Using molecular markers to characterize productivity in Chinese hamster ovary cell lines

Abstract

Selection of high producing cell lines to produce maximum product concentration is a challenging and time consuming task for the biopharmaceutical industry. The identification of early markers to predict high productivity will significantly reduce the time required for new cell line development. This study identifies candidate determinants of high productivity by profiling the molecular and morphological characteristics of a panel of six Chinese Hamster Ovary (CHO) stable cell lines with varying recombinant monoclonal antibody productivity levels ranging between 2 and 50 pg/cell/day. We examined the correlation between molecular parameters and specific productivity (qp ) throughout the growth phase of batch cultures. Results were statistically analyzed using Pearson correlation coefficient. Our study revealed that, overall, heavy chain (HC) mRNA had the strongest association with qp followed by light chain (LC) mRNA, HC intracellular polypeptides, and intracellular antibodies. A significant correlation was also obtained between qp and the following molecular markers: growth rate, biomass, endoplasmic reticulum, and LC polypeptides. However, in these cases, the correlation was not observed at all-time points throughout the growth phase. The repeated sampling throughout culture duration had enabled more accurate predictions of productivity in comparison to performing a single-point measurement. Since the correlation varied from day to day during batch cultivation, single-point measurement was of limited use in making a reliable prediction.

Figure 1. Cell growth, viability and MAb concentration.

Growth and viability curves (A) of GS-CHO cell lines analyzed by flow cytometry and (B) MAb concentration batch culture.

Figure 2. Specific productivities accounting for physiological characteristics.

Specific productivities calculated based on cell number (A; P-value = 0.053), total protein content (B; P-value<0.001), biomass (C; P-value<0.001), and cell volume (D; P-value<0.001). The data were normalized based on day two culture. Error bars symbolize the standard deviation as calculated between three biological replicates. P-values were calculated based on ANOVA at 95% confidence intervals.

Figure 3. Gene copy numbers of heavy chain (HC) and light chain (LC).

qPCR analysis of HC and LC gene copy numbers (GCNs) using SYBR Green chemistry at growth phase (day 3). Error bars symbolize the standard deviation calculated from three biological replicates. Student t-test analysis showed that the difference between the levels of HCs GCNs in the six cell lines was found to be insignificant (P-value = 0.05).

Figure 4. The relationship between mRNA and polypetides in 6 CHO cell lines.

HC and LC mRNA levels (A; P-values<0.001), intracellular HC and LC polypeptides levels (B; P-values<0.001), mRNA degradation constant, -k_D (D; P-values≤0.001) and intracellular MAb (E; P-value<0.001) at growth phase (day 3) of batch cultures. The correlation of mRNA level and polypeptides (C); (i) HC and (ii) LC. Error bars symbolize the standard deviation calculated from three biological replicates. P-values were calculated based on Student t-test analysis at 95% confidence intervals.

Figure 5. Flow cytometric analysis of endoplasmic reticulum (ER) and Golgi apparatus.

ER (A; P-value<0.001) and Golgi apparatus (B; P-value<0.001) levels at growth phase (day 3) normalized based on the cell size as measured by flow cytometry analysis. Error bars depict standard error calculated from three biological replicates. P-values were calculated based on Student t-test analysis at 95% confidence intervals.

Figure 6. Data distribution of molecular markers of six GS-C cell lines.

Data distribution collected throughout the growth phase (day 1 to day 5) of triplicate batch cultures in the six GS-CHO cell lines (A). P-values were calculated using one-pair ANOVA analysis. Boxplot edges represent quartiles of 25^th and 75^th. The median, where 50 percent of data are distributed is represented by the line between the 25^th and 75^th quartiles Boxplot whiskers indicate maximum and minimum values in the distributions. Dashed line represents the whole population median (B).