A slice of the plot at a fixed lactate concentration of 0.4 mM along the glucose axis yields the curve shown in Fig. 2B. On this constant lactate plane the ��switch-up�� and ��switch-down�� points can be seen. With increasing levels of lactate, the ��switch-up�� glucose concentration shifts to higher glucose levels. At very high lactate concentrations, the ��switch-up�� concentration will move to extremely high glucose concentrations that are not even seen in culture as they would represent near lethal high osmolality to cells. Even in this region of very high lactate, switching down from high state to low state is feasible if glucose concentration falls to very low levels. We performed a set of transient simulations to demonstrate different scenarios where cultures may have different metabolic fates. Cells are initially at a position on the high flux plane of high pAKT and move along a path with decreasing glucose concentration and increasing lactate concentration. As the growth rate decreases, the culture progression is depicted by a line in a surface plot corresponding to pAKT activation of 0.17. In the first case, the path encounters the receded section of the high flux plane DAPT leading to a switch down from the high flux plane to the low flux plane. In the other case, while moving along the path glucose concentration is increased due to glucose feeding in fed-batch culture. The feeding causes the path to shift and eventually leads to a trajectory that does not encounter the receded plane thus confining the cells to the high flux plane. In the first case, upon switching to a low flux state, an addition of the same amount of glucose to the culture will not cause the culture to switch back to a high flux state. Such a AMN107 difference in the timing of glucose feeding is not uncommon in industrial manufacturing, as observed in the archived manufacturing data and in laboratory practices. Even cultures under the same conditions are not exact replicas. Feeding of nutrient and glucose to a fed-batch culture typically is either prescribed as fixed time point or responding to a range of the controlled variables. A small difference in the timing of glucose feeding may cause cultures with very similar metabolic behavior to diverge to different outcomes. A recombinant CHO cell line was grown in four fed-batch cultures with initial sodium lactate concentrations of 0, 10, 25 and 40 mM. Osmolarity at inoculation was adjusted to 300 mOsm/ kg in all four conditions using appropriate amounts of sodium chloride. The concentration profiles of lactate and growth curves are shown along with the specific lactate production rates. Under all four different lactate concentrations, cells were in a high flux state during the growth phase.