The Molecular Mechanisms Underlying Prostaglandin D2-Induced Neuritogenesis in Motor Neuron-Like NSC-34 Cells.

Prostaglandins are a group of physiologically active lipid compounds derived from arachidonic acid. Our previous study has found that prostaglandin E2 promotes neurite outgrowth in NSC-34 cells, which are a model for motor neuron development. However, the effects of other prostaglandins on neuronal differentiation are poorly understood. The present study investigated the effect of prostaglandin D2 (PGD2) on neuritogenesis in NSC-34 cells. Exposure to PGD2 resulted in increased percentages of neurite-bearing cells and neurite length. Although D-prostanoid receptor (DP) 1 and DP2 were dominantly expressed in the cells, BW245C (a DP1 agonist) and 15(R)-15-methyl PGD2 (a DP2 agonist) had no effect on neurite outgrowth. Enzyme-linked immunosorbent assay demonstrated that PGD2 was converted to 15-deoxy-Δ12,14-prostaglandin J2 (15d-PGJ2) under cell-free conditions. Exogenously applied 15d-PGJ2 mimicked the effect of PGD2 on neurite outgrowth. GW9662, a peroxisome proliferator-activated receptor–gamma (PPARγ) antagonist, suppressed PGD2-induced neurite outgrowth. Moreover, PGD2 and 15d-PGJ2 increased the protein expression of Islet-1 (the earliest marker of developing motor neurons), and these increases were suppressed by co-treatment with GW9662. These results suggest that PGD2 induces neuritogenesis in NSC-34 cells and that PGD2-induced neurite outgrowth was mediated by the activation of PPARγ through the metabolite 15d-PGJ2.


Introduction
Neuritogenesis is an early event in the differentiation of neural progenitor cells into neurons and enables neurons to develop axons and dendrites with which they connect to other cells and receive and transmit electrical signals [1]. Spinal motor neurons innervate muscle contraction by extending long axons that project from the ventral horn of the spinal cord gray matter directly to peripheral skeletal muscle [2]. Degeneration and loss of spinal motor neurons cause progressive and fatal motor neuron diseases such as amyotrophic lateral sclerosis (ALS), primary lateral sclerosis, and spinal muscular atrophy. More recently, induced pluripotent stem cell-derived motor neurons

Neurite Outgrowth Assay
Phase-contrast micrographs were captured using an inverted microscope (IX70, Olympus, Tokyo, Japan) with i-NTER LENS (Microscope Network Co. Ltd., Saitama, Japan). Afterwards, 50 cells per condition were randomly chosen to estimate the number of neurite-bearing cells using ImageJ (National Institute of Health, Bethesda, MD, USA) [28]. The percentage of neurite-bearing cells was calculated as the ratio of cells bearing neurite processes >1 cell diameter in length according to the method described previously [16]. The neurite length on each neurite-bearing cell was measured as the distance from the soma to the end of the neurite using NeuronJ, an ImageJ plugin originally developed for neurite tracing and analysis [29].

Enzyme-Linked Immunosorbent Assay (ELISA)
Quantification of PGD 2 and 15d-PGJ 2 in cell-free culture media was performed according to previous studies [30,31]. To detect PGD 2 and 15d-PGJ 2 in the culture medium, freshly prepared 15 µM PGD 2 was made by diluting a stock solution in cell-free DMEM containing 10% FBS and 1% penicillin-streptomycin, which was incubated in a humidified atmosphere containing 5% CO 2 at 37 • C. The culture medium was collected at 0, 0.5, 1, 2, 4, and 24 h, homogenized in methyl acetate, and centrifuged at 10,000× g for 10 min at 4 • C. The supernatant was evaporated, and then the residue was reconstituted in the assay buffer. PGD 2 was measured using the Prostaglandin D 2 -MOX EIA Kit (Cayman Chemical, Ann Arbor, MI, USA) according to the manufacturer's protocol. The absorbance of samples was measured using the microplate reader SH-1000Lab (Corona Electric, Ibaraki, Japan) at a test wavelength of 412 nm. Similarly, 15d-PGJ 2 was measured according to the protocol of the 15-deoxy-∆ 12,14 -PGJ 2 ELISA kit (Enzo Life Sciences, Farmingdale, NY, USA). The absorbance of samples Cells 2020, 9, 934 4 of 16 was measured using SH-1000Lab (Corona Electric, Ibaraki, Japan) at test and reference wavelengths of 405 and 570 nm, respectively.

Live/Dead Assay
The amount of viable and dead cells was measured by LIVE/DEAD ® Viability/Cytotoxicity Kit for mammalian cells (Molecular Probes, Eugene, OR, USA) according to a previously described method [27]. The images were collected with an inversed fluorescence microscope (IX70, Olympus, Tokyo, Japan) and cell mortality was determined by calculating the percentage of EthD-1-positive cells of the total number of cells (the sum of calcein-positive live cells and EthD-1-positive dead cells).

Statistical Analyses
Data analyses were performed using GraphPad Prism 6.0 (GraphPad Software, San Diego CA, USA). Data are expressed as mean ± standard error of the mean (SEM) or standard deviation (SD). Statistical significance was assessed by Student's t-test or one-way analysis of variance (ANOVA) followed by post hoc Tukey's multiple tests. Differences at p < 0.05 were considered statistically significant.

DP1 and DP2 Agonists Do Not Affect Neurite Outgrowth
Next, we tested whether DP1 and DP2 were expressed in undifferentiated NSC-34 cells by Western blotting. Similar to the reported DP1 and DP2 expression in the mouse spinal cord [23], we detected protein expression of both DP1 and DP2 in undifferentiated NSC-34 cells (Figure 2). 0.00000543) compared to vehicle-treated cells (0.5% ± 0.3%) ( Figure 1B). The maximal effect of PGD2 was observed at 15 μM, with an increase of up to approximately 30%. Moreover, the average neurite length was significantly increased in PGD2 (15 μM)-treated cells (78.1 ± 7.9 μm, p = 0.0218) compared to vehicle-treated cells (43.6 ± 8.7 μm) ( Figure 1B). Moreover, we investigated the cytotoxicity of PGD2 for undifferentiated NSC-34 cells. As shown in Figure 1C, exposure of differentiated NSC-34 cells to 15 μM PGD2 for 24 h did not change the percentage of dead cells stained by EthD-1.

DP1 and DP2 Agonists do Not Affect Neurite Outgrowth
Next, we tested whether DP1 and DP2 were expressed in undifferentiated NSC-34 cells by Western blotting. Similar to the reported DP1 and DP2 expression in the mouse spinal cord [23], we detected protein expression of both DP1 and DP2 in undifferentiated NSC-34 cells ( Figure 2).  To characterize the subtype(s) of DP responsible for PGD2-induced neurite outgrowth in undifferentiated NSC-34 cells, we investigated the effects of two well-characterized DP agonists, BW245C (a DP1 agonist) and 15(R)-15-methyl PGD2 (a DP2 agonist). Phase-contrast microscopy revealed no neurite outgrowth in these agonist (15 μM)-treated cells for 24 h ( Figure 3A,B). Exposure to BW245C or 15(R)-15-methyl PGD2 exerted no significant effect on the percentage of neurite-bearing cells and the average neurite length within the concentration range tested (1−15 μM) ( Figure 3A,B). Moreover, co-treatment with MK0524 (a DP1 antagonist, 10 μM) or CAY10471 (a DP2 antagonist, 10 μM) and PGD2 (15 μM) had no effect on PGD2-induced neurite outgrowth ( Figure 4A). There were no significant differences in the percentage of neurite-bearing cells or neurite length between the cells co-treated with MK0524 or CAY10471 and PGD2, and the cells treated with PGD2 alone ( Figure 4B). To characterize the subtype(s) of DP responsible for PGD 2 -induced neurite outgrowth in undifferentiated NSC-34 cells, we investigated the effects of two well-characterized DP agonists, BW245C (a DP1 agonist) and 15(R)-15-methyl PGD 2 (a DP2 agonist). Phase-contrast microscopy revealed no neurite outgrowth in these agonist (15 µM)-treated cells for 24 h ( Figure 3A,B). Exposure to BW245C or 15(R)-15-methyl PGD 2 exerted no significant effect on the percentage of neurite-bearing cells and the average neurite length within the concentration range tested (1−15 µM) ( Figure 3A,B). Moreover, co-treatment with MK0524 (a DP1 antagonist, 10 µM) or CAY10471 (a DP2 antagonist, 10 µM) and PGD 2 (15 µM) had no effect on PGD 2 -induced neurite outgrowth ( Figure 4A). There were no significant differences in the percentage of neurite-bearing cells or neurite length between the cells co-treated with MK0524 or CAY10471 and PGD 2 , and the cells treated with PGD 2 alone ( Figure 4B).

Discussion
PGE 2 plays a role in the differentiation of various cell types, such as regulatory T cells [33], osteoblasts [34], and neurons [35]. Our previous study reported that PGE 2 , which is increased in the spinal cord of ALS patients [36] and model mice [30], induced morphological differentiation of undifferentiated NSC-34 cells [16]. Lipocalin-type PGDS is a major brain-derived protein abundant in human cerebrospinal fluid, second only to albumin [37]. Although PGD 2 may be present in the CNS in amounts equal to or greater than PGE 2 , the influence of PGD 2 in motor neuron differentiation remains unclear. In this study, we demonstrated that exposure of undifferentiated NSC-34 cells to PGD 2 not only increased the percentage of neurite-bearing cells but also elongated neurite length without affecting the number of EthD-1 cells PI-positive dead cells. This suggests that PGD 2 , similar to PGE 2 , induces neuritogenesis in undifferentiated NSC-34 cells without affecting the cell viability.
A previous study reported that DP1 and DP2 are expressed in motor neurons of the adult mouse spinal cord [23]. Consistent with this report, our study demonstrated that DP1 and DP2 are expressed in undifferentiated NSC-34 cells. These results suggest that the characteristics of the model cell used in this study are comparable to those of the motor neurons in mouse spinal cords. Therefore, we investigated whether DP1 and/or DP2 contribute to the effect of PGD 2 on neurite outgrowth using subtype-specific agonists, BW245C (a DP1 agonist) and 15(R)-15-methyl PGD 2 (a DP2 agonist). Neither of these two agonists affected neurite outgrowth. Moreover, we showed that a DP1-selective antagonist (MK0524) and a DP2-selective antagonist (CAY10471) were unable to suppress PGD 2 -induced neuritogenesis. These results suggest that PGD 2 -induced neuritogenesis is not mediated by DP1 or DP2 in NSC-34 cells. Our previous study reported that PGE 2 induced neurite outgrowth by activating EP2 that is coupled to Gs protein [16]. However, our present study indicated that the activation of DP1 did not induce neuritogenesis despite DP1 being the same Gs protein-coupled receptor as EP2. A previous study demonstrated that PGD 2 (1 µM) caused approximately 25% of DP1 expressed at the membrane level to be internalized in HEK293 cells, reaching a plateau after 120−240 min [38]. However, PGE 2 (1 µM) caused approximately 40% of EP4 to be rapidly internalized, whereas EP2 was not internalized after 60 min in HEK293 cells [39]. These results imply that one possible explanation for the difference between the effect of DP1 and EP2 on neuritogenesis is that DP1 is internalized and desensitized, whereas EP2 is not. PGD 2 is unstable and is spontaneously metabolized to J 2 prostaglandins including PGJ 2 , ∆ 12 -PGJ 2 , and 15d-PGJ 2 [32]. It has been reported that PGD 2 is converted to J 2 prostaglandins in culture medium in the presence of mouse primary cortical neurons, and that a similar pattern of PGD 2 metabolism is observed in the absence of neurons [40]. In particular, 15d-PGJ 2 plays an important role in neurite outgrowth of the human neuroblastoma cell line LA-N-5 [41] and rat embryonic midbrain cells [25]. We demonstrated the conversion of PGD 2 to 15d-PGJ 2 in culture medium in the absence of NSC-34 cells. Moreover, exposure of NSC-34 cells to 15d-PGJ 2 resulted in significant neurite outgrowth without affecting the number of EthD-1 cells PI-positive dead cells. These results suggest that the conversion of PGD 2 to 15d-PGJ 2 induces neuritogenesis, and also show that PGD 2 and 15d-PGJ 2 do not affect the cell survival at the concentration used as a neuritogenesis inducer.
The concentration of PGD 2 in the CNS is approximately 500 pmol/g tissue (500 pM) in rat hippocampus, 250 pmol/g tissue (250 pM) in rat cerebral cortex [42], and 40 pg/mg tissue (113 pM) in mouse spinal cord [43]. Although the concentration of 15d-PGJ 2 in the spinal cord remains unclear, the concentration of 15d-PGJ 2 in the hippocampus was demonstrated to be approximately 8 pmol/g tissue (8 pM) in juvenile rats [42]. These in vivo studies suggest that the concentrations of PGD 2 and 15d-PGJ 2 found in the CNS and spinal cord are in the pM range. However, our findings indicated that the effective concentrations of PGD 2 and 15d-PGJ 2 on neurite outgrowth in NSC-34 cells were in the µM range. Although the effective concentration of PGD 2 on neurite outgrowth has not been reported in other in vitro models, 15d-PGJ 2 has been reported to induce neurite outgrowth in rat embryonic midbrain cells at concentrations of ≥0.5 µM [25]. These results suggest that the necessary concentrations of PGD 2 and 15d-PGJ 2 in in vitro studies are higher than physiological concentrations in vivo. In addition, the measured concentrations in in vivo studies represent average tissue concentrations. Considering that prostaglandins act as autocrine or paracrine factors for their target cells, local cellular concentration is more important than average tissue concentration. However, to the best of our knowledge, the local cellular concentrations of PGD 2 and 15d-PGJ 2 in the spinal cord are unknown, and thus further studies are needed. Moreover, it is also important to understand how exogenously applied 15d-PGJ 2 was transported into the cells. Although a previous study reported that exogenously applied 15d-PGJ 2 was transported into the cells and accumulated in the nucleus [44], the underlying mechanism remains unknown. One possibility is that intracellular uptake was mediated by prostaglandin transporters, similar to that observed with other prostaglandins [43]. Further studies are needed to determine the underlying mechanism of transportation in neurons and its model cells.
15d-PGJ 2 binds to DP2 with an affinity similar to PGD 2 [45]. Furthermore, 15d-PGJ 2 enhancement of nerve growth factor-induced neurite outgrowth is inhibited by CAY10471 (a DP2 antagonist) in PC12 cells [46]. In this study, 15(R)-15-methyl PGD 2 (a DP2 agonist) did not induce neurite outgrowth, suggesting that DP2 activation is not involved in 15d-PGJ 2 -induced neuritogenesis in NSC-34 cells. Moreover, 15d-PGJ 2 is commonly known as an endogenous ligand of PPARγ [44]. Indeed, 15d-PGJ 2 induces differentiation of rat embryonic midbrain cells into dopaminergic neuronal cells in a PPARγ-dependent manner [25]. Nevertheless, 15d-PGJ 2 is also involved in PPARγ-independent pathways via direct covalent binding to proteins other than PPARγ [47]; 15d-PGJ 2 enhances nerve growth factor-induced neurite outgrowth via covalent binding to activator protein-1 [48] or the transient receptor potential vanilloid 1 [49] in PC12 cells. These studies suggest that 15d-PGJ 2 -induced neuritogenesis and neuronal differentiation involve pathways that are dependent or independent of PPARγ, according to the cell type. We demonstrated that simultaneous treatment with GW9662 (a PPARγ antagonist) and PGD 2 for 24 h suppressed PGD 2 -induced neuritogenesis, suggesting that 15d-PGJ 2 converted from PGD 2 induces neurite outgrowth via a PPARγ-dependent pathway in undifferentiated NSC-34 cells. 15d-PGJ 2 contains two electrophilic carbons: the 13-carbon position reacts with the sulfur atom of the cysteine residue in PPARγ, whereas the 9-carbon position within the cyclopentenone ring binds to the cysteine residue in other proteins including nuclear factor-κB, IκB kinase, and activator protein-1 [50]. Further studies are needed to determine how 15d-PGJ 2 targets proteins such as PPARγ.
Islet-1 is a homeodomain transcription factor that is expressed from the early stages of motor neuron differentiation [51]. The expression of Islet-1 is upregulated during the differentiation of human pluripotent stem cells into motor neurons, and Islet-1 depletion by shRNA prevents Islet-1 expression and motor neuron differentiation [52]. Islet-1 is also expressed in retinoic acid-induced differentiated NSC-34 cells [6]. Therefore, Islet-1 expression is a critical regulator of mature motor neuron formation. In this study, we evaluated Islet-1 expression to determine whether PGD 2 and 15d-PGJ 2 promote the maturation of NSC-34 cells into motor neurons. We demonstrated that PGD 2 and 15d-PGJ 2 significantly upregulate Islet-1 expression, suggesting that PGD 2 and 15d-PGJ 2 can potentiate neural conversion of mature motor neurons. Moreover, we observed that GW9662 (a selective PPARγ antagonist) dramatically inhibited the PGD 2 -and 15d-PGJ 2 -induced increase in Islet-1 expression in differentiated NSC-34 cells. Treatment with 15d-PGJ 2 , retinoic acid, and ciglitazone (a selective PPARγ agonist) suppressed Islet-1 gene expression in neural stem cells generated by culturing dissociated brain cells from newborn mice [53]. In contrast, ciglitazone induced Islet-1 gene expression in human brain tumor stem cell cultures [54]. These results suggest that PPARγ regulates Islet-1 expression at the gene level depending on the type of cell. Although the mechanism underlying the upregulation of Islet-1 warrants further investigation, our results indicate that the PGD 2 -and 15d-PGJ 2 -induced increase in Islet-1 expression was mediated by PPARγ activation.
Several studies have shown that the differentiation of NSC-34 cells leads to neurite outgrowth and the expression of motor neuron-specific proteins [5,6]; as such, these cells can be employed as an in vitro experimental model to study motor neuron dysfunction in response to neurotoxins such as staurosporine, thapsigargin, hydrogen peroxide, homocysteine [55], glutamate [56,57], and cerebrospinal fluid from sporadic ALS patients [58]. However, as pointed out by a previous study [59], the usefulness of differentiated NSC-34 cells has been questioned as these cells exhibit less cell death and a small unsustained increase in the concentration of intracellular calcium after exposure to glutamate compared to primary motor neurons. Our previous study has demonstrated that EP2 and EP3, but not EP1 and EP4, are dominantly expressed in differentiated NSC-34 cells as well as motor neurons in the mouse spinal cord [60]. Similar to the increase in EP2 expression in PGE 2 -treated differentiated NSC-34 cells, EP2 upregulation in motor neurons of ALS model mice was observed at the early symptomatic age [26]. Therefore, these cells are believed to be a suitable model for assessing the response to prostaglandins in motor neurons.
In conclusion, we demonstrated for the first time that exogenously applied PGD 2 induces neuritogenesis in NSC-34 cells. We suggest that the underlying mechanism involves nonenzymatic conversion of PGD 2 into 15d-PGJ 2 , which then activates PPARγ ( Figure 9). As such, both PGD 2 and 15d-PGJ 2 (converted from PGD 2 ) promote the differentiation of progenitor cells into motor neurons and could thus be useful for establishing a motor neuron model. This novel effect of PGD 2 on the differentiation of motor neuron-like cells may provide a potential target for regenerative therapies for motor neuron diseases. Several studies have shown that the differentiation of NSC-34 cells leads to neurite outgrowth and the expression of motor neuron-specific proteins [5,6]; as such, these cells can be employed as an in vitro experimental model to study motor neuron dysfunction in response to neurotoxins such as staurosporine, thapsigargin, hydrogen peroxide, homocysteine [55], glutamate [56,57], and cerebrospinal fluid from sporadic ALS patients [58]. However, as pointed out by a previous study [59], the usefulness of differentiated NSC-34 cells has been questioned as these cells exhibit less cell death and a small unsustained increase in the concentration of intracellular calcium after exposure to glutamate compared to primary motor neurons. Our previous study has demonstrated that EP2 and EP3, but not EP1 and EP4, are dominantly expressed in differentiated NSC-34 cells as well as motor neurons in the mouse spinal cord [60]. Similar to the increase in EP2 expression in PGE2-treated differentiated NSC-34 cells, EP2 upregulation in motor neurons of ALS model mice was observed at the early symptomatic age [26]. Therefore, these cells are believed to be a suitable model for assessing the response to prostaglandins in motor neurons.
In conclusion, we demonstrated for the first time that exogenously applied PGD2 induces neuritogenesis in NSC-34 cells. We suggest that the underlying mechanism involves nonenzymatic conversion of PGD2 into 15d-PGJ2, which then activates PPARγ ( Figure 9). As such, both PGD2 and 15d-PGJ2 (converted from PGD2) promote the differentiation of progenitor cells into motor neurons and could thus be useful for establishing a motor neuron model. This novel effect of PGD2 on the differentiation of motor neuron-like cells may provide a potential target for regenerative therapies for motor neuron diseases.