Sustained synchronized neuronal network activity in a human astrocyte co-culture system.

Characterization of hiPSC (-derived) cells during different steps of the differentiation protocol on laminin coated surface (a) Schematic overview of the differentiation protocol towards hiPSC-derived cortical neuronal cultures. (b) hiPSCs express pluripotency markers OCT4 and NANOG before the start of differentiation. (c) Neural precursor cells express neural stem cell markers PAX6, nestin, and OTX2 at the 25th day of the differentiation protocol. (d) Fully differentiated neurons express neuronal marker class III β-tubulin and cortical markers TBR1, CTIP2, and SATB2. However, due to the heterogeneous nature of the cultures, not all cells are immuno-positive for all markers (also undifferentiated neural precursor cells and potentially some astrocytes are present in the cultures).

Functional maturation of hiPSC derived cortical neurons in a short time frame via co-culture with astrocytes and treatment with DAPT (a) Cortical neurons are differentiated from NPCs after final plating on top of an astrocyte monolayer. During the first week of differentiation DAPT is added. Functional and morphological assays are performed at 2–8 weeks after final plating. (b) hiPSCs differentiated towards cortical neurons in co-culture with primary human astrocytes are stained with neuronal marker class III β-tubulin, astrocyte marker GFAP and nuclear marker DAPI. (c) Cortical fate of hiPSC-derived neurons grown in human astrocyte co-cultures is confirmed by immunocytochemistry for cortical markers TBR1, CTIP2, and SATB2 in combination with the nuclear marker DAPI. (d) Functional maturation of neurons analyzed using whole cell patch clamp. Traces of evoked potentials (protocol scheme included) show a clear difference between early (wk1 and wk2 respectively firing no action potentials or only one) and later time points (wk4-5 firing repetitive action potentials) of differentiation. The left graph shows a more negative resting membrane potential over time (One-way ANOVA, p = 0.0113). The middle graph shows an increasing maximum action potential frequency over time (One-way ANOVA, p = 0.0187). An increased rheobase is observed as well two and three weeks after final plating (One-way ANOVA, p = 0.0112). n ≥ 9 from ≥1 differentiation, mean + SEM, *p < 0.05.

Optimization of conditions for synchronized neuronal calcium oscillations (a) Image showing automated identification of FLUO-4 loaded cells by color-coded regions of interest (ROIs) and representative traces (each trace represents fluorescence of one cell) of co-cultures with DAPT two weeks after final plating (left, no synchronicity) and four weeks after final plating (right, highly synchronized calcium influxes). After 250 frames (61 frames per minute) 30 μM glutamate was added, resulting in a large calcium influx and used to distinguish neurons from astrocytes. (b) Co-culturing with primary human astrocytes and treatment with DAPT significantly increases the percentage of active neurons (Two-way ANOVA, p < 0.0001), bursting frequency (Two-way ANOVA, p < 0.0001) and synchronicity (Two-way ANOVA, p < 0.0001) compared to cultures without DAPT and astrocytes. n ≥ 4 from ≥2 differentiations, mean ± SEM, **p < 0.01,***p ≤ 0.0001. (c) Synchronized activity sustains up to 8 weeks after final plating in human astrocyte co-cultures with DAPT. n ≥ 5 from ≥2 differentiations, mean ± SEM. d) Limited FGF2 passaging significantly reduces the percentage of active neurons (Two-way ANOVA, p < 0.0001), bursting frequency (Two-way ANOVA, p < 0.0001) and synchronicity (Two-way ANOVA, p < 0.0001). n ≥ 4 from ≥2 differentiations, mean ± SEM, *p < 0.05, ***p ≤ 0.0001.

Synchronized calcium oscillations represent neuronal network activity (a) Representative traces of live cell calcium imaging recordings in 5 week old cortical neuronal co-cultures, FLUO-4 intensity is shown over time (61 frames per minute). After 200 frames cells are exposed to tetrodotoxin (0.1 μM) followed by addition of glutamate (30 μM) after 200 frames, resulting in a large calcium influx. (b) Representative traces of patch clamp recordings of sPSCs showing sparse sPSCs (upper trace), sparse bursts of sPSCs (middle trace) and frequent bursting (lower trace). (c) Co-culture with primary human astrocytes + DAPT increases the percentage of cells with sPSC bursts compared to control (no DAPT, no astrocytes). “No activity” represents up to five single sPCSs per minute, “sparse activity” reflects 5 or more single events or less than one burst per minute and cells bursting at a frequency higher than one burst per minute are labeled with “frequently bursting”. Number of cells per cell culture condition n ≥ 22, from ≥2 differentiations. (d) Frequency of sPSC bursts per minute in co-cultures with primary human astrocytes + DAPT or neuron-only cultures equals the calcium oscillation frequency (Two-way ANOVA, p = NS for the respective cell culture conditions), while culture conditions induce significantly different bursting frequencies (Two-way ANOVA, p ≤ 0.0001). sPSC frequency: number of patched cells per cell culture condition ≥ 22, from ≥2 differentiations, calcium oscillation frequency: n ≥ 6 from ≥2 differentiations. Mean + SEM, ***p ≤ 0.0001.

GABAergic and glutamatergic contribution to neuronal network function for cortical neurons in co-culture (a) Graphs showing a clear contribution of glutamatergic transmission to the calcium signal. Percentage of active neurons (One-way ANOVA, p < 0.0001), frequency (One-way ANOVA, p = 0.0002) and synchronicity (One-way ANOVA, p = 0.0004) significantly change compared to baseline. Mean + SEM, n ≥ 3, from 3 differentiations. *p < 0.05, ***p ≤ 0.0001 on the different time points. (b) Representative traces of live cell calcium imaging recordings in 5 week old cortical neuronal co-cultures, FLUO-4 intensity is shown over time (61 frames per minute). After 200 frames cells are exposed to DAP5 (50 μM) and CNQX (20 μM) followed by exposure to glutamate (30 μM) after 200 frames, resulting in a large calcium influx. (c) Graphs showing the contribution of GABAergic transmission to the calcium signal. The percentage of active neurons (One-way ANOVA, p = 0.0015), bursting frequency (One-way ANOVA, p = 0.0302) and burst amplitude (One-way ANOVA, p = 0.0125) are significantly changed after addition of picrotoxin. Mean + SEM, n ≥ 3, from 3 differentiations. *p < 0.05, **p < 0.01 on the different time points (d) Representative traces of calcium imaging. After 200 frames, the cells are exposed to the GABAergic inhibitor picrotoxin (50 μM), followed by glutamate exposure (30 μM) after 200 frames.

Astrocyte co-cultures are suitable for HCI and analyses (a) hiPSCs differentiated towards cortical neurons in co-culture with primary human astrocytes show expression of glutamatergic marker vGLUT1, GABAergic marker GAD65, and neuronal marker MAP2. (b) Image analysis based on raw data files from a plate scanner. Masks (in yellow) are drawn per channel to identify the number of nuclei (based on DAPI staining), neurite area (based on MAP2 staining) and the number of puncta per marker (either based on vGLUT1 or GAD65). (c) Quantification of glutamatergic marker vGLUT1 and GABAergic marker GAD65 per neurite area in human astrocyte co-cultures. Mean + SEM, n ≥ 4, from 3 differentiations.