Neural induction of porcine‐induced pluripotent stem cells and further differentiation using glioblastoma‐cultured medium

Abstract Prior to transplantation, preclinical study of safety and efficacy of neural progenitor cells (NPCs) is needed. Therefore, it is important to generate an efficient in vitro platform for neural cell differentiation in large animal models such as pigs. In this study, porcine‐induced pluripotent stem cells (iPSCs) were seeded at high cell density to a neural induction medium containing the dual Sma‐ and Mad‐related protein (SMAD) inhibitors, a TGF‐β inhibitor and BMP4 inhibitor. The dSMADi‐derived NPCs showed NPC markers such as PLAG1, NESTIN and VIMENTIN and higher mRNA expression of Sox1 compared to the control. The mRNA expression of HOXB4 was found to significantly increase in the retinoic acid‐treated group. NPCs propagated in vitro and generated neurospheres that are capable of further differentiation in neurons and glial cells. Gliobalstoma‐cultured medium including injury‐related cytokines treated porcine iPSC‐NPCs survive well in vitro and showed more neuronal marker expression compared to standard control medium. Collectively, the present study developed an efficient method for production of neural commitment of porcine iPSCs into NPCs.

and Development Center (GRDC) Program through the NRF funded by the Ministry of Education in Korea, Award/Award Number: 2017K1A4A3014959

| INTRODUCTION
Directed neural differentiation of pluripotent stem cells (PSCs) enables researchers to generate diverse neuronal cell sources for neural development study and regenerative medicine. [1][2][3] While many studies have used rodent models of neurogenesis, these models are not suitable for preclinical studies in regenerative medicine that include neural stem cell therapies or for studying human neurodegenerative diseases, because they are poor representations of the human neural system. 4 There is therefore a need for the generation of an efficient in vitro platform for neural cell production in large animal models such as the most closely related non-primate animal species, which is the pig. 5 For example, pigs have similar genomic, anatomical, immunological and physiological characteristics to those of humans, as well as a comparable organ size and lifespan. [6][7][8] Furthermore, porcine models are cost-effective as they are by-products from the abattoir, thus they are widely available compared to other large animal models. In particular, there are similarities with humans in the growth pattern of the neonatal brain and extent of peak brain growth at the time of birth. 9,10 Furthermore, pigs have gyrencephalic brain with similar grey and white matter composition and size that is more comparable to humans than are rodents. 11 Despite these advantages, there is lack of information about efficient direct neural differentiation in porcine model.
For the robust and reproducible neuronal differentiation of PSCs, the establishment of self-renewable neural progenitor cells (NPCs) is required. 12,13 In general, most of PSCs can differentiate into NPCs, despite significant variability in methods and efficiency for neural induction systems. Although some strategies have been used to differentiate porcine pluripotent cells into neuronal populations, little is currently known about neural development in pigs. Puy et al reported the differentiation of porcine neural cells derived from the inner cell masses (ICMs) of blastocyst-stage embryos. 14 Other studies have reported the derivation of NPCs from porcine isolated epiblasts from blastocysts, and evaluated their in vitro differentiation potential into neuronal and glial cells. 15,16 The majority of these initial reports have claimed that neural differentiation can be typically induced by the formation of the embryoid body (EB) 17 or that it uses undefined factors, such as animal feeder cell lines or components. 18 However, the EB formation method, which can form all three germ layers, produced only small numbers of neural commitment cells because of the mixed presence of other cell lineages, such as those of mesodermal or endodermal origin. Also, the inductive protocols for neuronal cells in such porcine research utilized a coculture system using a neural inducing feeder cell line called MS5. However, this is unsuitable for medical applications because the mechanism of neural induction remains unclear and animal sources pose a risk of xenogenetic pathogen transmission. This demonstrates the need to develop innovative strategies that can generate neural progenies from porcine PSCs.
Glioblastomas (GBM) are the most aggressive and malignant central nervous system (CNS) tumours. 19 In the tumour zone, cancer cells have a remarkable ability to alter their environment for their own advantage. These cells have significant effect on the surrounding vital normal cells such as astrocytes, neurons, endothelial cells and stem cells. GBM cells are known to secrete injury-related cytokines, including transforming growth factors (TGF-β and TGF-α), fibroblast growth factor (FGF), insulin-like growth factors (IGF), platelet-derived endothelial cell growth factor (PDGF) and vascular endothelial growth factor (VEGF). 20 These are involved in critical tissue functions, including control of early development, and survival and differentiation of neurons and astrocytes. [21][22][23] Especially, GBM is a grade IV fast-growing glioma originated from the transformation of the normal primitive precursors of glial cells (glioblasts). 24 Based on this aspect, the potential impact of the factors derived from GBM on maturation and differentiation of progenitor cells needs to be uncovered.
The aim of this study was thus to examine the effect of endogenous and exogenous factors in the differentiation of porcine iPSCs into NPCs. It has been known that dual small-molecule targeting of SMAD signalling efficiently neuralizes human PSCs by blocking both BMP and Activin/Nodal pathways. 25 By combining dual SMAD inhibitors (dSMADi) such as LDN193189 or SB431542 with the control of seeding cell density, we generated high-yield NPCs from porcine iPSCs. The neuronal fate specification can be recapitulated in vitro using regional specific factors. In addition, the effects of brain cancer cell line-derived conditioned medium (CM) on the differentiation of iPSC-NPCs into neuronal-like cells were also examined to establish an efficient system for the generation of neural cells from porcine iPSCs.

| Ethics statement
This study was carried out in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Veterinary and Quarantine Service. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Chungbuk National University (permit number: CBNUA-584-13-01). All animal subjects were killed using isoflurane anaesthesia, and all efforts were made to minimize suffering. Experimental procedures for establishing GBM cell lines were approved by the Ethics Committee and permission was granted from the institutional review board KIM ET AL.

| Chemicals
Consumables were purchased from Sigma-Aldrich Chemical Company (St. Louis, MO, USA) unless otherwise stated.

| Culture of porcine iPSCs
The porcine iPSCs used in this study were kindly provided by Pro-  (Gibco), was added in increasing 25% increments (25%, 50%, 75%) every other day starting on day 4 (100% N 2 on day 10). To initiate neuronal patterning and differentiation, we exposed porcine iPSCs treated with dSMADi to 50 ng/mL FGF8 or 0.5 µmol/L retinoic acid (RA) for 5 days with or without 500 ng/mL sonic hedgehog (SHH) in N2/B27 media. Neural expansion was performed in the presence of well-known mitogens such as 10 ng/mL epidermal growth factor (EGF), and 4 ng/mL basic fibroblast growth factor (bFGF, Invitrogen Corporation, Carlsbad, CA). 26-28

| Primary culture of glioblastoma (GBM) cells
Primary human GBM cells were established from fresh specimens, and verified by routine histopathological evaluation to be GBM. The GBM specimens from human patients were removed by craniotomy and transported to the laboratory after sampled in fresh Hank's Balanced Salt Solution (HBSS, Gibco, Carlsbad, CA) in a sterile tube at 4°C, washed five times with HBSS and cut into 1 mm pieces. Following digestion with 0.1% trypsin (Gibco) for 30 minutes at 37°C, these cells were washed in HBSS and seeded at a density of 5 × 10 5 cells/mL in 6-well plates. Cells then were cultured in N2/B27 medium supplemented with 1% non-essential amino acids (Gibco), 1% glutamine (Gibco), 0.1 mmol/L ß-mercaptoethanol (Gibco) and 1% antibiotic-antimycotic (Gibco), 10 ng/mL EGF and 4 ng/mL bFGF (Invitrogen Corporation) at a 3% O 2 and 5% CO 2 atmosphere. After achieving 80%-90% confluency, the cultures were passaged at 1:3.
The medium was changed twice a week. To prepare GBM-CM, GBM cells were cultured on 100 mm culture dishes until confluent, and then the medium was replaced with 8 ml of fresh medium. After 24 hours of incubation at 37°C, the medium was collected, filtered and stored at −80°C for experiments.

| Alkaline phosphatase (AP) activity detection
Porcine iPSCs were harvested, washed three times with PBS and fixed in 4% paraformaldehyde for 5-7 minutes at room temperature.

| Karyotyping
To check for cytogenetic abnormalities, karyotyping of porcine iPSC or GBM cells were performed in accordance with standard cytogenetic techniques. 29 Briefly, confluent monolayers of porcine iPSCs were treated with 10 µg/mL colcemid (Gibco, Carlsbad, CA) for 3-5 hours to induce metaphase arrest. The cells were then gently washed three times and resuspended in a pre-warmed hypotonic solution of 0.075 mol/L KCl (Merck, Darmstadt, Germany) and 1% sodium citrate (Merck, Darmstadt, Germany) and then incubated at 37°C for 25 minutes. Cells were then suspended by shaking the flask horizontally, transferred to a conical tube and then twice fixed at room temperature by resuspending in freshly made solution containing methane and acetic acid at 3:1 ratio (Merck). The fixed cells were subsequently added to cold wet slides, which were air-dried, treated with trypsin (Gibco) and stained with Giemsa (GTG-banding method).
Chromosomes were then counted under a standard bright field microscopy and checked for cytogenetic abnormalities.

| Gene expression analysis by real-time PCR or reverse transcription PCR
The porcine iPSC-derived NPCs or differentiated cells were analysed for the expression of neural and progenitor markers by RT-PCR or  Table S1. The fluorescence intensity was measured at the end of the extension phase of each cycle with threshold values set manually. Relative expression was determined by the 2 Ct method, with RN18S as a control. Experiments were repeated at least three times.

| Immunofluorescence
Immunofluorescence (IF) was performed as follows: Cells were washed with 1x PBS containing Ca 2+ and Mg 2+ and fixed with 4% paraformaldehyde. The cells were washed three times with PBS and permeabilized with 0.2% Triton X-100 for 5 minutes for intracellular markers analysis. The fixed cells were co-incubated with blocking solution (10% goat serum in PBS) and primary antibody overnight at 4°C. The primary antibodies used in this study are listed in Table S2.
The following day, cells were washed three times with washing medium (Tween-20, Triton X-100 and PBS) and incubated with appropriate secondary antibodies at room temperature for 1 hour. Nuclei were then stained with Hoechst 33342 and the stained cells examined using a confocal microscope and ZEN 2009 Light Edition software (Carl Zeiss, Oberkochen, Germany).

| Statistical analysis
Statistical analysis was performed using SPSS 17.0 (SPSS, Inc, Chicago, IL, USA). Results are expressed as the means ± SEM. One-way ANOVA was performed to test the null hypothesis of group differences, followed by Duncan's multiple range test or Student's t test.
P < 0.05 was considered statistically significant.

NPCs induced by dSMAD inhibition with high cell seeding density
Dual SMAD inhibition methods are known to promote neuralization of primitive ectoderm through BMP inhibition by LDN193189 and suppress mesendodermal fates by inhibiting endogenous activin and nodal signals using SB431542 (Figure 2A). Gene expression analysis confirmed that dSMADi-treated porcine iPSCs expressed markers such as PLAG1, Nestin and Vimentin at day 10 ( Figure 2B). There was no expression in these genes in the control group and EB group at day 10. In particular, the high cell density group revealed higher expression of the neural crest (NC) marker p75 and neuroectodermal marker Sox1 at day 10 of differentiation compared to those of the low-density group and control group ( Figure 2C). The expression of POU5F1, whose expression was maintained during pluripotent control group, decreased in both dSMADi-treated low-and high-density groups. In terms of protein levels, the expression of CNS marker,   Figure S2A. After neural induction using dSMADi, the high-density group showed a significantly higher number of colonies than the low-density group (0.8 ± 0.4 vs 10.2 ± 0.9, P < 0.01, Figure S2B).

| Patterning and further differentiation of porcine iPSC-NPCs
To investigate whether porcine iPSC-NPCs are responsive to instructive regionalization cues, we exposed porcine iPSCs treated with dSMADi to FGF8 or retinoic acid (RA) for 5 days with or without assay ( Figure 5F).

| Enhanced further differentiation of porcine iPSC-NPCs into progeny cells using GBM-CM
Although we obtained porcine iPSC-NPCs induced by dSMADi methods, they showed limited features in maturation and differentiation to neuronal progeny cells. Therefore, to obtain more robust and mature neuronal progeny cells from porcine iPSC-NPCs, we investi- Together, these data demonstrate that GBM cell-derived soluble factors may play a pivotal role in the extensive further differentiation of porcine iPSC-NPCs.

| DISCUSSION
In this study, we developed a novel platform for neural induction and differentiation of porcine NPCs and commitment progeny cells using dSMADi and patterning factors without EB formation or feeder cell coculture. Our system also included further differentiation from porcine iPSC-NPCs to neuronal cells using patient-derived GBM-CM.

Recent studies have indicated that the interactions between cells
and their microenvironment play critical roles in determining cell fate in vivo. 36,37 Cells are affected by signalling factors such as soluble cytokines, cell-cell contacts and insoluble extracellular matrices. 38 Therefore, it was thought that specific neuronal fates are determined during nervous system development through co-ordinated At the same time, the initial cell density, which is one of the endogenous factors, has an important impact in the commitment of PSCs into a particular cell lineage. 42 It is well known that the seeding cell density can result in variable terminal cell densities, giving rise to various outcome of differentiation. 25 In contrast to the MS5 protocol, which requires low plating density, the dSMADi method allowed for high plating densities. In the present study, the dSMAD inhibition group with low seeding density failed to initiate neural induction.
We found that the initial seeded cells with high density form-derived cells subsequently affected differentiation of the NE, the primordium of the nervous system. Therefore, the collaboration of dSMAD inhibition system and high cell density resulted in an efficient induction of early neural lineage from porcine iPSCs, which is consistent with a human study. 42 In terms of protein levels, the nuclear expression of the early NE marker, PAX6, was observed in the high cell density dSMADi group with the expression of CD133 and Vimentin.
Although the mRNA expression of placode marker, nerve growth factor receptor p75, was activated in the high cell seeding density group, there was no HNK1 protein, a marker for migrating NC cells at a detectable level. This indicates that neuroectodermal fate is enforced under the action of dSMAD inhibitors with high cell density under this protocol.
During embryonic developmental stages of body axis elongation, tissues express HOX genes along the rostrocaudal axis of the CNS. 43 The spatial varied expression of HOX genes diversifies the cell fates and restricts the differentiation of specific neural progeny cells. In this study, the treatment with RA, the biologically active form of vitamin A, caused changes of the posterior gene HOX in porcine iPSC-NPCs. Novel differentiation protocols have been reported that permit the controlled patterning towards regional-specific types of neuronal cells by exposing the NPCs to various signalling factors 43,44 ; RA has been reported to be involved in anterior versus posterior patterning during CNS development. 45   Although further studies are needed to definitively uncover this possibility, this is the first study to report the potential application of a differentiation-inducing effect of GBM-CM in porcine iPSC-NPCs.
The porcine iPSCs used in this study showed obvious expression in pluripotent markers such as POU5F1 and SSEA4, which are known to be characterized in other porcine iPSCs. 55-57 SSEA4+ porcine iPSCs are known to be highly capable of differentiating into neural cell types similar to in vivo-derived porcine NPCs. 58 However, one recent study has reported that the efficiency of deriving neural rosette (NR) structure, which consists of a radial arrangement of neuroepithelial cells, may vary for each cell line depending on the expression levels of POU5F1 and SSEA4. 59 According to this report, porcine iPSCs with low POU5F1 and high SEEA4 expression generated limited NR formation. However, we examined ZO-1 expression of NR structure in this study regardless of the relatively low expression of POU5F1.
In conclusion, we developed a robust and efficient platform to generate porcine NPCs from porcine iPSCs and commitment protocols that differentiate into neural and glial cells. The neural induction of porcine iPSCs into NPCs was based on the chemically defined

ACKNOWLEDG EMENTS
This work was supported, in part, by a grant from the "National