5-HT2 receptors mediate functional modulation of GABAa receptors and inhibitory synaptic transmissions in human iPS-derived neurons

Neural progenitors differentiated from induced pluripotent stem cells (iPS) hold potentials for treating neurological diseases. Serotonin has potent effects on neuronal functions through multiple receptors, underlying a variety of neural disorders. Glutamate and GABA receptors have been proven functional in neurons differentiated from iPS, however, little is known about 5-HT receptor-mediated modulation in such neuronal networks. In the present study, human iPS were differentiated into cells possessing featured physiological properties of cortical neurons. Whole-cell patch-clamp recording was used to examine the involvement of 5-HT2 receptors in functional modulation of GABAergic synaptic transmission. We found that serotonin and DOI (a selective agonist of 5-HT2A/C receptor) reversibly reduced GABA-activated currents, and this 5-HT2A/C receptor mediated inhibition required G protein, PLC, PKC, and Ca2+ signaling. Serotonin increased the frequency of miniature inhibitory postsynaptic currents (mIPSCs), which could be mimicked by α-methylserotonin, a 5-HT2 receptor agonist. In contrast, DOI reduced both frequency and amplitude of mIPSCs. These findings suggested that in iPS-derived human neurons serotonin postsynaptically reduced GABAa receptor function through 5-HT2A/C receptors, but presynaptically other 5-HT2 receptors counteracted the action of 5-HT2A/C receptors. Functional expression of serotonin receptors in human iPS-derived neurons provides a pre-requisite for their normal behaviors after grafting.


Induced neurons from neural stem cells of embryonic stem cell (ES) and induced pluripotent stem cells (iPS)
origins not only exhibited neuronal morphology with extensive axon and dendrites, but also possessed mature electrophysiological properties such as repetitive action potentials in response to current stimulation and integrated synaptic connections with host both in vitro and in vivo [1][2][3] . The iPS-derived neurons recapitulated symptoms of patients with neurological diseases at cellular and synaptic levels, providing a rather distinct insight into the underlying mechanisms 4,5 . Autologous neuronal progenitors obtained from iPS would provide enough source cells for transplantation, which hold great potentials to treat neurological disorders such as Parkinson's disease, schizophrenia and stroke [6][7][8][9][10][11][12][13] . Input-specific long-term potentiation 14 , homeostatic plasticity, as well as BDNF induced synaptic changes 15 were reported for ES-derived neurons, further physiologically supporting the functional integrity of transplanted progenitors.
The development and function of the cerebral cortex is subjected to massive monoaminegic modulation. Serotonin as one of important neuromodulators participates in many neural processes including synaptic transmission, mood, learning and memory [16][17][18][19] . Serotonergic fibers from raphe nucleus diffusely projecting to the brain exert its modulatory effect through multiple receptor subtypes (1-7) 19 . So far amounting evidences revealed the existence of glutamate and GABA receptors in neural stem cell derived neurons 1,2 and even the role of neuromodulators in the proliferation and differentiation of neural stem cells 20,21 ; however few studies addressed the neuromodulation mediated by serotonin and its receptors. The transplanted neurons have to communicate with host neurons and to be modulated normally to achieve satisfactory therapeutic efficacy. Activation of 5-HT 2 receptors has been reported to broadly regulate receptor function and neuronal excitability 7,[22][23][24][25][26] . Functional expression of serotonin receptors in human iPS-derived neurons would endow them with the ability to receive normal serotonergic modulation from the host after grafting, and to serve as a cell model to screen serotonin related neuropsychological drugs. Here we differentiated iPS into forebrain-like cortical neurons, and whole-cell patch-clamp recording technique was adopted to investigate the involvement of 5-HT 2 receptors in the modulation of GABAa receptor function and inhibitory synaptic transmissions.

Results
Acquisition of forebrain-like neural progenitor cells from human iPS. Neural progenitor cells (NPCs) were derived from human iPS reprogrammed from urine cells (UC5) or fibroblasts (GZ2) as described before 27,28 , as well as ES (H1) as a control. Unless otherwise specified, the majority of the results shown here were obtained from iPS line UC5, which was derived by using a feeder-free, serum-free and virus-free method without oncogene c-MYC, representing a more promising translational application for regenerative medicine. These NPCs expressed Sox1, Pax6 and Nestin of high percentages (Fig. 1a), without mesodermal and endodermal contamination (undetectable markers T and Sox17 by Q-PCR), whereas pluripotency markers including Oct4 and Nanog stopped to express (Fig. 1b,c). The NPCs can stably expand in both suspension and adhesion conditions in neural expanding medium containing EGF and bFGF. More importantly, these cells possessed dorsal forebrain regional identity, which maintained even after long-term passages (Fig. 1d). The NPCs were highly neurogenic, and showed typical neuronal morphology and gene expression profile upon differentiation and maturation (Fig. 1e).
Moreover, the puffed glutamate (1 mM) evoked an inward current (Fig. 3c) and spontaneous excitatory postsynaptic currents was blocked by DNQX (20 μ M), a selective glutamate receptor antagonist, which indicated that glutamate receptors mediated excitatory synaptic transmission (Fig. 3d). The iPS-derived neuron also responded to puffed 1 mM GABA with an inward current (Fig. 3e). The GABAergic spontaneous IPSCs (Fig. 3f) and GABA-activated currents reversed at 0.92 ± 2.33 mV (n = 4, Fig. 3g), which approximated the theoretical equilibrium potentials of Cl − with CsCl-based internal solutions 32 . Bicuculline (10 μ M), a selective GABAa receptor antagonist completely blocked GABA-activated current and spontaneous IPSCs (Fig. 3h). These indicated that GABAa receptors on the postsynaptic membrane mediated inhibitory synaptic transmission. These electrophysiological evidences suggested that iPS-derived neurons behaved as normal neurons communicating each other through fundamental synaptic connections. To minimize developmental difference of serotonin receptors, at least 4-week-old mature iPS-derived neurons were used to examine the serotonergic modulation.

5-HT 2A/C receptor mediated reduction of GABA-activated current. Serotonergic projections dif-
fusely innervate the central nervous system via multiple types of receptors, and most 5-HT receptors are metabotropic receptors that are involved in the regulation of neuronal function. Serotonin at 40 μ M significantly and reversibly reduced GABA activated currents (reduction%: 45.68 ± 6.14%, n = 5, p = 0.002, Student's t-test) (Fig. 4a,d) in the iPS-derived neurons. Since 5-HT 2 receptors were reported to responsibly modulate GABAa receptors in cortical neurons, we applied DOI (20 μ M), a 5-HT 2A/C receptor agonist to examine whether 5-HT 2A/C receptors accounted for the reduction. We found that DOI similarly reduced GABA-activated currents as serotonin (reduction%: 29.64 ± 2.65%, n = 10, p < 0.001, Student's t-test) (Fig. 4b,d). However α -methylserotonin, the agonist of 5-HT 2 receptor, did not cause any change of GABA-activated currents (reduction%: 0.80 ± 1.61%, n = 6, p = 0.612, Student's t-test) (Fig. 4c,d), suggesting a counteractive interaction between subtypes of 5-HT 2 receptors. The depressive effect of 5-HT and DOI was also replicated in the induced neurons reprogrammed from another cell line (GZ2) (reduction%: 5-HT: 52.87 ± 9.91%, n = 8, p = 0.001; DOI: 28.14 ± 5.37%, n = 6, p = 0.001, Student's t-test). The reducing effect of 5-HT and DOI on GABA current could be largely attenuated by ketanserin (40 μ M), a selective antagonist of 5-HT 2A/C receptor (reduction%: 5-HT: 14.59 ± 2.91%, n = 4; DOI: 5.04 ± 1.99%, n = 4) (5-HT + ketanserin vs. 5-HT, p < 0.005; DOI + ketanserin vs. DOI, p < 0.001, Student's t-test) (Fig. 4d), demonstrating that the reduction of GABA-activated current by serotonin could be majorly ascribed to the activation of 5-HT 2A/C receptors at the postsynaptic membrane in human iPS-derived neurons. Notably, the larger reduction caused by 5-HT than DOI, as well as the incomplete blocking by ketanserin, implied the contribution of other 5-HT subtype receptors. And we did observe that 8-hydroxy    G-protein is required for 5-HT 2A/C receptor mediated reduction of GABA-activated currents. Activation of G-protein is the prerequisite of the action of G-protein-coupled receptors (GPCR), and all 5-HT receptors except ionic 5-HT 3 were coupled to different G proteins 33 . We then examined whether activation of G-protein was the prerequisite of the action of 5-HT and DOI on GABA current. When dialysis of GDP-β -s (0.5 mM) constitutively prevented G-protein complex from dissociating into active subunits of G α and G βγ, the depressive effect of 5-HT (Fig. 5a) and DOI (Fig. 5b) on GABA activated currents was completely blocked (5-HT: 98.68 ± 0.49% of control, n = 6; DOI: 100.10 ± 3.21% control, n = 6) (5-HT + GDP-β -s vs. 5-HT, p < 0.005; DOI + GDP-β -s vs. DOI, p < 0.001, Student's t-test) (Fig. 5c). These data indicated that negative regulation of GABAa receptors by 5-HT 2 receptors was dependent on the activation of G-protein. In this and following recordings of applying intracellular blockers, it must be mentioned that at least 5-10 minutes' waiting period was needed to achieve a successful inhibition 34 . Calcium signaling is involved in the modulation of GABA current by 5-HT 2A/C receptor. Ca 2+ as an important second messenger participates in many physiological processes, and its rise triggered by IP3 would have a potential impact on GABAa receptors as reported before. BAPTA as a potential and rapid chelator of intracellular Ca 2+ was used to test the involvement of calcium signaling in the process of 5-HT 2A/C receptor  mediated reduction of GABA current. When BAPTA (10 mM) was included in the pipette solutions, interestingly both 5-HT (Fig. 6a) and DOI (Fig. 6b) lost their ability to change the amplitude of GABA currents (5-HT: 98.39 ± 1.66%, n = 7; DOI: 99.94 ± 1.32%, n = 5) (5-HT + BAPTA vs. 5-HT, p < 0.005; DOI + BAPTA vs. DOI, p < 0.001, Student's t-test) (Fig. 6c).   receptors. We examined whether the same mechanism occurred in iPS-derived neurons by intracellularly delivering correspondent inhibitors. As expected, when PLC activity was blocked with U-73122 (5 μ M) in recording pipette, GABA-activated currents remained unchanged in the presence of DOI (97.03 ± 0.73% control, n = 6) (DOI + U-73122 vs. DOI, p < 0.001, Student's t-test) (Fig. 7a,d), supporting the involvement of PLC in 5-HT 2A/C receptors mediated regulation of GABAa receptors.
Taken together, these data suggested that PLC and PKC act as downstream signaling cascade to achieve the modulation of GABAa receptors by 5-HT 2A/C receptors in human iPS-derived forebrain neurons.

Discussion
Our present study demonstrated presynaptic and postsynaptic 5-HT 2 receptors functionally regulated inhibitory synaptic transmissions in human iPS-derived neuronal networks, confirming that these induced neurons highly resemble those of naïve forebrain neurons in terms of physiological properties. Our major findings are: (1) Using our present neural induction protocol, NPCs can be stably and efficiently induced from human iPS with no manual selection needed; (2) These iPS-derived NPCs could be differentiated into morphologically and functionally mature cortical neurons; (3) Serotonin modulated GABAa receptors mainly through 5-HT 2A/C receptors; (4) Postsynaptically, 5-HT 2A/C receptors exerted its modulating effect through GPCR-PLC-PKC, and Ca 2+ signaling pathway; (5) Presynaptically, the inhibiting effect mediated by 5-HT 2A/C receptors was counteracted by other 5-HT 2 receptors.
Neuronal progenitors generated in vitro as donor cells for transplantation hold great promise to treat neurological diseases [6][7][8]11,13,35,36 . Neural stem cells from iPS have advantages over those from ES in term of less ethic concerns and immunity rejections 8,12 . Currently virus-free generation of iPS warranted the safety of iPS-derived neurons in translational application 1,35 . Increasing evidences indicated that transplanted progenitors could survive, migrate, differentiate, integrate into host neural circuitry, and even correct the behavioral deficits 2,3,6,7,13,36 . It has already been reported that ES-derived neurons exhibited homeostatic plasticity and BDNF-induced synaptic plasticity 15 , as well as input-specific long-term potentiation 14 . The brain circuitry is known to receive massive monoaminergic projecting fibers engaging in circadian rhythm, motor coordination , learning and memory 37 . The brain function can't be properly fulfilled without serotonin modulation 16 . GABAergic inhibition balances brain state by opposing glutamatergic excitation 38,39 . The crosstalk between GABA and serotonin systems has been justified in the forebrain 23 , where 5-HT 2 receptors mediate the regulation of GABAergic system both presynaptically and postsynaptically. Therefore the iPS-derived neurons for engrafting should be modulated as its host counterparts, otherwise new disorders might be caused by the transplantation.
Serotonin mainly activated postsynaptic 5-HT 2A/C receptors to reduce GABA -activated currents through G protein, PLC, PKC, as well as Ca 2+ signaling. DOI reduced inhibitory neurotransmitters activated currents in the cortical neurons either through receptor phosphorylation or trafficking 23,40 . We postulated that our human iPS-derived neurons likely shared the same mechanism. Presynaptic 5-HT 2 receptors were also reported to regulate transmitter release 22,25,41 . Serotonin significantly influenced mIPSCs through 5-HT 2 receptors in our iPS-derived neural network. The enhancement of mIPSCs by serotonin could be mimicked by 5-HT 2 receptor agonists. Conversely, a reduction of frequency and amplitude of mIPSCs was caused by DOI. The frequency of mIPSCs is closely related to the probability of transmitter release. The decreased frequency of mIPSCs by DOI means less transmitter release from presynaptic terminals, but this action can be counteracted by other 5-HT 2 receptors 42 . Overall our findings suggested that 5-HT 2A/C receptors not only downregulated postsynaptic GABAa receptors, but also attenuated presynaptic GABA release of the iPS-derived neurons.
Taken together, our findings demonstrate 5-HT 2 receptors functionally modulated GABAergic synaptic transmission in the neural networks composed of human iPS-derived neurons, suggesting that iPS-derived neurons would receive diffused serotonin neuromodulation after transplantation just like host cells, but also could sever as an ideal in vitro model for studying neurological diseases and screening serotonin related neuropsychological drugs.

Methods
All experiments were carried out in accordance with the guidelines of the Human Subject Research Ethics Committee at Guangzhou Institutes of Biomedicine and Health (GIBH), Chinese Academy of Sciences (CAS), and the Committee approved the experiments. Formal informed consent was obtained from all subjects. Cell culture and neural differentiation. Two human iPS lines from healthy human, UC5 (Passage [15][16][17][18][19][20][21][22][23][24][25] 28 and GZ2 (Passage 10-20) 27 , established and maintained in our laboratory and one human ESCs line, H1 (Passage 40-50, Wicell, Madison, WI, USA), were adopted in the present study. UC5 cell line was derived from urine cells using a feeder-free, serum-free and virus-free method without oncogene c-MYC. GZ2 cell line was reprogrammed from skin fibroblasts by using Yamanaka factors 43 . All these pluripotent cells were cultured as described elsewhere 44,45 on plates coated with Matrigel (BD Biosciences, San Jose, CA, USA) in mTesR1 medium (Stemcell Technologies, Vancouver, BC, Canada), and routinely passaged by EDTA (Ethylene Diamine Tetraacetic Acid, 0.5 mM) dissociation every 4-6 day.
Neural induction was performed as previously reported via a monolayer strategy 46 by dual inhibition of SMAD signaling with empirical modifications to get highly homogenous neural progenitor cells (NPCs) of dorsal forebrain identity. Briefly, once the pluripotent cells got 100% confluence (Day 0 of neural induction), the medium was changed to neural induction medium (NIM). NIM basically contained N2B27 medium (DMEM/ F12: Neurobasal (1:1), 0.5% N2, 1% B27, 1% Glutamax (GIBCO), 1% non-essential amino acid (GIBCO)), plus 2 inhibitors (5 μ M S431542 (Sigma), 5 μ M Dorsomophin (Sigma)). By the 8-9 th day post induction, when a uniform packed neuroepithelial layer appeared routinely, cell aggregates dissociated from the neuroepithelium by manually scratching and gentle pipetting with no selection, were replated on new Matrigel plate, and continually fed with NIM without inhibitors for another 8 days. Then NPCs were harvested by floating the neural rosettes with Dispase digestion without picking and further purified via 1-2 passage(s) with Accutase digestion into single cells. The dorsal forebrain identity of NPCs was confirmed by immunostaining and Q-PCR.
To directly activate GABAa receptors on the membrane of recorded iPS-derived neuron, GABA (1 mM) was locally applied with a pressure of 15 psi controlled by a custom-made puffing device 48 . An external solution flowed following the GABA delivery to reduce receptor desensitization, and neurons were generally held at −70 mV to record inward ion currents. Gamma-aminobutyric (GABA) receptor-mediated inhibitory postsynaptic currents (IPSCs) were recorded in a voltage-clamp mode. The external solutions including different drugs were exchanged to perfuse the neurons. 6,7-dinitroquinoxaline-2,3-dione (DNQX, 20 μ M) and R-2-amino-5phosphonopentanoate (APV, 50 μ M) were used to block glutamatergic transmission. 1 μ M TTX was routinely added in the external solution to prevent spontaneous action potentials of the neural networks. Serotonin and α -methylserotonin was purchased from Sigma-Aldrich, and other receptor agonist and antagonists were purchased from Tocris Cookson Ltd. Data analysis. Off-line data analysis was performed by using Clampfit 10.2 (Molecular Devices, USA).
The Minianalysis (Synaptosoft, USA) was used to extract events for analyzing the frequency and amplitude of spontaneous IPSCs and waveforms of action potentials. Processed data were further imported into Origin 8.0 (OriginLab Corporation, USA) for plotting graphs. Numerical data were reported as mean ± SE (standard error). Student's t-test and Kolmogorov-Smirnov test (K-S test) were used to evaluate significance level unless otherwise stated, and p < 0.05 was considered statistically significant.