Automation of Neuronal Differentiation of the Pluripotent Ntera-2 Cells

Pluripotent Stem cells (PSCs) are defined by their capacity for self-renewal and differentiation, making them an attractive platform for modelling human tissue formation in vitro. However, generating scalable, uniform differentiated cell types from PSCs can be challenging due to the limited control over the cell environment. While growth factors can direct differentiation, little is known about how their application dynamics affect PSC fate.

This thesis explores how application dynamics of agonists activating all-trans-retinoic acid
(atRA) and WNT signalling pathways in Ntera-2 differentiation. We applied atRA or
CHIR99021 either as pulses or through sustained replenishment at frequencies ranging from 1 to 48 hours. Differentiation was tracked using fluorescent labelling of surface markers indicating loss of pluripotency (TRA-1-60), and pro-neuronal commitment (A2B5), alongside morphological changes.

To enable these complex experiments, we developed an open-source liquid-handling and
imaging platform that requires minimal human input and operates at low cost. The platform can test over 60 conditions in a single experiment and collect data from more than 10 million cells, throughput that would be unfeasible manually.
Here we show that atRA dynamics and frequencies did not significantly alter A2B5 expression. In contrast, CHIR dynamics had a pronounced effect: pulsed CHIR led to higher A2B5 expression than sustained application. Shorter CHIR pulses (≤12 hr) produced circular cells with thin extrusions with glial-like morphology, while longer pulses (≥18 hr) generated elongated neuron-like cells.

In conclusion, our findings demonstrate that not only the ligand identity but also the timing and frequency of signalling influences cell fate. Automated liquid handling for differentiation enables scalable exploration of differentiation dynamics and provides framework for studying other key pathways in development.