Post-differentiation Replating of Human Pluripotent Stem Cell-derived Neurons for High-content Screening of Neuritogenesis and Synapse Maturation

The replating of hiPSCs-derived neurons that have been differentiated for multiple weeks offers several advantages. However, detaching and replating differentiated neurons that have long, interconnected dendrites and axons (Figure 1A) can result in a high fraction of irreversibly damaged neurons.

As described in the protocol section, we used incubation with a proteolytic enzyme to detach the neurons from the substrate. Typically, due to their thick meshwork of neurites (Figure 1A), the cells tend to detach all at once as a single syncytial-like layer, which often begins to float in the well (Figure 1C). This happens fairly quickly, and is perhaps why most laboratories tend to collect the cells after only 5 min of incubation with their proteolytic enzyme of choice, and to immediately break apart the sheet of cells by trituration (pipetting it up and down). However, this mechanical manipulation appears to result in high stress for neurons. We believe the degree of stress may be proportional to the area covered by the neuronal network, because we find that the apparent damage from inadequate protease incubation is greater for 10 cm or larger plates than for 35 mm or smaller plates. Thus, counterintuitively, we found that prolonging the enzymatic incubation to 40-45 min reduces cell death at the time of cell dissociation (Figure 1C) and permits more efficient recovery of live, healthy cells that emit processes over hours to days post-replating (Figure 1B). Presumably, the extended incubation time with a mild proteolytic enzyme allows the partial digestion of proteins that strongly adhere cells and their processes to one another and to the extracellular matrix. This allows the resuspension process to work efficiently to separate individual cells without the need for overly vigorous trituration.

To quantify the effectiveness of the extended enzyme incubation procedure, we evaluated viability of suspended cells immediately after trituration using trypan blue (Figure 1C), and later at 1, 3 and 7 days post-replating using an early cell death marker (Figure 2). Immediately after trituration, trypan blue staining indicated that there was substantially greater cell death after a 5 min incubation compared to a 45 min incubation with the proteolytic enzyme (Figure 1C). The two iPSC lines we used to illustrate this observation were differentiated into neurons from the neural progenitor stage using different protocols, and showed broadly different sensitivity to the replating procedure. Cultures from the WT126 line differentiated into NPCs using the "rosette selection method"⁷ were more sensitive to damage than cultures differentiated from the CVB WT24 line using a rapid differentiation method⁸. Nevertheless, for both lines the degree of immediate cell death was approximately halved by using the extended enzyme incubation procedure.

After replating, cultures exhibited an approximate doubling of cell viability over subsequent days using the extended enzyme incubation procedure, as determined by the density of DAPI-positive cells (Figure 2B, graph on the left). In addition, the extended enzyme incubation resulted in a lower density of dead or dying cells detected using a dye-exclusion viability kit (Figure 2B, graph on the right). The fraction of dead cells detected using the viability assay after 24 h post-replating was higher than that seen using trypan blue immediately following trituration. This probably reflects the accumulation of dead and dying cells over these first 24 h, although it is also possible that the viability assay more sensitively detects cell death than the trypan blue assay. Dead cells continued to be detected over 1-7 days post-replating (Figure 2B). The initial plating density for both experimental groups (5 min and 45 min) was identical and based on hemocytometer live cell counts of the post-trituration cell suspension. Importantly, at all time points post-replating, the number of live, healthy cells was approximately doubled following the 45 min enzyme incubation compared to the 5 min incubation. These observations were confirmed qualitatively using phase contrast microscopy to monitor cell morphology at every step: before replating, during replating, and after replating (Figure 1 and Figure 2).

One of the advantages of replating hiPSC-derived neurons after they have been differentiating for several weeks is that most of the cells will have exited the progenitor stage and become committed to a neuronal phenotype at the time of replating. The transition phase of neuronal differentiation from neural progenitors takes place over several days, with greater numbers of cells gradually expressing early-stage neuronal markers, such as beta-III tubulin (detected by antibody TuJ1) or MAP2. Gene expression evolves over several weeks, eventually resulting in expression of late-stage pan-neuronal markers, such as NeuN, or cell-type specific markers for cortical neurons such as CTIP2. Therefore, the replating of previously differentiated hiPSC-derived neurons allows one to study early and transient neuronal events, such as neurite initiation, in identified subtypes of neurons. Figure 3 illustrates the immediate presence of early neuronal markers in cultures that were replated after 4 weeks of pre-differentiation in a larger culture dish. Note that NeuN- and CTIP2-expressing neurons are readily identified within a few days post-replating (Figure 3). Characteristics and timing of neuronal differentiation can vary among hiPSC lines. For the WT 126 line and the CVB WT24 lines we describe here, 85 ± 12% (n = 2) and 52 ± 11% (n = 5) of the cells, respectively, expressed immunoreactivity for the βIII-tubulin marker TuJ1 at 4 days after replating using the extended protease incubation procedure. For both lines, 30-40% of the cells, and 80-90% of the neurons also express the layer 5/6 neocortical marker CTIP2. In addition to improved viability, the extended protease protocol also moderately enhanced neurite outgrowth (average neurite length per neuron), as shown in Figure 4. Both total neurite length (quantified using antibody Tuj1, which stains both axons and dendrites; Figure 4B, left graph) and dendrite growth (quantified using MAP2, which stains only dendrites; Figure 4B, center graph) were favored when using the extended protease incubation procedure. Note that the graph for DAPI-positive (DAPI (+)) cell counts in Figure 2 are based on the same image samples as the cell count graph in Figure 4, but in Figure 2 they were quantified using a non-automated method assisted by Fiji software, while in Figure 4 they were quantified using the automated image analysis software platform CellProfiler. These two image analysis approaches yielded similar results.

Replating is useful not only to study neurite outgrowth, but also to study synapses or other later-appearing biological features of neurons. Indeed, after just one week post-replating we observe markers for many presynaptic and postsynaptic proteins decorating the neuronal dendrites in a punctate pattern (Figure 5A). After 4 weeks we also begin to detect their colocalization, which is an indicator of the formation of functional synapses (Figure 5A). Moreover, electrical activity from spontaneous depolarization and synaptically driven currents is detectable using calcium imaging (Figure 5B) or multielectrode arrays (MEAs).

Figure 1: Superior preservation of neuronal viability after longer incubation with the proteolytic enzyme for cell dissociation. (A) Phase contrast image of NPC-derived neurons differentiated for 3 weeks. The boxed region in the left image is enlarged at right. At this stage neurons have long processes, which form a thick meshwork. Scale bars = 80 µm (left); 35 µm (right). (B) Selected images of hiPSC-derived neurons at 1 and 5 days post-replating. Scale bar = 100 µm. (C) Left: cells detach as a single sheet within a few minutes after initiating protease treatment. Scale bar = 200 µm. Right: after trituration cells are counted on the hemocytometer before replating. Graphs show quantification of non-viable, trypan-blue positive cells after 5 min or 45 min incubation with the proteolytic enzyme for two different lines CVB WT24 (*** p < 0.0003, unpaired t-test) and WT 126 (*** p < 0.0001, unpaired t-test). Values represent the mean ± S.E.M of 4 individual replicates. Please click here to view a larger version of this figure.

Figure 2: Neuronal viability after several days post-replating. (A) Images of WT126 NPC-derived neurons at 1 and 3 days post-replating using 5 min and 45 min protease incubation. Dying cells are labeled with the sensitive early cell death reporter. Corresponding brightfield images are shown on the left. Some cellular debris (blue arrow) is typically detected after replating, especially in the 5 min group. Scale bars = 150 µm. (B) Images of CVB WT24 NPC-derived cultures at 1, 3 and 7 days post-replating using 5 min and 45 min protease incubation. Dying cells are labeled with the sensitive early cell death reporter (red). All nuclei of living and dying cells are labeled with DAPI (cyan). The merged white signal represents DAPI and VivaFix-positive (+) cells. A few cells display VivaFix staining but lack DAPI staining (red cells in the combined image on the right); these are likely cells that were dead for many hours. Scale bars: 25 µm. Right graph shows quantification of the fraction of dead cells (VivaFix/DAPI) after different incubation time with the proteolytic enzyme (5 min vs 45 min: * p < 0.05 1 d and *** p < 0.001, 3 d; two-way ANOVA, followed by multicomparison Bonferroni post hoc test). Left graph shows changes in overall cell density (# DAPI(+) cells for 5 min vs 45 min: *** p < 0.0001 for all time points; two-way ANOVA, followed by multicomparison Bonferroni post hoc test). All values are shown as the mean ± S.E.M of 3 replicates, with a minimum of 1,500 cells scored per condition. Please click here to view a larger version of this figure.

Figure 3: Detectable expression of late-stage neuronal markers immediately after replating. Cultures of CVB WT24 hiPSC-derived cells imaged 4 days after replating, using 45 min protease incubation, from a 10 cm dish that had been differentiated from the NPC stage for 4 weeks. Examples shown were stained for the pan-neuronal markers TuJ1, MAP2, or NeuN; or the deep-layer cortical pyramidal neuron marker CTIP2. Note the presence of extensive neurites, and the robust expression of TuJ1 and MAP2. Note especially that a substantial fraction of the neurons also expresses late-stage markers NeuN and CTIP2. Scale bar = 16 µm. Please click here to view a larger version of this figure.

Figure 4: Neurite and dendrite growth over time following shorter and longer enzyme incubation during replating. (A) Replated CVB WT24 hiPSC-derived neurons quickly extend neurites, as detected using antibody Tuj1. Note that the extended protease incubation time promotes neuritogenesis. Scale bar = 25 µm. (B) Quantification of neurite and dendrite length, over 1, 3 and 7 days post-replating using either 5 min or 45 min protease. Total neurite length was quantified from TuJ1 (βIII-tubulin) staining; dendrite-specific staining was quantified from MAP2 staining snf corrected for the overall cell density (right graph). All values are shown as the mean ± S.E.M of 3 replicates. Neurite length 5 min vs 45 min: * p < 0.05 at 1 day and 7 days, ** p < 0.01 at 3 day; dendrite length 5 min vs 45 min: ** p < 0.01 at 1 day and 3 days, not significant (ns) at 7 day; # DAPI(+) 5 min vs 45 min: *** p < 0.0001 for all time points; two-way ANOVA, followed by multicomparison Bonferroni post hoc test. Please click here to view a larger version of this figure.

Figure 5: Synaptogenesis and spontaneous calcium transients in differentiated hiPSC-derived neurons after replating. (A) Left: at 4 weeks post-replating using the extended protease incubation protocol, the presynaptic marker synapsin 1 is detectable in punctate clusters along the MAP2 positive dendrites. Scale bars = 5 µm. Right: selected dendritic region showing colocalization between presynaptic and postsynaptic clusters (yellow arrow), an indication of potentially active synapses. Scale bars = 1.5 µm. (B) Left: hiPSC-derived neurons infected with AAV8-syn-jGCaMP7f-WPRE were used to monitor spontaneous calcium transients driven by network activity. The brightfield image is shown adjacent to the pseudocolor rendering of GCaMP7 fluorescence at a single time point in the time-lapse series. Scale bar = 25 µm. Colored numbers indicate the 4 selected cells in which GCaMP7 fluorescence was measured over time, as shown in the traces on the right. Each colored trace corresponds to one cell. The asterisk points to the time during the live recording to which the image on the left corresponds. Please click here to view a larger version of this figure.

Figure 6: Workflow of replating procedures to culture hiPSCs for high content screening and/or detailed studies of differentiation and neurite outgrowth or MEA recordings. Starting from a differentiated culture of neurons grown for 4 or more weeks in a larger format (e.g., a 10 cm culture dish), neurons are incubated with a protease for 40-45 min, triturated gently to dissociate and resuspend. Cells are then distributed into multi-wells compatible with HCS, replated onto substrates compatible with imaging growth cones (yellow arrow) or other structures, or onto multi-wells suitable for multielectrode array recordings. Scale bars = 27 µm, 5 µm, 2.5 µm, 130 µm (from left to right). Please click here to view a larger version of this figure.