Rapid and efficient differentiation of functional motor neurons from human iPSC for neural injury modelling

Primary rodent neurons and immortalised cell lines have overwhelmingly been used for in vitro studies of traumatic injury to peripheral and central neurons, but have some limitations of physiological accuracy. Motor neurons (MN) derived from human induced pluripotent stem cells (iPSCs) enable the generation of cell models with features relevant to human physiology. To facilitate this, it is desirable that MN protocols both rapidly and efficiently differentiate human iPSCs into electrophysiologically active MNs. In this study, we present a simple, rapid protocol for differentiation of human iPSCs into functional spinal (lower) MNs, involving only adherent culture and use of small molecules for directed differentiation, with the ultimate aim of rapid production of electrophysiologically functional cells for short-term neural injury experiments. We show successful differentiation in two unrelated iPSC lines, by quantifying neural-specific marker expression, and by evaluating cell functionality at different maturation stages by calcium imaging and patch clamping. Differentiated neurons were shown to be electrophysiologically altered by uniaxial mechanical deformation. Spontaneous network activity decreased with applied stretch, indicating aberrant network connectivity. These results demonstrate the feasi-bility of this rapid, simple protocol for differentiating iPSC-derived MNs, suitable for in vitro neural injury studies focussing on electrophysiological alterations caused by mechanical deformation or trauma.


Introduction
Supraphysiological elongation of peripheral nerves causes conduction blocks, resulting in chronic or temporary debilitating loss of motor function (Robinson, 2000;Miller, 1987). For example, brachial plexus birth palsy, initiated by excessive brachial plexus stretching during birth, causes motor and sensory impairment in 0.4% of US births (Abid, 2016). Conduction blocks implicate multiple nerve components including myelin (Sun et al., 2012), axonal paranodes (Rickett et al., 2011) and blood vessels (Ogata and Naito, 1986), but impairment of electrical transmission with strain has also been shown in isolated unmyelinated neuronal axons (Galbraith et al., 1993). Current studies on stretch-related conduction blocks are limited to animal tissue models (Sun et al., 2012;Rickett et al., 2011), and the majority of in vitro studies on the effect of mechanical injury on neurons are limited to using rodent primary neurons (Bar-Kochba et al., 2016;Kang et al., 2015), rodent neural stem cells (Kurtoglu et al., 2017), or immortalised cell lines (Skotak et al., 2012). Animal models provide simple, inexpensive models for neural damage studies, but concerns exist about the translational value of results from animal models to human conditions (Xiong et al., 2013). For example, candidate drugs for neuroprotection have failed human clinical trials after positive results in animal models (Marklund et al., 2006). The use of human induced pluripotent stem cells (iPSCs) provides physiologically relevant insights into the mechanics of human neural injury and motor function loss, overcoming potential issues of interspecies variability.
The induction of pluripotency in adult human cells provides an alternative cell source, both for in vitro studies and for clinical applications (Takahashi et al., 2007), avoiding ethical and regulatory complications of embryonic stem cell (ESC) use. Concerns about reproducibility and efficiency of iPSC differentiation, coupled with high costs and lengthy timeframes, have however limited their use (Kobold et al., 2015;Kim et al., 2014). Pluripotency induction by forced expression of defined factors generates iPSCs from adult cells, which can then be differentiated into adult cell lineages, including multiple neural fates such as dopaminergic (Zhang et al., 2014), GABAergic (Sun et al., 2016) and motor neurons (MNs) (Faravelli et al., 2014), by following developmental cues found in vivo (Tao and Zhang, 2016). iPSC-derived neurons have multiple uses, including disease modelling (Sances et al., 2016), in vitro neural injury modelling (Sherman et al., 2016) and transplantation trials (Kobold et al., 2015), for which differentiation protcol requirements vary. In modelling diseases such as amyotrophic lateral sclerosis (ALS), cells must acquire the diseased phenotype, and generally need to be kept in culture for extended periods of time (Sances et al., 2016;Dimos et al., 2008). In injury modelling, the emphasis is on large scale, rapid production of functional cells, which are only kept in culture until experiments are performed (Sherman et al., 2016).
Current human iPSC MN induction protocols vary in timescale and efficiency, with few protocols achieving both high efficiency and rapid MN generation. Faravelli et al. reviewed MN induction protocols from both human ESCs and iPSCs, showing high variability in efficiency and timescale of protocols, ranging from 30 days to more than six weeks (Faravelli et al., 2014). Measures of efficiency include specific marker expression, morphological analysis, electrophysiological analysis and animal transplantation, and depend on the intended use for iPSC-derived MNs (Faravelli et al., 2014). Sances et al. reviewed protocols used to derive MNs specifically for ALS modelling, reporting differentiation efficiencies from 20% to 95% and with timeframes ranging from 12 to 130 days, but without taking into account the quality of characterisation, the protocol difficulty, or the intended MN use (Sances et al., 2016).
In directed differentiation, iPSCs are induced to become neural progenitor cells (NPCs) through neuralisation, either by embryoid body (EB) formation, or by dual SMAD inhibition in monolayer cultures. EB formation, where suspension cells spontaneously arrange and grow in three-dimensional aggregates upon withdrawal of factors promoting pluripotency, is the most widely used neuralisation method (Shimojo et al., 2015;Maury et al., 2015;Salimi et al., 2014;Amoroso et al., 2013). However, EB formation requires lengthy protocols, the use of undefined culture conditions often including animal serum (Salimi et al., 2014), and is tedious to perform (Chambers et al., 2009;Surmacz et al., 2012). More recently, techniques for neural neuralisation in adherent monolayer cultures have been developed, using defined small molecules for dual-SMAD signalling inhibition: bone morphogenic protein (BMP) and TGF-β/nodal/activin inhibition has been shown to effectively convert human iPSCs to NPCs in adherent cultures (Chambers et al., 2009;Li et al., 2011).
Noggin, an endogenous polypeptide, was the first factor to be used for BMP inhibition (Chambers et al., 2009), but selective small molecule BMP inhibitors dorsomorphin and DMH1 have been shown to increase inibition efficiency (Neely et al., 2012;Zhou et al., 2010). Additionally, TGF-β/activin/nodal inhibition by small molecule SB431542 has widely been used in both EB and monolayer cultures to induce neuralisation (Shimojo et al., 2015;Chambers et al., 2009;Inman et al., 2002). Following the establishment of adherent monolayer neuralisation protocols, additional factors have been tested to enhance neuralisation. CHIR-99021, a Wnt activator by GSK-3 inhibition, has been shown to further increase neuralisation efficiency, and notch signalling inhibition by γ-secretase inhibitor compound-E has been found to accelerate neuralisation (Li et al., 2011;Du et al., 2015).
Here, we report the development of a fast induction protocol, aimed at rapid production of electrophysiologically active MNs for use in peripheral nerve mechanical injury studies. The aim was to rapidly produce adherent MN populations with high proportions of electrophysiologically active cells showing a mechanosensitive response to injury. Our protocol relies only on serum-free, adherent cultures, providing a facile method for the production of functional neurons. iPSCs monolayers are induced to NPCs in serum free conditions, by dual SMAD inhibition using dorsomorphin and SB431542, coupled with GSK-3 inhibition by CHIR-99021 and γ-secretase inhibition by compound-E. Following neuralisation, soluble RA, SHH, SAG, and purmorphamine are used to specify spinal MN fate. We assess MN maturity by whole cell patch clamping and calcium imaging, evaluating electrophysiological activity of differentiated neurons, and we characterise the mechanosensitivity of MNs generated by this protocol by measuring network activity modulation when subjected to uniaxial deformation. The differentiation protocol was validated in two iPSC lines independently reprogrammed from human dermal fibroblasts from unrelated control patients.

iPSC culture
All reagents were purchased from ThermoFisher and growth factors from Peprotech unless otherwise specified.
Two iPSC lines were used in this study. Cell culture at all stages was carried out on tissue culture plastic well plates and imaging dishes (IBIDI μ-dish), unless otherwise specified. Line SB-AD3-1 was generated by StemBANCC, Oxford. In brief, dermal fibroblasts were obtained from skin biopsy of healthy individuals following appropriate ethical consent. iPSC generation was performed using the CytoTune-iPS 2.0 Sendai Reprogramming Kit (Invitrogen). The cultures undergoing reprogramming were maintained at 37°C and 5% CO 2 in human ESC medium on feeder layers for two to three weeks or until colonies with typical human ESC morphology appeared. Individual colonies were mechanically dissected and plated onto fresh feeder plates for up to eight passages before being adapted to feeder free conditions which involved plating on Matrigel-coated plates (MG, Corning) with mTeSR1 media (Stem Cell Technologies), which was also the medium used in further experiments.
The second iPSC line (010S-1) was induced from a skin biopsy of a 18 year old female healthy subject, at the Highfield Unit, Warneford Hospital, Oxford. Skin fibroblasts were cultured from skin biopsy taken from the inner upper arm of the subject, and subsequently reprogrammed to iPSCs with CytoTune-iPS 2.0 Sendai Reprogramming Kit. The transfected cells were plated on mouse embryonic fibroblasts feeders to allow initial appearance of iPSC clones (about first 18 d). Picked clones were expanded, in a feeder-free manner, on Matrigel with mTeSR1 medium supplemented with 1% P/S. Colony 010S-1 was expanded to passage 13 and cryopreserved as the master stock. One vial from the master stocks was thawed in mTeSR medium on matrigel coated plates to perform quality control steps (data not shown) including normal speed of growth after thawing, CytoTune clearance assay by qPCR, flow-cytometry analysis of pluripotency markers NANOG and TRA-1-60, and SNP analysis of genomic integrity using an Illumina Human CytoSNP-12v2.1 beadchip array.
MG was diluted on ice in cold KnockOut-DMEM according to the batch-specific dilution factor, and added to well plates for 4-12 h before cell seeding. Thawed cells were plated in mTeSR supplemented with 1% Penicillin/Streptomycin and 1% RevitaCell (ROCK inhibitor and antioxidant compound) was added during the initial 24 h medium was replaced daily. Colonies were passaged by incubating for 3.5 min in versene (0.48 nM EDTA), lifted by washing with mTeSR medium using a large-bore pipette, and plated in mTeSR.

iPSC differentiation to NPC
The differentiation protocol ( Fig. 1) was performed using iPS line SB-AD3-1, and validated in line 010S-1 by evaluating differentiation end-points (neuron staining and electrophysiology).
NIM was replaced daily for six days, after which cells were dissociated with Accutase, and plated in NPC expansion medium, consisting of a 1:1 mix of KO-DMEM:F12 and NBM, supplemented with 1% P/S, 1% B27, 1% N2, 1% NEAA, 1% GlutaMAX, 0.1 mM L-AA, 10 ng/ mL bFGF and 10 ng/mL EGF. Cells were either expanded for banking, or kept for MN differentiation. One fifth of cells were retained at each passage for marker analysis. Cells were expanded for up to 4 passages.

Calcium activity
Calcium activity was imaged using Oregon Green 488 BAPTA-2 calcium indicator (Grienberger and Konnerth, 2012). Cells were incubated in a 4 μM dye solution in FluoroBrite-DMEM Imaging Medium for 30 min, washed twice in PBS, and further incubated in fresh FB-DMEM for 30 min. Cells were imaged using standard FITC filters at 10 fps for 4 min, and fluorescence intensities from individual segmented cells recorded in time. Intensity values were adjusted for fluctuations by subtracting average background, and expressed as = = I I t I t ( ) ( 0) .

Electrophysiology
A Digidata 1440A digitizer and a MultiClamp 700B amplifier piloted through pCLAMP 10 Software (all from Molecular Devices, CA) were used. Intracellular solution contained 140 mM KCl, 5 mM NaCl, 0.5 mM CaCl 2 , 2 mM MgCl 2 , 10 mM HEPES, 1 mM GTP, and 2 mM ATP, and was adjusted to 7.4 pH by KOH and 300 Osm L −1 by glucose addition. Bath solution contained 130 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM glucose, and 10 mM HEPES (all from Sigma-Aldrich), and was adjusted to 7.4 pH by NaOH and 290 Osm L −1 by glucose addition. Following gigaseal formation, whole-cell patch clamp was established. Initially, a higher range of voltages (−70 to +70 mV) was used to elicit a response and to establish conditions under which gigaseal formation occurred reliably. Ion currents were then recorded within a lower range to avoid cell injury by stimulation from −50 to +30 mV under voltageclamp recording, and action potentials (APs) were evoked in current clamp mode by stimulating current pulses from −50 pA to +30 pA.

Cell injury methods
Cell mechanosensitivity was tested by applying uniaxial strain to maturing cell networks. Previuosly, similar methods have been used to replicate in vitro traumatic brain injury, using rodent primary neurons (Magou et al., 2015;Pfister et al., 2003). A custom-built substrate deformation device, allowing simultaneous uniaxial cell deformation and high-frequency imaging was used to apply graded strains of up to 70% and 0.1 s -1 strain rate, with simultaneous network activity monitoring by calcium imaging. Cells were strained, and calcium flux was imaged for 4 min from cells kept in the strained state. MNs were seeded onto MG-coated custom-made rectangular deformable silicone wells with ultra-thin (50 μm) silicone culture surfaces measuring 10.0 mm × 6.5 mm at a density of 500 cells/mm 2 in maturation medium, and kept in culture two days before experiments.

Data analysis and statistics
Images were analysed using FIJI (Fiji Is Just ImageJ), and calcium imaging data was processed using a custom-written MATLAB code to filter data, extract and measure peaks. Statistical analysis was carried out using GraphPad PRISM v6. Fig. 1. Differentiation timeline. iPS colonies were neuralised by dual-SMAD, GSK-3 and γ-secratase inhibition for 7 days. Dissociated neural rosettes were expanded before differentiation to MN by caudalisation and ventralisation for 7 days.
MN specification induced a significant decrease in Pax6 and Sox1, whilst Nestin levels did not decrease significantly (Fig. 2). Olig2 expression after five days MN induction was significantly increased, with 88% ± 2% of cells positive for Olig2, indicating specification of MN fate. No significant Olig2 expression was detected in NPCs. Expression of Tuj1 also increased significantly, with 89% ± 4% of cells expressing β-III tubulin, showing induction to neural lineage (Fig. 3).

MN maturation
Cell morphology was altered during maturation, with reduced cell soma area and membrane spreading, and thin, extended projections connecting adjacent cells (Fig. 4 G). After 21 days in maturation medium, NPC marker expression was reduced significantly (Fig. 2), and neuron-specific structural proteins, Tuj1 (β-III Tubulin), NeuroFilament and MAP2, were expressed (Fig. 4). Synaptophysin was detected at cell junctions, suggesting synaptic connectivity. Positive staining for ChAT in somatas and HB9 in nuclei confirmed MN lineage for maturing cells (Fig. 4), with 73 ± 5% of cells positive for MN-specific HB9.

Efficiency of maturation to functional MNs
Neurons cultured in maturation medium for five days showed some calcium activity, with low spike magnitude and rate, and no bursting activity. Ion currents and induced APs were weak, indicative of successful differentiation but insufficient maturation (Fig. 5, A, B, D and E). Neurons cultured in maturation medium for 21 days showed faster rising spikes, higher spiking frequency, some bursting behaviour, defined ion currents, and induced APs, all indicative of neurons successfully maturing towards active cells (Fig. 5). After 21 days maturation, 75% ± 3% of cells (N = 8 independent culture dishes) were active, significantly more than after day 5 of maturation (Fig. 5, C). The average magnitude of calcium spikes, measured as a percentage of the baseline value, was 4.8% at day 5, and 15.2% at day 21 of maturation. The ratio of rise to fall time of peaks was 0.42 and 0.29 and the average rise slope was 0.035 and 0.095, at day 5 and day 21 respectively.

iPSC-derived MNs form functional connections
Network connectivity at maturation day 21 was assessed from calcium imaging recordings. Spontaneous oscillations in intracellular calcium have been widely used as an indirect reporter of neural activity, as well as of functional synaptic connectivity (Grienberger and Konnerth, 2012;Wang and Gruenstein, 1997). Detection of concurrent calcium  spikes in distinct neuron somatas was used as a marker of connectivity, as previously used in whole-brain network reconstruction (Mann et al., 2017). Maturing neurons successfully formed synaptic connections. Connected cells showed synchronised action potential firing, traceable along the connecting axon (Fig. 5). Cell connectivity was observed in all samples of maturing MNs tested (N = 8), indicating that neurons formed functional networks. No connectivity was observed after 5 days of maturation, indicating network maturation in time. To further confirm this, positive synaptophysin puncta staining was observed in neural cultures after 21 days in maturation medium (Fig. 4).

Validation of differentiation to maturing MN
Successful differentiation was validated by repeating the protocol using iPSC line 010S-1, and evaluating end-point functionality and neural marker expression as 010S-1 cells differentiate to motor neurons. After 21 days in maturation medium, cells show extensive neurite networks, and electrophysiological activity indicative of maturation. Voltage-clamp experiments show induced action potentials (Fig. 6 (A)), including AP train firing (Fig. 6 (B)). Calcium imaging analysis shows functional network connectivity, as previously shown in cell line SB-AD3-1 (Fig. 6 (C)). Marker staining shows positive expression of NPC markers after seven days neuralisation (Fig. 6 (D-G)), and positive expression of MN markers after 21 days maturation (Fig. 6 (H-K)). Qualitatively, cells tend to become more confluent during the same culture time, and a higher percentage of cells starts expressing Tuj1 early, compared to the same timepoint staining of NPC derived from SB-AD3-1 (Fig. 6 (E) vsFig. 3 (C)).

Deformation injury affects network activity
Uniaxial stretch of cultured neurons has been previously used as a model for traumatic brain inury (TBI), where primary embryonic rat hippocampal neurons have been shown to respond to stretch by a marked decrease in spontaneous activity (Magou et al., 2015). However, inter-species differences in response to stretch injury possibly decrease the translational value of these results, and provide motivation to develop human iPSC-based in vitro models (Sherman et al., 2016). In order to test the response of iPSC-derived MN spontaneous calcium oscillations to uniaxial stretch, MNs matured for 21 days were subjected to increasing levels of stretch (25%, 45% and 75% strain at 0.1 S -1 strain rate), during which calcium imaging recordings were performed. Network activity was measured by calcium imaging for 4 min before deformation, and at each sequential deformation level. The number of active cells, measured as the percentage of cells within the field of view exhibiting spontaneous calcium spikes, decreases linearly with applied deformation, and does not recover immediately after stretch release (Fig. 7). A significant decrease in active cells was observed at each level, both compared to the un-stretched (control) condition, and compared to the previous stretch level (incremental effect).

Discussion
The establishment of simple, reproducible protocols for successful differentiation of human iPSC is paramount for facilitating more widespread use of iPS-derived cells for human in vitro MN injury modelling. Here, we aimed to develop a simple, rapid protocol for the efficient directed differentiation of electrophysiologically active, MNs from human iPSCs, aimed at producing active cultures for short-term neural injury studies. Efficiency of differentiation was measured by specific marker expression, and number of electrophysiologically active cells, measured in synaptically active networks, indicating cell populations suitable for neural injury experiments. The mechanosensitivity of cells was evaluated by exposing spontaneously active networks to uniaxial strain, and recording changes in network activity.
Adherent culture differentiation by dual-SMAD inhibition eliminates the need for EB formation, a complex and inefficient culture technique widely used in iPSC neural differentiation protocols (Shimojo et al., 2015;Maury et al., 2015;Salimi et al., 2014;Zhou et al., 2010), and increases efficiency of neuralisation (Faravelli et al., 2014;Sances et al., 2016;Chambers et al., 2009). Here, we combine dual-SMAD inhibition with wnt activation and γ-secretase inhibition, to further increase efficiency, resulting in 83% PAX-6 and 76% Sox1 positive NPCs after 6 days. Previous use of four neuralisation factors involved either EB formation or suspension culture step following neuralisation, and did not demonstrate efficiency in terms of electrophysiological maturation of cells and networks (Shimojo et al., 2015;Du et al., 2015). NPCs are induced to MNPs by dual HH activation, through addition of Fig. 5. Immature MNs showed low magnitude, low frequency calcium activity. Insert shows characteristic peak shape (A). Cells cultured in maturation media for 21 days showed higher frequency firing, with some bursting behaviour. Inset shows characteristic peak shape (B). Maturation for 21 days induced 75% of cells to active neurons, compared to 7% at day 5 (C, n = 8, **** = p < 0.0001 by unpaired t-test). Patch clamping current-clamp recordings showed immature neurons at day 5 (D) and maturing neurons at day 21 (E). Neurons formed connected networks, with signals propagating down axons between cells (F).
soluble SHH and direct smoothened activators purmorphamine and SAG. This results in highly pure MNP populations, with 88% of cells positively expressing OLIG2, in line with the highest values reported in literature (Hu and Zhang, 2010;Karumbayaram et al., 2009).
After 21 days in maturation medium, 89% of cells positively expressed Tuj1, 73% of cells positively expressed MN-specific HB9, and 75% of cells differentiated to electrophysiologically functional MNs within five weeks from iPSC thawing. Other protocols validated by electrophysiology range between 20 and 70% efficiency, and between 5 and 18 weeks duration (Faravelli et al., 2014;Sances et al., 2016). After 21 days of maturation, we show functional network formation by calcium imaging, positive expression of punctuated synaptophysin suggesting synapse formation, and positive expression of motor-neuron specific factors HB9 and cHAT, combined with neural-specific markers Tuj1, MAP2 and NeuroFilament. Moreover single-cell electrophysiology shows how, after 21 days, induced APs can be recorded, and waveform analysis of calcium spikes indicates maturation, with an increase in spike magnitude, decrease in rise and fall time, and increase in rise rate (Biffi et al., 2013). The same protocol was validated in 010S-1 cell line, a human dermal fibroblast derived iPSC line, unrelated to the SB-AD3-1  With increased applied substrate deformation, the percentage of active cells decreases. n = 7 independent samples. **** = p < 0.0001, ** = p < 0.01, * = p < 0.05, by one way paired ANOVA with Tukey's multiple comparisons. Cell bodies and neurites (b) are subjected to deformation (c). Arrow indicates direction of stretch. Scale bars = 20 μm. line, showing electrophysiological characteristics after 21 days maturation, and expression of neural progenitor and neural markers, confirming successful differentiation. For both cell lines, videos of calcium activity can be seen in supplementary material V1 and V2.
Maturing MNs forming spontaneously active networks were stretched to evaluate cell mechanosensitivity. Uniaxial strain can be used to model the mechanical environment of the peripheral nerve under physiological and traumatic conditions, where movement induces elongation along the nerve axis deforming tissues and neural fibres, and can be used to model peripheral nerve traumas such as increased strains due to carpal tunnel syndrome, joint dislocation, or iatrogenic aberrant limb positioning (Bueno and Shah, 2008;Topp and Boyd, 2006). A significant decrease in cells exhibiting spontaneous calcium spiking was observed during deformation, and no recovery of activity was measured following stretch release. This is consistent with results reported for primary embryonic and neonatal cortical and hippocampal rat neurons subjected to stretch (Magou et al., 2015;Prado et al., 2005;Goforth et al., 2011), and confirms that human iPSC-derived MNs are suitable for use in neural injury experiments. Videos of calcium activity in networks before and after stretch can be seen in supplementary materials V3 and V4. We present the first human MN-specific in vitro injury model, focussing on electrophysiological alterations caused by mechanical deformations, to provide a human relevant cell model for the study of mechanical deformation and injury.
Using MN-only cultures allows the functional response of neurons to stretch to be isolated. To further develop this as an in vitro model of human PNI, co-culture with Schwann cells could provide myelination, a key component of the nervous system which is disrupted by mechanical trauma (Sun et al., 2012), and that enhances neuron maturation and long-term survival. To maintain the premise of physiological relevance to human PNI, human stem cell-derived Schwann cells should be used. Although human iPSC-derived neurons have been myelinated by rodent primary Schwann cells (Clark et al., 2017), and human fibroblasts have been successfully reprogrammed to myelinating Schwann cells (Mazzara et al., 2017), to the best of our knowledge no co-culture model exists that would allow the development of a stem-cell derived human myelinated PNI injury model.

Conclusions
In summary, we present a protocol to facilitate the production of electrophysiologically active MNs from human iPSCs, for short-term cultures, aimed at in vitro human neural injury modelling. In 5 weeks from iPSC culture establishment, we show efficient differentiation of electrophysiologically active neurons, with 75% of cells showing spontaneous calcium activity, and expressing neural and MN markers. Compared to other protocols showing electrophysiological maturity, this represents an improvement in terms of time as well as in efficiency. We show that human iPSC-derived MNs are suitable for in vitro neural injury modelling by measuring functionality alterations in networks exposed to uniaxial strain. This protocol improves efficiency of rapid MN directed differentiation, producing populations of highly active neurons expressing MN specific markers within 5 weeks, providing an accessible, easily implementable alternative to animal cells. Despite maintaining advantages in terms of cost, complexity and range of applications, concerns raised over the translational value of rodent models in neurotrauma, including differences in nociception (Rostock et al., 2017) and response to traumatic injury (Povlishock et al., 1994), indicate the importance of developing in vitro alternatives using human cells.