Characteristic analyses of a neural differentiation model from iPSC-derived neuron according to morphology, physiology, and global gene expression pattern

Induced pluripotent stem cells (iPSCs) can differentiate into neural progenitor cells (NPC) under proper conditions. NPC can be used as a model and is a useful tool for disease mechanism exploration and drug screening. However, the characteristics of the cells in various stages from NPC to functional neurons have not been fully described. This study investigated the characteristics of iPSC-derived NPCs during differentiation. Morphological characteristics of the NPCs, including soma area, neurite length, and the number of neurite branches, were examined on selected differentiation days. Physiological functions were assessed by recordings of sodium current, spontaneous excitatory postsynaptic current (sEPSC), and spontaneous inhibitory postsynaptic current (sIPSC). Furthermore, gene expression patterns were assessed with RNA-seq. We found that NPCs derived from iPSCs can be differentiated into glutamatergic and gabaergic neurons. Cell growth peaked during differentiation day 7–12, as the soma area decreased after day 12, growth cone and the number of branches peaked at day 9 and decreased afterwards; whereas a functional synapse formed after day 23. RNA-seq analysis found that a differential expression pattern emerged by day 7. Overall, the study provides a framework for the differentiation process of hiPSC-derived NPCs.

importantly, previous studies have also shown that physiologically functional neurons, characterized by synaptic transmission and generation of action potentials, can be differentiated from iPSCs or fibroblast-direct conversion, indicating the neuronal cells induced from iPSCs are likely to be functional [20][21][22][23][24][25][26][27] . However, many limitations still affect the application of this technology in personalized medicine in a clinical setting. One of the main limitations is that the characteristic parameters of the differentiation cells in different stages have not been clearly described to date. In our study, we examined the transcriptome phenotype coupled with functional neuron mature process assessed by both morphology and electrophysiological analyses.

Results
In vitro neuronal progenitor cell model. We first established an in vitro neuronal progenitor cell (NPC) model by culturing hiPSCs with a two-inhibitor culture system. At the end of the culture period, the treated hiPSCs were stained for neuroectodermal stem cell markers including NESTIN, PAX6, and SOX2. We found that the majority of the treated cells stained positive for these markers, indicating that most of the treated hiPSCs differentiated into NPCs (Fig. 1).
We next examined the differentiation potential of these NPCs. The NPCs were cultured in a neuron differentiation media system (N 2 B 27 + 20 ng bdnf + 1 μM dibutyryl-cAMP) for 21 days. The cells were then stained for TuJ1, a neuronal cell marker, and GFAP, an astrocyte marker. We found that both the neuronal marker and the astrocyte marker were expressed in the cultured cells (Fig. 1). These data indicated that NPCs derived from hiPSCs could differentiate into neuronal cells as well as astrocytes, and could be used as an in vitro model of neural differentiation. Furthermore, the neuronal cells stained positive for GABA, Glu1R, tyrosine hydroxylase (TH), and synapsin 1, indicating that the NPCs can differentiate into different types of mature neurons ( Supplementary Fig. S1). Further analyses found that in differentiated cells, 54.9% were gabaergic neurons, 17.3% were TH-positive neurons, and 10.7% were glutamatergic neurons ( Supplementary Fig. S1). The composition of neuronal cells did not change over the 15-day differentiation period.
Neuronal growth profile. We next investigated the morphological characteristics of these neurons. The somatic area of the neuronal cells and neurite length were measured, and the number of branches was counted in differentiated cells. The area of the somatic region increased significantly from D3 to D12. However, when assessed on D15, the somatic area decreased (Fig. 2). Time-lapse analysis showed that both the secondary and tertiary branches of the dendrites increased from D3 to D15, while the number of primary branches reached a peak on D9 of differentiation and then decreased to a level similar to that on D3 (Fig. 3). The lengths of the axons of the cultured neurons were significantly longer when measured on D9, D12, and D15 compared with the length on D3. We also noted that although the number of primary branches of dendrites reached a peak on D9, the length of the neurites continued to steadily increase throughout the 15 days of the differentiation protocol. However, the length of the neuritis did not increase significantly after D3, as no significant differences in neurite lengths were detected between neurons when assessed on D6, D9, D12, and D15 despite a trend toward an increase throughout the differentiation days. Overall, we found that the neurites of cultured neurons grow in length from D3 to D9; however, when assessed on later development days, both the area of the soma and the length of neurites decreased (Fig. 3).
We next characterized the development of the growth cone. The growth cones are categorized into three subtypes: blunt-ended (collapsed with no visible filopodia or lamellipodia), filopodial (growth cones with numerous filopodia and a small or absent lamellipodial veil), or lamellipodial (well spread growth cones with elaborate lamellipodia). The total growth cone area peaked on D9 and then decreased sharply (Fig. 4). In addition, we found that the ratio of the lamellipodia growth cone was the highest among the three subtypes during the 15-day differentiation period (Fig. 4).
The pattern of gene expression during the course of differentiation. We analyzed the gene expression profile of NPCs throughout the differentiation stages by time-lapse analysis using STEM software. We found 16 modules of genes including more than 30 000 genes that were enriched in NPCs. The 16 modules can be divided into three groups by trends of expression level. Of the 16 modules, 8 showed statistically a significant difference. The three groups were a decreasing (2 modules) group, increasing (4 modules) group, and disorder-associated (2 modules) group (Fig. 7). Modules 11,12,13, and 15 showed a pattern of increasing expression, and GO pathway analysis showed that the increasing group of genes primarily included the genes that decided neuronal fate and the genes that are involved in the anatomic structure of neural cells morphology and function (Table S1). Specifically, module 11 included genes that regulate neuron projection and soma to dendritic compartmentalization, module 12 included genes that regulate cytoplasmic vesicle and extracellular structure organization, module 13 included genes that regulate ion channel activity and transmembrane transporter activity, and module 15 included genes that regulate potassium ion transmembrane transport, synaptic transmission. We next annotated the enriched genes with KEGG in the early differentiation period from D0 to D7 (Table 1). From D0 to D7, the genes in the signaling pathways that regulate the general biological proliferation of "stem cells", such as DNA replication, cell cycle, spliceosome, and others, were down regulated (Table 1), whereas the genes in the pathways that regulate neuronal maturity, including the MAPK signaling pathway, neuroactive ligand-receptor interaction, Wnt signaling pathway, and others were up-regulated ( Table 1). The disorder groups were module 14 and module 1. In these groups, the modulate synapse and neuron projection in module 14 showed a pattern of increased expression from D0 to D28.
It is worth noting that the genes in the WNT-associated pathways presented a unique pattern. The expression of this group of genes was not uniform, with some genes being upregulated during D0-D7, some genes being upregulated during D7-28, and some genes being down regulated ( Table 2). We found that the group of genes that were up regulated during D0-7 and genes that were up regulated during D7-28 did not overlap; however, PRKCA was found to be upregulated during D0-7 and down regulated during D7-28.
We further analyzed the expression pattern of WNTs (Supplementary Table S2). We found that WNT5A gene was down regulated, whereas WNT7A gene was up regulated in the early differentiation period (D0 to D7). In contrast, from D7 to D28, WNT5A gene was up regulated whereas WUT7A gene was down regulated. We then confirmed the expression of WNT2B, WNT5A, and WNT7A using qRT-PCR. The results were consistent with the findings when RNA-seq was used to assess gene expression (Fig. 8).

Discussions
Since the development of the methods that make generating stem cells from adult somatic cells possible, the field of stem cell research has been fueled with the potential that these methods can offer. Of particular interest, stem cell therapies have shown promise in treating neurological disorders. However, the differentiation stages of the NPCs that are generated from iPSCs are not described in detail. In the present study, we analyzed the differentiation processes of hiPSC-derived NPCs using multiple analyses. We first assessed the morphological changes and then examined the physiological function of hiPSC-derived NPCs during cell maturation. Lastly, we examined the whole cell gene expression pattern.
We found that the growth of the somatic area of the hiPSC-derived NPCs followed an increasing and then decreasing pattern, in which the soma area increased significantly to day 12 and then plateaued. The growth of neurites followed a similar pattern, with the number of the primary, secondary, and tertiary branches peaking at differentiation day 3 to day 9 and the length of axons and dendrites peaking at differentiation day 9. The growth cone is the key factor of the guidance and motility of axonal growth in the early stage of differentiation and neuron regeneration 28 . Our study found that growth cones also showed a pattern of rapid growth up to differentiation day 9, followed by a significantly decrease. The bundled F-actin in blunt and filopodia growth cone transport cargo both anterogradely and retrogradely. In contrast, F-actin in the lamellipodium growth cone forms a dendritic network through the action of dendritic nucleator proteins and capping proteins. It is not surprising that the growth pattern of the growth cone coincides with that of the dendrites and the axon. Overall, morphologically, the growth of NPCs peaked at differentiation days 9 to 12, indicating that the cells at this stage have reached a stable differentiated stage.
The physiological functions of the NPCs were assessed with recordings of sodium current, sEPSC, and sIPSC. As the NPCs stained positive for Glu1, GABA, and synpasin 1, it is highly likely that the NPCs can signal through synapses with each other. Notably, sEPSC and sIPSC were both present in the NPCs. The amplitude and frequency of both EPSC and IPSC increased with cultured days, indicating the development of mature physiological function. This observation was consistent with previous findings that NPCs generated from iPSCs responded to glutamate or GABA stimulation after two months of neuronal differentiation 29 and that sodium channel expression progressively increased in the iPSC-derived neurons, which eventually led to the generation of an action potential 30 . Overall, our results indicated that morphological maturation precedes functional maturation in these cells.
We next assessed the gene expression pattern over the differentiation stages and found that genes that are characteristic of "stem cells" were down-regulated, whereas genes in the neuronal cell induction pathway were up-regulated during differentiation day 0 to 7. These observations indicated that neural induction-associated pathways have emerged earlier in the timeline than the peak morphological changes and the development of physiological functions.
Of note, we observed that different genes in the WNT family present the opposite expression trend. WNT7A shows a significant increase in the early stage from D0 to D7, coupled with noncanonical Wnt signaling pathway genes, PRKCA and NLK; however, WNT5a shows a decreased expression pattern coupled with canonical Wnt signaling pathway genes FZD and β-Catenin in the early differentiation stage, which then increased. It was reported 31 that the canonical Wnt signaling pathway is involved in regulating neuronal function maturation, while the noncanonical Wnt signalling pathway is related with cell fate and migration. In the present study, cell fate-related noncanonical Wnt signaling pathway emerged earlier than the neuronal function maturation-related canonical Wnt signaling pathway, which is in accordance with the development from fate to function.
Given the need for a model of iPSC-derived functional neuronal culture, our study confirmed that neuronal cells can be induced from iPSCs using a two-inhibitor induction system. The induced cells lost their pluripotency in the early differentiation stage. This study found that the method of developing neuronal cells from iPSC using a two-inhibitor induction system provides an in vitro model of valid neuronal differentiation. Using this model, we can perform drug screening on neuronal cells at the appropriate differentiation stages based on the morphological and physiological characteristics of the cells. In the present study, we provide a framework for the differentiation process of hiPSC-derived NPCs. Compare to some researchers 19 , our neural differentiation model has higher effectiveness in converting iPSC into neuron and costs less time. The stage of NPC suspension culture can improve the purity and differentiation efficiency of neuron. Specially, we observed that functional synapse formed after day 23 by electrophysiological recording as well as a differential expression pattern emerged by day 7 by RNA-seq analysis, confirming by the expression of Wnt signaling pathway genes. Furthermore, neuron derived from patient iPSC can be used in drug screening and disease mechanism exploration, especially neurological diseases and neurodevelopment diseases, such as autism spectrum disorders, RETT syndrome, Alzheimer's disease and schizophrenia. In 2013, Shcheglovitov 19 proved the dysfunction of the ratio of excitation and inhibition neurons derived from PMDS patient iPSC and the synaptic deficits can be rescue by IGF1. What is more, translational research (such as cell transplantation) and genetic correction by CRISPR-Cas9 hold great promise for therapy of neurological diseases. Researchers had successfully corrected the defect gene form patient iPSC 32 , which can significantly promote the progress of efficient treatments for neurological diseases and neurodevelopment diseases.   Selleck) so as to induce neuronal differentiation. Cells were mechanical divided into fragments after 8 days of neuronal differentiation induction. The fragments in one well of 12 well plate were replanted into 2 wells of a 6-well matrigel-coated plate, and cultured in neural proliferation system I (N 2 B 27 , Thermo Fisher Scientific) medium.
On the 8th day of proliferation, cell clones form one well of the 6-well plate were mechanical scraped into floating fragments, and the floating fragments were plated into non-coated T25 flask for floating culture with neural proliferation system II (N 2 B 27 + 20 ng/mL bFGF + 20 ng/mL EGF, Thermo Fisher Scientific). Neurospheres usually formed on the second day of culture in neural proliferation system II. Accutase (StemPro ® Accutase ® Cell Dissociation Reagent, A1110501) was then used to detach neurosphere, and the single cells were plated in matrigel-coated dishes with neuron differentiation medium (N 2 , Thermo Fisher Scientific; B 27 , Thermo Fisher Scientific; 1 μM dibutyryl-cAMP, Sigma-Aldrich; 20 ng bdnf, PeproTech).

Morphology analysis.
The cultured cells were then infected with 5 mL of Lenti Virus (6 × 10 8 TU/mL) packed with a tetracycline-controlled red fluorescent protein (RFP) expression sequence per well. The cells with RFP signals were then cultured at a density of 100 000 cells per well on a matrigel-coated coverslip that was placed at the bottom of a well of a 24-well plate for neuronal development for six hours, and then the vector was washed out. RFP expression was induced by addition of doxycycline to the medium after wash-off of the vector. The first day of induction on cover-slip was defined as day-0. Cells were used for morphological and functional analysis from D0. Cells grown on coverslips were collected on culture day 0, 3, 6, 9, and 15; and their morphological characteristics were analyzed. The cells were fixed with 4% paraformaldehyde dissolved in phosphate based saline (PBS, pH 7.4) for 30 minutes. Cells were imaged with a microscope (Imager Z2, Zeiss) and the images were acquired with Zeiss camera (Axiocam 506 mono, Zeiss). The soma area, neurite length, and branches were analyzed with ImageJ Eletrophysiological recordings were made with an amplifier (Axon MultiClamp 700B, Molecular Devices), signals were filtered at 3 kHz (low pass) and digitized at 20-kHz sampling frequency (DigiData 1550 A, Molecular Devices), and acquired and analyzed with pClamp10 (Molecular Devices). Neurons were first recorded in current-clamp mode, and the responses to 1 second current steps (−10 pA, 2 pA, or 3 pA steps) were recorded to determine the input resistance of the cells. Resting membrane potential (RMP) was not corrected for a liquid junction potential. Following current-clamp recording, neurons were recorded in voltage-clamp mode. The cells were hold at −70 mV, and the sodium currents were elicited by 5 mV voltage steps (200 ms) from -20 to 50 mV. Spontaneous excitatory postsynaptic currents (sEPSC) and spontaneous inhibitory postsynaptic currents (sIPSC) were recorded at holding potentials of −70 mV and 0 mV respectively. In some experiments, kynurenic acid (3 mM KyA, sigma) or SR95531 (10 µM, Tocris) was used to block sEPSCs or sIPSCs. The data were analyzed with Clampfit.