell line produced higher levels of Man5, G1F, G2F, and G2FS1 than the 1030 cell line (Figure 4). However, regardless of culture conditions, the 1030 cell line consistently produced more Man7 than the 4384 cell line. Furthermore, the A+FA8 cultures produced less Man5 and A1G0F but more G1F and G2F.

Figure 3

Figure 4

Cell pellets from A+FA8 and B+FB4 D2 and D11 were collected for Western blot analysis. The expression levels of GnT1 and the UDP-GlcNAc transporter were analyzed to gain insight into the first step of glycan chain elongation in the Golgi apparatus, during which UDP-Glc is transported into the organelle and subsequently added onto the Man5 glycan structure. As shown in Figure 5, the expression of GnT1 at both time points appeared to be independent of the culture conditions used for the 1030 cell line. However, distinct differences in the expression of the UDP-GlcNAc transporter were observed in the 1030 cell line. In the 1030 cell line, A+FA8 exhibited baseline levels of UDP-GlcNAc transporter expression at D2, which then decreased by 50% by D11. By contrast, the 1030 cell line B+FB4 showed slightly lower levels at D2, but these levels significantly increased (by a factor of two) by D11. Regardless of the culture conditions, the expression profiles of GnT1 and the UDP-GlcNAc transporter in the 4384 cell line were similar. Additionally, it was noted that at both time points, the GnT1 levels in B+FB4 were lower than those in A+FA8.

Figure 5

Experimental conclusion

When transitioning from the B+FB4 condition in the 1030 and 4384 cell lines to the A+FA8 condition, due to a more balanced amino acid concentration in the culture medium, the glucose consumption rate increases, leading to faster cell growth, higher IVC and IgG yields, and more mature glycosylation patterns.

When comparing B+FB4 with A+FA8, the higher Ala content in the positive sample, along with higher Ile and Leu consumption rates and elevated Glu levels as well as increased consumption rates of Pro, His, and Arg, indicate that the cells exhibit higher transaminase activity, thereby allowing more intermediates to enter the TCA cycle in the B+FB4 culture. The high accumulation of NH4+ is thought to inhibit cell growth; compared to A+FA8, this could be one of the significant drawbacks of B+FB4. In extreme cases, in A+FA4, the imbalanced nutrient supply due to cellular activity leads to extremely high concentrations of lactate and NH4+, which in turn suppress cell growth.

Ser and Leu are the two amino acids with the highest consumption rates in B+FB4 medium. The rapid consumption of Ser and Leu can be attributed to their high abundance in the amino acid composition of the antibodies produced. It has been reported that Arg has a negative impact on mAb production in CHO DG44 cell cultures, and its concentration in A+FA8 is lower than that in B+FB4. In B+FB4, more amino acids are consumed via catabolism compared to A+FA8, thereby replenishing the energy derived from glucose. Consequently, fewer raw materials may be available for IgG production, potentially negatively affecting specific productivity in B+FB4.

In the 4384 cell line, the accumulation of Man5 may be attributable to bottlenecks in the availability of UDP-GlcNAc and the expression of GnT1. β-1,2-N-acetylglucosaminyltransferase I (GnT1) is one of the key enzymes involved in the N-linked glycan processing of mammalian cell glycoproteins; it catalyzes the conversion of mannose-type glycans into hybrid-type glycans. The 4384 cell line was found to have extremely low intracellular concentrations of UDP-GlcNAc. As shown in Figure 5, this cell line exhibited lower levels of GnT1 expression at both D2 and D11 time points in the B+FB4 condition. Compared with 4384 A+FA8, the 4384 B+FB4 cells accumulated substantial amounts of Man5. The lower intracellular UDP-GlcNAc concentration in the former may result from reduced biosynthesis of this nucleotide sugar, whereas the lower UDP-GlcNAc concentration observed in 4384 A+FA8 could be due to its higher consumption. The limited biosynthesis of UDP-GlcNAc in B+FB4 was confirmed by the low availability of extracellular glutamine under these culture conditions (Figure 1D). As widely reported, the limited availability of this nutrient restricts the biosynthesis of UDP-GlcNAc.

However, cell line 1030 behaved quite differently. In B+FB4 medium, the 1030 cell line produced 81.1% ± 3.5% more Man5 than in A+FA8 medium—a result that sharply contrasts with the intracellular UDP-GlcNAc concentration data (Figure 3). Therefore, the UDP-GlcNAc concentration may not be the limiting factor for glycosylation processing in this cell line. If the rate of nucleotide sugar consumption falls below its production rate, it could lead to an accumulation of high levels of UDP-GlcNAc and a corresponding increase in Man5 secretion. The reduced UDP-GlcNAc consumption might be attributed to insufficient expression of either GnTI or the UDP-GlcNAc transporter. However, Figure 5 shows that GnTI was expressed at similar levels under both culture conditions in cell line 1030, suggesting that enzyme expression is unlikely to be the limiting factor. Based on this analysis, the only remaining plausible explanation for the reduced UDP-GlcNAc consumption involves the activity of either GnTI or the UDP-GlcNAc transporter. Furthermore, it was found that the high accumulation of extracellular ammonia in this cell line correlated with elevated Man5 secretion. Previous reports have suggested that ammonia accumulation might raise the pH within the Golgi apparatus, thereby inhibiting the activity of glycosyltransferases. Other studies have also indicated that an increase in Golgi pH can cause mislocalization of glycosyltransferases.

The experimental results indicate that the balance between glucose and amino acid concentrations in the culture medium is crucial for cell growth, IgG titer, and N-glycosylation. Imbalances in nutrient supply due to cellular activity can lead to the accumulation of byproducts such as NH4+ and lactate, which in turn inhibit cell growth. The levels of Leu and Arg—amino acids closely associated with cell growth and IgG productivity—must be carefully controlled in the culture. The amino acids with the highest consumption rates are also those most abundantly present in the produced IgG; therefore, their availability needs to be sufficient throughout the cultivation process. In some cases, the presence of Man5 oligosaccharides may be linked to the limitation of UDP-GlcNAc biosynthesis caused by insufficient extracellular Gln. However, under different culture conditions, high Man5 levels might also result from low GnTI activity and reduced UDP-GlcNAc transporter activity, possibly attributable to elevated NH4+ concentrations in cell cultures. Case-by-case analyses are essential for understanding how medium composition and process optimization affect glycosylation. The impact of medium and process optimization on glycosylation should be interpreted in a context-specific manner, as it arises from the interplay among protein processing rates, cellular metabolism, and the expression and activity of Golgi-resident proteins. Given the diversity of mechanisms involved, systems biology approaches could play a pivotal role in the future by helping us better understand protein glycosylation and further facilitating efforts toward glycoform control.