The brain’s serotonergic system centrally regulates several physiological processes and its dysfunction has been implicated in the pathophysiology of several neuropsychiatric disorders. While in the past our understanding of serotonergic neurotransmission has come mainly from mouse models, the development of pluripotent stem cell and induced fibroblast-to-neuron (iN) transdifferentiation technologies has revolutionized our ability to generate human neurons in vitro. Utilizing these techniques and a novel lentiviral reporter for serotonergic neurons, we identified and overexpressed key transcription factors to successfully generate human serotonergic neurons. We found that overexpressing the transcription factors NKX2.2, FEV, GATA2 and LMX1B in combination with ASCL1 and NGN2 directly and efficiently generated serotonergic neurons from human fibroblasts. Induced serotonergic neurons (iSNs) showed increased expression of specific serotonergic genes that are known to be expressed in raphe nuclei. iSNs displayed spontaneous action potentials, released serotonin in vitro and functionally responded to selective serotonin reuptake inhibitors (SSRIs). Here, we demonstrate the efficient generation of functional human serotonergic neurons from human fibroblasts as a novel tool for studying human serotonergic neurotransmission in health and disease.
Serotonin, found in the central nervous system as well as peripherally in the blood and gut, is an important neurotransmitter with a diverse set of physiological functions. In the central nervous system, serotonin is produced by a relatively small number of serotonergic neurons located in the raphe nuclei. Serotonergic neurons extensively innervate nearly all brain regions, regulating cognition as well as emotion.1, 2 Serotonergic neurotransmission is modulated by factors such as certain genetic polymorphisms, environmental stress, nutrition and certain drugs of abuse.3, 4, 5 Serotonergic dysfunction has been implicated in neurological disorders such as major depression, generalized anxiety, schizophrenia and autism, suggesting a central role for serotonergic neurotransmission in healthy brain functioning.6, 7, 8 Thus far, most of our understanding of the cellular and molecular underpinnings of serotonergic neurotransmission comes from animal models that only partially resemble human systems. Advances in human pluripotent stem cell9, 10 and transdifferentiation11, 12 technologies offer new approaches for generating and examining human neurons in vitro.13 The neuronal population in the brain is heterogeneous, with different neuronal subtypes utilizing different neurotransmitters such as glutamate, GABA, dopamine, serotonin, acetylcholine and norepinephrine. Currently, a major challenge in the field is generating and enriching for these diverse human neuronal subtypes in vitro.
Recent research work has enabled the generation of specific neuronal subtypes such as GABAergic14, 15, 16 and dopaminergic neurons.17, 18 These methodologies are based on utilizing growth factors (for example, Sonic hedgehog (Shh) for GABAergic neurons) or transcription factors (for example, Ascl1, Lmx1a and Nurr1 for dopaminergic neurons) that are important during the development and specification of these neuronal subtypes in vivo. In the case of raphe serotonergic neurons, the patterning factors FGF-8, FGF-4 and Shh, which are released from the midbrain–hindbrain organizer, the primitive streak and the notochord, respectively, are crucial for serotonergic fate specification.4, 19 Mouse knockout studies have revealed that multiple transcription factors downstream of these patterning molecules are required for the development of serotonergic neurons in rostral or caudal raphe clusters.20 Among others, the transcription factors Ascl1, Nkx2.2, Pet-1 (FEV in humans), Lmx1b, Gata2 and Gata3 have been shown to be crucial for the specification and maturation of serotonergic neurons in the rodent midbrain dorsal raphe nuclei.21, 22, 23, 24 In the adult, serotonergic neurons have dense ascending projections to structures such as the cortex, hippocampus and amygdala.20 In contrast, the caudal raphe nuclei have descending fibers projecting to the brainstem and the spinal cord.20 So far, studies have highlighted the role of FGF-8, Shh and FGF-4 as well as certain transcription factors in enhancing serotonergic fate differentiation from mouse embryonic stem cells (ESCs).25, 26, 27, 28, 29 However, methods for efficiently generating human serotonergic neurons in vitro have been more difficult to develop.30, 31, 32
In our study, we sought to examine which transcription factors were involved in serotonergic specification of human neurons and whether they could be utilized for directed serotonergic differentiation from human fibroblasts in vitro. Overexpression of key transcription factors under defined conditions enables the transdifferentiation of human fibroblasts to neurons without reversion to the pluripotent stem cell stage.11, 33, 34 Thus far, studies have shown successful transdifferentiation of fibroblasts to oligodendrocyte progenitor cells,35 dopaminergic neurons,17 motor neurons,36 and cholinergic neurons.37 In our study, using a novel lentiviral reporter for serotonergic neurons, we were able to enrich serotonergic neurons from pluripotent human stem cells in vitro. We observed an increased expression of the serotonergic transcription factors (S4F) NKX2.2, FEV, GATA2 and LMX1B. Overexpression of these S4F along with the neuronal transcription factors (ASCL1, NGN2) in human fibroblasts induced serotonergic transdifferentiation with high efficiency. Induced serotonergic neurons (iSNs) expressed key markers of serotonergic neurons, fired action potentials and released serotonin in the culture medium, which could be modulated by treatment with selective serotonin reuptake inhibitors (SSRIs). Together, our results provide evidence for a novel strategy for efficiently generating and studying human serotonergic neurons in vitro.
Materials and methods
Culture of neural precursor cells and neurons from human ESCs
Human ESC (hESC) (line H9) colonies were maintained on Matrigel (BD, San Jose, CA, USA) coated plates in TeSR medium (Stemcell Technologies, Vancouver, Canada), passaged using Dispase (Millipore, Billerica, MA, USA) and neuralized using a modified version of our previously described midbrain differentiation protocol (schematic in Supplementary Figure 1a).38 Briefly, H9 colonies were detached with Collagenase (1 mg ml−1 in DMEM) and transferred to ultra-low attachment plates (Corning, NY, USA) in neural induction medium (DMEM-F12/Glutamax, 1 × N2 supplement (Invitrogen, Thermo Fisher Scientific, Waltham, MA USA), 1 × B27 supplement (Invitrogen, Thermo Fisher Scientific), 100 ng ml−1 Noggin, 100 nM LDN and 10 μM SB431542. After 7–10 days, the free-floating embryoid bodies were transferred to poly-ornithine (PORN)- and laminin-coated plates in neural precursor cell (NPC) medium containing either FGF-2 (20 ng ml−1) for ‘pan-neuronal' cells or FGF-8 (100 ng ml−1) plus Shh (200 ng ml−1) for midbrain specification, and Laminin (1 μg ml−1) in DMEM/F12-Glutamax+N2+B27 supplements. Rosette-forming embryoid bodies were manually picked and dissociated using Accutase (Millipore) and plated on PORN–laminin-coated plates to generate NPCs. NPCs were maintained at high density as monolayers on PORN–laminin-coated plates in NPC medium and were plated at lower densities for neuronal differentiation (for 3 and 8 weeks); the medium was then switched to neural differentiation medium (NDM) containing brain-derived neurotrophic factor (20 ng ml−1, Peprotech, Rocky Hill, NJ, USA), glial cell–derived neurotrophic factor (20 ng ml−1, Peprotech), dibutyrl-cyclicAMP (1 mM, Sigma-Aldrich, St. Louis, MO, USA), ascorbic acid (200 nM, Sigma-Aldrich) and laminin (1 μg ml−1) in DMEM/F12-Glutamax+N2+B27. The medium was changed every third day for 3 and 8 weeks, as per the timeline of the experiment. All human cell line experiments were performed in accordance with protocols approved by the Salk Institute for Biological Studies.
Culture and induction of human fibroblasts to neurons
Primary human dermal fibroblasts were established from skin biopsies from healthy donors and cultured in DMEM containing 15% fetal bovine serum (FBS) and 0.1% non-essential amino acids (all from Gibco, Thermo Fisher Scientific, Waltham, MA, USA). Fibroblast cell lines were obtained from the Coriell Institute, ATCC GM22159 and University Hospital in Erlangen, Germany (details listed in Supplementary Table 1). Lentiviral particles for pLVX-EtO,39 ASCL1/NGN2 (AN) and the iSN factors were incubated with fibroblasts for 20–24 h. Forty-eight hours following transduction, transgenic induced neuron (iN)- and iSN-competent fibroblasts were further passaged in selective media in the presence of G418 (200 μg ml−1; Gibco, Thermo Fisher Scientific) and puromycin (1 μg ml−1; Sigma-Aldrich) in tetracycline-free FBS-containing medium. Direct neuronal transdifferentiation was performed using a modified AN-based protocol that has been previously described.39 Fibroblasts were seeded on tissue culture grade plastic petri dishes/slides/plates (Ibidi, Germany) and, after 24 h, the medium was changed to induced neuron conversion medium (NCM) for 3 weeks. NCM has a 50:50 base of DMEM:F12/Neurobasal containing the following supplements: N2 supplement, B27 supplement (both 1 × ; Gibco, Thermo Fisher Scientific), doxycycline (2 μg ml−1, Sigma-Aldrich), Laminin (1 μg ml−1, Life Technologies, Carlsbad, CA, USA), dibutyryl cyclic-AMP (500 μg ml−1, Sigma-Aldrich), human recombinant Noggin (150 ng ml−1; Peprotech), LDN-193189 (0.5 μM; Cayman Chemical Co, Ann Arbor, MI, USA), A83-1 (0.5 μM; Stemgent, Lexington, MA, USA), CHIR99021 (3 μM, LC Laboratories, Woburn, MA, USA), Forskolin (5 μM, LC Laboratories) and SB-431542 (10 μM; Cayman Chemical Co). The medium was changed every third day. For further maturation after 3 weeks, induced neurons (iNs) were switched to neural maturation medium (NMM), also a 50:50 base of DMEM:F12/Neurobasal containing glial cell–derived neurotrophic factor, brain-derived neurotrophic factor (both 20 ng ml−1, R&D Systems, Minneapolis, MN, USA), dibutyryl cyclic-AMP (500 μg ml−1, Sigma-Aldrich), doxycycline (2 μg ml−1, Sigma-Aldrich) and laminin (1 μg ml−1, Life Technologies). For electrophysiology, iNs were dislodged using TrypLE, replated on a feeder layer of mouse astrocytes and cultured in NMM for an additional 2–3 weeks (Figure 2a).39 The timeline of the methodology is shown in the schematic in Figure 2a. Supplementary Tables 1 and 2 provide details of the fibroblast and hESC lines used, viral transduction scheme, differentiation protocol used, time points analyzed, maximum passage numbers and total number of transdifferentiation events done per line.
Cloning and generation of lentiviral particles
A lentiviral transcriptional reporter vector encoding EGFP was designed to label TPH2+ neurons in culture. The human Tryptophan Hydroxylase 2 (TPH2) RefSeq gene is located on chromosome 12, where it spans approximately 93.5 kb of DNA with a 12-kb intergenic region separating it from the upstream gene TBC1D15 (Supplementary Figure 1c). A 2-kb DNA proximal region was used as a promoter.40 The sequence −1992 to −1 relative to the translational start codon, including a 5’UTR, was amplified by PCR from male genomic DNA (Promega, Madison, WI, USA) using Phusion High-Fidelity DNA Polymerase (New England Biolabs, Ipswich, MA, USA) and a forward primer (5′-IndexTermgtcaATCGATgcttaaggttcagctttccca-3′) adding a ClaI site, and reverse primer (5′IndexTerm-gcatGGATCCcggtgtaatattctttctctg-3′) adding a BamHI site. The PCR product was subjected to restriction digestion and introduced as a promoter into lentiviral vector pCSC-SP-PW-EGFP containing an optimal Kozak sequence. The construct (pCSC-TPH2-EGFP) was verified by DNA sequencing. The human Synapsin::DsRed and Synapsin::GFP reporter constructs have been previously tested and described.41 Neurons were transduced with the TPH2::GFP and/or Synapsin::DsRed reporter lentiviral particles approximately 4–6 days before processing for further analysis.
Coding sequences for the human iSN factors NKX2.2, FEV, GATA2 and LMX1B were amplified by PCR and cloned into the pLVX-Tight-Puro construct (Clontech, Mountain View, CA, USA). The pro-neuronal factors ASCL1 and NGN2 were linked by a 2A peptide sequence before insertion into the pLVX-Tight-Puro vector (Supplementary Figure 2a). Lentiviral particles were prepared as previously described.39 Briefly, lentivector plasmid and lentiviral packaging plasmids were introduced to 293 T human embryonic kidney cells using polyethyleneamine and viral particles were concentrated by ultracentrifugation, aliquoted, titered and frozen for future use.
Immunocytochemistry, immunohistochemistry and western blotting
Cells were washed with Tris-buffered saline and fixed with 4% paraformaldehyde for 10 min, at room temperature (RT), blocked with 3% donkey or horse serum in Tris-buffered saline, incubated with the primary antibodies overnight at 4°C, washed, incubated with secondary antibodies (45 min, RT), counterstained with DAPI (4',6-diamidino-2-phenylindole, Sigma-Aldrich) and mounted with PVA-DABCO mounting solution (Sigma-Aldrich). The following antibodies were used: GFP (chicken, 1:500, Aves Labs, Tigard, OR, USA), beta-III-tubulin (rabbit, 1:3000, Millipore), hTau (PHF1, mouse, 1:500, kind gift from Peter Davies), Vimentin (goat, 1:500, Millipore), NeuN (mouse, 1:100, Millipore), MAP2ab (chicken, 1:1000, Abcam, Cambridge, UK), Nkx2.2 (mouse, 1:100, DSHB, Iowa City, Iowa), Fev (goat, 1:100, Santa Cruz, Dallas, TX, USA), Gata2 (rabbit, 1:500, Abcam), Lmx1b (mouse, 1:50, DSHB), TPH (sheep, 1:500, Millipore), 5-HT (rabbit, 1:1000, Protos BioTech, New York, NY, USA), SERT (rabbit, 1:800, Abcam), GABA (rabbit, 1:1000, Millipore), TH (mouse, 1:500, Millipore), GFAP (chicken, 1:1000, Millipore) and TPH2 (rabbit, 1:500, Abcam).
Immunohistochemistry of mouse brain sections was performed as previously described.42 Briefly, floating sections were incubated with TPH primary antibody in blocking solution over two nights at 4 °C on a shaker, washed in Tris-buffered saline, incubated with secondary antibodies for 2 h at RT, counterstained with DAPI, washed, mounted on slides, coverslipped using PVA-DABCO (Sigma-Aldrich), and dried overnight protected from light.
For western blotting, cells were scraped from the dishes in RIPA buffer (Sigma-Aldrich) containing protease inhibitors (Complete ULTRA™, Roche, Switzerland) and incubated for 15 min on ice. Lysates were clarified by centrifugation (16 000 g; 15 min; 4 °C) and western immunoblotting was performed following standard procedures using the iBlot gel and transfer system (Life Technologies). Primary antibodies for FEV, Nkx2.2 and GAPDH were incubated (overnight at 4 °C) and horseradish peroxidase–coupled secondary antibodies (anti-rabbit and anti-mouse) were added (1 h, RT) and detected sequentially using ECL luminescence solution (Millipore) as per the manufacturer’s instructions.
Whole transcriptome RNA sequencing and analysis
RNA was prepared into RNA-Seq libraries using the TruSeq Stranded mRNA Sample Prep Kit according to the manufacturer’s instructions (Illumina, San Diego, CA, USA) and was reverse transcribed into cDNA using SuperScript II reverse transcriptase (Invitrogen, Thermo Fisher Scientific). Stranded cDNA sequencing libraries were generated according to Illumina’s procedures. Total (mRNA) RNA-Seq libraries were sequenced single-end 50 bp using the Illumina HiSeq 2500 (San Diego, CA, USA) platform according to the manufacturer’s specifications. Low quality ends and adapter removal/trimming were performed using cutadapt (http://journal.embnet.org/index.php/embnetjournal/article/view/200). The trimmed reads were mapped using STAR43 and normalized counts and differential expression were calculated using the DESeq R package.44 Normalized count numbers were processed using R software to generate MA plots and scatter plots for each group. DAVID (http://david.ncifcrf.gov/) was used to perform gene ontology (GO) annotation analysis. Gene lists for human midbrain raphe nucleus-enriched genes were obtained from the Allen Brain Atlas (http://human.brain-map.org). Microarray expression data in the human MBRa (midbrain raphe nuclei) was compared with human cortex and significantly MBRa-enriched probes were chosen for comparison between our fluorescent activated cell sorting-purified iSNs and iNs (greater than two fold). The RNA-Seq data have been deposited at the European Bioinformatics Institute: www.ebi.ac.uk/arrayexpress and can be accessed using the accession code: E-MTAB-3680. Statistical analyses show P-values that are not false discovery rate corrected. GO analysis of iSN-enriched genes was performed using a cutoff of more than twofold significantly upregulated genes in iSNs vs iNs. Figures 2a and c show gene reads of a one-to-one sample-specific comparison of fibroblasts vs iNs and iNs vs iSNs derived from one individual (iN/iSN#1). Figures 2d and f are a comparison of iSNs and iNs (independent duplicate runs per sample), with neurons derived from fibroblasts from two people (iN/iSN#1 and iN/iSN#2).
Whole-cell patch recordings were performed on 6-week-old iSNs and iNs (n=6–10 neurons in total) co-cultured with wild-type mouse astrocytes on acid-etched coverslips and typically transduced with a lentiviral Synapsin::GFP reporter to label all neurons. Two to three weeks before patch-clamping experiments, the cells were transferred in a neurophysiological custom-made basal medium (BrainPhys basal) with supplements.45 Two neurons were found to be spontaneously active (traces shown in Figure 4h). Individual coverslips were transferred into a heated recording chamber and continuously perfused (1 ml per minute) with artificial cerebrospinal fluid bubbled with a mixture of CO2 (5%) and O2 (95%) and maintained at 25 °C. The composition of artificial cerebrospinal fluid was adjusted to match the inorganic salt concentration and osmolarity of BrainPhys. Artificial cerebrospinal fluid contained (in mM) 121 NaCl, 4.2 KCl, 1.1 CaCl2, 1 MgSO4, 29 NaHCO3, 0.45 NaH2PO4-H2O, 0.5 Na2HPO4 and 20 glucose (all chemicals from Sigma-Aldrich).
For targeted whole-cell recordings, we used a 40 × water-immersion objective, differential interference contrast filters (all from Olympus, Waltham, MA, USA), an infrared digital camera (Rolera XR-Qimaging, Surrey, BC, Canada), a digidata 1440A/Multiclamp 700B and Clampex 10.3 (Molecular Devices, Sunnyvale, CA, USA). Patch electrodes were filled with internal solutions containing 130 mM K-gluconate, 6 mM KCl, 4 mM NaCl, 10 mM Na-HEPES, 0.2 mM K-EGTA; 0.3 mM GTP, 2 mM Mg-ATP, 0.2 mM cAMP, 10 mM d-glucose, 0.15% biocytin and 0.06% rhodamine. The pH and osmolarity of the internal solution were close to physiological conditions (pH 7.3, 290–300 mOsmol). Data was all corrected for liquid junction potentials (10 mV). The resistance of the patch electrodes was around 5 MOhm. Electrode capacitances were compensated online in cell-attached mode (~7 pF). Recordings were low-pass filtered at 2 kHz, digitized and sampled at intervals of 50 ms (20 kHz). To control the quality and the stability of the recordings throughout the experiments, access resistance, capacitance and membrane resistance were continuously monitored online and recorded. The access resistance of the cells in our sample was ~48 MOhm.
Time-lapse calcium imaging
Six-week-old iN and iSN cultures maintained on primary rat astrocytes were loaded with the calcium-sensitive dye Fluo-4. Fluo-4 was prepared according to the manufacturer’s instructions from the Fluo-4 calcium imaging kit (Life Technologies). Briefly, iN or iSN cultures were incubated with Fluo-4 (1 μM in NMM) for 20 min at 37 °C in an incubator. Neurons were washed with NMM once and incubated for 20 min before imaging. Calcium imaging was performed using the Zeiss spinning disk confocal microscope (Carl Zeiss, Germany). Images were obtained using a standard Alexa Fluor 488 filter and were taken every 131 ms, using the 20x objective for a total time of 120 s. Areas with neurons/serotonergic neurons were selected based on lentiviral reporter expression (Synapsin::DsRed/TPH2::TdTomato) in the red channel. Neurons were outlined as regions of interest and analyzed using ImageJ (total number of neurons traced: iN=18; iSN=20; from two separate individuals). For each region of interest, change in fluorescence intensity over time (Delta F) was plotted. Values were normalized to the minimum fluorescence (F min) for each individual region of interest and representative traces have been presented as Delta F/Fmin over time (Figure 4k).
Drug treatment and enzyme-linked immunosorbent assay for 5-HT
Six-week-old iSN cultures were treated with the SSRI Citalopram (5 μM, Sigma-Aldrich), tryptophan (10 μM, Sigma-Aldrich) or vehicle (phosphate-buffered saline) daily for 5 days. All drugs were diluted to their final concentration in NMM, in which cells had been maintained after the 3-week time point. NMM is a 50:50 base of DMEM-F12:Neurobasal media, containing glial cell–derived neurotrophic factor, brain-derived neurotrophic factor (both 20 ng ml−1, R&D), dibutyryl cyclic-AMP (500 μg ml−1, Sigma-Aldrich), doxycycline (2 μg ml−1, Sigma-Aldrich) and laminin (1 μg ml−1, Life Technologies). Five-day-old culture medium was collected before (5- day-old medium) and after SSRI treatment (5 days after treatment) and frozen at −20 °C for enzyme-linked immunosorbent assay (ELISA) together with all other samples. Three experiments for each iSN line (two different donor lines) were done and medium samples were collected for 5-HT ELISA. 5-HT concentrations in culture media of iNs and iSNs were quantified using a competitive ELISA kit (Enzo Life Sciences, Farmingdale, NY, USA) according to the manufacturer’s instructions. Briefly, pure 5-HT standards were serially diluted and the percentage of displaced 5-HT was measured. The percentage of still bound 5-HT was inversely proportional to the concentration of 5-HT in our samples. A standard curve was generated and an equation (Supplementary Figure 5) for the trend line was used to determine the concentration of 5-HT in our samples (from three runs per condition per neuronal type, derived from two individual lines—iSN#1 and iSN#2, described in Supplementary Table 1).
Quantitative data were generated in independent triplicates, from at least two donor fibroblast lines, and in independent triplicates from one hESC line. Averaged data, normalized means, variance, s.d. and s.e.m. were computed using Microsoft Excel and Prism software (GraphPad, La Jolla, CA, USA) from independent repetitions of the experiments. Bar graphs represent means with error bars showing s.e.m. Student’s t-test was performed to determine whether a significant difference existed between groups (*P<0.05).
Generation and isolation of hESC-derived serotonergic neurons in vitro
To examine human serotonergic neurons, we started by neuralizing hESCs using a modified midbrain differentiation protocol. NPCs, pre-patterned with FGF-8 and Shh, were differentiated to neurons for 3–4 weeks in medium containing brain-derived neurotrophic factor, glial cell–derived neurotrophic factor and cAMP as previously described.38, 46 To assess serotonergic differentiation in these culture conditions, we examined the percentage of neurons that were positive for TPH. TPH is the rate-limiting enzyme in the serotonin biosynthetic pathway whose expression is restricted to serotonergic neurons, as observed by immunohistochemistry of the mouse raphe (Figure 1a). TPH is encoded by two genes, TPH1 and TPH2, of which TPH2 is specifically expressed the central nervous system.24 Using the midbrain differentiation protocol (FGF-8+Shh), we found that approximately 8–10% of all MAP2ab+ neurons were TPH+ (Figures 1b and c), as compared with around 2% in the commonly used FGF-2 or ‘pan-neuronal’ conditions at 3 weeks (Figures 1b and c). To ascertain the serotonergic identity of a TPH+ cell, we co-stained neurons for serotonin (5-HT), GABA, tyrosine hydroxylase (TH) and TPH2. We found that all 5-HT+ neurons were TPH+ (Figure 1d) and that 5-HT+ neurons formed a subset of TPH+ neurons (around 30%), suggesting that TPH might be an early marker (3 weeks) for serotonergic neurons in our culture system (Figure 1e). We extended this analysis to a later time point of 8 weeks and found similar percentages of TPH+ and 5-HT+ neurons as observed at 3 weeks (Figures 1d and f). As a negative control, we detected no overlap between TPH-, GABA- and TH-expressing neurons (Supplementary Figure 1b). As a positive control, we observed that all TPH-expressing neurons were specifically positive for TPH2, as confirmed by immunostaining with a TPH2-specific antibody (Supplementary Figure 1b). Together, these results suggest that FGF-8 and Shh conferred a midbrain identity to human NPCs and generated a subpopulation of TPH-expressing serotonergic neurons upon differentiation.
Next we asked which transcription factors might determine the fate of the small percentage of human serotonergic neurons we observed in vitro. Despite an increase in serotonergic neurons in our culture conditions, serotonergic neurons constituted less than 10% of all MAP2ab+ neurons, making it necessary to isolate the serotonergic population for further studies. To this end, we isolated TPH-expressing cells using a novel lentiviral reporter. We designed a lentiviral construct that expressed EGFP or TdTomato under a short 5’ proximal promoter of the TPH2 gene (TPH2::GFP or TPH2::TdT; Supplementary Figure 1c). Three-week-old neurons were transduced with the TPH2::GFP lentivirus and quantification revealed that approximately 65% of GFP+ cells were indeed TPH+ and exhibited neuronal morphologies (Figures 1g and h); also, a majority of TPH+ neurons were GFP+, indicating that they were labeled by the reporter lentivirus (Figure 1h). These findings show that a majority (65%) of neurons isolated using the TPH2 reporter lentivirus were serotonergic, suggesting that this tool could be used for enriching for serotonergic neurons. To purify TPH-expressing neurons, we combined the TPH2::GFP reporter lentivirus with a Synapsin::DsRed reporter lentivirus that specifically labels neurons.41 Using fluorescent activated cell sorting, at the 3-week time point we isolated TPH2-expressing serotonergic neurons (GFP+/DsRed+) and non-serotonergic neurons (GFP−/DsRed+) for RNA collection and QPCR analysis (Supplementary Figures 1d and e). This time point was chosen based on the fact that there was a stable percentage of TPH+ and 5-HT+ neurons at this time, as we did not observe a significantly higher percentage of TPH+ neurons even with 8 weeks of differentiation (Figure 1f). It also becomes increasingly challenging to dissociate neurons as they mature past 3 weeks, likely due to the physical characteristic of a complex neuronal network between cells and the somewhat sticky nature of astrocytes. GFP+/DsRed+ cells indeed showed around 80-fold higher mRNA levels of TPH2 (Figure 1i), confirming the suitability of this reporter-based system for the isolation of TPH-expressing cells. We examined transcript levels of four transcription factors known to be involved in mouse serotonergic fate specification: NKX2.2, FEV, GATA2 and LMX1B. We found a substantial upregulation (between 1.5- and 1000-fold) of each of the S4F but not of the GABAergic transcription factor Lhx6. All of these transcription factors are thought to be important for serotonergic neuron specification during embryonic development. Hence, our results suggested that, using this method, we had likely isolated immature serotonergic neurons still undergoing differentiation and maturation (Figure 1i). Interestingly, all S4F were upregulated to varying degrees, hinting that they may all be instrumental in serotonergic fate specification in vitro.
Overexpression of S4F in human fibroblasts
Based on these findings, we next hypothesized that, using a transdifferentiation strategy for inducing neurons,47 we could use the S4F found in our screen—NKX2.2, FEV, GATA2 and LMX1B—for direct conversion of human fibroblasts to serotonergic neurons (Figure 2a, Supplementary Figure 2a). Recent studies have shown that forced expression of lineage-specific transcription factors can induce cell-fate changes in different human somatic cell types, including fibroblasts.48, 49, 50 Using this strategy, we next sought to examine whether overexpression of S4F, in combination with the pro-neuronal transcription factors ASCL1 and NGN2 (AN), could induce serotonergic transdifferentiation (Figure 2a). We first transduced human skin fibroblasts with the tetOn protein rtTA and a tetracycline-inducible and 2A sequence-linked NGN2-ASCL1 lentivirus to generate doxycycline-inducible AN fibroblasts (Supplementary Figure 2a). For direct conversion into iNs, AN-transduced fibroblasts were transferred to NCM containing doxycycline and small molecular enhancers (Figures 2a and b).37, 39 Within 3 weeks of the start of neuronal conversion, around 60% of the fibroblasts adopted a neuronal fate (Figures 2b, c and e), as seen by the expression of neuronal markers such as βIII-tubulin and NeuN in high percentages (βIII-tubulin—58.4±4.2%) (Figure 2e; Supplementary Figures 2d and e). The majority of transdifferentiated neurons were glutamatergic (Supplementary Figure 2f), a smaller percentage was GABAergic (Supplementary Figure 2g) and 5-HT+ cells were only sporadically observed (Supplementary Figure 2g) (<1% of βIII-tubulin). We checked whether neuronal transdifferentiation efficiencies changed with increasing passaging of the fibroblasts and found no significant differences between passages 6 and 18 (Supplementary Figure 2i). We next generated separate inducible lentivectors for each of the S4F and transduced AN fibroblasts to also overexpress NKX2.2, FEV, GATA2 and LMX1B. We then proceeded with neuronal conversion by maintaining cells in doxycycline-containing NCM (Figure 2a). Treatment with doxycycline-containing medium increased expression of the S4F within 48 h, but no leaky expression was observed in the absence of doxycycline (Supplementary Figures 2b and c). Three weeks after neuronal induction, we observed a substantial increase in the percentage of Map2ab+ neurons (61±15%) that were TPH+ in the S4F group, indicating their serotoninergic identity as compared with control iNs (8±4%) (Figures 2d and f). Similar to what was done for hESC-derived serotonergic neurons, we confirmed that converted neurons specifically expressed TPH2 using the TPH2-specific antibody (Supplementary Figure 2h). To further examine the robustness of this protocol, we overexpressed S4F in fibroblasts derived from a third individual (Supplementary Table 1) and examined the percentage of TPH+ neurons over multiple viral transduction events. We found that S4F-transduced neurons had a high percentage of TPH+ neurons compared with AN from the same line (Supplementary Table 3). Interestingly, the variability in TPH expression correlated with viral transduction events rather than with the transdifferentiation events. These data showed that the transduction of S4F was sufficient for robustly inducing TPH expression in transdifferentiated human neurons.
Whole transcriptome analysis of iSNs from human fibroblasts
To further examine the identity of TPH-expressing neurons, we performed whole transcriptome RNA sequencing of transdifferentiated iSNs, purified using the TPH2::GFP and Synapsin::DsRed lentiviral reporters. Control AN-expressing iNs were sorted using only the Synapsin::DsRed reporter. We specifically examined regulation of genes known to be upregulated or downregulated in human serotonergic neurons (raphe nuclei expression, Allen brain). Transcriptome analysis confirmed that iSNs showed marked enrichment for essential neuronal genes such as SYN1, SYP, DCX and TAU, and Na+/K+ channel-associated genes and a downregulation of fibroblast-specific genes such as COL8A1, CCNA2, DMKN, KRT19 and KRT34 (Figure 3a). Further, we observed high expression of TPH2 along with the six overexpressed transcription factors (Supplementary Figure 3a), confirming both the accuracy of the TPH reporter and the immunocytochemistry results. We examined serotonergic gene expression, in iSN vs iN from the same fibroblast line, and observed higher expression of genes known to be upregulated in human midbrain raphe nuclei, obtained from Allen Brain Institute, including GATA3, TAC1, GCH1, GLRA2 and NKX6-1 (Figure 3b). Concomitantly, we observed low expression of genes associated with other neuronal subtypes, such as NPY, PVALB, NEUROD1, NEUROD6 and PCP4, that are known to be expressed in GABAergic, cortical and Purkinje neurons (Figures 3b and d). To confirm that observed upregulated serotonergic genes were not due to generalized midbrain specification-related changes, we also examined regulation of other monoaminergic genes, for example those involved in dopaminergic specification. We found only low levels of expression of dopaminergic genes such as PITX3, TH, LMX1A, ALDH1A1, BNC2 and GFRA1 in iSNs (Figure 3c). These data showed that iSNs upregulated the expression of several genes associated with serotonin biosynthesis.
For a more unbiased approach for determining the identity of iSNs, we next performed statistical and GO analyses on expression data from multiple conversion experiments of iSNs vs iNs derived from fibroblasts from two donors. We first examined the overlap between statistically significantly upregulated genes in human raphe vs human cortex (436) and iSNs vs iNs (312) and found 32 overlapping genes (Figure 3d), some of which are known to be involved in serotonergic specification or biosynthesis. Other genes such as RGS9, SCN9A, FXYD5, CRYBA2 and EDNRB have been shown to be specifically expressed in raphe nuclei of mice but do not have clear ties with serotonin biosynthesis.51, 52 Of all significantly upregulated genes in iSNs vs iNs, we have listed fold change and P-values of select genes known to be expressed in serotonergic neurons of mice51, 52 and humans (Allen Brain Atlas) (Figure 3e). Noteworthy genes include TPH1, TPH2, the synthesis enzymes for serotonin; GCHFR, a feedback regulator of GCH1 that generates the TPH-cofactor tetrahydrobiopterin; RGS-9, the regulator of G-protein signaling; TAC3, the tachykinin peptide also known as Neurokinin B; and GABAB receptors 1 and 2. As expected, we also observed significantly higher levels of the overexpressed S4F. In an effort to examine the identity of iSNs in an unbiased manner, we next performed GO analysis for genes significantly upregulated more than twofold in iSNs vs iNs. To detect a meaningful set of GO categories for enriched iSN genes, we filtered for categories based on non-false discovery rate corrected significance cutoff of P<0.01 (Figure 3f). Using this method we found significant GO categories such as ‘regulation of neurotransmitter levels,’ ‘regulation of vesicle-mediated transport,’ ‘serotonin biosynthetic process’ and ‘transmission of nerve impulse’ (Figure 3f). Our target-directed as well as unbiased approaches for analyzing our transcriptomic data provided evidence for a bona fide serotonergic identity of iSNs.
Functional characterization of human iSNs
Interestingly however, despite our results suggesting serotonergic specification, at the 3-week time point we found a low number of 5-HT-immunopositive neurons (data not shown). We reasoned that this low number might be due to the immature state of iSNs at this time point and hypothesized that prolonged maturation might yield more mature serotonergic neurons capable of producing and releasing serotonin. Indeed, with 6 weeks of extended maturation of iSNs on a feeder layer of rat astrocytes, we observed a substantial increase in the percentage of 5-HT+ neurons (38±2%) and TPH+ neurons (Figure 4c) as compared with control iNs (0.2±0.06%) (Figures 4a and b). This increase was not due to serotonin potentially being generated by astrocytes, as we observed no 5-HT or TPH immunoreactivity in primary rat astrocytes when cultured alone (Supplementary Figure 4a). Examination at higher magnification revealed what appeared to be 5-HT+ vesicle-like structures in the somata, dendrites and axons of TPH+ iSNs (Figure 4d). We also found punctate SERT staining in the cell soma and in the processes of some TPH+/Map2ab+ neurons (Figure 4e). We extended our analysis to a later time point of 9 weeks and found a similar percentage of serotonergic neurons as compared with the 6-week time point (Figure 4f; Supplementary Figure 4d). Given that TPH was expressed by 3 weeks, we hypothesized that serotonergic fate specification must have already occurred by then and that doxycycline may not be continually required for serotonin production. To test this possibility, we transdifferentiated S4F fibroblasts as described above but withdrew doxycycline from the media after 4 weeks. We found similar percentages of 5-HT+ neurons in both conditions (Supplementary Figures 4b and c). Our results suggest that doxycycline is not continually required and that 4 weeks of conversion with doxycycline is sufficient for serotonergic fate specification.
To assess the activity of iSNs, we next analyzed their electrophysiological characteristics, such as evoked and spontaneous action potentials. Patch-clamp recordings confirmed mature neuronal properties, including Na+/K+ channel-mediated inward/outward currents and the generation of multiple action potentials (Figures 4g–i). Using this methodology, we also observed spontaneous action potentials in iSNs, but at low frequencies (Figure 4i). We confirmed that neurons from which electrical recordings were performed were serotonergic, as they were immunopositive for TPH+ (Figure 4g). While patch-clamp recordings ascertained mature neuronal properties of individual iSNs, we next sought to examine patterns in network activity by observing calcium transients of iNs and iSNs. We performed time-lapse calcium imaging of iNs and iSNs cultured on rat astrocytes until 6 weeks (Figures 4j and k). We first labeled iNs and iSNs with the lentiviral reporters Synapsin::DsRed and TPH2::TdT, respectively, and then loaded cells with the calcium-sensitive dye Fluo-4 and proceeded with time-lapse imaging for approximately 2 min (Figure 4j). First, we observed that both iNs and iSNs exhibited spontaneous calcium transients, suggesting that they were part of an active network (Supplementary Videos 1–4). Second, we observed a variety of patterns of calcium transients in iSNs and iNs (Figure 4k) but no noticeable differences in the patterns of activity of neurons between the two groups. These data demonstrate that iSNs were spontaneously active and were part of an active neuronal network in vitro. Collectively, these results indicate that iSNs are functional neurons, capable of generating intracellular serotonin as well as induced and spontaneous electrical activity.
Drug treatments and extracellular serotonin measurement from iSNs
Given that iSNs were immunopositive for 5-HT and were able to fire spontaneously, we next asked whether they were capable of releasing serotonin in the culture medium. Using a competitive ELISA method (Supplementary Figure 5), we found higher serotonin (7.5±1.0 ng ml−1) in the culture medium of iSNs as compared with control iN medium (1.3±0.4 ng ml−1) (Figure 5a). These data showed that, at this time point, iSNs released detectable levels of 5-HT in our culture conditions. Based on this information, we next asked whether iSNs were responsive to drugs that target serotonergic neurons for regulating extracellular and synaptic 5-HT concentrations. For this experiment, we utilized a drug that binds to SERT, which is expressed by serotonergic neurons. SSRIs, a commonly prescribed class of antidepressants, prevent 5-HT reuptake by blocking SERT, thereby increasing extracellular 5-HT levels. We hypothesized that, if iSNs were functionally active in releasing serotonin and degrading 5-HT via reuptake, then treatment with SSRIs might increase extracellular 5-HT concentration in the culture medium. To test this hypothesis, we collected 5-day-old culture medium before and after treatment of 6- to 8-week-old iSNs with Citalopram (SSRI), daily for 5 days. ELISA measurements revealed a less than twofold increase in 5-HT concentration after 5 days of SSRI treatment (Figures 5b and c). Our results suggest but do not confirm that the observed increase in serotonin concentration with Citalopram treatment is necessarily via SERT blockade, as SSRIs may have non-SERT related effects.53, 54 Furthermore, iSNs respond in a predictable manner to serotonergic drugs such as Citalopram. Tryptophan is the required precursor for serotonin biosynthesis via TPH, and increasing the precursor is thought to boost serotonin production. To test this in our culture system in vitro, we added tryptophan daily to iSN culture medium and collected medium 5 days later. We found nearly a twofold increase in extracellular serotonin concentrations after treatment with tryptophan vs before treatment in both tested iSN lines (Figure 5c). Our results demonstrate that iSNs are responsive to SSRIs and tryptophan in a predictable manner, suggesting that this platform could be utilized for examining the serotonin-releasing action of other relevant compounds.
Given the heterogeneity of the neuronal subtypes in the brain and their specific contributions to brain function, a major scientific focus has been to generate and examine specific human neuronal subtypes in vitro. In our study, we observed that four transcription factors associated with serotonergic differentiation—NKX2.2, FEV, GATA2 and LMX1B—were endogenously upregulated during serotonergic specification of human pluripotent stem cells in vitro. These serotonergic transcription factors, in combination with the neuronal transcription factors ASCL1 and NGN2 (AN), were sufficient for the targeted transdifferentiation of human fibroblasts directly into functional serotonergic neurons. iSNs not only expressed the required enzymatic machinery for serotonin biosynthesis but also upregulated other genes known to be expressed in the human and mouse raphe. Further, iSNs exhibited action potentials and released serotonin in vitro. As proof of principle, we also demonstrated that iSNs could be utilized for examining responses to serotonin-boosting drugs, such as SSRIs and tryptophan. Previous reports have indicated the utility of Shh and FGF-8 patterning molecules in increasing the yield of mostly dopaminergic neurons and, to a smaller extent, serotonergic neurons from ESCs in vitro.27, 55, 56 Despite the substantial increase in percentage of serotonergic neurons, as shown in previous reports and in our own, the total number of serotonergic neurons remains small in comparison with the other neuronal subtypes, making it technically challenging to examine serotonergic neurons accurately. We have overcome this challenge by using a novel tool, a TPH2::GFP lentiviral reporter that labels serotonergic neurons at around 70% accuracy. Utilizing this reporter, we were able to label and isolate serotonergic neurons from a mixed population and found enrichment of transcription factors NKX2.2, FEV, GATA2 and LMX1B. Our results corroborate in vivo mouse studies that confirm the expression and importance of these transcription factors in the developing raphe nuclei.2 While, in hindsight, it is not surprising that NKX2.2, FEV, GATA2 and LMX1B were upregulated, our study provides empirical data for the enrichment of these transcription factors in hESC-derived serotonergic neurons in vitro. A technical point of contention with utilizing fibroblasts as a source for transdifferentiation is that they have limited capacity for self-renewal. However, this is not an issue until they reach senescence, which typically occurs following 50 passages or more. A key aspect of our methodology is transducing and selecting fibroblasts at earlier passages, allowing us to harness the proliferative potential of fibroblasts for several transdifferentiation events before senescence is reached. Furthermore, we did not observe significant differences in transdifferentiation efficiencies over multiple passages and could efficiently generate serotonergic neurons from fibroblasts up to 35 passages.
Several recent studies have shown that forced expression of specific transcription factors may serve as a powerful tool for inducing transdifferentiation of fibroblasts to neurons.12 In accordance with our observations, previous studies showed that ASCL1- and/or NGN2-based neuronal transdifferentiation protocols have low overall yield of monoaminergic neurons,37, 39 whereas dopaminergic fate can be induced by the addition of region-specific transcription factors.17, 57, 58 Analogously, we show that co-expression of S4F is sufficient for substantially increasing serotonergic yield of AN-based neuronal conversion of human fibroblasts. Here, AN may act as the dominant pro-neuronal factors that determine neuronal identity, whereas S4F act to induce serotonergic subtype and some degree of regional specificity.59, 60, 61, 62 This concept is supported by the fact that the overexpression of S4F does not increase overall neuronal transdifferentiation efficiencies of an AN-based system. Interestingly, overexpression of NKX2.2 and FEV with AN was not sufficient for serotonergic fate conversion (data not shown), suggesting that all four transcription factors may have distinct roles in serotonergic fate specification. Our data support previous reports showing that overexpression of three transcription factors together but not singly (Lmx1b, Nkx2.2, Pet-1) was required for serotonergic fate specification in the chick embryo.59
Interestingly, we observed that TPH expression preceded serotonin immunoreactivity in neurons 3 weeks after the start of conversion. One explanation may be the immaturity of TPH+ neurons at this time point. Support for this idea comes from the facts that genes involved in the packaging and transport of serotonin, such as the vesicular momoamine transporter (SLC18A2) and serotonin transporter (SERT, SLC6A4), were not highly expressed at 3 weeks (Figure 3), and that an additional 3 weeks of maturation on astrocytes was enough to increase the percentage of 5-HT+ neurons as well as serotonin transporter expression (Figure 4). These results, along with the fact that other raphe-associated genes were upregulated at 3 weeks, suggest that TPH is likely an early/immature marker for serotonergic neurons in vitro.
In our time-lapse calcium imaging data, we observed a diverse set of activity patterns, such as single spiking and regular spiking in iNs and iSNs. Interestingly, adult rodent raphe serotonergic neurons also exhibit diverse activity patterns in vivo.63 Some serotonergic neurons exhibited slow, regular firing bursts,64 some single, regular and slow spiking,65 and others fast firing bursts that synchronized with hippocampal theta rhythm.65, 66 While it is possible that the observed firing patterns in raphe serotonergic neurons are cell autonomous, it is likely that the observed activity patterns are a result of serotonergic neurons being part of a mature and complex neuronal circuit. Network partners heavily influence activity patterns and electrophysiological properties of serotonergic neurons in the raphe,55 and given the stark differences between in vitro and in vivo environments, it is challenging to accurately compare iSNs in vitro to raphe serotonergic neurons in vivo.67
Serotonergic dysfunction has been implicated in several neuropsychiatric disorders such as major depression, schizophrenia and autism. Several classes of neuropsychiatric drugs have been shown to modulate serotonergic drive in the brain, for example, antidepressants and atypical antipsychotics.2, 68, 69, 70 The primary focus of our drug-testing experiments was to test whether iSNs were able to respond in a predictable manner to serotonin modulators. In an effort to ascertain the utility of iSNs in vitro, we found a near twofold increase in extracellular serotonin concentrations after treatment with tryptophan and the SSRI Citalopram. SSRIs such as Citralopram have a high affinity for SERT and have been shown to increase extracellular serotonin concentrations in vivo and in vitro.71 While this increase in extracellular concentration for serotonin has been proposed as a mechanism through which SSRIs mediate their antidepressant effects in vivo, mounting evidence suggests that SSRIs may have other non-SERT related targets that are therapeutically relevant.53, 54, 72 It is entirely possible that the observed increase in extracellular serotonin concentrations in iSNs treated with Citalopram was not directly via SERT but through indirect mechanisms. Nonetheless, directly and/or indirectly, both SSRIs and tryptophan increased serotonin concentrations, suggesting that this step may be relevant for their mood-elevating properties. Our data suggest that tryptophan may serve as a more potent molecule for boosting extracellular concentrations as compared with SSRIs, at least in vitro. Although it is not yet clear whether and how much these increases in serotonin correlate with therapeutically relevant concentrations, our results fit with studies showing that increasing dietary tryptophan boosts serotonin levels in vivo.73 Results from these experiments provide impetus for further studies characterizing the action, therapeutic dosage and timing of a wider range of serotonergic drugs in vitro. Further, we showed that fibroblasts derived from adults, 29 years of age, underwent transdifferentiation successfully. This age falls within the range of typical adult-onset psychiatric disorders such as bipolar disorder (at least 50% of cases before age 25), schizophrenia (average age between 18 to 35) and major depression (average age of onset at 32 and most prevalent between ages of 25–44) (https://www.nimh.nih.gov/health/topics/). However, only one of our lines falls within this age range and therefore, these results must be extrapolated with caution.
Taken together, our data demonstrate a unique strategy that enables the efficient generation of human serotonergic neurons, raising the possibility of studying cellular and molecular aspects of human serotonergic neurotransmission in the context of neurological disorders and drug response.
Lesch KP, Waider J . Serotonin in the modulation of neural plasticity and networks: implications for neurodevelopmental disorders. Neuron 2012; 76: 175–191.
Deneris ES, Wyler SC . Serotonergic transcriptional networks and potential importance to mental health. Nat Neurosci 2012; 15: 519–527.
Daubert EA, Condron BG . Serotonin: a regulator of neuronal morphology and circuitry. Trends Neurosci 2010; 33: 424–434.
Gaspar P, Cases O, Maroteaux L . The developmental role of serotonin: news from mouse molecular genetics. Nat Rev Neurosci 2003; 4: 1002–1012.
Benekareddy M, Vadodaria KC, Nair AR, Vaidya VA . Postnatal serotonin type 2 receptor blockade prevents the emergence of anxiety behavior, dysregulated stress-induced immediate early gene responses, and specific transcriptional changes that arise following early life stress. Biol Psychiatry 2011; 70: 1024–1032.
Vaidya VA, Vadodaria KC, Jha S . Neurotransmitter regulation of adult neurogenesis: putative therapeutic targets. CNS Neurol Disord Drug Targets 2007; 6: 358–374.
Lin SH, Lee LT, Yang YK . Serotonin and mental disorders: a concise review on molecular neuroimaging evidence. Clin Psychopharmacol Neurosci 2014; 12: 196–202.
Mosienko V, Beis D, Pasqualetti M, Waider J, Matthes S, Qadri F et al. Life without brain serotonin: reevaluation of serotonin function with mice deficient in brain serotonin synthesis. Behav Brain Res 2015; 277: 78–88.
Ming GL, Brustle O, Muotri A, Studer L, Wernig M, Christian KM . Cellular reprogramming: recent advances in modeling neurological diseases. J Neurosci 2011; 31: 16070–16075.
Gage FH, Temple S . Neural stem cells: generating and regenerating the brain. Neuron 2013; 80: 588–601.
Pang ZP, Yang N, Vierbuchen T, Ostermeier A, Fuentes DR, Yang TQ et al. Induction of human neuronal cells by defined transcription factors. Nature 2011; 476: 220–223.
Yang N, Ng YH, Pang ZP, Sudhof TC, Wernig M . Induced neuronal cells: how to make and define a neuron. Cell Stem Cell 2011; 9: 517–525.
Marchetto MC, Brennand KJ, Boyer LF, Gage FH . Induced pluripotent stem cells (iPSCs) and neurological disease modeling: progress and promises. Hum Mol Genet 2011; 20: R109–R115.
Nicholas CR, Chen J, Tang Y, Southwell DG, Chalmers N, Vogt D et al. Functional maturation of hPSC-derived forebrain interneurons requires an extended timeline and mimics human neural development. Cell Stem Cell 2013; 12: 573–586.
Maroof AM, Keros S, Tyson JA, Ying SW, Ganat YM, Merkle FT et al. Directed differentiation and functional maturation of cortical interneurons from human embryonic stem cells. Cell Stem Cell 2013; 12: 559–572.
Liu Y, Weick JP, Liu H, Krencik R, Zhang X, Ma L et al. Medial ganglionic eminence-like cells derived from human embryonic stem cells correct learning and memory deficits. Nat Biotechnol 2013; 31: 440–447.
Caiazzo M, Dell'Anno MT, Dvoretskova E, Lazarevic D, Taverna S, Leo D et al. Direct generation of functional dopaminergic neurons from mouse and human fibroblasts. Nature 2011; 476: 224–227.
Kriks S, Shim JW, Piao J, Ganat YM, Wakeman DR, Xie Z et al. Dopamine neurons derived from human ES cells efficiently engraft in animal models of Parkinson's disease. Nature 2011; 480: 547–551.
Ye W, Shimamura K, Rubenstein JL, Hynes MA, Rosenthal A . FGF and Shh signals control dopaminergic and serotonergic cell fate in the anterior neural plate. Cell 1998; 93: 755–766.
Kiyasova V, Gaspar P . Development of raphe serotonin neurons from specification to guidance. Eur J Neurosci 2011; 34: 1553–1562.
Craven SE, Lim KC, Ye W, Engel JD, de Sauvage F, Rosenthal A . Gata2 specifies serotonergic neurons downstream of sonic hedgehog. Development 2004; 131: 1165–1173.
Hendricks TJ, Fyodorov DV, Wegman LJ, Lelutiu NB, Pehek EA, Yamamoto B et al. Pet-1 ETS gene plays a critical role in 5-HT neuron development and is required for normal anxiety-like and aggressive behavior. Neuron 2003; 37: 233–247.
Zhao ZQ, Scott M, Chiechio S, Wang JS, Renner KJ, Gereau RWt et al. Lmx1b is required for maintenance of central serotonergic neurons and mice lacking central serotonergic system exhibit normal locomotor activity. J Neurosci 2006; 26: 12781–12788.
Alenina N, Bashammakh S, Bader M . Specification and differentiation of serotonergic neurons. Stem Cell Rev 2006; 2: 5–10.
Barberi T, Klivenyi P, Calingasan NY, Lee H, Kawamata H, Loonam K et al. Neural subtype specification of fertilization and nuclear transfer embryonic stem cells and application in parkinsonian mice. Nat Biotechnol 2003; 21: 1200–1207.
Nefzger CM, Haynes JM, Pouton CW . Directed expression of Gata2, Mash1, and Foxa2 synergize to induce the serotonergic neuron phenotype during in vitro differentiation of embryonic stem cells. Stem Cells 2011; 29: 928–939.
Lee SH, Lumelsky N, Studer L, Auerbach JM, McKay RD . Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells. Nat Biotechnol 2000; 18: 675–679.
Shimada T, Takai Y, Shinohara K, Yamasaki A, Tominaga-Yoshino K, Ogura A et al. A simplified method to generate serotonergic neurons from mouse embryonic stem and induced pluripotent stem cells. J Neurochem 2012; 122: 81–93.
Dolmazon V, Alenina N, Markossian S, Mancip J, van de Vrede Y, Fontaine E et al. Forced expression of LIM homeodomain transcription factor 1b enhances differentiation of mouse embryonic stem cells into serotonergic neurons. Stem Cells Dev 2011; 20: 301–311.
Kumar M, Kaushalya SK, Gressens P, Maiti S, Mani S . Optimized derivation and functional characterization of 5-HT neurons from human embryonic stem cells. Stem Cells Dev 2009; 18: 615–627.
Tailor J, Kittappa R, Leto K, Gates M, Borel M, Paulsen O et al. Stem cells expanded from the human embryonic hindbrain stably retain regional specification and high neurogenic potency. J Neurosci 2013; 33: 12407–12422.
Malchenko S, Xie J, de Fatima Bonaldo M, Vanin EF, Bhattacharyya BJ, Belmadani A et al. Onset of rosette formation during spontaneous neural differentiation of hESC and hiPSC colonies. Gene 2014; 534: 400–407.
Kim KS . Converting human skin cells to neurons: a new tool to study and treat brain disorders? Cell Stem Cell 2011; 9: 179–181.
Pfisterer U, Wood J, Nihlberg K, Hallgren O, Bjermer L, Westergren-Thorsson G et al. Efficient induction of functional neurons from adult human fibroblasts. Cell Cycle 2011; 10: 3311–3316.
Najm FJ, Lager AM, Zaremba A, Wyatt K, Caprariello AV, Factor DC et al. Transcription factor-mediated reprogramming of fibroblasts to expandable, myelinogenic oligodendrocyte progenitor cells. Nat Biotechnol 2013; 31: 426–433.
Son EY, Ichida JK, Wainger BJ, Toma JS, Rafuse VF, Woolf CJ et al. Conversion of mouse and human fibroblasts into functional spinal motor neurons. Cell Stem Cell 2011; 9: 205–218.
Liu ML, Zang T, Zou Y, Chang JC, Gibson JR, Huber KM et al. Small molecules enable neurogenin 2 to efficiently convert human fibroblasts into cholinergic neurons. Nat Commun 2013; 4: 2183.
Boyer LF, Campbell B, Larkin S, Mu Y, Gage FH . Dopaminergic differentiation of human pluripotent cells. Curr Protoc Stem Cell Biol 2012; Chapter 1: Unit1H 6.
Ladewig J, Mertens J, Kesavan J, Doerr J, Poppe D, Glaue F et al. Small molecules enable highly efficient neuronal conversion of human fibroblasts. Nat Methods 2012; 9: 575–578.
Gentile MT, Nawa Y, Lunardi G, Florio T, Matsui H, Colucci-D'Amato L . Tryptophan hydroxylase 2 (TPH2) in a neuronal cell line: modulation by cell differentiation and NRSF/rest activity. J Neurochem 2012; 123: 963–970.
Marchetto MC, Carromeu C, Acab A, Yu D, Yeo GW, Mu Y et al. A model for neural development and treatment of Rett syndrome using human induced pluripotent stem cells. Cell 2010; 143: 527–539.
Vadodaria KC, Brakebusch C, Suter U, Jessberger S . Stage-specific functions of the small Rho GTPases Cdc42 and Rac1 for adult hippocampal neurogenesis. J Neurosci 2013; 33: 1179–1189.
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S et al STAR: ultrafast universal RNA-seq aligner Bioinformatics 2013; 29: 15–21.
Anders S, Huber W . Differential expression analysis for sequence count data. Genome Biol 2010; 11: R106.
Bardy C, van den Hurk M, Eames T, Marchand C, Hernandez RV et al. Neuronal medium that supports basic synaptic functions and activity of human neurons in vitro. Proc Natl Acad Sci U S A 2015; 112: E2725–E2734.
Chambers SM, Fasano CA, Papapetrou EP, Tomishima M, Sadelain M, Studer L . Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat Biotechnol 2009; 27: 275–280.
Vierbuchen T, Ostermeier A, Pang ZP, Kokubu Y, Sudhof TC, Wernig M . Direct conversion of fibroblasts to functional neurons by defined factors. Nature 2010; 463: 1035–1041.
Ambasudhan R, Talantova M, Coleman R, Yuan X, Zhu S, Lipton SA et al. Direct reprogramming of adult human fibroblasts to functional neurons under defined conditions. Cell Stem Cell 2011; 9: 113–118.
Broccoli V, Caiazzo M, Dell'Anno MT . Setting a highway for converting skin into neurons. J Mol Cell Biol 2011; 3: 322–323.
Wapinski OL, Vierbuchen T, Qu K, Lee QY, Chanda S, Fuentes DR et al. Hierarchical mechanisms for direct reprogramming of fibroblasts to neurons. Cell 2013; 155: 621–635.
Spaethling JM, Piel D, Dueck H, Buckley PT, Morris JF, Fisher SA et al. Serotonergic neuron regulation informed by in vivo single-cell transcriptomics. FASEB J 2014; 28: 771–780.
Dougherty JD, Maloney SE, Wozniak DF, Rieger MA, Sonnenblick L, Coppola G et al. The disruption of Celf6, a gene identified by translational profiling of serotonergic neurons, results in autism-related behaviors. J Neurosci 2013; 33: 2732–2753.
Diaz SL, Doly S, Narboux-Neme N, Fernandez S, Mazot P, Banas SM et al. 5-HT(2B) receptors are required for serotonin-selective antidepressant actions. Mol Psychiatry 2012; 17: 154–163.
Richardson-Jones JW, Craige CP, Guiard BP, Stephen A, Metzger KL, Kung HF et al. 5-HT1A autoreceptor levels determine vulnerability to stress and response to antidepressants. Neuron 2010; 65: 40–52.
Studer L . Derivation of dopaminergic neurons from pluripotent stem cells. Prog Brain Res 2012; 200: 243–263.
Sanchez-Pernaute R, Studer L, Bankiewicz KS, Major EO, McKay RD . In vitro generation and transplantation of precursor-derived human dopamine neurons. J Neurosci Res 2001; 65: 284–288.
Pfisterer U, Kirkeby A, Torper O, Wood J, Nelander J, Dufour A et al. Direct conversion of human fibroblasts to dopaminergic neurons. Proc Natl Acad Sci USA 2011; 108: 10343–10348.
Liu X, Li F, Stubblefield EA, Blanchard B, Richards TL, Larson GA et al. Direct reprogramming of human fibroblasts into dopaminergic neuron-like cells. Cell Res 2012; 22: 321–332.
Cheng L, Chen CL, Luo P, Tan M, Qiu M, Johnson R et al. Lmx1b, Pet-1, and Nkx2.2 coordinately specify serotonergic neurotransmitter phenotype. J Neurosci 2003; 23: 9961–9967.
Ding YQ, Marklund U, Yuan W, Yin J, Wegman L, Ericson J et al. Lmx1b is essential for the development of serotonergic neurons. Nat Neurosci 2003; 6: 933–938.
Scott MM, Krueger KC, Deneris ES . A differentially autoregulated Pet-1 enhancer region is a critical target of the transcriptional cascade that governs serotonin neuron development. J Neurosci 2005; 25: 2628–2636.
Krueger KC, Deneris ES . Serotonergic transcription of human FEV reveals direct GATA factor interactions and fate of Pet-1-deficient serotonin neuron precursors. J Neurosci 2008; 28: 12748–12758.
Marinelli S, Schnell SA, Hack SP, Christie MJ, Wessendorf MW, Vaughan CW . Serotonergic and nonserotonergic dorsal raphe neurons are pharmacologically and electrophysiologically heterogeneous. J Neurophysiol 2004; 92: 3532–3537.
Hajos M, Gartside SE, Villa AE, Sharp T . Evidence for a repetitive (burst) firing pattern in a sub-population of 5-hydroxytryptamine neurons in the dorsal and median raphe nuclei of the rat. Neuroscience 1995; 69: 189–197.
Allers KA, Sharp T . Neurochemical and anatomical identification of fast- and slow-firing neurones in the rat dorsal raphe nucleus using juxtacellular labelling methods in vivo. Neuroscience 2003; 122: 193–204.
Kocsis B, Varga V, Dahan L, Sik A . Serotonergic neuron diversity: identification of raphe neurons with discharges time-locked to the hippocampal theta rhythm. Proc Natl Acad Sci USA 2006; 103: 1059–1064.
Varga V, Szekely AD, Csillag A, Sharp T, Hajos M . Evidence for a role of GABA interneurones in the cortical modulation of midbrain 5-hydroxytryptamine neurones. Neuroscience 2001; 106: 783–792.
Albert PR, Benkelfat C . The neurobiology of depression—revisiting the serotonin hypothesis. II. Genetic, epigenetic and clinical studies. Philos Trans R Soc Lond B Biol Sci 2013; 368: 20120535.
Charney DS . Monoamine dysfunction and the pathophysiology and treatment of depression. J Clin Psychiatry 1998; 59: 11–14.
Meltzer HY . The role of serotonin in antipsychotic drug action. Neuropsychopharmacology 1999; 21: 106S–115S.
Nagayasu K, Kitaichi M, Nishitani N, Asaoka N, Shirakawa H, Nakagawa T et al. Chronic effects of antidepressants on serotonin release in rat raphe slice cultures: high potency of milnacipran in the augmentation of serotonin release. Int J Neuropsychopharmacol 2013; 16: 2295–2306.
Komlosi G, Molnar G, Rozsa M, Olah S, Barzo P, Tamas G . Fluoxetine (prozac) and serotonin act on excitatory synaptic transmission to suppress single layer 2/3 pyramidal neuron-triggered cell assemblies in the human prefrontal cortex. J Neurosci 2012; 32: 16369–16378.
Fernstrom JD . Role of precursor availability in control of monoamine biosynthesis in brain. Physiol Rev 1983; 63: 484–546.
KCV was supported by the Swiss National Science Foundation (SNSF) outgoing postdoctoral fellowship. JM and CB were supported by fellowships Glenn Center for Aging Research and the FP7 Marie Curie, respectively. This research was supported by Lynn and Edward Streim, the Robert and Mary Jane Engman Foundation, the JPB Foundation, and the Leona M and Harry B Helmsley Charitable Trust grant #2012-PG-MED002. We thank Dr Juergen Winkler for a line of primary human fibroblasts, Dr Manching Ku for help with RNA sequencing, Jamie Simon for help with illustrations, Dr Himanish Ghosh for help with semi-automated image analysis and Mary Lynn Gage for editorial comments.
The authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Molecular Psychiatry website
About this article
Cite this article
Vadodaria, K., Mertens, J., Paquola, A. et al. Generation of functional human serotonergic neurons from fibroblasts. Mol Psychiatry 21, 49–61 (2016). https://doi.org/10.1038/mp.2015.161
Molecular Psychiatry (2021)
TransSynW : A single‐cell RNA ‐sequencing based web application to guide cell conversion experiments
STEM CELLS Translational Medicine (2021)
Frontiers in Pharmacology (2021)
Serotonin-specific neurons differentiated from human iPSCs form distinct subtypes with synaptic protein assembly
Journal of Neural Transmission (2021)
Stem Cell Reports (2021)