The effect of rho kinase inhibition on morphological and electrophysiological maturity in iPSC-derived neurons

Abstract

Induced pluripotent stem cell (iPSC)-derived neurons permit the study of neurogenesis and neurological disease in a human setting. However, the electrophysiological properties of iPSC-derived neurons are consistent with those observed in immature cortical neurons, including a high membrane resistance, depolarized resting membrane potential, and immature firing properties, limiting their use in modelling neuronal activity in adult cells. Based on the proven association between inhibiting rho kinase (ROCK) and increased neurite complexity, we sought to determine if short-term ROCK inhibition during the first 1–2 weeks of differentiation would increase morphological complexity and electrophysiological maturity after several weeks of differentiation. While inhibiting ROCK resulted in increased neurite formation after 24 hours, this effect did not persist at 3 and 6 weeks of age. Additionally, there was no effect of ROCK inhibition on electrophysiological properties at 2–3, 6, or 12 weeks of age, despite an increase in evoked and spontaneous firing and a more hyperpolarized resting membrane potential over time. These results indicates that while there is a clear effect of time on electrophysiological maturity, ROCK inhibition did not accelerate maturity.

Introduction

Induced pluripotent stem cell (iPSC) derived neuron technology provides a tractable model system to study the neurobiological consequences of genetic disorders in a human cell setting. Unlike embryonic stem cells, iPSCs allow the comparison of neurotypic control cells with those from a patient with a neurological disease. Since the first description of iPSCs (Takahashi et al. 2007), there has been great impetus to use the approach to model human neurological disorders including schizophrenia, autism, and epilepsy (Bellin et al. 2012; Brennand et al. 2015).

Despite the identification of drug-sensitive, disease-relevant phenotypes, the electrophysiological behavior of iPSC-derived neurons appears immature relative to the neural activity observed in rodents or human resected tissue (Bradford and McNutt 2015; Randall 2016). A depolarized resting membrane potential and high membrane resistance suggest that iPSC-derived neurons are comparable to neurons resected from human fetal tissue. Moreover, iPSC-derived neurons often produce low-amplitude, abortive action potentials that are similar to those frequently observed in the human fetal cerebral cortex (Livesey et al. 2016; Moore et al. 2009; Pre et al. 2014; Song et al. 2013). More thorough electrophysiological characterization has demonstrated that iPSC-derived neurons often develop, albeit slowly, more mature functional properties when cultured for months; nonetheless, these neurons remain immature relative to adult human neurons (Belinsky et al. 2014; Chinchalongporn et al. 2015; Lam et al. 2017; Livesey et al. 2016; Moe et al. 2005; Nicholas et al. 2013; Pre et al. 2014; Weick 2016). Moreover, as long-term culturing procedures are labor-intensive and costly, there exists great interest to accelerate neuronal maturity.

Neuronal branching complexity is associated with more mature electrophysiological properties such as robust firing activity (Bardy et al. 2016). Neurite branching is in part regulated by the Rho kinase pathway (Luo 2000). ROCK is a member of the Ras family of GTPases, which critically regulate cell adhesion and cytoskeletal rearrangement (Hall 1998). ROCK is activated by RhoA and phosphorylates myosin light leading to myosin activation and contraction of actin filaments, ultimately resulting in neurite retraction (Amano et al. 2000). In contrast, activation of the Rac1 pathway polymerizes actin, leading to neurite extension and synapse formation (Machesky and Hall 1997) (Fig. 1a).

Fig. 1.

Experimental setup and assessment of Y-27632 short-term effects. a The RAC1 and RHOA pathways regulate cytoskeletal rearrangement to cause neurite extension and retraction. ROCK is a substrate of RHOA, and can be specifically blocked by the drug Y-27632. This decreases myosin phosphorylation leading to subsequent neurite retraction. b. Neurotypic iPSCs are differentiated into a line of neural progenitor cells (NPCs). These are plated on coverslips and switched to neuron media (see methods). Neurons will be treated with or without Y-27632 in neuron media for the first 1–2 weeks after plating. c-e. Neurons were treated with 0, 5,10, 25, or 50 μM Y-27632 for 24 hours, stained with βIII-tubulin and DAPI and were automatically analyzed for neurite properties. The average number of neurites per cell (c, p<0.0001; n=42,747; 51,526; 56,953; 61,661; 57,269), average neurite length (d, p<0.0001; n=22,022, 34,613, 41,918, 44,864, 57,269), and average number of branch points per neurite (e, p<0.0001; n=22,022, 34,613, 41,918, 44,864, 57,269) are significantly increased with all 4 tested concentrations compared to control cells. f. Representative images of 24 hour post- differentiation neurons treated with (f’, f’’’) or without (f, f’’) 10 μM Y-27632, stained with neurite marker βIII- tubulin and DAPI. Bottom images (f’’, f’’’) are magnified views of insets shown in top images (f, f’). g. Representative image (g’) and quantification (g’’) of western blot of NPCs cultured with or without 10 μM Y-27632 for 24 hours and probed for phosphorylated myosin regulatory light chain (pMLC) and tubulin as a loading control. Levels of pMLC relative to tubulin are decreased in Y-27632-treated cells (n=3,3; p<0.0001). p<0.05 **p<0.001 by 1-way ANOVA with post-doc HSD; Data plotted as mean ± standard error of the mean (SEM); see Table 1 for complete statistics

ROCK inhibitors such as Y-27632 can be used to enhance the survival of embryonic stem cells and iPSCs by protecting against dissociation-induced apoptosis (Kurosawa 2012; Watanabe et al. 2007). ROCK inhibition enhances neurite outgrowth in multiple in vitro contexts, including cultured mouse neural stem cells (Gu et al. 2013; Jia et al. 2016), human N-TERA-2 cells (Lingor et al. 2007; Roloff et al. 2015), human PC12 cells (Minase et al. 2010; Yang et al. 2010) and cultured dorsal root ganglion neurons from chicks and mice (Fournier et al. 2003; Yang et al. 2010). In vivo, ROCK inhibition promotes axonal outgrowth following CNS injury and contributes to injury recovery (Chan et al. 2005; Fournier et al. 2003; Lingor et al. 2007; Minase et al. 2010), leading to speculation that ROCK inhibition may have clinical potential in promoting nerve regeneration (Kubo et al. 2007; Tan et al. 2011).

While the effects of ROCK inhibition on neuron morphology are well-established, little information exists regarding how ROCK inhibition affects the electrophysiological properties of neurons. Given the correlation between morphological complexity and electrophysiological maturity seen in iPSC-derived neuron cultures, we sought to assess how ROCK inhibition during terminal differentiation of iPSC-derived neurons effects both morphological and electrophysiological properties. To this end, we plated iPSC-derived neurons in a 2D culture system and measured neurite branching, membrane properties, and evoked and spontaneous firing activity of neuron cultures at different time points with and without Y-27632 treatment in early differentiation (Fig. 1b).

Materials and Methods:

Astrocyte Culture.

Human cortical astrocytes were obtained from Sciencell (Carlsbad, CA, USA) and cultured in astrocyte media supplemented with astrocyte growth serum, FBS, and pen/strep (Sciencell, Carlsbad, CA, USA). Astrocytes were passaged and split onto cell culture plates coated with 0.1% gelatin (Bio-Rad, Hercules, CA, USA). All neurons were cultured on coverslips on top of astrocyte-coated plates excluding the 24-hour morphology experiments.

Generation of NPC lines.

iPSCs (line 9319 from Brennand et al, 2011) were maintained under standard conditions and dissociated from plates using collagenase type IV (Invitrogen, Carlsbad, CA, USA) to form floating embryoid bodies, which were cultured in DMEM/F12 + glutamax (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 1% N2 (Thermo Fisher Scientific, Waltham, MA, USA), 2% B27 without vitamin A (Thermo Fisher Scientific, Waltham, MA, USA), SB 431542 (Tocris, Bristol, UK) and LDN193189 (Stemgent, Cambridge, MA, USA)). After 7 days, the embryoid bodies were re-plated onto polyornithine (Sigma- Aldrich,St. Louis, MP, USA) and laminin (Thermo Fisher Scientific, Waltham, MA, USA) coated plates and cultured in the same media for an addition 7 days until neural rosettes formed. Neural rosettes were dissociated using neural rosette selection reagent (Stemcell, Vancouver, Canada) and replated onto Corning™ matrigel (Fisher Scientific, Hamptom, NH, USA))-coated plates to form populations of neural progenitor cells (NPCs).

Neuron Differentiation:

NPCs were grown until confluent and then plated at a density of 200,000 cells/well of a 6-well plate for 24-hour neurite branching analysis, 100,000 cells per well of a 12-well plate for 3 and 6 week morphological analysis, and 50,000, 75,000, or 100,000 cells per well of a 12-well plate for electrophysiological analysis. For electrophysiological measurements, density did not have a significant effect on any measure (data not shown) so density groups were combined. For 24-hour neurite analysis, NPCs were changed to neuron media (see below) for 24 hours with Y-27632 (Selleck Chemicals, Houston, TX, USA) added as indicated, then fixed for immunohistochemistry. For 3- and 6-week morphological and all electrophysiological assays, NPCs were plated onto polyornithine/laminin-coated 18 mm coverslips and kept in NPC media for 48 hours before being transferred onto astrocyte-coated plates in neuron media consisting of DMEM/F12 + glutamax supplemented with 1% N2 (Thermo Fisher Scientific, Waltham, MA, USA), 2% B27 (Thermo Fisher Scientific, Waltham, MA, USA), 20 ng/mL BDNF (Shenandoah Biotechnology, Warwick, PA, USA), 20 ng/mL GDNF (Shenandoah Biotechnology, Warwick, PA, USA), cyclic AMP (Sigma-Aldrich, St. Louis, MO, USA), ascorbic acid, pen/strep (Thermo Fisher Scientific, Waltham, MA, USA), and laminin (Thermo Fisher Scientific, Waltham, MA, USA) (morphology experiments only). 10 µM Y-27632 (Selleck Chemicals, Houston, TX, USA) was added for one to two weeks where indicated. Neurons were kept in culture for 3, 6, or 12 weeks as indicated. Media was changed every 5–7 days.

Western blot:

NPCs were cultured for 24 hours with or without 10 µM Y-27632 before being dissociated with Accutase (Innovative cell technologies, San Diego, CA, USA) , pelleted, and lysed in RIPA buffer consisting of 50 mM Tris HCl (pH 8.0), 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, and 0.1% SDS, 10% NaF, and 1 tablet of protease inhibitor (Sigma, 11836145001). Protein was quantified using the Pierce BCA assay kit (ThermoFisher Scientific #23225) added to 2x Laemmli buffer (4% SDS, 10% 2-mercaptoethanol, 20% glycerol, .004% bromophenol blue, 0.125M Tris HCl, pH 6.8) and boiled for 10 minutes at 95 degrees. For immunoblotting, 20 ug of total cell lysate was resolved by 4–15% Mini-PROTEAN precast gradient gel (Biorad #4561086). Proteins were transferred to a PVDF membrane for 90 minutes at 100V and incubated at room temperature for 2 hours in 3% BSA in TBS-T with antibodies anti-myosin pS19/pS20 (Rockland Immunochemicals, Pottstown, PA, USA; Cat # 600–401-416) and anti-alpha tubulin (Thermo Scientific, Cat #: MS581P1). Blots were visualized with SuperSignal™ West Pico PLUS Chemiluminescent Substrate(ThermoFisher Scientific, #34577) and a Bio-Rad ChemiDoc imaging system.

24-hour neuron staining and morphological analysis:

After 24 hours in neuron media containing 0, 5, 10, 25, or 50 µM Y-27632, cells were fixed with 4% formaldehyde with 4% sucrose. Cells were permeabilized for 2 minutes in 0.1% Triton X-100 (Sigma Aldrich, St. Louis, MO, USA) in PBS then incubated overnight at 4º with a primary antibody for βIII-tubulin (1 µg/mL, R&D systems, Minneapolois, MN, USA; Cat #: MAP 1195) in 2% goat serum-PBS, then incubated at room temperature for one hour with a secondary antibody (4µg/mL, Thermo Fisher Scientific, Waltham, MA, USA; Alexa Fluor goat anti-mouse 568) in 2% goat serum-PBS. Glass coverslips were mounted onto the wells using Vectashield Mounting medium with DAPI (Vector labs, Burlingame, CA, USA). Images were taken on a EVOS FL Auto 2 (Thermo Fisher Scientific, Waltham, MA, USA) at 10X magnification and neurites were assessed using the CellInsight CX5 High Content Screening Platform (Thermo Fisher Scientific, Waltham, MA, USA) and quantified for the number, length, and branching of neurites per cell (100 fields per well).

3- and 6-week neuron staining and morphological analysis:

After 1 week of differentiation, neurons were transfected with an AAV inducing mCherry fluorophore expression driven by a synapsin promotor (AAV-hSyn-mCherry, Deisseroth in-stock AAV vectors, UNC vector core, Chapel Hill, NC, USA). At 3 and 6 weeks of age, cells were fixed in 4% PFA for 15 minutes and transferred into PBS until the time of staining. Cells were washed in 0.5% triton/PBS followed by permeabilization in 3% bovine serum albumin (BSA, Thermo Fisher Scientific, Waltham, MA, USA) in 0.1% triton in PBS (PBS-T) for 1 hour at room temperature. Cells were left overnight at 4°C in a primary antibody (1:1000, chicken anti-mcherry, Novus Biologicals Littleton, CO, USA; Cat # NBP2–25158). The following day, cover slips were washed with PBS-T before applying a secondary antibody (1:1000, goat anti-chicken 555, Thermo Fisher Scientific, Waltham, MA, USA Alexa Fluor A-21437) for one hour at room temperature. Cells were washed again with PBS-T and then incubated at room temperature in DAPI (1:1000, Thermo Fisher Scientific, Waltham, MA, USA) for 5 minutes before being transferred into water. Coverslips were mounted using prolong gold antifade mountant (Thermo Fisher Scientific, Waltham, MA, USA). Cells were then imaged at 63x using the Neurolucida system (MBF Bioscience, Williston, VT) with an Axioskope microscope driven stage and an AxioCam MRc camera (Zeiss Microscopy, Oberkochen, Germany). Neurites were manually traced and analyzed using Neurolucida and Neurolucida Explorer software.

Whole-cell patch-clamping.

Neurons were recorded in artificial cerebrospinal fluid (ACSF) containing 0.25 mM potassium chloride, 1 mM glucose, 12.6 mM sodium chloride, 0.125 mM sodium phosphate, 0.1 mM magnesium sulfate, 0.2 mM calcium chloride, and 2.6 mM sodium bicarbonate, adjusted with water to an osmolarity of ~300. Whole-cell patch clamping was performed with pipettes filled with potassium gluconate internal containing 100 mM potassium-gluconate, 9 mM magnesium chloride, 13 mM potassium chloride, 0.07 mM calcium chloride, 10 mM Hepes, 10 mM EGTA, 2 mM ATP, and 0.5 GTP adjusted to an osmolarity of ~285.

Resting membrane potential was assessed by recording a 50-second long trial in current clamp with no applied current and averaging the measured membrane potential. Results shown are adjusted for liquid junction potential. Membrane resistance was assessed by voltage clamping the cell at −70 mV and applying a 50 ms, 15 mV square pulse during each of 100 trials, and then dividing the averaged resultant steady state current by the voltage step Evoked and spontaneous activity were recorded in current clamp with a holding current to maintain ~−65 mV resting membrane potential. The cell was run through a ten-trial protocol, in which every trial contained a one second, +25 pA depolarizing pulse to evoke firing activity.

Electrophysiology analysis:

All electrophysiology recordings were performed and analyzed using pClamp™ and Clampfit™ software (Molecular Devices, LLC, San Jose, CA, USA). Action potentials were manually measured from the trough of each spike to the peak (see Fig. 2c); events greater than 10 mV in amplitude were considered action potentials. Action potentials that occurred within the 10 second +25 pA pulse were counted as evoked activity, while those occurring at other times in the 10 second trials were considered spontaneous.

Fig. 2.

Morphological effects of Y-27632 treatment at 3–6 weeks of age. a Representative images of neurons at 3 weeks of age stained for mcherry (a, a’’) and corresponding cell traces (a’, a’’’) with (a’’, a’’’) and without (a, a’) treatment with 10 µM Y-27632. b The average soma size of neurons increased significantly over time (p=0.0043) but did not vary based on treatment (p=0.8303). c There is no significant effect of of age (p=0.0730) or treatment (p=0.6132) on the average number of neurites per cell. d There is no significant effect of of age (p=0.3371) or treatment (p=0.4833) on the average neurite length, but there is an interaction effect (p=0.0416). e There is no significant effect of of age (p=0.7137) or treatment (p=0.6222) on the average number of branch points per neurite, but there is an interaction effect (p=0.0442). All graphs: Control n=17, 13; Y-27632: n=14, 15

p<0.05 **p<0.001 by 2-way ANOVA with post-hoc HSD; Data plotted as mean ± standard error of the mean; see Table 2 for complete statistics

Statistics:

Unless otherwise indicated, p-values are the result of a 1- or 2-way ANOVA for unbalanced design followed by post-hoc Tukey HSD corrections for individual group comparisons and data is shown as mean ± standard error of the mean (SEM). For action potential amplitude, significance was assessed using a Wilcoxon signed-rank test. Statistics, data analysis and figure generation were performed using Matlab (Natick, MA, USA) and CorelDRAW (Corel, Ottawa, Canada).

Results:

Short-term ROCK inhibition increases neurite formation during the first 24 hours of neuronal differentiation.

To determine if ROCK inhibition increases initial neurite formation in iPSC-derived neuron cultures, neural progenitor cells (NPCs) were plated for terminal differentiation in neuron media containing 0, 5, 10, 25, or 50 µM Y-27632. After 24 hours, cells were fixed and stained for DAPI and β-III-Tubulin (Fig. 1d). Automated morphological analysis using the CellInsight CX5 Screening Platform revealed that all of the treatments increased the number of neurites per cell (p<0.0001, see Table 1 for full statistics), average neurite length (p<0.0001), and average number of branch points per neurite (p<0.0001) compared to control cells (Fig. 1c–f’’’).

Table 1.

Statistics for 24-hour morphology experiments

Measurement Associated Figure	Group	N	Mean	SEM	Statistics
					One way ANOVA p-value	F (df)	Post-hoc Tukey HSD p-value
							Group	vs. Control	vs. 5 µM
Number of Neurites Fig. 1e	Control	42,747	0.75	0.0035	<0.0001 **	1,762.92 (4)	Control
	5 µM	51,526	1.00	0.0036			5 µM	<0.0001 **
	10 µM	56,953	1.15	0.0037			10 µM	<0.0001 **	<0.0001 **
	25 µM	61,661	1.14	0.0038			25 µM	<0.0001 **	<0.0001 **
	50 µM	57,269	1.17	0.0038			50 µM	<0.0001 **	<0.0001 **
Average neurite length (µm) Fig. 1f	Control	22,002	38.64	0.2038	<0.0001 **	241.17(4)	Control
	5 µM	34,613	46.26	0.2446			5 µM	<0.0001 **
	10 µM	41,918	50.19	0.2652			10 µM	<0.0001 **	<0.0001 **
	25 µM	44,864	49.74	0.2666			25 µM	<0.0001 **	<0.0001 **
	50 µM	42,537	51.50	0.2757			50 µM	<0.0001 **	<0.0001 **
Average branch points per neurite Fig. 1g	Control	22,002	0.5325	0.0054	<0.0001 **	144.13 (4)	Control
	5 µM	34,613	0.6941	0.0065			5 µM	<0.0001 **
	10 µM	41,918	0.7531	0.0067			10 µM	<0.0001 **	<0.0001 **
	25 µM	44,864	0.7658	0.0067			25 µM	<0.0001 **	<0.0001 **
	50 µM	42,537	0.7856	0.0068			50 µM	<0.0001 **	<0.0001 **

Rock inhibition by Y-27632 is expected to block neurite retraction, in part, through diminished myosin phosphorylation. Western blot analyses demonstrated that, relative to controls, NPCs treated for 24 hours with 10 µM Y-27632 had significantly decreased levels of phosphorylated myosin regulatory light chain (pMLC) relative to tubulin (Fig. 1g–g’; Control: pMLC::tubulin=0.746 ± 0.0130, n=3; Y-27632: 0.5651 ± 0.0104, n=3, 3; t-test p<0.0001), thereby confirming that the inhibitor downregulates phosphorylation of known ROCK substrates in this cell line (Newell-Litwa et al. 2015).

Short-term ROCK inhibition does not cause a prolonged change in neurite morphology.

To evaluate if early ROCK inhibition produced lasting morphological changes, neurons were cultured with or without 10 µM Y-27632 in the neuron media for the first two weeks of differentiation and then examined at two later time points. Cell morphology was examined by labeling cells with mCherry via viral transfection at one week of age. Cells were fixed at either three or six weeks of age, and manually traced with Neurolucida software (Fig. 2a–a’’’).

Soma size of iPSC-derived neurons increases with time in culture and correlates well with increased functional maturity (Bardy et al. 2016; Nicholas et al. 2013). Consistent with these findings, we observed that soma size was larger in six week-old cells compared to three week-old cells (Fig. 2b, p=0.0043, 3 week: 214.67 ± 16.42 µm², n=31; 6 week: 296.60 ± 21.70 µm², n=28; see Table 2 for 3 and 6 week morphology statistics). Nonetheless, Y-27632 treatment did not affect soma size (p=0.8303).

Table 2.

Statistics for 3 and 6 week morphology experiments

Measurement Associated Figure	Group	Age (wk)	N	Mean	SEM	Statistics
						Comparison	2-way ANOVA p-value	F (df)	Post-hoc Tukey HSD p-value
Soma Size (µm2) Fig. 2b	Control	3	17	217.36	26.27	Control vs. Y-27632	0.8303	0.05 (1)
		6	13	287.08	19.26
		All	30	247.57	22.69
	Y-27632	3	14	211.40	18.55	Age	0.0043*	8.85 (1)	3 vs.6	0.0043*
		6	15	304.85	17.93
		All	29	259.74	17.74
	All 3 wk		31	214.67	16.42	Interaction	0.6669	0.19 (1)
	All 6 wk		28	296.60	21.70
Number of neurites Fig. 2c	Control	3	17	4.35	0.27	Control vs. Y-27632	0.6132	0.26 (1)
		6	13	5.69	0.34
		All	30	4.93	0.47
	Y-27632	3	14	4.43	0.33	Age	0.0730	3.34(1)
		6	15	5.07	0.32
		All	29	4.76	0.27
	All 3 wk		31	4.39	0.21	Interaction	0.5195	0.42 (1)
	All 6 wk		28	5.36	0.51
Average neurite length (µm) Fig. 2d	Control	3	17	234.96	45.10	Control vs. Y-27632	0.4833	0.50 (1)	C3 vs. Y3	0.7495
		6	13	413.93	46.97				C6 vs. Y6	0.2284
		All	30	312.51	42.67
	Y-27632	3	14	315.85	45.26	Age	0.3371	0.94 (1)	C3 vs. C6	0.1447
		6	15	250.35	43.72				Y3 vs. Y6	0.8600
		All	29	281.97	42.12
	All 3 wk		31	271.49	32.40	Interaction	0.0416*	4.35 (1)	C3 vs. Y6	0.9974
	All 6 wk		28	326.30	51.66				R3 vs. C6	0.6683
Average number of branch points per neurite Fig. 2e	Control	3	17	1.84	0.35	Control vs. Y-27632	0.6222	0.25 (1)	C3 vs. Y3	0.2604
		6	13	3.08	0.70				C6 vs. Y6	0.7030
		All	30	2.38	0.35
	Y-27632	3	14	3.14	0.67	Age	0.7137	0.14 (1)	C3 vs. C6	0.3209
		6	15	2.28	0.65				Y3 vs. Y6	0.6353
		All	29	2.70	0.38
	All 3 wk		31	2.43	0.37	Interaction	0.0442*	4.24 (1)	C3 vs. Y6	0.9191
	All 6 wk		28	2.65	0.36				Y3 vs. C6	0.9998

The average number of neurites per cell did not change based on drug treatment (p=0.6132) or age group (p=0.0730) (Fig. 2c). While neither drug treatment (p=0.4833) nor age (p=0.3371) alone caused a change in average neurite length, there was a significant interaction between the two factors based on a multi-way ANOVA (p=0.0416; control n=17, 13; Y-27632 n=14, 15) (Fig. 2d). A similar pattern was observed with the average number of branch points per neurite: neither treatment (p=0.6222) nor age alone (p=0.7137) affected the parameter, but a significant interaction between the two was observed (p=0.0442) (Fig. 2e).

The significant p-value for the interaction of the two variables, condition and age, indicates that there may be a difference in how the values change with age based on the condition, even though none of the individual comparisons yielded a significant change. For example, while neither neurite length nor number of branch points per neurite showed any significant differences between individual age and drug treatment groups, in both cases, control cells showed a slight increase over time (neurite length: 234.96 ± 45.10 µm to 413.93 ± 46.97 µm; number of branch points: 1.84 ± 0.35 to 3.08 ± 0.70), while Y-27632-treated neurons showed a slight decrease over time (neurite length: 315.85 ± 45.26 µm to 250.35 ± 43.72 µm; number of branch points: 3.14 ± 2.51 to 2.28 ± 0.65). While neither trend was statistically significant, this result suggests that cultures may develop neurite length and complexity differently over time depending on whether ROCK inhibition had been applied early on in development.

Inhibition of ROCK does not affect electrophysiological properties of neurons.

We next assessed electrophysiological properties of iPSC-derived neurons treated with 10 µM Y-27632 for the first week of differentiation. Electrophysiological measurements of iPSC-derived neurons were performed at three time points (2–3, 6, and 12 weeks). While Y-27632 had no effect on either resting membrane potential (Fig. 3a), or membrane resistance (Fig. 3b), the resting membrane potential of both treatment groups progressively hyperpolarized over time (p=0.0181; control: n=21, 29, 11; Y-27632: n=29 25 14; see Table 3 for full electrophysiology statistics), increasing from −45.51 mV ± 0.95 at the 2–3-week time point to −51.34 mV ± 2.33 at 12 weeks. Membrane resistance tended to decrease with age (p=0.0938; control: n=21, 29, 11; Y-27632: n=29, 25, 14), with all groups averaging between 1 and 2 GΩ; however, this trend may partially result from the increased soma size observed over time (see Fig. 2b). The number of action potentials did not differ between control and Y-27632-treated neurons at any time point for either evoked or spontaneous activity (p=0.7548, 0.6897; control: n=21, 29, 11; Y-27632: n=29, 25, 13; Fig. 3d, g). However, the number of action potentials in both evoked (p<0.0001) and spontaneous (p<0.0001) activity increased significantly with age (Fig. 3d, g). The increase in average number of action potentials per trial is also reflected by an increase in the number of responsive cells (cells that fired at least one during the testing period) over time (Fig. 3j, k).

Fig. 3.

Electrophysiological effects of Y-27632 treatment at 2–3, 6, and 12 weeks of age. a The resting membrane potential (RMP) of neurons does not differ between treatment groups (p=0.4948), but does become more hyperpolarized between 2–3 weeks of age and 6/12 weeks of age (p=0.0181; post-hoc HSD: 3 vs 6 week p=0.0046; 3 vs 12 week p=0.0426; Control: n= 25, 28, 12; Y-27632: n=35, 27, 18). b Membrane resistance does not change significantly by treatment group (p=0.5594), or age (p=0.0938; Control: n=25, 28, 12; Y-27632: n=35, 27, 15). c Representative traces at 2 weeks (top) and 12 weeks (bottom) of age in response to +25 pA current pulse (bars). Amplitude of APs is measured from AP peak to trough (line, top). d Evoked APs per depolarizing pulse did not differ between treatment group (p=0.7548) but does increase over time (p<.0001; post-hoc HSD: 2–3 vs 6 week p=0.0186, 6 vs 12 week p=0.0032, 2–3 vs 12 week p<0.0001; Control n=21, 29, 11, Y-27632 n=29, 25, 13). e Histogram of evoked AP amplitudes at 6 weeks of age shows no difference between treatments (Wilcoxon rank-sum p=0.9363, Control p=25; Y-27632 p=20). f Histogram comparing evoked AP amplitudes at 3 and 12 weeks of age shows a highly significant increase over time (Wilcoxon rank-sum test p<0.0001; Control n=12, 11; Y-27632 n=8, 14). g Spontaneous APs(APs occurring outside of the 1 second depolarizing current pulse) did not differ between treatment group (p=0.6897) but does increase significantly between 2–3/6 and 12 weeks of age (p<0.0001; post-hoc HSD: 6 vs 12 week p<0.0001; 3 v 12 week p<0.0001; Control n=21, 29, 11; Y-27632 n=29, 25, 14). h A histogram of spontaneous AP amplitudes at 12 weeks of age shows no difference between treatments (Wilcoxon rank-sum p=1; Control n=8; Y-27632 n=12). i A histogram comparing spontaneous AP amplitudes at 6 and 12 weeks of age shows no change over time (Wilcoxon rank-sum test p=0.6151). j Percent of cells exhibiting evoked activity (responsive) versus cells that do not (non-responsive) for each treatment and time point. k Percent of cells exhibiting spontaneous activity (responsive) versus cells that do not (non-responsive) for each treatment and time point. *p<0.05; ** p<0.001 by 2-way ANOVA with post-hoc HSD; Data in a, b, d, and g is presented as mean ± standard error of the mean; see Table 3 for complete statistics

Table 3.

Statistics for electrophysiology experiments.

Measurement Associated Figure	Group	Age (wk)	N	Mean	SEM	Statistics
						Comparison	2-way ANOVA p value	F (df)	Post-hoc Tukey HSD p value
Resting Membrane Potential Fig. 3a	Control	2–3	25	−45.90	1.2998	Control vs. + Y-27632	0.4948	0.47 (1,136)
		6	28	−49.60	2.5933
		12	12	−54.06	3.5366
		All	65	−49.00	1.4130
	Y-27632	2–3	35	−45.23	1.3516	Age groups	0.0181*	4.13 (2,136)	2–3 vs 6	0.0446*
		6	27	−51.19	2.3238				6 vs 12	0.8801
		12	15	−49.17	3.0757				2–3 vs 12	0.0426*
		All	77	−48.09	1.2062	Interaction	0.4437	0.82 (2,136)
	All 2–3 wk		60	−45.51	0.9499
	All 6 wk		55	−50.38	1.7320
	All 12 wk		27	−51.34	2.3257
Membrane Resistance Fig. 3b	Control	2–3	25	2.09	0.1839	Control vs. + Y-27632	0.5594	0.34 (1,136)
		6	28	1.91	0.1739
		12	12	1.56	0.1967
		All	65	1.91	0.1102
	Y-27632	2–3	35	1.88	0.1815	Age groups	0.0938	2.41 (2,136)
		6	27	1.88	0.1369
		12	15	1.52	0.1634
		All	77	1.81	0.1009	Interaction	0.8526	0.16 (2,136)
	All 2–3 wk		60	1.97	0.1303
	All 6 wk		55	1.89	0.1102
	All 12 wk		27	0.54	0.1236
Measurement Associated Figure	Group	Age (wk)	# of cells	Mean	SEM	Statistics
						Comparison	2-way ANOVA p-value	F (df)	Post-hoc Tukey HSD p-value
Evoked Activity (Number of action potentials per pulse) Fig. 3d	Control	2–3	21	1.92	0.8218	Control vs. + Y-27632	0.7548	0.7548 (1,123)
		6	29	3.15	0.6637
		12	11	5.09	1.3175
		All	61	3.08	0.4972
	Y-27632	2–3	29	0.65	0.3159	Age groups	<0.0001 **	14.57 (2,123)	2–3 wk vs 6 wk	0.0186 *
		6	25	3.23	0.7002				6 wk vs 12 wk	0.0032 *
		12	14	6.91	1.3718				2–3 wk vs 12 wk	<0.0001 **
		All	68	2.89	0.4898	Interaction	0.2082	1.59 (2,123)
	All 2–3 wk		50	1.18	0.3962
	All 6 wk		54	3.19	0.7733
	All 12 wk		25	6.11	0.9608
Spontaneous Activity (Number of action potentials off- pulse) Fig. 3g	Control	2–3	21	0.04	0.0245	Control vs. Y-27632	0.6897	0.16 (1,123)
		6	29	0.43	0.1754
		12	11	1.71	0.6573
		All	64	0.53	0.1597
	Y-27632	2–3	29	0.00	0	Age groups	<0.0001 **	14.39 (2,123	2–3 wk vs 6 wk	0.4934
		6	25	0.29	0.1841				6 wk vs 12 wk	<0.0001**
		12	14	2.24	1.0122				2–3 wk vs 12 wk	<0.0001**
		All	68	0.57	0.2375	Interaction	0.6471	0.44 (2,123)
	All 2–3 wk		50	0.02	0.0106
	All 6 wk		54	0.36	0.1262
	All 12 wk		25	2.00	0.6268
Measurement Associated Figure	Group	Age (wk)	# of cells	Median	Interquartile Range	Statistics
						Comparison	Wilcoxon rank-sum test
Evoked Activity (Action potential amplitude) Fig. 3e/f	Control	2–3	12	32.93	9.13	Control vs. + Y-27632	2–3 wk	0.6441
		6	25	50.09	24.97		6 wk	0.9363
		12	11	46.93	29.43		12 wk	0.2180
		All	48	44.29	24.93
	Y-27632	2–3	9	33.80	15.31	Age groups	2–3 wk vs 6 wk	<0.0001 **
		6	20	52.95	21.23		6 wk vs 12 wk	0.0955
		12	14	61.93	16.83		2–3 wk vs 12 wk	<0.0001 **
		All	29	41.60	25.09
	All 2–3 wk		21	33.16	12.10
	All 6 wk		45	51.42	23.28
	All 12 wk		25	56.88	25.84
Spontaneous Activity (Action potential amplitude) Fig. 3 h/i	Control	2–3	3	35.86	14.86	Control vs. + Y-27632	2–3 wk	N/A
		6	11	51.90	19.88		6 wk	0.0782
		12	8	43.64	19.84		12 wk	1.0000
		All	22	43.52	21.59
	Y-27632	2–3	0	0.00	0.00	Age groups	2–3 wk vs 6 wk	0.2040
		6	6	31.12	14.07		6 wk vs 12 wk	0.6151
		12	12	47.37	15.07		2–3 wk vs 12 wk	0.0913
		All	18	45.13	24.99
	All 2–3 wk		3	35.86	19.82
	All 6 wk		17	44.37	25.41
	All 12 wk		20	45.61	20.05

Finally, Y-27632 treatment did not alter the amplitude of either evoked or spontaneous action potentials (Fig. 3e, h). However, the amplitude of evoked action potentials increased with age (2–3 week: 33.16 mV ± 12.10; 6 weeks: 51.24 mV ± 23.28; 12 weeks: 56.88 mV ± 25.84). The observed increase in action potential amplitude largely occurred between three and six weeks of age (Wilcoxon rank-sum, p<0.0001). The amplitude of spontaneous action potentials did not change with age (Fig. 3i), possibly due to a lower sample size; very few cells produced spontaneous action potentials at three weeks of age (Fig. 3k).

Discussion

The ROCK inhibitor Y-27632 is routinely used in hiPSC culture, notably to improve cell survival after passaging. Given that it is well-tolerated, and previous studies show that ROCK inhibition can promote neurite outgrowth (Newell-Litwa et al. 2015), we reasoned that Y-27632 treatment might accelerate the development of electrophysiological activity in hiPSC-derived neurons. As expected, we show that 24-hour Y-27632 treatment increases neurite length, number, and branching of iPSC-derived neurons (Fig. 1e–g’, Table 1). This effect does not persist at 3 or 6 weeks of age (Fig. 2c–e, Table 2); however, we find that there is trend towards an increase in neurite length and branching over time in control cells that is not present in the Y-27632-treated cells. This is not due to a difference in the overall neuron size, since the cell body area was unaffected by treatment (Fig. 2b). While the reason for this discrepancy is unclear, it is possible that short-term ROCK inhibition provides an early increase in morphological properties, but this effect subsides after it is removed, while the control cells continued to develop. This interpretation is supported by the fact that at 3 weeks of age, the Y-27632-treated neurons exhibited a slightly increased number of neurites, neurite length, and branch points per neurite compared to control neurons at 3 weeks of age. While not statistically robust, it is possible that a larger sample of cells would better test this possibility. Additionally, treating cell cultures with Y-27632 for longer than one week may show results that more closely align with the morphological changes seen at 24 hours.

Electrophysiological assays show no effect of Y-27632 treatment on either evoked or spontaneous firing properties (Fig. 3, Table 3). While ROCK inhibition did not affect the parameters we measured, we did observe increased electrophysiological maturity with age. While some neurons at every time point exhibited low levels of activity, we observed an increase in the total number of responsive neurons at later time points in both the control and Y-27632-treated cultures (Fig. 3j–k). Additionally, the number of action potentials observed in in both evoked and spontaneous activity increased with age, as did the amplitude of evoked action potentials (Fig. 3d, f, g). The average resting membrane potential of cells also hyperpolarized over time (Fig. 3a).

In this study, we show patch-clamp data from a total of 129 cells at three different ages. Numerous other studies reporting patch-clamp measurements from more than 100 neurons describe changes in functional maturity due to age or culture conditions (Bardy et al. 2015; Bardy et al. 2016; Belinsky et al. 2014; Bilican et al. 2014; Nicholas et al. 2013; Rushton et al. 2013). Despite significant differences in differentiation and electrophysiological protocols as well as the cell lines and ages tested, there are key commonalities that are observed in many of these reports that also align with our data. For example, reported values for resting membrane potential and membrane resistance are widely variable, but there is consistently a more depolarized resting membrane potential and higher membrane resistance than is physiologically typical in mature neurons (Bardy et al. 2015; Belinsky et al. 2014; Bilican et al. 2014; Nicholas et al. 2013; Rushton et al. 2013). When reported, there is often a more hyperpolarized resting membrane potential and decreased membrane resistance in older cultures (Bilican et al. 2014; Nicholas et al. 2013; Rushton et al. 2013). Additionally, the majority of these studies demonstrate increased proportions of neurons exhibiting evoked or spontaneous activity in older cultures (Bardy et al. 2016; Bilican et al. 2014; Nicholas et al. 2013; Rushton et al. 2013). While Belinsky et al report only a weak developmental trend, the authors address that this is likely due to factors such as cell death in older cultures and the contribution of inactive neural precursors and neural progenitors at all ages. This latter phenomenon is also consistent with our data, as there are neurons in every age group that exhibit a low level of activity.

While there exists great heterogeneity among neuron types in the human brain, several electrophysiological features are generally consistent across subtypes. First, the resting membrane potential of mature cortical neurons is hyperpolarized and averages between −60 and −80 mV (Bean 2007; McCormick et al. 1985). This membrane potential is achieved by a highly regulated process of ion channel expression, particularly potassium channels, and hyperpolarization occurs gradually over time as these channels are expressed (Swayne and Wicki-Stordeur 2012). This observation is correlated with an age-dependent decrease in membrane resistance. These electrophysiological changes are observed in both rodents and in human tissue (Moore et al. 2009; Picken Bahrey and Moody 2003; Spitzer 2006; Swayne and Wicki-Stordeur 2012). The development of electrical activity is also a hallmark for the maturation of neuronal function, and is key in many other aspects of early neurodevelopment (Spitzer 2006). This process is largely controlled by the development of voltage-gated ion channels (Picken Bahrey and Moody 2003; Song et al. 2013).

Overall, our study suggests that iPSC-derived neurons possess electrophysiological properties that are consistent with 2^nd trimester fetal tissue, when compared with the properties of human fetal cerebral cortex (Moore et al. 2009; Moore et al. 2011). Given the relatively brief culture durations, as compared with the development of the human nervous system, the conclusion that iPSC-derived neurons are electrophysiologically immature is not necessarily surprising. Herein, we attempted to accelerate maturation by manipulating processes known to regulate morphological maturation. Many alternative strategies have been employed towards a similar end, including astrocyte co-culture (Johnson et al. 2007; Kuijlaars et al. 2016; Tang et al. 2013), 3D culturing (Lancaster et al. 2013; Pasca et al. 2015; Yan et al. 2016), and specialized culture media that is formulated to promote neuron activity or mimic the early-stage in vivo environment more closely (Bardy et al. 2015; Kemp et al. 2016).

While we did not succeed in accelerating the timeline of electrophysiological or long-term morphological maturity, we reaffirmed the efficacy of inhibiting ROCK activity as a means of enhancing initial neurite formation, and it is possible that including a ROCK inhibitor long-term during cell culture would result in a sustained effect on morphological, and perhaps electrophysiological, properties. This study also reaffirms many functional phenotypes that are shared in the existing literature, including underlining the importance of culture duration in neuronal properties. Moving forward, it will be critical to consider these factors when using iPSC-derived neurons as a model system.