Comprehensive characterization of glutamine synthetase-mediated selection for the establishment of recombinant CHO cells producing monoclonal antibodies

Abstract

To characterize a glutamine synthetase (GS)-based selection system, monoclonal antibody (mAb) producing recombinant CHO cell clones were generated by a single round of selection at various methionine sulfoximine (MSX) concentrations (0, 25, and 50 μM) using two different host cell lines (CHO-K1 and GS-knockout CHO). Regardless of the host cell lines used, the clones selected at 50 μM MSX had the lowest average specific growth rate and the highest average specific production rates of toxic metabolic wastes, lactate and ammonia. Unlike CHO-K1, high producing clones could be generated in the absence of MSX using GS-knockout CHO with an improved selection stringency. Regardless of the host cell lines used, the clones selected at various MSX concentrations showed no significant difference in the GS, heavy chain, and light chain gene copies (P > 0.05). Furthermore, there was no correlation between the specific mAb productivity and these three gene copies (R² ≤ 0.012). Taken together, GS-mediated gene amplification does not occur in a single round of selection at a MSX concentration up to 50 μM. The use of the GS-knockout CHO host cell line facilitates the rapid generation of high producing clones with reduced production of lactate and ammonia in the absence of MSX.

Figure 1

A schematic diagram of the process for mAb producing clone generation and long-term culture for testing the production stability.

Figure 2

(A) Selection stringency and (B) mAb concentration of the culture supernatant in the wells with a cell pool at various MSX concentrations. Cell pools were derived either from CHO-K1 (light gray) or from GS KO (dark gray). The numbers in the parenthesis above the bar graphs indicate the number of wells with a cell pool. The box plot shows the mAb concentration of the culture supernatant in the wells with a cell pool. The dotted line in the box indicates the mean value, and the solid line in the box indicates the median value.

Figure 3

Correlation between (A) the relative mRNA level of GS and q_mAb, (B) relative mRNA level of HC and q_mAb, (C) relative mRNA level of LC and q_mAb, (D) relative gene copy number of GS and q_mAb, (E) relative gene copy number of HC and q_mAb, and (F) relative gene copy number of LC and q_mAb. The closed circles indicate the clones derived from CHO-K1, and the open circles indicate the clones derived from GS KO.

Figure 4

Specific consumption or production rate of (A) glucose, (B) lactate, (C) ammonia, (D) arginine, (E) glutamate, and (F) glycine. Clones from different selection conditions (0, 25 and 50 μM of MSX) are shown in the box plots. The dotted line indicates the mean value, and the solid line indicates the median value.

Figure 5

Feasibility test of GS-mediated gene amplification by stepwise selection for resistance to increasing concentrations of MSX. Viability change during the selection at (A) 25 μM MSX and at (B) 50 μM MSX. K1-0 1–1 (closed circles), K1-0 1–2 (open circles), K1-0 2-1 (closed triangles), K1-0 2–3 (open triangles), KO-0 1-1 (closed squares), KO-0 1–3 (open squares), KO-0 2–1 (closed diamonds) and KO-0 2–2 (open diamonds). Change in (C) q_mAb, (D) relative gene copy number of GS, (E) relative gene copy number of HC, and (F) relative gene copy number of LC after further selection at 25 or 50 μM MSX. The gene copy number and mRNA level are relative to the values of the clones before further MSX selection. (a) K1-0 1–1, (b) K1-0 1–2, (c) K1-0 2-1, (d) K1-0 2–3, (e) KO-0 1–1, (f) KO-0 1–3, (g) KO-0 2–1 and (h) KO-0 2–2. Before selection (black), after selection at 25 μM MSX (gray) and after selection at 50 μM MSX (dark gray). Error bars represent the standard deviation determined by duplicate experiments. Asterisks (*) indicate a significant difference at P < 0.05.

Figure 6

The μ of the mAb producing clones during long-term cultures with MSX (closed circle) or without MSX (empty circle). (A) K1-0 1–1, (B) K1-0 1–2, (C) K1-0 2-1, (D) K1-0 2–3, (E) KO-0 1–1, (F) KO-0 1–3, (G) KO-0 2-1, (H) KO-0 2–2, (I) K1-25 1–1, (J) K1-25 1–2, (K) K1-25 2-1, (L) K1-25 2–2, (M) KO-25 1–1, (N) KO-25 1–3, (O) KO-25 2-1, (P) KO-25 2–2, (Q) K1-50 1–2, (R) K1-50 1–3, (S) K1-50 2–1, (T) K1-50 2–2, (U) KO-50 1–1, (V) KO-50 1–3, (W) KO-50 2-1, and (X) KO-50 2–2.

Figure 7

The q_mAb of the mAb producing clones during long-term cultures with MSX (closed circle) or without MSX (empty circle). (A) K1-0 1–1, (B) K1-0 1–2, (C) K1-0 2-1, (D) K1-0 2–3, (E) KO-0 1–1, (F) KO-0 1–3, (G) KO-0 2-1, (H) KO-0 2–2, (I) K1-25 1–1, (J) K1-25 1–2, (K) K1-25 2-1, (L) K1-25 2–2, (M) KO-25 1–1, (N) KO-25 1–3, (O) KO-25 2-1, (P) KO-25 2–2, (Q) K1-50 1–2, (R) K1-50 1–3, (S) K1-50 2–1, (T) K1-50 2–2, (U) KO-50 1–1, (V) KO-50 1–3, (W) KO-50 2–1, and (X) KO-50 2–2.

Figure 8

Correlation between (A) % change in the GS gene copies and % change in the q_mAb (all clones), (B) % change in the HC gene copies and % change in the q_mAb (all clones), (C) % change in the LC gene copies and % change in the q_mAb (all clones), (D) % change in the GS gene copies and % change in the q_mAb (selected clones), (E) % change in the HC gene copies and % change in the q_mAb (selected clones) and (F) % change in the LC gene copies and % change in the q_mAb (selected clones) during long-term cultures with MSX (closed circle) or without MSX (empty circle).