Improved Proliferative Capacity of NP-Like Cells Derived from Human Mesenchymal Stromal Cells and Neuronal Transdiferentiation by Small Molecules

Alejandro Aguilera-Castrejon1 · Herminia Pasantes-Morales1 · Juan José Montesinos2 · Lorena V. Cortés-Medina1 ·
Marta E. Castro-Manrreza2,4 · Héctor Mayani3 · Gerardo Ramos-Mandujano1

Received: 6 June 2016 / Revised: 13 October 2016 / Accepted: 20 October 2016 © Springer Science+Business Media New York 2016

Abstract Neural progenitors (NP), found in fetal and adult brain, diferentiate into neurons potentially able to be used in cell replacement therapies. This approach how- ever, raises technical and ethical problems which limit their potential therapeutic use. Alternately, NPs can be obtained by transdiferentiation of non-neural somatic cells evad- ing these diiculties. Human bone marrow mesenchymal stromal cells (MSCs) are suggested to transdiferentiate into NP-like cells, which however, have a low prolifera- tion capacity. The present study demonstrates the requisite of cell adhesion for proliferation and survival of NP-like cells and re-evaluates some neuronal features after difer- entiation by standard procedures. Mature neuronal markers, though, were not detected by these procedures. A chemi- cal diferentiation approach was used in this study to con- vert MSCs-derived NP-like cells into neurons by using a cocktail of six molecules, CHIR99021, I-BET151, RepSox, DbcAMP, forskolin and Y-27632, deined after screening

Electronic supplementary material The online version of this article (doi:10.1007/s11064-016-2086-7) contains supplementary material, which is available to authorized users.

* Gerardo Ramos-Mandujano [email protected]

1División de Neurociencias, Instituto de Fisiología Celular, Universidad Nacional Autónoma de México, 04510 Mexico City, DF, Mexico
2Mesenchymal Stem Cells Laboratory, Oncology Research Unit, Oncology Hospital, National Medical Center, IMSS, Mexico City, Mexico
3Hematopoietic Stem Cells Laboratory, Oncology Research Unit, Oncology Hospital, National Medical Center, IMSS, Mexico City, Mexico
4CONACYT-Children Hospital of Mexico Federico Gómez, Mexico City, Mexico
combinations of 22 small molecules. Direct transdiferen- tiation of MSCs into neuronal cells was obtained with the small molecule cocktail, without requiring the NP-like intermediate stage.

Keywords Mesenchymal stromal cells · Neural progenitor cells · Transdiferentiation · Cell proliferation ·
Neuronal reprogramming · Small molecules

Introduction

Human neural progenitor/stem cells (NPCs) have poten- tial interest for cell-replacement therapy in neurological disorders [1, 2]. NPCs can be isolated from fetal or adult brain tissue [3, 4], or diferentiated from embryonic stem cells [5]. NPCs can be subsequently diferentiated into functional neurons. However, problems of histocompat- ibility, technical limitations or ethical issues raised by the use of these cells restrict the potential of NPCs as a suitable option for cell replacement [6]. Reprogramming of somatic cells is an experimental approach eluding these diiculties [7, 8]. The direct reprogramming option permits genera- tion of NPCs or mature neurons, avoiding the disadvantage associated with an intermediate pluripotent state [9]. Ei- cient cell reprogramming has been generally accomplished by viral-based expression of exogenous genes [10]. This procedure, though, has the risks and limitations inherent to the genetic manipulation [10]. Small molecules modulat- ing signaling pathways and/or epigenetic mechanisms have emerged as valuable tools to manipulate cell fate [11]. A chemical reprogramming of somatic cells has been recently developed, which represents a useful alternative to genetic intervention [12]. By screening of a large number of small molecules, speciic combinations (cocktails) are selected

1 3

to reach the optimal conditions for reprogramming. This approach has been tested in the direct conversion of human ibroblasts or astroglial cells into functional neurons [13, 14].
In the present study, we used human bone marrow mesenchymal stromal cells (MSCs) to generate cells with traits similar to NPCs, which can be later converted into neuronal-like cells by the standard diferentiation proto- col. MSCs are considered as a convenient cell source to generate neuronal cells for therapeutic strategies. They are easy to obtain from postnatal and adult tissues with- out raising ethical conlicts, and its transdiferentiation into ectodermic lineages has been suggested [15–21]. MSCs cultured under neural stem cell-speciic condi- tions on low-attachment culture surfaces generate neuro- sphere-like loating aggregates, formed by cells showing an increase in neural progenitor markers, which have the potential to diferentiate into cells positive for neuronal markers [15, 16, 18, 20]. These cells have been named NP-like cells and are intermediate cells in the neuronal transdiferentiation of MSCs [15]. Generation of NP-like cells from MSCs has been extensively investigated, but many aspect of the process remain controversial. Prolif- eration and self-renewal, which are characteristic features of neural progenitors (NP), have not been conclusively demonstrated. The proliferative ability of NP-like cells has been questioned. The loating structures formed by NP-like cells increase in size with time in culture, and this has been interpreted as evidence of cell prolifera- tion [22]. However, at present there is no agreement as to whether there is an increase with time in the number of cells in cultures, as result of a proliferative activity. While some reports describe cell number increases dur- ing the serial passages [15, 16], others conclude that cells in the aggregates are quiescent [18] or even that their number decreases during the time in culture [20, 21]. Proliferation markers as Ki67 and BrdU are commonly used to evaluate cell proliferation, but unequivocal pres- ence of these markers in cultures of NP-like cells has not been yet established [18, 19, 23]. Inconsistencies are found also regarding neuronal diferentiation of NP-like cells. In some studies NP-like cells do not develop a neu- ronal morphology and in others the neuron-like appear- ance has been related to a cytoskeleton disorganization which mimics the neuronal phenotype [15, 24, 25]. Some markers of immature neurons such as β-III tubulin are consistently found, but detection of mature neuron mark- ers is still controversial [15, 18, 26]. These issues were re-evaluated in the present study by exploring conditions improving the proliferation capacity of NP-like cells and deining the features of the induced neuronal diferen- tiation of these cells. Since we could not detect mature neuronal markers under the standard diferentiation

conditions, we investigated whether a chemical, small molecule-based transdiferentiation protocol may convert MSCs-derived NP-like cells into cells with a neuronal phenotype, a neuronal gene expression proile, and posi- tive to neuronal markers. We also investigated whether MSCs could be directly transdiferentiated into neurons without the NP-like cell intermediate stage.

Materials and Methods

Reagents and Antibodies

Neurobasal medium, low glucose DMEM, DMEM/F-12, FBS, L-glutamine, Glutamax, antibiotics, trypsin/EDTA, human growth factors, N2, B27 all were obtained from Gibco™, Thermo Fisher Scientiic Inc. (MA, USA). Sec- ondary antibodies and annexin V-FITC were obtained from Molecular Probes, Thermo Fisher Scientiic Inc. (MA, USA). Fibronectin, heparin, poly-D-Lysine, gelatin, non- essential aminoacids, BrdU, Hoechst 33258, propidium iodide and RNase were purchased from Sigma-Aldrich (St. Louis, MO, USA). The antibodies that were used included: rat anti-BrdU antibody from Accurate Chemical (NY, USA), mouse anti-Nestin and rabbit anti-Ki67 from Gene- Tex (TX, USA), and mouse anti-βIII tubulin and rabbit anti- Map2 from Santa Cruz Biotechnology (CA, USA). Small molecules, including A83-01, Ascorbic acid, BAYK-8644, BIX01294, CHIR99021, DAPT, Dibutyryl cAMP, Dor- somorphin, Forskolin, GÖ6983, I-BET151, PD0325901, Purmophamine, RepSox, Retinoic acid, RG108, Sodium butyrate, SP600125, Valproic acid and Y-27632 were from Sigma-Aldrich. ISX9 were from Tocris BioScience and SB203580 from Calbiochem.

Isolation and Culture of Human MSCs Derived from Bone Marrow

Human MSCs were isolated from three healthy donors (males, 15–30 years-old) according to the ethical guide- lines of the Villacoapa Hospital, Mexican Institute for Social Security (IMSS), and characterized as previously described [27] (online resource 1). The cells were cultured in low glucose DMEM that was supplemented with 10 % fetal bovine serum (FBS), 4 mM L-glutamine, 50 U/mL of penicillin, 50 µg/mL of streptomycin and 50 µg/mL of gentamicin. The medium was changed every 2 days and cells were subcultured using trypsin/EDTA (0.05 % trypsin, 0.48 mM EDTA) when they reached 90 % conluence. All experiments were performed with MSC between the fourth to sixth passages.

Generation of Neural Progenitor-Like Cells Derived from Human MSC

Human MSCs were plated on ultra-low attachment (Corning, ME, USA) or standard plastic tissue culture plates at a density of 1.5 × 104 cells/cm2 and cultured for 9 days in two diferent serum-free neural stem cell cul- ture media: DMEM/F-12 supplemented with 1 × N2 and 1 × B27 without vitamin A (N2/B27) or human NeuroC- ult™ NS-A Proliferation Medium (hNCult) (StemCell™ Technologies, BC, CA) supplemented with 20 ng/mL each human epidermal growth factor (EGF) and human ibroblast growth factor 2 (FGF2) plus 5 µg/ml heparin. For ibronectin-coated plates (2 µg/cm2), the initial cell density was 3 × 103 cells/cm2 and the same procedure was followed. Cells were incubated at 37 °C in humidi- ied 5 % CO2/95 % air atmosphere. Growth factors were added every 3 days and half medium was changed at day six. To measure cell number after culture, either aggre- gates or monolayer cells were dissociated with trypsin/
EDTA and counted directly on a Neubauer chamber. In all cultures, viability was determined by a trypan blue exclusion assay.
For neuronal diferentiation, dissociated MSCs or NP-like cells were plated at a concentration of 1 × 103 cells/cm2 on poly-D-Lysine/ibronectin coated coverslips and incubated by 12 days with Neurobasal medium plus 1 × N2, 0.5 % FBS and 10 ng/mL human brain derived neurotrophic factor (hBDNF) (NB:N2:FBS:BDNF).

Generation of Chemically Induced Neurons

Human MSCs or NP-like cells were seeded onto ibronectin(2 µg/cm2)/0.1 % gelatin-coated plates at a density of 20 × 103 cells/cm2 and cultured in MSCs growth medium or N2/B27, respectively, for 1 day. To chemical induction, cells were washed with basal medium and then cultured with neuronal induction medium (50 % Neurobasal medium, 50 % DMEM/F12 with 1 × N2 and 1 × B27 with vitamin A, 1 × Glutamax, 1 × nonessential aminoacids, and 20 ng/mL FGF2 plus 5 µg/mL heparin) plus deined chemical cocktails for
8days at 37 °C and 5 % CO2. Concentrations of mol- ecules for the ICFRYA cocktail were 1 µM I-BET151, 20 µM CHIR99021, 50 µM Forskolin, 1 µM RepSox,
5µM Y-27632 and 100 µM dbcAMP. Neuronal induc- tion medium was replaced at day 4 (prepare neuronal induction medium with small molecules freshly before induction). Information about all chemicals used in this study can be found in the online resource 5.

BrdU Incorporation and Immunocytochemical Detection

For BrdU incorporation assays at 3 and 6 days, pulses of BrdU (10 µM) were applied to cultures 72 h before ixa- tion. Cells were collected, washed, resuspended with culture medium and plated on poly-D-Lysine/ibronectin treated coverslips. 2 h after plating, cells were ixed with
4% cold paraformaldehyde for 15 min, washed (3 times,
5min each) with PBS/0.1 % BSA, incubated with pre- heated PBS-2N HCl solution (37 °C, 10 min), washed and permeabilized/blocked using PBS/0.1 % BSA + 10 % GS (goat serum) + 0.3 % Triton X-100 (45 min at room temperature). Then, cells were incubated overnight at 4 °C with rat anti-BrdU antibody (1:400) followed by 1 h incubation with Alexa Flour 488-conjugated goat anti-rat IgG (1:200). Nuclei were counterstained with 2 µg/mL Hoechst 33258 diluted in PBS. Microphotographs were obtained by direct epiluorescence microscope Olympus IX71 using the QCapture Pro 6.0 software and data ana- lyzed with ImageJ software. The percentage of BrdU-pos- itive cells was calculated over the total of cells counted.

Immunocytochemistry

For immunocytochemical analysis cultured cells were ixed, permeabilized/blocked, and then incubated overnight at 4 °C with the following antibodies: mouse anti-Nestin (1:500), rabbit anti-Ki67 (1:50), mouse anti-β-III tubulin (1:1000), rabbit anti-Map2 (1:500) followed by 1 h incu- bation with goat anti-mouse Alexa 488 or goat anti-rabbit Alexa 568. Images were obtained using an epiluorescence Olympus IX71 microscope with a ×10 objective and ana- lyzed using ImageJ software. At least 500 nuclei from ive randomly selected ields were counted to calculate the per- centage of positive cells.

Quantiication of Apoptotic and Necrotic Cells

Cells were harvested, washed with cold PBS, and then stained for detection of apoptotic or necrotic cells. Propidium iodide (1 μg/mL) was used to label necrotic cells. To label apoptotic cells, exposure of phosphatidylserine on the outer cell mem- brane was detected using annexin V-FITC prepared in annexin- binding bufer (10 mM HEPES, 140 mM NaCl, and 2.5 mM CaCl2, pH 7.4) according to the manufacturer´s instructions. A two-color luorescence analysis was conducted using an Attune Acoustic Focusing Cytometer (BL-1/BL-2 channels) and data were analyzed using Attune Cytometric Software v1.2.5.

In Situ Analysis of Cell Viability

Cells were incubated for 30 min with 8 µg/mL Hoechst. After the incubation period, propidium (1 µg/mL) iodide was added. Microphotographs were obtained using a direct epiluorescence microscope Olympus 1 × 71 with ×10 mag- niication and the QCapture Pro 6.0 software.

LDH Release Assay

Necrotic cells were quantitatively assessed by the measure- ment of lactate dehydrogenase (LDH) released in culture medium at 3, 6 and 9 days. Culture supernatants were col- lected and centrifuged (300 g, 5 min) to remove cell debris. LDH release was quantiied using the Pierce LDH Cytotox- icity Assay Kit (Thermo Fisher Scientiic Inc., MA, USA), according to the manufacturer’s instructions and measuring absorbance at 490 nm using a Synergy™ HT plate reader (Biotek, VT, USA). To calculate the percentage of LDH release in each condition, data were normalized to a back- ground LDH control from culture medium without cells and a 100 % LDH content in 0.1 % Triton X-100 lysates.

MTT Assay

Cell viability was evaluated by adding 20 µL per well of il- tered MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra- zolium bromide) (12 mM stock in PBS) and incubating at 37 °C per 2 h. The inal volume of culture medium in each well was 0.2 mL. After the incubation period, media was removed and 100 µL of DMSO were added to dissolve the formazan. The absorbance was measured on a Synergy™ HT plate reader at 570 nm.

Real-Time Quantitative PCR (qPCR)

RNA samples from cultures were extracted with the RNe- asy Mini Kit (Qiagen, MD, USA) according to the manu- facturer’s instructions. The concentration of RNA samples was quantiied in a NanoDrop spectrophotometer (Thermo scientiic). Complementary DNA (cDNA) was synthesized from 1 μg RNA with the iScript cDNA Synthesis Kit (Bio- Rad Laboratories Inc., CA, USA) and a Veriti® 96-Well Thermal Cycler (Applied Biosystems). PCR primers were designed using the last reference sequence (RefSeq) ver- sion of each gene with the PrimerBlast software. Online resource 3 includes a list of primers used in this study. qPCR reactions were conducted using QuantiFast SYBR Green PCR Kit (Qiagen, MD, USA), according to the manufacturer’s instruction, and run on a StepOnePlus Real- Time PCR system (Applied Biosystems). After ampliica- tion, melting curves of the RT-PCR products were acquired to demonstrate product speciicity. PCR eiciency was

optimal and ranged from 90 to 100 % in the diferent target gene qPCR assays (online resource 3, 4A). For selection of the best reference gene, a comparison of the transcription variation for diferent housekeeping genes in response to conditions was performed. The gene with the least CT vari- ation between treatments and control cultures was selected (online resource 4B, C). Thereby, gene expression was normalized to TATA-Box binding protein (TBP) mRNA. The relative quantiication of mRNA levels was performed using the Pfal method [28].

Statistical Analysis

All results are expressed as the mean ± standard error (SEM) of three biological samples. The statistical analy- ses were performed using the Student’s t test for between- group comparisons and a one-way ANOVA and the Tukey test for multiple comparisons. GraphPad Prism software for Windows version 6 (La Jolla, CA, USA) was used for all statistical procedures. Diferences with p < 0.05 were con- sidered statistically signiicant.

Results

MSCs Generate Neurosphere-Like Structures
with High Nestin Reactivity But Without Proliferation Capacity

Human MSCs were isolated and characterized from three bone marrow samples (“Materials and Methods” section; online resource 1a, b). MSCs showed the typical ibroblas- tic morphology and weak nestin reactivity (Fig. 1a, upper panel). MSCs proliferation, as shown in Fig. 1b, was main- tained during the 9 days in culture (online resource 1c). To generate NP-like cells, MSCs were seeded on low attach- ment plates with N2/B27 neural medium containing hEGF/
hFGF. As previously reported [15–17, 19], this condition induced formation of neurosphere-like structures with high nestin reactivity (Fig. 1a, medium panel). The number of cells from disaggregated neurospheres did not increased after 9 days in culture (Fig. 1b). Seeding MSCs in human NeuroCult™ NS-A proliferation medium (hNCult), a spe- ciic medium to expand human neural stem cells, similarly induced the formation of neurosphere-like structures, but also failed to increase the cell number (Fig. 1a, lower panel, b). These results suggest that the proliferation rate and/or cell viability were afected during the generation of loating neurosphere-like structures. To clarify this point we next compared proliferation, viability and other properties of NP-like cells generated as loating structures with those in monolayer, in the presence or absence of the cell adhesion molecule ibronectin.

Fig. 1 Generation of neural progenitor (NP)-like cells from human mesenchymal stromal cells (MSCs). a Bright ield and nestin immunoluores- cence images of MSCs and neurosphere-like structures obtained after 9 days of culture on low-attachment plates, with
the neural expansion media N2/
B27 or hNCult. b Cell number in controls and in neurosphere- like structures. The dotted line indicates the initial cell number. Values are the mean ± SEM of the three independent sam-
ples. *p < 0.01 between MSCs control and NP-like cells. Images were obtained using an Olympus IX71 microscope with a ×10 objective. Nuclei were counterstained with Hoechst 33258. Scale bars represent
50 µm

NP-Like Cells Grown as Monolayers on a Fibronectin-Coated Surface Show Restored Cell Proliferation and Sustained Viability

NP-like cell cultures, grown either as loating neuro- sphere-like structures or as monolayer on standard plates, showed no increase in cell number with time (Fig. 2a). In contrast, cultures on ibronectin-coated plates (2 µg/cm2) showed a progressive increase in cell number, up to ivefold after 9 days in culture (Fig. 2a). Cell proliferative capac- ity was assayed by BrdU incorporation at 3 and 6 days in neurosphere-like, standard or ibronectin-coated surface cultures. At 3 days, BrdU+ cells were found in all condi- tions (Fig. 2b), but the number of positive cells in neuro- sphere-like or standard surface cultures was only 5 and
6% respectively, whereas in the ibronectin-coated condi- tion this proportion increased up to 50 % (Fig. 2b). After
6days in culture, only ibronectin-coated cultures showed BrdU+ cells (27 %) (Fig. 2b). Fibronectin concentrations ranging 0.5–4 µg/cm2 or addition of poly-D-Lysine (5 µg/
cm2) did not modify the percentage of BrdU incorporation (online resource 2a, b). The cell cycle marker Ki67 was used to further characterize NP-like cell proliferation. In neurosphere-like structures or standard dishes, few or none positive cells were detected (Fig. 2c), whereas in ibronec- tin-coated cultures, the percentage of Ki67+ cells was 13 % at day 3 and 3 % at day 6 (Fig. 2c). Analysis of cell cycle phases also indicated that the percentage of cells in the S and G2/M phases was signiicantly higher in the ibronectin condition (data no showed).
Evaluation of cell viability by low cytometry at 6 days in culture showed that ibronectin-adhesion decreased sig- niicantly the proportion of apoptotic cells, without show- ing an efect on necrotic death (Fig. 3a). These assays were made on trypsin-disaggregated cells from neurospheres or monolayers; however, a dark core indicative of cell death in the neurosphere-like structures is detected at 6-culture days, suggesting that a signiicant number of death cells are discarded during the trypsin treatment. To better evalu- ate necrotic death, propidium iodide staining was used to visualize necrosis in situ. As shown in Fig. 3b, a high num- ber of necrotic cells were found in neurosphere-like struc- tures, few in standard plates while no signal of necrosis in ibronectin-coated cultures. Necrosis in the three types of cultures was also evaluated by LDH release. The necrosis percentage in low-attachment or standard plates, increased with time, attaining 18 and 14 % respectively, at 9 days in culture (Fig. 3c). In ibronectin-coated cultures the necrotic percentage was remarkably lower, of only 3 % of necrosis at
9days in culture (Fig. 3c).

NP-Like Cells Express Neural Progenitor Markers, But Did Not Diferentiate into MAP2+ Cells

The presence of the NP marker nestin was detected in the non-adherent, standard and ibronectin-coated cultures (Fig. 4a). The percentage of nestin+ cells increased pro- gressively and was similar in the three types of cultures (Fig. 4b). Changes in the expression of genes associated to neural stem/progenitor cells (Nestin, Sox2, Sox1, and

Fig. 2 Cell number and proliferative markers in NP-like cells cul- ▸ tured on low-attachment, standard or ibronectin-coated plates. a Fold change in cell number. b BrdU incorporation. c Ki67 reactivity. Rep- resentative images correspond to 3-day cultures. BrdU incorporation was measured after a 72 h BrdU pulse. Data are the mean ± SEM of three independent samples. *p < 0.01 between the diferent culture conditions. Scale bars 50 µm

Pax6) or to mesenchymal stromal cells (CD73, CD90, and CD105) were investigated in the NP-like cells (Fig. 4c). In all culture conditions cells up-regulated the neural progeni- tor genes Nestin, SOX2, SOX1 and PAX6 genes and down- regulated the mesenchymal markers CD105 and CD90 (Fig. 4c).
Neuronal diferentiation was induced in NP-like cells using neurotrophic factors, according to previous reports [15]. In MSCs control cultures, the diferentiation proce- dure did not induced morphological changes nor increased reactivity to neuronal (Fig. 4d) or glial markers (data not shown). NP-like cells from neurosphere-like struc- tures, standard plate and ibronectin condition all showed neuronal-like morphology and reactivity to β-III tubulin (Fig. 4d). However, reactivity to the mature neurons marker MAP2 (Fig. 4d) or to the astrocyte marker GFAP (data not shown) was not observed in any condition.

Neuronal Transdiferentiation of MSCs by Small Molecules

Neuronal transdiferentiation was conducted either directly in MSCs or in intermediate NP-like cells by chemical induc- tion with a number of the compounds generically known as small molecules. These are molecules acting through activation/inhibition of signaling pathways and epigenetic modiiers. The molecules were selected from those known to regulate signaling pathways implicated in neuronal spec- iication and inhibition of mesoderm diferentiation. Other molecules targeting epigenetic modiications were chosen to increase cell reprogramming eiciency (online resource 5). The 22 small molecules selected were: A83-01, ascorbic acid, BAYK-8644, BIX01294, CHIR99021, DAPT, Dibu- tyryl cAMP, dorsomorphin, forskolin, GÖ6983, I-BET151, ISX9, PD0325901, purmophamine, RepSox, retinoic acid, RG108, sodium butyrate, SB203580, SP600125, valproic acid and Y-27632. The analysis was irst performed in MSCs. Diferent protocols, testing numerous combinations (cocktails) of small molecules were assayed to ind those better inducing neuronal morphology and speciic markers for immature or mature neurons. Some cocktails not induc- ing neuronal-like morphology nor increasing Tubulin β-III reactivity, or those afecting cell viability were discarded (Electronic supplementary material 6). Pre-induction with epigenetic modiiers or changing the cocktail of molecules

Fig. 3 Cell viability analysis in NP-like cells cultured on low-attach- ment, standard or ibronectin-coated plates. a Percentage of apop- totic (Annexin V+/PI-) and necrotic (PI+/Annexin V-) cells meas- ured by low cytometry at 6-day cultures. b Representative images of necrotic cells visualized by PI staining. Images were obtained by direct epiluorescence using a ×10 objective with an Olympus IX71 microscope. Nuclei were counterstained with Hoechst. c Quantiica-

tion of necrosis by LDH release. Results are expressed as percentage of necrosis in the indicated conditions. Values are the mean ± SEM of three independent samples. Scale bar 100 µm. In c signiicant dif- ferences between ibronectin-coated vs. low-attachment and stand- ard plate (*p < 0.01), and between low-attachment vs. standard plate (+p < 0.01). PI propidium iodide, LDH lactate dehydrogenase

during the time in culture was not eicient (Electronic sup- plementary material 6). The best results were obtained with a cocktail containing I-BET151 (inhibitor of the BET pro- teins family), CHIR99021 (Glycogen synthase kinase 3-β inhibitor), forskolin (cAMP agonist), Repsox (TGF-βRI inhibitor), Y-27632 (inhibitor of Rho-associated kinase), and DbcAMP (Fig. 5a). In the presence of this six mole- cule cocktail (ICFRYA) we observed cells with neurite-like prolongations (Fig. 5b), strong Tubulin β-III reactivity, and cells positive to the mature neurons marker MAP2 (Fig. 6). Each molecule alone failed to induce neuronal properties (Electronic supplementary material 6). By withdrawing one or more of the six molecules from the cocktail, we

identiied I-BET151 and CHIR99021 as the most efective neural induction molecules (Fig. 5c; online resource 7); and RepSox, Y-27632, forskolin and DbcAMP as those which contribute to maintain cell viability (Fig. 5c, d). The cock- tail vehicle DMSO alone had no efects on cells (Fig. 5b). As expected, neuronal induction results in the arrest of cell proliferation. This was conirmed by using the prolifera- tion marker Ki67 and incorporation of BrdU. After 2 days of chemical induction, no positive cells to Ki67 were found and only a low percentage (about 8 %) of cells were posi- tive to BrdU (online resource 8).
The efect on neuronal diferentiation of the small mol- ecule cocktail observed in MSCs was compared to that in

Fig. 4 Neural progenitor markers and neuronal dif- ferentiation of NP-like cells. a Representative immuno-
luorescence images for nestin reactivity in NP-like cells cultured in low attachment, standard or ibronectin-coated plates at 9-days. b Percentage of nestin+ cells quantiied at 3, 6 and 9 days in culture. c Gene expression analysis by qRT- PCR in NP-like cells cultured in low-attachment, standard
or ibronectin-coated plates at 6-day cultures. Bars represent
fold-changes relative to control (MSCs). Values were normal- ized to TBP expression. d β-III tubulin/MAP2 immunoluores- cence after neuronal induc-
tion of MSCs control cultures and NP-like cells cultured
in low attachment, standard or ibronectin-coated plates. Values are represented by the
mean ± SEM (n = 3). Scale bars represent 50 µm (a) and 100 µm (d)

intermediate NP-like cells obtained as neurosphere-like aggregates or as ibronectin-coated monolayers. By this comparison we would be able to demonstrate whether cells with neuronal morphology and showing mature neuron markers can be obtained directly from MSCs or if the intermediate step of NP-like cells is necessary. Fol- lowing chemical induction with the ICFRYA cocktail, the NP-like cells grown in low attachment or ibronectin- coated plates showed neuronal morphology, β-III tubu- lin+ and MAP2+ cells (Fig. 6a). The comparative results between MSCs and NP-like cells are shown in Fig. 6b, c. The percentage of cells positive to β-III tubulin ranged 65–72 % in MSCs and in NP-like cells, regardless of the adherence conditions in cultures (Fig. 6b). A proportion of 12–15 % of β-III tubulin+ cells co-expressed MAP2 (Fig. 6c). Therefore, no signiicant diferences in neu- ronal induction by ICFRYA were found between MSCs,
neurosphere-like or ibronectin cultures. Neurite-like structures increased in length during culture time, and this increase was no diferent between MSCs and NP- like cells (Fig. 6d). To further characterize the small molecule-induced neuron-like cells, changes the expres- sion of genes associated with neuronal, glial and mes- enchymal/mesodermal lineages was evaluated in both MSCs and intermediate NP-like cells. In the two cell types, after 8 days of neuronal induction by the cocktail, the expression of the neuronal genes MASH1, NCAM, TUB β-III, DCX, NEUROD1, NEFH, MAP2 and NeuN was up-regulated (Fig. 7a) and the expression of the mes- enchymal genes (CD73, FN1,CD90,CD105) was down- regulated (Fig. 7b). The expression of the glial genes (GFAP, S100B, MBP, CNP) remained mainly unchanged (Fig. 7c).

Fig. 5 Neuronal morphology induced by a speciic cocktail of small molecules. a Deined combination of six chemi-
cal compounds (ICFRYA). b Representative phase contrast images of MSCs treated with vehicle or with the ICFRYA cocktail for 8 days. Efect of removing chemicals from
the ICFRYA cocktail on the percentage of cells with neu- ronal morphology (c) and on cell viability assayed by MTT (d). Data are means ± SEM
of three independent samples. Signiicant diferences between a deined combination with respect to the ICFRYA cocktail (+p < 0.05; *p < 0.01). Scale bar 100 µm

Discussion

The present results described that NP-like cells derived from MSCs growing as neurospheres have a low prolif- erative capacity, thus lacking one of the most distinctive features of NP. The marginal increase in cell number in cultures as well as the low percentage of cells positive to the proliferation markers Ki67 and BrdU suggest that the increase in size of neurosphere-like structures is a consequence of cellular coalesce or aggregation, rather than of cell proliferation. We found here that adhesion was a crucial element for the proliferative process since a robust cell proliferation occurs in cultures grown as mon- olayers on ibronectin-coated plates, but not on standard plates. The ibronectin-adherent condition also maintains high viability of NP-like cells, which contrasts with the signiicant percentage of necrotic cells found in the neu- rosphere-like cultures. These results are not unexpected, since adhesion molecules, such as ibronectin stimulate signals regulating cell proliferation, migration and sur- vival [29, 30]. Other attributes of NP-like cells such as nestin reactivity and increase in the expression of neural progenitor genes were unchanged in the loating aggre- gates or monolayer cultures, suggesting that these fea- tures are independent of cell adhesion.
NP-like cells diferentiated by the common protocols showed neuronal-like morphology and reactivity to β-III tubulin when cultured either as neurosphere-like struc- tures or on monolayers with or without ibronectin, sug- gesting that adhesion has no inluence on these processes. Cells obtained after the diferentiation procedure, showed a deined neuronal morphology, but we could not detect the mature neuronal marker MAP2, under any condition. This was possible only after replacing the diferentiation protocol by a chemical diferentiation procedure based on addition of small molecules. A cocktail of six speciic small molecules (ICFRYA) was selected after screening combinations of 22 small molecules. By treatment with ICFRYA, NP-like cells could diferentiate into neuronal cells as deined by their cell phenotype, gene expression proile and the presence of neuronal markers including MAP2.
The small molecule cocktail used successfully to achieve the NP-like neuronal diferentiation was also suitable to accomplish a direct chemical transdiferentia- tion of MSCs. The irst studies demonstrating a chemical neuronal reprogramming based on the screening of dif- ferent combinations of small molecules were conducted in mouse embryonic ibroblasts [31], adult human ibro- blasts [13], and later on astroglial cells [14]. The small

◂Fig. 6 Mature neuron markers and cell size in MSCs and NP-like cells treated with the ICFRYA cocktail. a Immunostaining for β-III tubulin and MAP2 in MSCs control or NP-like cells from neuro- sphere-like and ibronectin cultures induced for 8 days with the six molecules cocktail. Quantiication of β-III tubulin positive cells (b) and β–III tubulin/MAP2 double positive cells (c) induced by small molecules. d Cell length measurements after 1, 4 and 8 days of ICFRYA. Data are the means ± SEM of n ≥ 20 random ields of a ×10 objective from three independent samples. Scale bar 100 µm

molecule cocktail used in our study for chemical direct reprogramming of MSCs was also based on this selection procedure.

Cells obtained by MSCs reprogramming in our study by treatment with ICFRYA show deined neuronal fea- tures. They exhibit a neuronal phenotype, with progres- sive neurite-like outgrowth, requiring some time in culture to get established, consistent with a true acquisition of a neuronal phenotype and not with cytoskeletal disorgani- zation and cell shrinkage. The direct neuronal reprogram- ming typically bypasses the intermediate proliferative stage [13, 31, 32]. This proliferation arrest was observed in our study by Ki67 and BrdU after cell exposure to the small molecule cocktail, in further support of the direct neuronal

Fig. 7 Gene expression proile of chemically induced neuron-like Bars represent fold-changes relative to that of MSCs control, which

cells. Relative expression of neuronal (a), glial (b) and MSCs (c)
is considered as 1 (dotted line). C
T
values were normalized to TBP

markers at day 8 of chemical induction, as measured by qRT-PCR. expression. Data are the mean ± SEM of three independent samples

MSCs transdiferentiation by this treatment. MSCs were converted into cells with reactivity to immature/mature neuronal markers, which up-regulated neuronal genes and down-regulated mesenchymal/mesodermal genes. All these features support the conversion of mesenchymal into neu- ronal cell lineage.
The molecule cocktail tested in our study contains six molecules, CHIR99021, I-BET151, RepSox, DbcAMP, for- skolin and Y-27632. CHIR99021 and particularly I-BET151 were determinant for the neuronal transdiferentiation of MSCs. CHIR99021 has been commonly used to induce neuronal diferentiation from pluripotent stem cells [33] and has been included in small molecule cocktails for chemical neuronal reprogramming of ibroblasts and astrocytes [13, 14, 31]. The efect of CHIR99021 on neuronal transdiferen- tiation is suggested to occur by activating the Wnt pathway through inhibition GSK-3β [34]. In our study I-BET151 was a crucial molecule for the MSCs chemical neuronal transdif- ferentiation since its withdrawal from the cocktail prevents completely the neuronal induction. The key role of I-BET151 has been stressed in the chemical neuronal reprogramming of embryonic ibroblasts reported previously by Li et al. [31]. As discussed in the later study, I-BET151 may act by inter- fering with proteins of the BET family, involved in maintain- ing the gene expression pattern of the initial cell identity. This disruption of the cell-fate gene expression could be a prereq- uisite for the successful MSCs chemical neuronal reprogram- ming. The other molecules of the cocktail used in our study, DbcAMP, forskolin, Y-27632 and RepSox, appear to be more relevant for maintaining cell viability. Such protective action has been described at least for Y-27632 via the Rho/ROCK signaling pathway which is important for neuronal survival [35, 36].
Results of the present study identiied a cocktail of small molecules for the chemical neuronal transdiferentiation of human bone marrow MSCs. Our indings indicate that involvement of an intermediate NP-like cell is not increas- ing the eiciency of neuronal conversion as compared to the direct reprogramming of MSCs. Further research is con- ducted to optimize the neuronal induction and generate trans- diferentiated neurons with the biophysical and biochemical features of functional neurons. Also, studies are in progress in our laboratory to comparatively examine the MSC neuronal reprogramming here shown for bone marrow cells, in other adult or neonatal sources. This may be an important issue in terms of eiciency as well as on the accessibility of the source. It would be also of interest is to assay diferent combi- nations of small molecule cocktails directed to obtain diverse neuron subtypes for potential cell replacement therapy in neurological dysfunctions.

Acknowledgments This work was supported by the Dirección General de Asuntos del Personal Académico (DGAPA), Universidad

Nacional Autónoma de México (UNAM) (H.P.M. and G.R.M., Grant No. IN205916). We are indebted to the National Council of Science and Technology (CONACYT) for support to J.J.M.M. (Grant No. 258205), IMSS support to J.J.M.M. (Grant No. 1311), and CONA- CYT Red temática células troncales y medicina regenerative (Grant No. 271609). A.A.C. acknowledge the inancial support provided by the “Ayudante de Investigador Nacional Emérito” fellowship of CON- ACYT (Núm. Exp. 10802). We thank QFB Carlos Castellanos Barba from the low-cytometry unit at the Instituto de Investigaciones Bio- médicas for his expert technical assistance in low cytometry proto- cols. We also thank Dr. Laura Ongay Larios from the molecular biol- ogy unit at the Instituto de Fisiología Celular for her assistance with qPCR.

Author Contributions A.A.C. conception and design, collec- tion and/or assembly of data, data analysis and interpretation, igure preparation, manuscript writing and inal approval of manuscript. H.P.M. Conception and design, inancial support, manuscript writing, and inal approval of manuscript. L.V.C.M. collection and/or assem- bly of data and data analysis and interpretation. J.J.M. Provision of study material, inancial support, and inal approval of manuscript; M.E.C.M. Provision of study material and collection and assembly of data; H.M. Financial support and inal approval of manuscript; G.R.M. conception and design, collection and/or assembly of data, data analysis and interpretation, igure preparation, inancial support, manuscript writing and inal approval of manuscript.

Compliance with Ethical Standards

Conlict of Interest The authors declare that they have no conlict of interest.

References

1.Martino G, Pluchino S (2006) The therapeutic potential of neu- ral stem cells. Nat Rev Neurosci 7(5):395–406. doi:10.1038/
nrn1908
2.Harris L, Zalucki O, Piper M et al (2016) Insights into the biol- ogy and therapeutic applications of neural stem cells. Stem Cells Int 2016:9745315. doi:10.1155/2016/9745315
3.Hernandez-Benitez R, Vangipuram SD, Ramos-Mandujano G et al. (2013) Taurine enhances the growth of neural precursors derived from fetal human brain and promotes neuronal speciica- tion. Dev Neurosci 35(1):40–49. doi:10.1159/000346900
4.Reynolds BA, Weiss S (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255(5052):1707–1710
5.Zhang SC, Wernig M, Duncan ID et al (2001) In vitro dif- ferentiation of transplantable neural precursors from human embryonic stem cells. Nat Biotechnol 19(12):1129–1133. doi:10.1038/nbt1201-1129
6.Barker RA, de Beaufort I (2013) Scientiic and ethical issues related to stem cell research and interventions in neurode- generative disorders of the brain. Prog Neurobiol 110:63–73. doi:10.1016/j.pneurobio.2013.04.003
7.Prasad A, Manivannan J, Loong DT et al (2016) A review of induced pluripotent stem cell, direct conversion by trans-difer- entiation, direct reprogramming and oligodendrocyte diferentia- tion. Regen Med 11(2):181–191. doi:10.2217/rme.16.5
8.Takahashi K, Tanabe K, Ohnuki M et al (2007) Induction of pluripotent stem cells from adult human ibroblasts by deined factors. Cell 131(5):861–872. doi:10.1016/j.cell.2007.11.019

9.Xu J, Du Y, Deng H (2015) Direct lineage reprogramming: strat- egies, mechanisms, and applications. Cell stem cell 16(2):119– 134. doi:10.1016/j.stem.2015.01.013
10.Takahashi K, Yamanaka S (2016) A decade of transcription fac- tor-mediated reprogramming to pluripotency. Nat Rev Mol Cell Biol 17(3):183–193. doi:10.1038/nrm.2016.8
11.Li W, Li K, Wei W et al (2013) Chemical approaches to stem cell biology and therapeutics. Cell Stem Cell 13(3):270–283. doi:10.1016/j.stem.2013.08.002.
12.Hou P, Li Y, Zhang X et al (2013) Pluripotent stem cells induced from mouse somatic cells by small-molecule compounds. Sci- ence 341(6146):651–654. doi:10.1126/science.1239278
13.Hu W, Qiu B, Guan W et al (2015) Direct conversion of normal and Alzheimer’s disease human ibroblasts into neuronal cells by small molecules. Cell Stem Cell 17(2):204–212. doi:10.1016/j. stem.2015.07.006
14.Zhang L, Yin JC, Yeh H et al (2015) Small molecules eiciently reprogram human astroglial cells into functional neurons. Cell Stem Cell 17(6):735–747. doi:10.1016/j.stem.2015.09.012
15.Hermann A, Gastl R, Liebau S et al (2004) Eicient generation of neural stem cell-like cells from adult human bone marrow stromal cells. J Cell Sci 117(Pt 19):4411–4422. doi:10.1242/
jcs.01307
16.Fu L, Zhu L, Huang Y et al (2008) Derivation of neural stem cells from mesenchymal stemcells: evidence for a bipoten- tial stem cell population. Stem Cells Dev 17(6):1109–1121. doi:10.1089/scd.2008.0068
17.Harris VK, Faroqui R, Vyshkina T et al (2012) Characterization of autologous mesenchymal stem cell-derived neural progeni- tors as a feasible source of stem cells for central nervous sys- tem applications in multiple sclerosis. Stem Cells Transl Med 1(7):536–547. doi:10.5966/sctm.2012-0015
18.Heo JS, Choi SM, Kim HO et al (2013) Neural transdiferentia- tion of human bone marrow mesenchymal stem cells on hydro- phobic polymer-modiied surface and therapeutic efects in an animal model of ischemic stroke. Neuroscience 238:305–318. doi:10.1016/j.neuroscience.2013.02.011
19.Ma K, Fox L, Shi G et al (2011) Generation of neural stem cell- like cells from bone marrow-derived human mesenchymal stem cells. Neurol Res 33(10):1083–1093. doi:10.1179/17431328 11Y.0000000053
20.Balasubramanian S, Thej C, Venugopal P et al (2013) Higher propensity of Wharton’s jelly derived mesenchymal stromal cells towards neuronal lineage in comparison to those derived from adipose and bone marrow. Cell Biol Int 37(5):507–515. doi:10.1002/cbin.10056
21.Chung CS, Fujita N, Kawahara N et al (2013) A comparison of neurosphere diferentiation potential of canine bone marrow- derived mesenchymal stem cells and adipose-derived mesenchy- mal stem cells. J Vet Med Sci 75(7):879–886
22.Huat TJ, Khan AA, Pati S et al (2014) IGF-1 enhances cell prolif- eration and survival during early diferentiation of mesenchymal

stem cells to neural progenitor-like cells. BMC Neurosci 15:91. doi:10.1186/1471-2202-15-91
23.Yang E, Liu N, Tang Y et al (2015) Generation of neuro- spheres from human adipose-derived stem cells. Biomed Res Int 2015:743714. doi:10.1155/2015/743714
24.Neuhuber B, Gallo G, Howard L et al (2004) Reevaluation of in vitro diferentiation protocols for bone marrow stromal cells: disruption of actin cytoskeleton induces rapid morpho- logical changes and mimics neuronal phenotype. J Neurosci Res 77(2):192–204. doi:10.1002/jnr.20147
25.Hermann A, Liebau S, Gastl R et al (2006) Comparative analysis of neuroectodermal diferentiation capacity of human bone mar- row stromal cells using various conversion protocols. J Neurosci Res 83(8):1502–1514. doi:10.1002/jnr.20840
26.Chen G, Wang Y, Xu Z et al (2013) Neural stem cell-like cells derived from autologous bone mesenchymal stem cells for the treatment of patients with cerebral palsy. J Transl Med 11:21. doi:10.1186/1479-5876-11-21
27.Montesinos JJ, Flores-Figueroa E, Castillo-Medina S et al (2009) Human mesenchymal stromal cells from adult and neonatal sources: comparative analysis of their morphology, immunophe- notype, diferentiation patterns and neural protein expression. Cytotherapy 11(2):163–176. doi:10.1080/14653240802582075
28.Pfal MW (2001) A new mathematical model for relative quanti- ication in real-time RT-PCR. Nucleic Acids Res 29(9):e45
29.Pankov R, Yamada KM (2002) Fibronectin at a glance. J Cell Sci 115(Pt 20):3861–3863
30.Sottile J, Hocking DC, Swiatek PJ (1998) Fibronectin matrix assembly enhances adhesion-dependent cell growth. J Cell Sci 111(Pt 19):2933–2943
31.Li X, Zuo X, Jing J et al (2015) Small-molecule-driven direct reprogramming of mouse ibroblasts into functional neurons. Cell Stem Cell 17(2):195–203. doi:10.1016/j.stem.2015.06.003
32.Vierbuchen T, Ostermeier A, Pang ZP et al (2010) Direct con- version of ibroblasts to functional neurons by deined factors. Nature 463(7284):1035–1041. doi:10.1038/nature08797
33.Li W, Sun W, Zhang Y et al (2011) Rapid induction and long-term self-renewal of primitive neural precursors from human embryonic stem cells by small molecule inhibitors. Proc Natl Acad Sci USA 108(20):8299–8304. doi:10.1073/
pnas.1014041108
34.Kim H, Wu J, Ye S et al (2013) Modulation of beta-catenin func- tion maintains mouse epiblast stem cell and human embryonic stem cell self-renewal. Nature Commun 4:2403. doi:10.1038/
ncomms3403
35.Lamas NJ, Johnson-Kerner B, Roybon L et al (2014) Neuro- trophic requirements of human motor neurons deined using ampliied and puriied stem cell-derived cultures. PloS One 9(10):e110324. doi:10.1371/journal.pone.0110324
36.Xi G, Hu P, Qu C et al (2013) Induced neural stem cells gener- ated from rat ibroblasts. Genomics Proteomics Bioinformatics 11(5):312–319. doi:10.1016/j.gpb.2013.09.003E-616452