The Efficiency of Direct Maturation: the Comparison of Two hiPSC Differentiation Approaches into Motor Neurons

Motor neurons (MNs) derived from human-induced pluripotent stem cells (hiPSC) hold great potential for the treatment of various motor neurodegenerative diseases as transplantations with a low-risk of rejection are made possible. There are many hiPSC differentiation protocols that pursue to imitate the multistep process of motor neurogenesis in vivo. However, these often apply viral vectors, feeder cells, or antibiotics to generate hiPSC and MNs, limiting their translational potential. In this study, a virus-, feeder-, and antibiotic-free method was used for reprogramming hiPSC, which were maintained in culture medium produced under clinical good manufacturing practice. Differentiation into MNs was performed with standardized, chemically defined, and antibiotic-free culture media. The identity of hiPSC, neuronal progenitors, and mature MNs was continuously verified by the detection of specific markers at the genetic and protein level via qRT-PCR, flow cytometry, Western Blot, and immunofluorescence. MNX1- and ChAT-positive motoneuronal progenitor cells were formed after neural induction via dual-SMAD inhibition and expansion. For maturation, an approach aiming to directly mature these progenitors was compared to an approach that included an additional differentiation step for further specification. Although both approaches generated mature MNs expressing characteristic postmitotic markers, the direct maturation approach appeared to be more efficient. These results provide new insights into the suitability of two standardized differentiation approaches for generating mature MNs, which might pave the way for future clinical applications.


Introduction
Human-induced pluripotent stem cells (hiPSC) are of great importance for regenerative medicine, as a wide variety of cell types can be generated, avoiding the controversially discussed use of embryonic stem cells (ESC). Transplantations with a low-risk of immunological rejection are made possible with autologous hiPSC-derived cells. Furthermore, reprogramming of cells from diseased individuals to hiPSC allows the analysis of pathomechanisms, e.g., of Alzheimer's disease, to be investigated and sets the course for specific drug discovery. With patient-derived hiPSC, it may soon be possible to predict the patient's individual response to drug-specific therapy [1].
Currently, there is no appropriate therapy available for amyotrophic lateral sclerosis (ALS), a neuromuscular disease in which degeneration of motor neurons (MNs) occurs. Replacement of those degraded MNs would be desirable for the treatment of ALS but also other diseases involving degenerated, dysfunctional, or missing MNs. hiPSC are a promising tool for motor neuron (MN) generation, which exhibit the same MN differential potential in comparison to ESC. For clinical application, the generation of human mature and functional MNs needs to underly ensures quality and reproducibility with minimum use of xenogeneic material [2]. Thus, the selection of hiPSC induction method is a critical factor. Existing MN differentiation protocols use, e.g., genomic-integrating lenti- [3] or retroviruses [4,5] for hiPSC reprogramming which is questionable because of their mutagenic potential [6]. Other protocols are performed with, e.g., nongenomic-integrating sendai virus [7] but apply feeder-cells or antibiotics for maintaining hiPSC, having a not yet fully investigated impact on hiPSC differentiation potential [8]. In addition, the selection of reprogramming factors and type of cells used for hiPSC reprogramming is decisive. Reprogramming factors like C-myc, being described as oncogenic, need to be replaced by factors with comparable reprogramming efficiency such as L-myc [9]. The epigenetic memory of reprogrammed cells also has to be considered as it might affect the differentiation potential of generated hiPSC [10]. The process of hiPSC differentiation into MNs still seems to be challenging as there are no defined differentiation media in certified, ready to use-quality available, and postmitotic, xeno-free hiPSC-derived MNs were not purchasable at the beginning of this study.
Various hiPSC differentiation protocols aim to mimic motor neurogenesis in vitro. Bianchi et al. [11], e.g., pursue a three-step differentiation strategy to generate functional MNs, starting with neural induction via inhibition of BMP and TGFβ/Activin/Nodal pathways (dual-SMAD inhibition) followed by MN differentiation via, i.e., retinoic acid (RA) and sonic hedgehog (SHH) exposure, and subsequent MN maturation. The use of feeder cells limits this differentiation protocol in its implementation in regenerative medicine. A similar three-step sequence is used in the protocols of Stemcell Technologies (Vancouver, Canada) applying standardized, defined, and antibiotic-free culture media for neural induction, differentiation, and maturation. Interestingly, previous studies demonstrated that neurons with caudal positional identity can be generated by inhibition of TGFβ/ Activin/Nodal-signalling, which acquire electrophysiological competence in the further course of cultivation [12,13]. This might be achieved by STEMDiff™ SMADi Neural Induction Kit and subsequent maturation.
This study is aimed at comparing this two-step direct maturation approach, named approach A, with the three-step protocol, named approach B, in terms of its efficiency to generate mature MNs from hiPSC. To approach the requirements of clinical application, an adapted reprogramming cocktail and virus-, feeder-, and footprint-free method were used for hiPSC generation. Serum-and antibiotic-free culture medium manufactured under clinical good manufacturing practice (cGMP) was applied for maintaining hiPSC. In the further course of differentiation, chemically defined and standardized culture media were used to ensure a robust and reproducible protocol. Differences on morphological, genetic, and protein levels were investigated, which might provide new insights into a more efficient differentiation method that approaches applications in regenerative medicine.

Generation of hiPSC via Nucleotransfection of Human
Neonatal Dermal Fibroblasts. Human neonatal dermal fibroblasts (hPDF) (ATCC, Manassas, Virginia, USA) were cultured in Fibroblast Growth Kit-Low serum (FGM) (ATCC, Manassas, Virginia, USA) until maximum passage 3 accord-ing to manufacturer's protocol. On day 0, episomal nucleotransfection with three plasmids encoding reprogramming factors and one plasmid encoding Green Fluorescent Protein (GFP) (see Supplementary Method 1 and 2) as control (Addgene, Watertown, Massachusetts, USA) was performed with Amaxa® Nucleofactor II Device (Lonza, Basel, Switzerland) due to manufacturer's instructions. Briefly, 1 × 10 6 hPDFs were transfected with 1 μg DNA of each of the four plasmids (total 4 μg DNA). As a transfection control, 2 μg of pmaxGFP®-Vector (Lonza, Basel, Switzerland) was introduced into the cells according to the manufacturer's directions. After changing the media on day 1, transfected cells were captured by ZEISS LSM 710 confocal microscope (ZEISS, Oberkochen, Germany). On day 2, cells were passaged by washing once with PBS, adding Gibco® Trypsin 0.05%-EDTA (Thermo Fisher, Waltham, Massachusetts, USA) and incubating for 5 min at 37°C in a humidified atmosphere containing 5% CO 2 . Trypsination was stopped by adding medium. Cells were centrifuged for 5 min at 526 × g and plated 1 : 6. On day 4, cells were passaged as described before and plated on Corning® Matrigel® precoated plates. For maintaining hiPSC, cells were cultivated for 30 days in mTeSR™1 (produced under cGMP by Stemcell Technologies, Vancouver, Canada) with a medium change every other day. Passaging was performed with Gentle Cell Dissociation Reagent (GCDR, produced under cGMP by Stemcell Technologies, Vancouver, Canada) every 5 to 7 days. In brief, cells were washed once with PBS and incubated for 10 min with GCDR. Supernatant was removed and mTeSR™1 supplemented with 10 μM Y-27632 (Stemcell Technologies, Vancouver, Canada) was added. hiPSC were collected by rinsing in a circular pattern and were plated 1 : 4 on Corning® Matrigel® precoated plates. The identity of hiPSC was confirmed by alkaline phosphatase live staining, immunofluorescence, and gene expression analysis (see Supplementary Method 3-6). Approach B was first treated with STEMDiff™ Neuron Differentiation Medium for 7 days with a daily medium change. Cells were then exposed to STEMDiff™ Neuron Maturation Medium (Stemcell Technologies, Vancouver, Canada) for 21 days with a change of medium every other day.

Differentiation of hiPSC into
2.4. Differences in Neuronal Clustering Captured by the IncuCyte®. 500,000 cells of approach A and B were plated per well on poly-L-ornithine/laminin precoated 6-well plates. To observe the differences in the capability of forming neuronal networks, the plates were placed in the IncuCyte® S3 system (Sartorius, Göttingen, Germany). Cells were imaged two times a day over a period of 28 days. The experiment was performed in triplicate.

Test for Expression of Neurofibrils.
For staining of neurofibrils with Bielschowsky (Bio-Optica, Milan, Italy), NPC and cells differentiated over approaches A and B for 28 days were seeded onto poly-L-ornithine/laminin (Sigma-Aldrich, St. Louis, Missouri, USA) precoated Nunc™ Lab-Tek™ II Chamber Slides™ (Thermo Fisher, Waltham, Massachusetts, USA), fixed with 4% w/v PFA in PBS by incubation for 20 min at 4°C and washed twice with PBS. According to the manufacturer's protocol, the slide was washed twice with ultrapure water and was incubated with 10 drops of Reagent A for 15 min at 40°C. After washing two times with ultrapure water, 10 drops of Reagent B were added following incubation for 20 min at 50°C. The supernatant was discarded, and the slide was treated with reduction solution (20 drops Reagent C, 8 drops each of Reagents D, E, and F in 50 mL ultrapure water) for 2 min. Cells were washed twice with ultrapure water and subsequently incubated with 10 drops of Reagent G for 3 min. Before dehydrating the slide with ascending concentration of ethanol and treating twice with Xylene (Sigma-Aldrich, St. Louis, Missouri, USA), it was washed two times with ultrapure water. Finally, the slide was embedded in Entellan Neu (Merck, Darmstadt, Germany) following incubation for 30 min at room temperature to dry. The stained neuronal structures were imaged with Nikon Eclipse TS100 microscope (Nikon, Minato, Japan).

Flow Cytometric Analysis with Neuronal Markers to
Determine Differentiation Status. For fluorogenic staining, 250,000 NPC and cells of approaches A and B differentiated for 28 days were fixed by carefully resuspending in 300 μL of 4% w/v PFA in PBS and incubating for 20 min at 4°C. Subsequently, 1 mL perm solution (0,1% v/v Triton X-100 and 0,3% w/v sodium dihydrogen citrate ((Sigma-Aldrich, St. Louis, Missouri, USA) in ultrapure water) was added and cells were centrifugated for 5 min at 296 × g. The supernatant was discarded and permeabilization was performed by careful resuspension in 300 μL perm solution and incubation for 2 min at 4°C. Cells were pelleted (5 min at 296 × g), resuspended in 45 μL FACS solution (PBS with 2% v/v FBS), and subsequently stained with the following labelled antibodies by incubation for 30 min in the absence of light: Nestin PerCP-Cy™5.5 (5 μL, Becton Dickinson, Franklin Lakes, USA), ChAT-APC (5 μL of 1 : 500 in FACS solution, abcam, Cambrigde, UK), and DAPI (5 μL of 1 : 1,000 in FACS solution, Biolegend, San Diego, California, USA). Unstained cells and stained cells with the following antibodies for fluorescent isotypes were taken along as controls: PerCP-Cy™5.5 (5 μL, Becton Dickinson, Franklin Lakes, USA), IgG APC (5 μL of 1 : 500 in FACS solution, abcam, Cambrigde, UK). After washing three times with 300 μL FACS solution, 10,000 cells were analyzed using the Amnis® ImageStream® X Mk II instrument with ISX Software (Luminex Corporation, Austin, USA). The experiment was performed three times independently and the data was processed and presented with the IDEAS® Software (Luminex Corporation, Austin, USA).

Western
Blot. The method of Western Blot analysis used here is based on the protocol recently published by Horn et al. [14] and was carried out as follows. NPC and cells differentiated over approaches A and B for 28 days, cultured in 6-well plates, were washed once with PBS (Gibco, Thermo Fisher, Waltham, Massachusetts, USA), detached by incubation with StemPro™ Accutase™ Cell Dissociation Reagent (Thermo Fisher, Waltham, Massachusetts, USA) for 5 min at 37°C in humidified atmosphere containing 5% CO 2 , and centrifuged for 5 min at 296 × g. 2 × 10 6 cells of each cell type were resuspended in 400 μL of Tris-EDTA-Triton X-100 extraction buffer (6.25 mM TRIS, 12.5 mM NaCl, 2.5 mM EDTA, 1.5% Triton X-100, one Complete Mini Inhibitor Cocktail (Roche, Basel, Switzerland) and one PhosSTOP (Roche, Basel, Switzerland) in 10 mL ultrapure water) and were incubated on ice for 15 min. Cells were disrupted using an ultrasonic homogeniser (Bandelin Electronic, Berlin, Germany) by pulsing three times for 10 s, 0.3 interval, and 30% intensity, following incubation on ice for 1 h, with vortexing every 15 min. Cell debris was removed by centrifuging for 10 min at 17,186 × g at 4°C. Supernatant was transferred to a 0.5 mL Protein LoBind® Tube 1 : 1,000. Subsequently, the membranes were washed two times for 10 min with washing buffer (PBS with 0.1% v/v Tween® 20). Secondary antibody incubation was performed for 90 min under light exclusion with IRDye® 800CW Goat Anti-Mouse or IRDye® 800CW Goat Anti-Rabbit (LI-COR Biosciences, Lincoln, Nebraska, USA) 1 : 8,000 in antibody diluent. Membranes were washed again two times for 10 min with washing buffer and rinsed two times with ultrapure water. For normalization with the house-keeping protein glyceraldehyde-3-phosphate dehydrogenase (GAPDH), the primary antibody recombinant Rabbit Anti-GAPDH (ab181602, abcam, Irvine, California, USA) 1 : 10,000 in antibody diluent was added and detected by secondary antibody Goat anti-Rabbit IRDye® 680RD (LI-COR Biosciences, Lincoln, Nebraska, USA) 1 : 8,000 in antibody diluent. The experiment was carried out three times independently.
2.9. Analysis of Gene Expression. The method of gene expression analysis used here is based on the protocol described previously by Horn et al. [14] and was performed as follows. NPC and cells of approach A and B differentiated for 28 days were washed once with PBS and were obtained by using a cell scraper (Greiner AG, Kremsmünster, Austria) directly into RNA later (Qiagen, Hilden, Germany). RNA extraction was performed according to the manufacturer's instructions using RNeasy Mini Kit (Qiagen, Hilden, Germany). Briefly, RNA later was discarded, and cells were lysed under repeated resuspension with 600 μL buffer RLT. The cell lysate was transferred to a QIAshredder spin column (Qiagen, Hilden, Germany) and centrifuged for 2 min at 21,135 × g. The eluate was mixed with 600 μL of 70% ethanol. 700 μL of this mixture was pipetted into a spin column and centrifuged for 30 s at 13,250 × g. The eluate was discarded, and the column was washed three times. Once with 700 μL buffer RW1 (30 sec, 13,250 × g) and two times with 500 μL buffer RPE (2 min, 21,135 × g). To dry the column, it was centrifuged at 21,135 × g for 1 min. RNA was eluted with 30 μL nuclease free water (Qiagen, Hilden, Germany) by spinning 1 min at 21,135 × g in a Biopur® 1.5 mL tube (Eppendorf AG, Hamburg, Germany).
The NanoQuant Plate™ with Plate Reader infinite M200 Pro (Tecan Group AG, Männedorf, Switzerland) was used for RNA concentration determination. The amount of RNA for cDNA synthesis was set to 500 ng per qRT-PCR plate. cDNA synthesis was performed with the RT 2 First Strand Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. In brief, for elimination of genomic DNA, 2 μL of buffer GE, RNA, and RNase free water were mixed to a total volume of 10 μL and incubated for 5 min at 42°C followed by cooling on ice for at least 1 min. 10 μL reverse transcription mix consisting of 4 μL 5× buffer BC3, 1 μL control P2, 2 μL RE3 reverse transcriptase mix, and 3 μL RNase free water was added to the genomic elimination mix. Reverse transcription was performed by incubation for 15 min at 42°C, followed by incubation for 5 min at 95°C to stop the process. The Mastercycler® nexus GX2 (Eppendorf AG, Hamburg, Germany) was used for all incubation steps. 91 μL RNase free water was added to the preparation. The cDNA mix was stored at -20°C for a maximum of 7 days.
For the preparation of the detection mix and the pipetting of qPCR plates, the Freedom Evo automated pipetting machine controlled by EVOware™ Standard Software (Tecan Group AG, Männedorf, Switzerland) was used. To minimize any contamination, the tips were washed intensively with 7% v/v sodium hypochlorite (Carl Roth, Karlsruhe, Germany) in ultrapure water and ultrapure water before and after the pipetting steps. According to the manufacturer's instructions, 102 μL of the cDNA mix were diluted in 1,248 μL RNase free water and then placed in the Freedom Evo automated pipetting machine for adding 1,350 μL of 2x RT 2 SYBR® Green ROX qPCR Mastermix (Qiagen, Hilden, Germany). The mixture was mixed by gentle inversion to avoid bubbles. 25 μL of this mix were pipetted in each well of the following 96-well RT 2 Profiler™ PCR Arrays: Human-Induced Pluripotent Stem Cells, Human Neurogenesis, Human Neurotransmitter Receptors, Human Neutrophins and Receptors, Human Neuronal Ion Channels, and a custom-designed plate for analysis of genes involved in motoneurogenesis (Qiagen, Hilden, Germany). Each plate was sealed with Optical Thin-Wall 8-Cap Strips (Qiagen, Hilden, Germany). qRT-PCR analysis was performed with Eppendorf Mastercycler® epgradient S realplex 2 with Mastercycler ep realplex software (Eppendorf AG, Hamburg, Germany) by activating polymerase for 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C. After each run, the threshold was adjusted to 200 and drift correction was enabled for better comparison. All experiments were repeated three times independently.
For data analysis, C t values were exported to GeneGlobe (http://www.qiagen.com/geneglobe) using C T cut-off of 35, RPLP0 housekeeping gene for normalization, fold regulation cut-off of 2.0, and p value cut-off of 0.05. The calculation of p value is based on an unpaired Student's t-test. The software RStudio (version 4.0.3 (2020-10-10)) was used for graphic presentation (RStudio Inc., Boston, USA).

Morphological Pattern of the Two Differentiation
Approaches. A scheme of 28 days NPC differentiation by direct maturation, approach A, and with a preceding differentiation step, approach B, is shown in Figure 2(a). Analyzed by IncuCyte®, both approaches yielded branched networks with MN-type clusters after 28 days of differentiation in comparison to NPC (Figure 2(b)). These clusters were confirmed as neuronal networks by Bielschowsky silver staining (Figure 2(c)) and immunostaining for the neuronal marker Tubb3 (Figure 2(d)). Approach B produced denser filaments than approach A, particularly evident in Tubb3 staining.

Variation in Neuronal Gene Expression Tested via qRT-PCR.
To further characterize the neurons derived from 28 days differentiation via approaches A and B, the expression of various genes characteristic of motor neurogenesis [19], neuronal ion channels [20], neurotrophins [21], and neurotransmitters [22] was determined. As shown in Figure 3, genes were significantly differentially expressed in cells of both approaches A and B compared to NPC (fold regulation ≥ 2:0 and ≤ -2.0 in combination with p < 0:05).
In approach A, some genes were selectively downregulated, including Sex Determining Region Y Box 2 (Sox2;

Expression of Early Differentiation Markers Pax6 and
Nestin. To evaluate whether both approaches A and B were able to generate mature neurons after 28 days of differentiation, the expression levels of the early neuron differentiation markers Pax6 and Nestin were determined in comparison to NPC. As Nestin is described to be only expressed in uncommitted NPC and Pax6 is crucial for the generation and maintenance of NPC, both need to be downregulated to allow motoneuronal maturation [25,26]. Pax6 was found downregulated, both at the mRNA level (A/-13.07; B/-8.72; p < 0:05) (Figure 4(b)) and the protein level (A/0:36 ± 0:13; B/0:29 ± 0:02) (Figure 4(c)). Nuclear staining for Pax6 was only observed in NPC but not in cells of approach A and B (Figure 4(a)), thus confirming the results of qRT-PCR and Western Blot.    GRM8  HCRT  HCRTR2  HTR2A  IL1R1  IL6R  KCND2  KCNH1  KCNH2  KCNJ3  KCNJ5  NGF  NGFRAP1  NPY  NPY1R  NPY2R  PTGER2  SCN3A  SCN9A  TACR2  TP53  TRPC1  TRPM2   A B Approach  (Figure 4(f)). These results confirm the quality of the generated NPC and suggest a successful differentiation of approaches A and B towards mature neurons.

Expression of the Motoneuronal Marker MNX1.
To verify the generation of motoneuronal populations during 28-day differentiation of both approaches, the expression of MNX1, an early marker protein of postmitotic spinal MNs [27], was investigated. By immunofluorescence, MNX1 was detected in the cell nuclei of all three approaches ( Figure 5(a)). These findings were confirmed by Western Blot results (Figure 5 (Figure 5(b)). Furthermore, MNX1 was significantly 6.67-fold higher expressed (p < 0:05) in cells of approach A compared to cells of approach B. These results suggest that motoneuronal precursor cells are also present in the early neuronal differentiation stage of NPC and indicate that SHH-and RA-signalling pathways have been activated during dual-SMAD inhibition.

Postmitotic Motoneuronal Markers ChAT and Islet-1.
To validate both differentiation approaches in terms of their efficiency to generate mature MNs, cells differentiated by approaches A and B for 28 days were tested for postmitotic motoneuronal markers ChAT and Islet-1 (ISL1) in comparison to NPC. By immunofluorescence and flow cytometry, all approaches were positively stained for ChAT (Figure 6(a)), an enzyme catalysing the synthesis of the neurotransmitter acetylcholine that is specifically found in cholinergic neurons as MNs of the spinal cord [3]. Images of cells of approach A and B, captured during flow stream, showed much higher ChAT-APC (Allophycocyanin) staining intensity in comparison to NPC, indicating that only cells of approach A and B should be considered positive. Thus, 71.5% of approach A and 65.6% of approach B were detected positive for ChAT-APC compared to 15.9% among NPC (Figure 6(b)).
ISL1, a transcription factor that is essential for the generation of mature and functional human MNs [3], was found in nuclei of cells of approach A and B but not in NPC detected by immunofluorescence (Figure 6(c)). In qRT-PCR, ISL1 was also significantly upregulated in both approaches (A/17.94; B/30.59; p < 0:05) in comparison to

10
Stem Cells International NPC, confirming the suitability of both approaches to form mature MNs (Figure 6(d)).

Neuronal Proteins Essential for Synaptic Transmission.
To test if both approaches A and B express proteins associated with mature MNs and neuronal functionality after 28 days of differentiation, the presence of Synapsin I (Syn I) Peripherin, and Neurofilament Heavy Polypeptide (Smi-32) was examined by immunofluorescence. Syn I belongs to the phosphoprotein family of synapsins, which play a key role in neurite outgrowth, synapse formation, and synaptic transmission [28] and is a characteristic marker for mature neurons [29]. Positive staining for Syn I in the form of puncta, which are presumed to be synaptic vesicles, was found in approach A and B, but not in NPC (Figure 7(a)), indicating maturity and assuming electrophysiological activity of differentiated cells of both approaches [30]. Peripherin expression was observed in neurons of the spinal cord [31] and is considered to be involved in neurite elongation during neuronal development [32]. Immunofluorescent results show Peripherin-positive neuronal filaments in approach A and B (Figure 7(b)). As a type III neurofilament protein, the function of Peripherin is affiliated with the postmitotic marker Smi-32 to form filament networks [33]. As illus-trated in Figure 7(c), nuclei and axons of approach A and B were positively stained for Smi-32 indicating that functional networks were generated.

Discussion
This study gives insight into the potential of two differentiation approaches A and B using standardized, chemically defined, and ready-to-use culture media to generate mature MNs from foodprint-and integration-free derived hiPSC. Both differentiation approaches avoid the use of antibiotics or feeder cells, thus the derived MNs may have the potential to be implemented in the therapy of motor neurodegenerative diseases. Since hiPSC differentiation into MNs is a multiple-step process including the formation of NPC, the identification and characterization of these generated NPC is essential. NPC were differentiated successfully under serum-free conditions from hiPSC, expressing early neuron differentiation markers such as Nestin, Sox2, Pax6, FABP7, and Tubb3 assessed by qRT-PCR, Western Blot, immunocytochemistry, and flow cytometry [34]. Nestin, a class VI intermediate filament protein and Sox2 are proteins characteristically occurring in NSC [35]. Expression levels of Pax6, FABP7, 11 Stem Cells International and Tubb3 are increased in the developing spinal cord and are early induced in differentiating MNs [18]. In particular, Pax6 plays a key role in patterning neuroectoderm cells into ventral spinal progenitors [36]. Interestingly, MNX1, a spe-cific marker for pre-and postmitotic MNs [37] and described as being expressed in the developing spinal cord [38], was detected in NPC indicating that it was possible to generate NPC with a spinal fate after dual-SMAD inhibition.

Stem Cells International
Jordan et al. [37] observed MNX1 expression after treatment with bFGF, which is well-known to induce proliferation of NSC and might be a supplement of the standardized NPC expansion medium used here. Surprisingly, ChAT was detected at a low level in NPC by flow cytometry and immunofluorescence. This also confirms the neural conversion towards MNs and might also be related to growth factors as Nistor et al. [39] found an increased number of ChATexpressing NPC after treatment with bFGF and FGF8. Maturation of MNs was achieved via both differentiation approaches A and B. Downregulation of Pax6 and Nestin is described in the course of MN maturation [25,40] and was observed in both approaches, as shown in the results of qRT-PCR, Western Blot, immunofluorescence, and flow cytometry. Concerning MN morphology, a hallmark in the evaluation of MNs [41], approach A and B can be distinguished by their level of maturity. MN maturation is morphologically associated with extending long projections [3] and an increase in complexity in neurite outgrowth [42]. Although this was confirmed in both approaches by the Incu-Cyte® and neuron-specific staining for Tubb3, the approach A showed less dense neuronal filaments and less intense brown

Stem Cells International
Bielschowsky staining compared to approach B which could be an indicator of a lower degree of maturation.
Concerning the expression profile, both approaches differ in aspects of neurotrophic proteins (e.g., GFRA2, BDNF, NTF3, and FGF1). These are involved in the regulation of neuronal survival, axon outgrowth, dendritic pruning, and synaptic plasticity [43,44] and therefore are responsible for the development and maintenance of neuronal functionality. As GFRA2 is downregulated in postnatal spinal MNs [43], reduced GFRA2 expression indicates successful MN maturation. Furthermore, a reduced MN survival rate was found to be related to loss of BDNF, NTF3, and NTF4 [43]. As BDNF is downregulated in approach B and NTF3 is upregulated in approach A, this could indicate that MNs generated by approach A possess a higher level of robustness. Investigations on the NTF4 expression level would be needed for clarification. Concerning neurotrophin FGF1, Renaud et al. [44] described that increased expression levels can be found in MNs, which reach a maximum in adult neuronal tissues. As FGF1 is selectively upregulated in approach B, this may confirm approach B being more mature. Another protein involved in neuronal plasticity and long-term potentiation is S100b. It prevents developmental cell death in MNs [45] and was found upregulated in approach B. Those findings put the assumption that the approach A may be more robust into perspective.
Characteristic proteins essential for the identification of postmitotic and functional MNs are MNX1, ISL1, and ChAT [46]. MNX1 (also known as Hb9) promotes the specification of MNs and is considered to be a marker for lower MNs [47,48]. Neurons obtained via approach A showed increased MNX1 mRNA level compared to approach B, suggesting that a higher amount of lower MNs was generated. Investigation on MNX1 protein level revealed no significant difference between both approaches, which relativises this assumption. In the further course of MN-specification, the expression of MNX1 determines the expression of the LIM homeodomain transcription factor ISL1 in postmitotic MNs [49]. ISL1 is one of the first motoneuronal genes expressed in postmitotic MNs and plays a key role in the specification of functional MNs by being involved in neurotransmitter expression and MN migration [3,50]. ISL1 gene and protein expression was significantly increased in both approaches, confirming the generation of putative functional MNs. Furthermore, ISL1 regulates the expression of cholinergic genes such as VAChT, CHT, and ChAT [51]. ChAT is crucial for the synthesis of the neurotransmitter acetylcholine and is selectively expressed in MNs of the spinal cord [3]. Several studies found that ChAT-positive neurons possess electrophysiological activity [27,46,[52][53][54]. On protein level, ChAT was detected in both approaches by flow cytometry and immunohistochemistry. In contrast, ChAT gene expression levels were decreased in both approaches, which may be due to posttranscriptional regulation in process of maturation [55]. These findings were also observed in rats by Corsetti et al. [53]. As ChAT is not expressed in the upper MNs, this may suggest successful NPC differentiation into mature and functional lower MNs [52,56].
Genes coding for neuronal ion channels being upregulated in approaches A and B can be considered for the func-tional classification of mature neurons [57]. KCNH1, encoding for the voltage-gated potassium channel Kv10.1, is associated with neurotransmitter release and synaptic transmission in MNs [58,59]. ADRA1A, coding for adrenergic receptor, and GRIA2, coding for glutamate ionotropic receptor, are involved in neurotransmitter release of catecholamines and glutamate and are essential for maintenance of functional MNs [60,61]. Increased expression of KCNH1, ADRA1A, and GRIA2 in both approaches indicates successful maturation which is further underlined by decreased expression of genes coding for SLC17A6 and DRD2 [61,62]. Further proteins associated with neural electrophysiological activity are Smi-32, Peripherin, and Syn I. Expression of Smi-32 is specific for MNs of the spinal cord and was found in cytoplasm and dendrites of both approaches [3]. Structural support and transport of nutrients are the main functions of neurofilament Smi-32. Smi-32 and Peripherin, a type III intermediate filament protein appearing during MN development, are functionally interdependent and assure normal conduction velocity [31,33]. Thus, both proteins are essential for the formation of functional MNs. The detection of scattered synaptic puncta on dendrites is characteristic for immunostaining of Syn I, a synaptic vesicle protein implicated in synapse formation and synaptic transmission [28,63]. Syn I is selectively expressed in mature neurons and was found to increase the amplitude of evoked synaptic currents [28]. Thus, the detection of Smi-32, Peripherin, and Syn I confirms maturation and suggests functional activity of both approaches [33,64,65].
To clarify if the generated MNs possess functionality, not only investigations on electrophysiological activity using, e.g., patch-clamp or multielectrode-array system need to be performed but also coculture experiments with skeletal muscle cells are required as the proper synaptic and signalling context is provided by neuromuscular junctions [27]. Considering therapy of motoneuronal degeneration diseases, Trawczynski et al. [2] suggest that particularly promising results are achieved by transplanting MN precursor cells which form appropriate synapses and exhibit better functional recovery compared to postmitotic MNs. By using xeno-free matrices, the method presented here for generating MN precursor cells constitutes a promising tool for regenerative medicine. In the future, it will be interesting to sort differentiated MNX1-positive motoneuronal precursor cells and to investigate their ability to differentiate into mature and functional MNs in vivo.

Conclusion
This study demonstrated that MNX1-and ChAT-positive MN precursors can be differentiated from integration-, feeder-, serum-, and antibiotic-free generated hiPSC by dual-SMAD inhibition and subsequent expansion. The derived MN precursor cells form mature and putative functional MNs after application of chemically defined, standardized, and antibiotic-free culture media. The quality of hiPSC, MN precursors, and mature MNs was continuously confirmed in terms of their morphology, protein, and gene expression, ensuring a robust and reproducible protocol.

Stem Cells International
Comparison of differentiation approaches A and B revealed a significant difference in MNX1 gene expression. Thus, the approach A appears to be a more efficient differentiation method that might set the course for clinical translation if the entire process of hiPSC generation and differentiation, including all culture media and matrices used, complies with cGMP requirements.

Data Availability
Data may be available from the corresponding author upon reasonable request.

Conflicts of Interest
The authors declare that they have no conflict of interest.