A phenotypic small-molecule screen identifies an orphan ligand-receptor pair that regulates neural stem cell differentiation

Abstract

High-throughput identification of small molecules that selectively modulate molecular, cellular, or systems-level properties of the mammalian brain is a significant challenge. Here we report the chemical genetic identification of the orphan ligand phosphoserine (P-Ser) as an enhancer of neurogenesis. P-Ser inhibits neural stem cell/progenitor proliferation and self-renewal, enhances neurogenic fate commitment, and improves neuronal survival. We further demonstrate that the effects of P-Ser are mediated by the group III metabotropic glutamate receptor 4 (mGluR4). siRNA-mediated knockdown of mGluR4 abolished the effects of P-Ser and increased neurosphere proliferation, at least in part through upregulation of mTOR pathway activity. We also found that P-Ser increases neurogenesis in human embryonic stem cell-derived neural progenitors. This work highlights the tremendous potential of developing effective small-molecule drugs for use in regenerative medicine or transplantation therapy.

Figure 1. P-Ser enhances neuronal differentiation of neural stem cells

(a) Scheme of small molecule screen. The SEM micrograph is from the embryo images collection at www.med.unc.edu/embryo_images, used with permission. (b) Secondary E12.5 spheres differentiated with or without P-Ser and stained for Tuj1 and propidium iodide, scale bar = 300μm. (c) Neuron counts of differentiated secondary E12.5 spheres, p=0.0013. (d) Neuron counts of differentiated secondary P1 cortical spheres, p=0.0169. (e) Dose-response of P-Ser concentration vs. number of neurons generated from differentiated E12.5 spheres. Data is normalized so that untreated is set to 100%. (f) RT-PCR analysis of differentiated E12.5 spheres. –RT, no reverse transcriptase control. For densitometry, data is normalized so that untreated is set to 100%. 10μM P-Ser was used in all experiments. Data are represented as mean ± SEM in c-e.

Figure 2. P-Ser slows progenitor proliferation and enhances neurogenic commitment

(a) Cells were plated at 1k/mL and grown with or without 10μM P-Ser for one week, then trypsinized and counted by trypan blue exclusion, p=0.0293. (b) Histogram of clonal neurosphere size after one week in culture. (c) BrdU labeling index; dissociated primary neurospheres plated were plated on poly-ornithine-fibronectin-coated coverslips and grown for four days with or without 10μM P-Ser, then pulsed with 5μM BrdU for two hours and fixed, p=0.009, scale bar = 100μm. (d) Number of tertiary clonal neurospheres derived from P-Ser treated (10μM) or non-treated secondary neurospheres. Data is normalized to untreated (100%), p=1.91×10^-5. (e) Analysis of Tuj1+ cells present in acutely dissociated neurospheres grown with or without 10μM P-Ser, p=0.00137. (f) Neurospheres were grown continually with or without 20μM P-Ser for the indicated number of passages. Neurospheres were then harvested for RNA for RT-PCR analysis of NGN2 expression, or differentiated and stained for Tuj1, p(2°)=0.0031, p(6°)=0.00031. (g) Neurospheres were continually grown with or without 20μM P-Ser over multiple passages. Aliquots of each passage were differentiated and stained for Tuj1, p(3°)=.00295, p(4°)=0.0163, p(5°)=0.0111, p(6°)=0.0075. Multiple comparisons across passages were not made. (h) Tertiary neurospheres were derived from P-Ser treated (20μM) or non-treated secondary spheres, differentiated for five days, and stained for Tuj1, p=2.86×10^-5. Data are represented as mean ± SEM in b, d-h.

Figure 3. P-Ser improves survival in differentiating neuron cultures

(a) Three-day old E14.5 cortical neurons cultured with or without 10μM P-Ser. TUNEL assay (green) was performed on cultures, followed by staining for Tuj1 (red), scale bar = 25μm. (b) Number of Tuj1+ cells in one or three day E14 cortical neuron cultures untreated or with 10μM P-Ser. 1dayCtrl vs. 1dayP-Ser p=0.0931, 3dayCtrl vs. 3dayPSer p=0.0346, 1dayCtrl vs. 3dayCtrl p=0.0136, 1dayPSer vs. 3dayPSer p=0.0284. (c) Number of TUNEL+ cells in one or three day E14 cortical neuron cultures untreated or with 10μM P-Ser. 1dayCtrl vs. 1dayP-Ser p=0.0719, 3dayCtrl vs. 3dayP-Ser p=0.0138, 1dayCtrl vs. 3dayCtrl p=0.025, 1dayPSer vs. 3dayPSer p=0.1877. (d) Number of TUNEL+Tuj1 double-positive cells as a percentage of total Tuj1+ cells in three day E14 cortical neuron cultures untreated or with 10μM P-Ser, p=0.0013. All data are presented as mean ± SEM.

Figure 4. The effects of P-Ser are mediated by metabotropic glutamate receptor 4

(a) Histogram of secondary sphere sizes, cells plated at 1k/mL and grown untreated or with 10μM P-Ser, 20μM L-AP4, 15μM UBP1112, or 10μM P-Ser + 15μM UBP1112. (b) P1 secondary cortical spheres differentiated untreated or with 20μM L-AP4, 10μM P-Ser, 10μM P-Ser + 15μM UBP1112, or 15μM UBP1112, then stained for Tuj1 expression. Ctrl vs. L-AP4 p=0.0034, Ctrl vs. P-Ser p=0.0041, P-Ser vs Pser+UBP p=0.0065. Presented as mean±SEM. (c) RT-PCR analysis of group III metabotropic glutamate receptor expression in primary (1°) and secondary (2°) neurospheres, and postnatal day 8 mouse brain (P8, positive control). The retina-specific mGluR6 was not detected in any experiments (data not shown). (d) RT-PCR analysis of cells transfected with negative control siRNA (siNC) or duplexes targeting mGluR4 (simGluR4). (e) Histogram of neurosphere size following transfection with siNC or simGluR4 and treatment with water or 10μM P-Ser. Similar results were obtained with both duplexes; for clarity, only results from duplex #2 (simGluR4-2) are shown. –RT, no reverse transcriptase. (f) Western blot analysis of E12.5 adherent progenitors cultured for three days post-transfection with either siNC, simGluR4-1, or simGluR4-2, or no transfection (NT) (g) Chemical structures of L-Serine, P-Ser, L-Glutamate, L-AP4, and UBP1112.

Figure 5. mGluR4 mediates the effects of P-Ser on differentiation

(a) mGluR4 knockdown abolishes P-Ser-induced neuronal differentiation. E12 progenitors were transfected with negative control siRNA (siControl) or duplex targeting mGluR4 (simGluR4) and differentiated for three days in the absence or presence of 20μM P-Ser, and stained for Tuj1. siCtrl UT vs. siCtrl P-Ser p=0.0018, siCtrl P-Ser vs simGluR4 Pser p=0.0021. (b) mGluR4 knockdown increases oligodendrocyte differentiation. Cells were treated as in (a) and stained for the oligodendrocyte marker O4. siCtrl UT vs. simGluR4 UT p=0.0374, siCtrl P-Ser vs. simGluR4 P-Ser p=0.0184. (c) E12 progenitors transfected with siControl or simGluR4, differentiated in the absence of P-Ser, and stained for Tuj1 (green) and O4 (blue), counterstained with propidium iodide (red). Scale bar = 50 μm.

Figure 6. P-Ser increases neuronal differentiation of human embryonic stem cells

(a) Human embryonic stem cell-derived neural progenitors (hES-NPCs) derived from line HSF-1 were differentiated with or without 10μM P-Ser for six days, then stained for Tuj1 expression. Left panels, Tuj1 (green); right panels, Tuj1 (green) and propidium iodide (red). Scale bar = 100μm. (b) hES-NPCs derived from line HSF-1 were differentiated as above and the percentage of Tuj1+ cells was determined, p=0.000862. (c) hES-NPS derived from line HSF-6 were differentiated as above and the percentage of Tuj1+ cells was determined, p=0.0158. Data are represented as mean ± SEM.