IPSC derived neuronal progenitors of cortical lineage
We differentiated human iPSCs into neuronal progenitors and cortical neurons using previously established methodology11. The differentiated neural cells were characterised using immunofluorescence. First, we demonstrated that our iPSC lines retained pluripotency through the expression of OCT4 and NANOG (supplementary figures 1 and 2). Second, we confirmed the appearance of primary and intermediate neuronal progenitor cells using PAX6 and TBR2 respectively (figure 1a and 1b). Third, we demonstrated that these neuronal progenitor cells gave rise to both CTIP2-expressing deep cortical neurons and BRN2-expressing upper-layer cortical neurons (figure 1c-1f)
Figure 1
Figure 1: iPSC derived cortical linage neuronal progenitor cells and cortical neurons. A: PAX6 expressing primary neuronal progenitors. B: TBR2 expressing intermediate neuronal progenitor cells. C and D: CTIP2 expressing deep cortical layer neurons. D and E: BRN2 expressing upper-layer cortical neurons.
We used neuronal progenitor cells from day twenty-five of differentiation to make neurospheres. The neuronal progenitor cells were first dissociated into single cells and subsequently allowed to reaggregate in 96-well V-bottom plates coated with non-adhesive compounds. This resulted in the formation of three dimensional neurospheres that were transferred onto a matrigel matrix after 24 hours. We characterised the neurospheres using immunofluorescence to confirm the presence of intermediate progenitor cells and cortical neurons (figure 2). Over time, we observed the evolution of neurospheres (figure 2a and 2b). Specifically, we found an increasing number of TBR2 expressing intermediate progenitor cells appearing over time (figure 2b). Similarly, an increasing number of nascent CTIP2 expressing cortical neurons migrated along the radial fibres over time (figure 2a).
Figure 2
Figure 2: Evolution of neurospheres over three time points 6-, 12-, and 24-hours using immunofluorescence. Top panel: Appearance of CTIP2 expressing deep cortical layer neurons over time. 3X and 14X zoom panels show the migration and differentiation of CTIP2 expressing neurons as they reach the outer edge of neurospheres. Bottom panel: Appearance of TBR2 expressing intermediate neuronal progenitor cells over time. 3X and 14X zoom panels show the appearance of TBR2 expressing intermediate neuronal progenitors within the radial extensions of neurospheres over time.
Live cell imaging of neurospheres and automated analysis
Time-lapse images of neurospheres were acquired using fluorogenic live cell probes. We used SiR-tubulin, a highly specific marker of microtubules, to identify radial glia fibres. Whereas SPY555-DNA, a highly specific DNA probe, was used to identify the nuclei of migrating neural cells. Images were taken every 30 minutes over a period of 48 hours using SpinSR microscope system (Olympus). A total depth of 140um was imaged in the z-plane at 10um intervals using a 10x objective. We captured the growth of radial fibres away from the central mass of cells (figure 3, bottom panel) and migration of neural cells along the radial fibres (figure 3, middle panel) (supplementary movie 1).
Figure 3
Figure 3: Evolution of a single neurosphere over time using live cell imaging. The top panel shows combined radial glia (Tubulin) and nuclei of migrating neural cells (DNA) at four selected time points 0, 12, 24, 36 and 48 hours. The middle and bottom panels show the breakdown of nuclei of migrating neural cells (DNA) and radial glia (Tubulin) respectively at 0, 12, 24, 36 and 48 hours. The middle and bottom panels show the breakdown of nuclei of migrating neural cells and radial glia respectively at time points 0, 12, 24, 36 and 48 hours.
To robustly quantify the dynamic nature of neurospheres over time, we developed an automated pipeline for image analysis using the Arivis Vision4D image analysis platform. The projected images were processed using closing and blurring algorithms. An intensity threshold was then applied to define areas of migrating neural cells and radial fibre extensions using nuclear and tubulin staining respectively (figure 4 and supplementary movies 2 and 3) Finally, radial fibre and neural cell migration areas were calculated after each time point. At 48 hours, the average radial fibre area was 4.3mm2 (SEM: 0.74mm2) whereas the neural migration area was 1.78mm2 (SEM: 0.09mm2).
Figure 4:
Figure 4: Automated analysis pipeline. Top panel shows the temporal evolution of a single neurosphere. The middle panel shows how the area of migrating neural cells is tracked over time. The bottom panel shows how the area of radial glia growth is quantified over time.
Effects of viral infection on cortical development
Given the emerging role of in utero viral infections in schizophrenia and autism12, 13, we investigated the effects of viral infection on radial glia growth and neural cell migration using our neurosphere assay. We used polyinosinic:polycytidylic acid (polyI:C), a well-established viral mimic, to recapitulate infection twelve hours prior to live imaging. The neurospheres were constructed and live imaging was carried out over a period of 48 hours as described above. The areas encompassing radial fibres and migrating neural cells were calculated using the automated pipeline. Linear mixed effects model showed that radial glia growth was greater in unstimulated neurospheres relative to stimulated neurospheres by 0.028 mm2/hour (SE:0.0046mm2/hour, p=1.78e-05) (figure 5d). A likelihood-ratio test indicated that a model that included the interaction between time and stimulation status (i.e. allowing different rates of growth based on stimulation status) was better fit for radial glial growth data (χ2 = 18.9, p=1.33e-05). Also, linear mixed effects model identified that neural cell migration was greater in unstimulated neurospheres by 0.011 mm2/hour (SE:0.0025 mm2/hour, P=0.0009) (figure 5e). Similarly, a model that included the interaction between time and stimulation status (i.e. allowing different rates of growth based on stimulation status) was a better fit for neural migration data (χ2 = 11.5, p=0.0007). Furthermore, we replicated the effects of polyI:C on radial glia extension and neural migration using immunofluorescence data. Similar to our findings from live cell imaging, we found that polyI:C stimulation resulted in significant reductions in radial glia growth (p=0.0069) and neural migration (p=0.0033) at 24 hours (figure 5b and 5c).
Figure 5:
Figure 5: The effect of viral infection on cortical development. A: Images of fixed polyI:C stimulated and unstimulated neurospheres after 24 hours. B and C: Analysis of immunofluorescence data from fixed neurospheres B: Comparison of radial fibre extension between polyI:C stimulated and unstimulated neurospheres using immunofluorescence. C: Comparison of neural migration area between polyI:C stimulated and unstimulated neurospheres using immunofluorescence. D and E: Analysis of live imaging data. D: Mixed effects linear models showing neural migration based on stimulation status over time. E: Mixed effects linear models showing radial glia extension based on stimulation status over time.