Human foreskin fibroblasts were obtained from Millipore (Catalog no. SCC058). Cells were maintained as per the protocol in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA), 10% fetal bovine serum (FBS) (Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA), and 1% penicillin-streptomycin (P/S) (Welgene, Inc., Gyeongsan, Korea).

Glass coverslips were pretreated as previously described by Richner et al. (16). In brief, the coverslips were washed with detergent followed by washing with distilled water (DW) six times for 30 mins each. After washing, coverslips were dehydrated using 95% ethanol for 2 hrs before overnight treatment with nitric acid. Thereafter, the cover slips were washed 6 times with DW, 30 mins each, and were UV sterilized. Before their use for cell culture, glass coverslips were coating with 1μg/ml laminin (Sigma-Aldrich, Inc., Saint Louis, MO, USA) in 1×PBS for 4 hrs.

Human fibroblasts were cultured on pretreated and laminin-coated 12 mm coverslips at a seeding density of 25,000 to 30,000 cells per well of a 24-well plate in DMEM, 10% FBS, and 1% P/S, one day prior to neuronal induction. For the direct conversion of HF to neuron-like cells, three different protocols were generated and tested. In the protocol 1 of neuronal conversion, the seeded cells in the first step were treated with a chemical cocktail containing small molecules 0.5 μM JQ-1(+) (MedChem-Express, Monmouth Junction, NJ, USA), 10 μM CHIR99021 (MedChemExpress, Monmouth Junction, NJ, USA), 5 μM RepSox (MedChemExpress, Monmouth Junction, NJ, USA), 12.5 μM forskolin (MedChemExpress, Monmouth Junction, NJ, USA), 10 μM Y276332 (MedChemExpress, Monmouth Junction, NJ, USA), and 0.3 μM trichostatin A (TSA) (MedChemExpress, Monmouth Junction, NJ, USA) in Neuro-basal Plus media (Invitrogen, Thermo fisher, Waltham, MA, USA). Additionally, the media was supplemented with 10 ng/ml of brain-derived neurotrophic factor (BDNF) (PeproTech, Cranbury, NJ, USA), 10 ng/ml of Glial Cell Line-Derived Neurotrophic Factor (GDNF) (PeproTech, Cranbury, NJ, USA), 10 ng/ml of Neurotrophin-3 (NT3) (PeproTech, Cranbury, NJ, USA) and 0.5×B27 (Invitrogen, Thermo fisher, Waltham, MA, USA). Two days of treatment during the first step was followed by the next maturation step which contained 12.5 μM forskolin, 5 μM CHIR99021, and 10 μM Y27632, and growth factors such as 10 ng/ml basic fibroblast growth factor (bFGF) (Pepro-Tech, Cranbury, NJ, USA), 10 ng/ml BDNF, 10 ng/ml GDNF, 10 ng/ml NT3, 0.5×B27 and 1×N₂ (Invitrogen, Thermo fisher, Waltham, MA, USA) in Neurobasal Plus media for four days.

The protocol 2 for the induction of neuronal program and generation of homogenous neuron-like MAP²⁺ cells involved a two-step epigenetic resetting process. The first step of neuronal induction (NI) was similar to the first step of protocol 1 with BDNF, GDNF and 0.5×B27 replaced by 2 μM retinoic acid (RA) (MedChemExpress, Monmouth Junction, NJ, USA) and 10 ng/ml bFGF. After two days, the cells were treated with extended epigenetic regulation (EER) media containing Neurobasal Plus supplemented with small molecules such as 0.3 μM JQ-1(+), 5 μM CHIR99021, and 10 μM Y27632, and growth factors which included 50 ng/ml Insulin Growth Factor 1 (IGF-1) (PeproTech, Cranbury, NJ, USA), 10 ng/ml bFGF, 20 ng/ml NT3, 100 μM N⁶, 2’-O-Dibutyryladenosine 3’, 5’-cyclic mo-nophosphate (dbcAMP) (Sigma-Aldrich, Inc., Saint Louis, MO, USA), and 1×N₂ for four days.

Protocol 3 was designed to generate MAP2⁺/NeuN⁺/vGLUT1⁺ neuron-like cells from MAP2⁺/NeuN⁻ cells generated from protocol 2. After the two-step strategic epigenetic regulation, the MAP2⁺/NeuN⁻ cells neuron like cells on day 6 were treated with neuronal maturation (NM) media comprising Neurobasal Plus media supplemented with small molecules such as 5 μM forskolin, 5 μM Y27632, 2 μM SP600125 (MedChemExpress, Monmouth Junction, NJ, USA), 0.5 μM dorsomorphin (MedChemEx-press, Monmouth Junction, NJ, USA), 0.5 μM LDN193189 (MedChemExpress, Monmouth Junction, NJ, USA), and growth factors such as 50 ng/ml IGF-1, 20 ng/ml BDNF, NT3 10 ng/ml, GDNF 20 ng/ml, 100 μM dbcAMP, 1× L-Glutamax (Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA), 200 mM Vitamin C (Sigma-Aldrich, Inc., Saint Louis, MO, USA), Laminin 1 μg/ml (Sigma-Aldrich, Inc., Saint Louis, MO, USA) and 1×N₂ supplement.

Conversion efficiency was calculated as previously described (15). Briefly, we randomly selected 4∼6 view fields for each sample on day two post neuronal induction using an Olympus IX-51 microscope and counted the total number of TUJ1⁺/DAPI and DCX⁺/DAPI. The conversion efficiency was calculated as the ratio of TUJ1⁺/DAPI cells and DCX⁺/DAPI to the initial number of seeded cells in each visual field. The neuronal purity represents the percentage of MAP2⁺ cells on day six in total cells stained by DAPI (15). Quantitative data are presented as the mean± SEM of at least three independent experiments.

Cells were harvested from three wells of 24-well plates at the indicated time points, and total RNA was extracted using RNeasy Mini Kit (QIAGEN, Hilden, Germany), for a total of 50∼100 ng/μl of pure RNA. The isolated RNA had an A260/A280 ratio between 1.8 to 2.1, which indicates RNA purity. The isolated RNA was stored at −80℃. For cDNA synthesis, a One-Script^Ⓡ cDNA Synthesis Kit (abm, Vancouver, Cananda) was used to convert 2 μg of RNA to cDNA in a total reaction volume of 20 μl.

RT-qPCR was performed using SYBR Green PCR Master Mix (Bio-Rad, Bio-Rad Laboratories, Inc., Hercules, CA, USA) on a Bio-Rad Prime PCR instrument. The qRT-PCR conditions were 40 cycles of 30 s at 95℃, 15 s at 60℃, and 15 s at 72℃. The primers used in these studies are listed in Supplementary Table S1.

For cell staining, cultures were washed twice with 1× phosphate-buffered saline (PBS) (Welgene, Republic of Korea) and fixed in 4% formaldehyde (Sigma-Aldrich, Inc., Saint Louis, MO, USA) in PBS for 15 min at room temperature. The cells were then washed thrice with 1× PBS and permeabilized with 0.1% Triton-X-100 (USB Corporation, OH, USA) in PBS for 10 min at room temperature. The cells were washed thrice with 1×PBS and blocked with a blocking solution containing 1% bovine serum albumin (BSA) (Amresco, France), 22.52 mg/ml glycine (Affymetrix, CA, USA), and 0.1% Tween 20 (Affy-metrix, CA, USA) in PBS for 60 mins. Subsequently, the cells were incubated with primary antibodies diluted in dilution solution containing 1% BSA (Amresco, France) and 0.1% Tween 20 (Affymetrix, CA, USA) overnight at 4℃. The cells were then washed thrice with 1× PBS containing 0.1% Tween-20 (PBST) and incubated with secondary antibodies for 2 h at room temperature. Cells were then washed thrice with PBST and incubated with 1 μg/ml DAPI (Sigma-Aldrich, MO, USA) for 5 min at room temperature to stain the nuclei. The samples were subsequen-tly visualized using a fluorescence microscope (IX71S1F3, Olympus, Japan). All antibodies used in this study are listed in Supplementary Table S2.

Calcium imaging was performed on MAP2⁺/NeuN⁺/vGLUT1⁺ neuron-like cells derived from human fibroblast using Rhoda-2-AM (Sigma-Aldrich, Inc., Saint Louis, MO, USA) at a concentration of 2 μM. Imaging was performed in Live-Imaging Solution (Invitrogen, Thermo Fisher Scie-ntific, Inc., Waltham, MA, USA). Images were acquired at 30 frames/s using a scientific CMOS camera. The microscope was controlled by Micro-Manager software and the image processor ImageJ. The samples were visualized using fluorescence microscopy (IX71S1F3, Olympus, Tokyo, Japan). Changes in fluorescence were measured for individual cells, and the average of the first 10 time-lapse images for each region of interest (ROI) was defined as the initial fluorescence (F₀).

Current protocols for generating neurons from fibroblasts exploit the reprogramming potential of epigenetic modifiers using inhibitors against HDAC or BET alone (7, 17-21). The use of dual HDAC/BET inhibitors instead of using one inhibitor for reprogramming human somatic cells into neurons has been demonstrated to induce epigenetic plasticity in human adult astrocytes, which resulted in improved neuronal conversion (15). Therefore, we hypothesized that human fibroblasts that resist neuronal reprogramming like human adult astrocytes do, could show improved neuronal efficiency upon application of dual HDAC/BET inhibitors.

For our study, we used human fibroblast (HF) cells, the identity of which was confirmed by in-house characte-rization. The staining results indicated that the HF cells did not contain any neuronal contaminants (Supplementary Fig. S1A and S1B). To prove our hypothesis regarding the rapid and efficient generation of neurons by applying the dual epigenetic regulation strategy, we developed a small-molecule chemical cocktail with dual HDAC/BET inhibitors, comprising HDAC inhibitor trichostatin A (TSA), and BET inhibitor JQ-1(+). In addition, we supplemented these epigenetic modifiers with other small molecules reported to promote neuronal reprogramming (7, 18), such as forskolin (cAMP activator), CHIR99021 (WNT activator), Y27632 (ROCK inhibitor), and RepSox (TGF-b inhibitor), constituting a cocktail referred to as 6C (Fig. 1A). HF cells treated with 6C cocktail along with growth factors BDNF, GDNF, NT3, and 0.5×B27 (Neuronal induction of protocol 1, Fig. 1A) generated a morphological change over a period of two days, showing a shrinkage of the soma and loss of fibroblast-like morphology (Day 2, Fig. 1B).

The ability of these six small molecules to induce neuronal reprogramming was confirmed by the presence of the early neuronal markers TUJ1 and DCX on day two with an efficiency of 88±4% and 71±3% respectively (Fig. 1C and 1D). Furthermore, real-time quantitative PCR (RT-qPCR) analysis showed upregulation of genes related to neuronal reprogramming and downregulation of fibroblast-associated markers (Fig. 1E and Supplementary Fig. S1C).Taken together, our data demonstrate that the usage of two epigenetic modifiers together in combination with small molecules to modulate signaling pathways could rapidly and efficiently induce neuronal programming.

The use of 6C with dual HDAC/BET inhibitors instigated rapid and efficient induction of TUJ1⁺/DCX⁺ cells after two days of treatment. To highlight the importance of this temporal dual epigenetic regulation used in the first step, we compared the arrival of DCX expression after two days of treatment with the complete cocktail to the cocktail with only one or none of the epigenetic modifiers. Only bFGF was supplemented to the media to ensure the cell survived under the exposure of aforementioned small molecules (22). The removal of any of the epigenetic regulators significantly affected the neuronal induction efficiency of the cocktail as evident by the number of DCX⁺ cells and the change in the morphology of cells generated after two days (Fig. 2A and 2B). TSA was found to be the key regulator as its removal resulted to a significant drop in neuronal reprogramming efficiency with only 8±4% cells showing DCX expression. The removal of JQ-1(+) from the 6C cocktail resulted to a reprogramming efficiency of 34±9% which is nearly half of DCX⁺ cells generated by 6C cocktail, 67±2% (Fig. 2C), when treated only in the presence of bFGF. Furthermore, the neuronal reprogramming efficiency in the absence of both the epigenetic modifiers was found to be 11±5% and the number of DCX⁺ cells in the media control was negligible. Taken together, our results highlight the importance of each of the epigenetic modifiers in the cocktail and demonstrated a preliminary level of evidence of the synergistic effect of dual HDAC/BET inhibition for the rapid and efficient induction of neuronal program.

Upon TUJ1 and DCX expression after two days of 6C treatment, we introduced another step of neuronal maturation for the maintenance and maturation of induced early neurons (Fig. 1A). In an attempt to generate matured neurons from TUJ1⁺/DCX⁺ induced early neurons, we adopted a reduction approach by removing the chemicals from the 6C cocktail. In addition to removing the epigenetic modifiers because of their ability to induce toxicity, we removed RepSox because of its known role in promoting Sox2 expression during the formation of a pluripotent stage (23). The resultant maturation cocktail comprising forskolin, CHIR99021, and Y27632 (FCY) promoted cell survival, with few cells showing typical neuron-like morphology (D6, Fig. 1B) with an abundance of non-reprogrammed, fibroblast-like cells. The neuronal purity after six days, defined by the percentage of MAP2⁺, was found to be approximately 19±5% along with the expression of pan-neuronal marker TUJ1 (Fig. 1F and 1G). However, the expression of another mature marker, NeuN, was negligible after six-day treatment (Fig. 1H). A detailed morphological analysis of the cells at day 6 revealed that most MAP2⁺ cells lacked proper neurite extension (Fig. 1G). Moreover, an extended incubation with FCY beyond day 6 did not improve the neuronal morphology (Supplementary Fig. S1D). These results suggested the requirement for a modification in the chemical cocktail to generate a homogenous population of mature neuron-like cells.

Since the FCY treatment in the second step of protocol 1 generated neurons with very low purity, we investigated numerous other chemical combinations for their ability to maintain neuronal maturation and purity beyond the induction state (Supplementary Fig. S2). Interestingly, one condition containing JQ-1(+) was found to be effective in promoting neurite extension and maintaining the purity. These results suggested the requirement of an extended epigenetic remodeling (ERR) for the sustenance of the neuronal program and preventing the generation of non-reprogrammed cells. Hypothetically, the epigenetic regulation in the first induction step of two days in the protocol 1 was sufficient to trigger gene expression responsible for neuronal induction, but those involved in the neuronal maturation were not optimally expressed and the epigenetic resetting favoring neuronal-lineage was not stable. We found that a cocktail containing JQ-1(+), CHIR99021 and Y27632 (JCY) resulted in the generation of neuron-like cells having proper soma with well-defined elongated neurites which is a typical neuronal morphology (24). In addition, the presence of non-reprogrammed fibroblast-like cells was minimal as compared to the other condition showing excellent morphological homogeneity (Supplementary Fig. S2); however, with a certain level of toxicity. We reduced this JCY-induced toxicity using appropriate growth factors which included dbCAMP, IGF-1, NT3, bFGF and N₂ during the two-step epigenetic regulation (Fig. 3A and 3B). Moreover, we further improved the induction protocol by replacing GDNF, BDNF and B27 used during the first two days with Retinoic acid (RA) and bFGF. The introduction of RA and bFGF during the first two days resulted in the brightening of the cell soma which resembles to the reported neuronal morphology (7); however the induction efficiency remained similar (TUJ1⁺ cells 91±2% and DCX⁺ 75±7%) in comparison to Protocol 1 (TUJ1⁺ cells 88±4% and DCX⁺ cells 71±3%) (Fig. 3B, 3D and 1D). Finally, we standardized a new two-step protocol comprising of neuronal induction for two days followed by the maintenance of neuronal program using an extended epigenetic regulation approach (Protocol 2, Fig. 3A).

The neuron-like cells generated by our new two-step protocol (Protocol 2) showed obvious morphological differences from those induced by protocol 1 (Fig. 3B Day 6 vs. Fig. 1B Day 6). By applying the new protocol, we were able to generate a large pool of cells with a high neuronal purity of 91±3% (MAP2⁺ cells) (Fig. 3E) co-expressing TUJ1 and MAP2 (Fig. 3F). However, the expression of the matured marker NeuN was still lacking (Fig. 3G). Taken together, our results demonstrated the need of an extended BET inhibition for the maintenance of the neuronal program and generation of a homogenous neutron-like population for their further maturation.

The continued exposure of the cocktail JYC beyond six days did not promote NeuN expression but resulted in cell toxicity and loss of neuronal morphology (Supplementary Fig. S3A and S3B). Moreover, the application of neuronal maturation cocktail from protocol 1 resulted to similar outcome as seen with JYC (Supplementary Fig. S3A and S3C). One of the possible explanations for the neurons not showing NeuN expression upon extended JYC treatment is the role of JQ-1(+) in suppressing nuclear factor kappa B (NF-κB) which is important for neuronal development and synaptic activation (25, 26). To circumvent this issue and generate fully matured neurons, we included an additional neuronal maturation step after EER with a chemical cocktail comprising forskolin, Y27632, LDN193189, dorsomorphin and SP600125 (FYLDSp), based on the small molecules used in previously reported protocols (7, 19), along with dbcAMP, BDNF, NT3, GDNF, IGF-1, 1× L-Glu-taMAX, Vitamin C, Laminin and 1×N2 supplement (27) (Protocol 3, Fig. 4A). The addition of the new maturation step enabled cell survival for more than 10 days, allowing them to acquire proper neuronal morphology (Fig. 4B and Supplementary Fig. S3A and S3D).

Moreover, the maturation cocktail from day six promoted NeuN expression which was observed by immunostaining on day 10 (Fig. 4C). Furthermore, along with the presence of punctuated synaptic vesicles evident by Syn1 staining of neurites (Fig. 4D), we also found that nearly 64% of the cells surviving on day 10 were vGLUT1-positive (Fig. 4E).

Our maturation step using the cocktail FYLDSp generated TUJ1⁺/MAP2⁺/NeuN⁺/vGLUT1⁺ neuron-like cells, indicating that the challenges posed by JQ-1(+) during neuronal maturation can be addressed. To this end, we investigated whether these reprogrammed cells attained neuronal functionality by analyzing the calcium influx/outflow, which is one of the functional properties of neurons (28, 29). Incubating the cells with Rhod 2-AM, a high affinity Ca²⁺ indicator, showed spontaneous calcium influx followed by a gradual outflow of calcium over period of 50 sec. A detailed analysis of neuron-like cells’ signal pattern revealed that intracellular calcium influx and outflow are heterogeneous. We observed that 7.8% of the neuron-like cells showed sudden rise of calcium influx followed by gra-dual outflow after 20 seconds, which were comparable to previously described neurons showing action potential (Fig. 4F and ROI 1, ROI 2 and ROI 3 respectively in Fig. 4G) (28). However, a large number of the neuron-like cells mai-ntained a passive influx and outflow of calcium signaling, indicating the developmental phase of the reprogrammed cells which are yet to attain full functionality (30) (ROI 4, ROI 5 and ROI 6 respectively in Fig. 4G).

To sum up, the immunostaining results showing the presence of mature neuronal markers along with the spontaneous influx and outflow of calcium demonstrate the efficacy of the maturation cocktail to generate mature neuron-like cells.