Lysophospholipids are signaling molecules that play broad and major roles within the nervous system during both early development and neural injury. We used neural differentiation of human embryonic stem cells (hESC) as an in vitro model to examine the specific effects of lysophosphatidic acid (LPA) at various stages of neural development, from neural induction to mature neurons and glia. We report that LPA inhibits neurosphere formation and the differentiation of neural stem cells (NSC) toward neurons, without modifying NSC proliferation, apoptosis, or astrocytic differentiation. LPA acts through the activation of the Rho/ROCK and the phosphatidylinositol 3-kinase/Akt pathways to inhibit neuronal differentiation. This study is the first demonstration of a role for LPA signaling in neuronal differentiation of hESC. As LPA concentrations increase during inflammation, the inhibition of neuronal differentiation by LPA might contribute to the low level of neurogenesis observed following neurotrauma.
Disclosure of potential conflicts of interest is found at the end of this article.
Neural stem cells (NSC) have been extensively studied with the aim of using endogenous and/or donor NSC to replace neurons and restore circuitry in a neurodegenerative microenvironment. One of the most challenging aspects of NSC biology is that despite showing efficient neural differentiation in vitro, their differentiation potential in vivo tends to be biased toward glial lineages, particularly during degeneration/injury, when there is extensive and ongoing inflammation. It is therefore pertinent to study how NSC differentiation may be influenced by factors released during injury. In addition, the understanding of the downstream signaling pathways involved in this process provides an important step toward refining potential therapeutic strategies.
Lysophosphatidic acid (LPA) is a bioactive lysophospholipid that is released by activated platelets and constitutively present in serum . It is thus a significant factor contributing to an inflammatory response during neurotrauma following the impairment of the blood-brain barrier. LPA acts mainly through binding to its specific G-protein-coupled receptors: LPA1/Edg-2/rec.1.3/vzg-1/Gpcr26/Mrec1.3, LPA2/Edg-4, LPA3/Edg-7/RP4–678I3/HOFNH30, LPA4/P2Y9/GPR23, and LPA5/GPR92 (reviewed in ). It can also act intracellularly through the specific binding of the transcription factor peroxisome proliferator-activated receptor-γ (PPARγ) . In the central nervous system, LPA has been shown to target microglia , astrocytes [5, –7], oligodendrocytes , and neurons . LPA also plays an important role during neurogenesis in establishing the cerebral cortex , whereby it has a range of effects on neural stem/progenitor cells, including morphological rearrangement and differentiation (reviewed in ).
Given the essential role of lysophospholipids in the nervous system, both during early mammalian embryogenesis and in pathological states, we examined the biological effects of LPA on stem/progenitor cells at various stages of neural differentiation to mature cell types. For these studies, we used neural differentiation of human embryonic stem cells (hESC) as an in vitro model of analyzing various stages of neural development. Following protocols developed by Reubinoff et al.  and by Pera et al.  that allow progressive differentiation from hESC to mature postmitotic cells, the effects of LPA were examined at different stages of neural differentiation. We found that LPA specifically inhibited hESC-derived NSC differentiation into neurons and demonstrated the signaling pathways involved in this inhibition. These studies shed new light on how LPA may influence NSC differentiation in vivo, which is relevant not only during neurogenesis but also during neural injury.
Materials and Methods
Culture and Neuronal Induction of hESC
HES-2, HES-3, and HES-4 cells (WiCell Research Institute, Madison, WI, http://www.wicell.org) were cultured as previously described . Neuronal induction by noggin (500 ng/ml; R&D Systems Inc., Minneapolis, http://www.rndsystems.com) was performed as described . Noggin-treated cells were harvested after 14 days using dispase and were further subcultured in suspension in neural basal media (NBM) together with basic fibroblast growth factor (bFGF) (20 ng/ml; R&D Systems) and epidermal growth factor (EGF) (20 ng/ml; R&D Systems) as neurospheres . After 2 weeks of growth in suspension culture, the neurospheres were plated, as previously described, onto laminin- or fibronectin-coated dishes in NBM (lacking growth factors when plated onto laminin and in the presence of bFGF and EGF when plated onto fibronectin) , allowed to attach, and incubated in the presence or in the absence of LPA (complete with any inhibitor used) for 5 days, unless otherwise stated. In some experiments, neurospheres were preincubated with pertussis toxin (PTX) (10 ng/ml) for 18 hours prior to plating onto laminin-coated dishes. For each type of experiment, after attachment, medium was changed every 2nd day. Dilutions of LPA were made in 0.1% fatty acid-free bovine serum albumin (BSA) (final concentration, 0.01% BSA; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). LPA (BIOMOL International LP, Plymouth Meeting, PA, http://www.biomol.com) was used at concentrations up to 10 μM, the LPA1/LPA3 antagonists VPC32183 (Avanti Polar Lipids, Inc., Alabaster, AL, http://www.avantilipids.com) and Ki16425 (Sigma-Aldrich) at 10 μM, U0126 (Promega, Madison, WI, http://www.promega.com) at 30 μM, LY294002 (Promega) at 10 μM, Y27632 (BIOMOL) at 1 μM, PTX (BIOMOL) at 10 ng/ml, H-89 dihydrochloride (Calbiochem, San Diego, http://www.emdbiosciences.com) at 100 nM, and GW9662 (Cayman Chemicals, Ann Arbor, MI, http://www.caymanchem.com) at 1 μM. VPC32183, Ki16425, U0126, LY294002, Y27632, and GW9662 were preincubated for 30–60 minutes prior to LPA.
Neurosphere Formation Assay
Noggin-treated cells were harvested after 14 days using dispase and were further subcultured in suspension in NBM together with fibroblast growth factor (FGF) and EGF (20 ng/ml each) as neurospheres, in the presence or in the absence of LPA (10 μM). Medium was changed every 2nd day. After 11 days, the number of neurospheres was counted in each condition. In some experiments, noggin-treated cells cultivated in the presence or in the absence of LPA (10 μM) for 14 days were harvested as described above and subcultured in suspension in NBM with FGF and EGF; assessment of neurosphere formation was performed after 7 days.
Reverse Transcription-Polymerase Chain Reaction
mRNA was extracted from the hESC cell lines HES-2, HES-3, and HES-4, as well as from the neurospheres derived from the cell lines HES-2, HES-3, and HES-4, using the Dynabeads mRNA direct micro kit (Dynal Biotech, Carlsbad, CA, http://www.invitrogen.com/dynal) following the supplier's protocol. Reverse transcription was performed using Superscript III (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Standard polymerase chain reaction (PCR) was carried out using Taq DNA polymerase (Bioline, London, http://www.bioline.com), gene-specific primers (Geneworks Pty Ltd, Hindmarsh, SA, Australia, http://www.geneworks.com.au), and optimal annealing temperatures (Table 1), with amplification for 30–35 cycles. LPA5 primers were designed using Primer3 (open source, http://frodo.wi.mit.edu/primer3/input.htm). PCR products were resolved by agarose gel electrophoresis. Molecular sizes (in base pairs) were calculated using 1 kilobase plus DNA ladder markers. The amplicons were purified and sequenced. The analysis showed that sequenced PCR products obtained corresponded to those of the expected human cDNA.
Table Table 1.. Sense and antisense primers
NSC and differentiated cells were cultured on glass chamber slides (seeded with the appropriate matrix) prior to fixation with 4% paraformaldehyde or ethanol (Oct-4 and Pax-6). Neurospheres were fixed in 4% paraformaldehyde and embedded in Tissue-Tek O.C.T. Compound (Sakura Finetek, Torrance, CA, http://www.sakura.com). Frozen sections (10 μm) and/or fixed cells were blocked in 10% fetal calf serum-PBS-0.1% Tween 20 and immunostained using antibodies against the following: LPA1 (Cayman Chemicals), EDG4/LPA2 (Abcam, Cambridge, MA, http://www.abcam.com), LPA3 (Cayman Chemicals), LPA4/P2Y9/GPR23 (Abcam), LPA5/GPR92 (Imgenex, San Diego, http://www.imgenex.com), Oct-4 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), Pax-6 (Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/∼dshbwww), glutamate aspartate transporter (GLAST) (Abcam), S100 protein (Abcam), A2B5, glial fibrillary acidic protein (GFAP), and β-tubulin (all from Chemicon, Temecula, CA, http://www.chemicon.com). Cells were then immunostained with the appropriate conjugated secondary antibodies (Alexa Fluor 568 or 488; Molecular Probes, Eugene, OR, http://probes.invitrogen.com). Nuclei were counterstained with Hoechst 33342 (Sigma-Aldrich). Specificity of the staining was verified by the absence of staining in negative controls consisting of the appropriate negative control immunoglobulin fraction (Dako, Glostrup, Denmark, http://www.dako.com) (supplemental online Fig. 2).
Quantification of Neuron-Forming Spheres
Quantification was done by counting the number of spheres from which neuronal outgrowth was observable. In some cases, neurospheres failed to attach, independently of the treatments, and these floating neurospheres were not taken into consideration for quantification.
Dissociated neurospheres plated onto laminin were labeled with 5-bromo-2′-deoxyuridine (BrdU Labeling and Detection Kit; Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com) for the last hour of incubation. Cells were fixed in ethanol/sodium acetate, and cellular DNA was denaturated by 4 N HCl. After pH neutralization, cells were immunostained with anti-5-bromo-2′-deoxyuridine (BrdU)-FLUOS (Roche Diagnostics) followed by goat anti-mouse IgG-Alexa Fluor 568 secondary. Cell nuclei were counterstained with Hoechst 33342. For quantification, cell proliferation was quantified by counting at least 1,900 cells per sample manually.
Quantification of Apoptosis
Cell apoptosis was quantified by measuring numbers of condensed nuclei with Hoechst 33342 and terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) immunocytochemistry. TUNEL analysis was performed using the In Situ Cell Death Detection Kit (Roche Diagnostics) following the manufacturer's instructions. Briefly, dissociated neurospheres plated onto laminin were fixed in 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 before being immunostained with a fluorescein-conjugated TdT enzyme. Cell nuclei were counterstained with Hoechst 33342. Specificity of the staining was verified by the absence of staining in negative controls without the TdT enzyme. Cell apoptosis was quantified by counting at least 1,900 cells per sample manually.
All sets of experiments were performed at least three times in triplicate, unless specified (n refers to the number of independent experiments performed on different cell cultures). All data are expressed as mean ± SEM. Significance of the differences was evaluated using the t test or the one-way analysis of variance followed by the Newman-Keuls test for multiple comparisons. Statistical significance was established at *, p < .05, **, p < .01, and ***, p < .001.
LPA Receptors Are Expressed by hESC-Derived NSC
We previously showed that hESC express the mRNA for LPA1–3 . However, LPA4 and LPA5 have since been identified as LPA receptors, and their presence on hESC was therefore checked. Here we report the expression of LPA4 and LPA5 mRNA in hESC (Fig. 1A). Expression analysis of the LPA receptors within the neurosphere indicates the presence of the mRNA transcripts and the corresponding proteins for LPA1–5 (Fig. 1). LPA1–5 proteins were found to be homogeneously present in the neurosphere (Fig. 1). These data thus demonstrate that the five LPA receptors are present both in undifferentiated hESC and during their differentiation toward neuronal stem/progenitor cells, which suggests that LPA receptors are not differentially regulated upon neural differentiation of hESC.
LPA Does Not Modify Neural Induction
When hESC are cultivated in a serum-containing medium and in the presence of noggin (500 ng/ml) for 14 days, they give rise to NSC, which can be evidenced by the upregulation of the transcription factor Pax-6 . We assessed the effect of LPA on neural induction by incubating hESC with noggin (500 ng/ml), with or without LPA, for 2 weeks. In these conditions, we observed that cells cultivated in the presence of LPA (10 μM) expressed Pax-6, as in the control conditions (Fig. 2A, 2B). Furthermore, when LPA-treated cells were subcultured in suspension in NBM together with bFGF and EGF (20 ng/ml each), in the absence of LPA, they formed neurospheres to the same extent as non-LPA-treated cells (data not shown). These data suggest that LPA does not modify the very early stages of hESC differentiation toward NSC.
LPA Inhibits Neurosphere Formation
We then investigated the effect of LPA on neurosphere formation. For this purpose, noggin-treated cells were incubated in the presence of LPA while being subcultured in suspension in NBM together with bFGF and EGF (20 ng/ml each) for 11–14 days. The neurosphere formation rate was then assessed and compared with that of the cells grown in the absence of LPA. Our data indicate that LPA (10 μM) inhibits the formation of neurospheres. Indeed, 13.47% ± 6.94% of cultures formed neurospheres in the presence of LPA, whereas 48.60% ± 8.15% were observed under control conditions (Fig. 2C–2E; n = 3; *, p < .05). These data thus suggest that LPA inhibits the ability of NSC to form neurospheres, even in the presence of bFGF and EGF stimulation.
LPA Inhibits Neuronal Differentiation
We next assessed the effect of LPA on an additional differentiation step: NSC toward mature cells. When plated onto laminin and cultured with NBM, neurospheres predominantly differentiate into neurons . Although we observed the formation of elongated cells positive for β-tubulin in the control conditions, NSC incubated in the presence of LPA did not differentiate into elongated cells (Fig. 2F–2H; Fig. 3). Indeed, in the presence of LPA (10 μM), we observed very few, if any, β-tubulin positive cells, generally localized within the neurospheres. In most experiments, neurospheres plated in the presence of LPA (10 μM) did not give rise to neuronal cells (Fig. 2F–2H; Fig. 3). This effect is rapid, as it was already observable after 3 days (data not shown). In control conditions (no LPA), 94.64% ± 3.79% of plated spheres gave rise to neurons (neuron-forming spheres), whereas only 24.40% ± 7.28% spheres gave rise to neurons in the presence of LPA (10 μM; n > 3; p < .001), an effect not observed at lower LPA concentrations (1 μM LPA, 87.50% ± 12.50%; 0.1 μM LPA, 72.92% ± 17.80% of neuron-forming spheres; n > 3; p > .05; Fig. 2H). LPA's effect is reversible, as addition of LPA to the culture medium for 3 days, followed by its removal, was followed by the ability of neurospheres to form neurons (n = 3; supplemental online Fig. 1). Furthermore, LPA did not appear to have an effect on differentiated neurons. Indeed, when LPA (10 μM) was added to differentiated neurons (generated from neurospheres plated onto laminin and cultivated in NBM for 5 days), we still observed the presence of neurons after 3–7 days of LPA-treatment (supplemental online Fig. 1).
Our data thus indicate that LPA exerts a potent effect on neuronal differentiation. This effect could be due to an inhibition of differentiation, to an effect on the cell cycle, or to an increase in cell death. To address these possibilities, we subjected neurospheres plated on laminin to LPA (10 μM) for 18 hours and compared proliferation and apoptosis with the control conditions (no LPA). BrdU incorporation indicated that LPA does not affect DNA synthesis (control, 21.07% ± 1.16%; 10 μM LPA, 21.99% ± 0.86% of BrdU incorporation; n = 3; p > .05; Fig. 3E), whereas quantification of condensed nuclei with Hoechst 33342 and TUNEL analysis (data not shown) revealed that LPA does not modify apoptosis of NSC (control, 13.23% ± 1.01%; 10 μM LPA, 11.23% ± 2.47% of condensed nuclei; n = 3; p > .05; Fig. 3F).
LPA Maintains Astrocytic Differentiation
Plating neurospheres onto fibronectin in the presence of bFGF and EGF encourages the differentiation of the NSC toward glia but maintains some neuronal differentiation (Fig. 2I). Indeed, in the control conditions, NSC differentiated into both astrocytes and neurons, as assessed by the presence of the astrocytic markers A2B5 and GLAST, as well as of the neuronal marker β-tubulin (Fig. 4). In the presence of LPA (10 μM) for 5 days, cells also expressed A2B5 and GLAST (Fig. 4), as well as GFAP, an additional astrocytic marker not expressed in the control conditions at this time point (Fig. 4). As observed on laminin, LPA-treated neurospheres showed very little, if any, differentiation into neurons, as indicated by the absence of β-tubulin staining (Fig. 4). Our data shows that in neurospheres plated on fibronectin, LPA maintains the differentiation of NSC toward astrocytic lineages and inhibits neuronal differentiation, whereas in the control conditions (absence of LPA), a mixed population of astrocytes and neurons was observed. These data thus further confirm the very potent role of LPA on NSC differentiation through inhibition of neuronal differentiation.
Signaling Pathways Involved in LPA's Inhibition of Differentiation
We next assessed the signaling pathways involved in the LPA-induced inhibition of neuronal differentiation. For this purpose, NSC plated onto laminin and pretreated with the LPA1/LPA3 antagonist VPC32183 (10 μM) were incubated in the presence of LPA (10 μM) for 5 days. Under these conditions, we observed that the inhibitory effect of LPA on neuronal differentiation was partially blocked, with the formation of neurons from 58.33% ± 16.67% (n > 3; p < .05) of neurospheres (Fig. 5A; supplemental online Fig. 1). Similar trends were obtained with another LPA1/LPA3 antagonist, Ki16425 (10 μM, 46.67% ± 13.33% of neuron-forming neurospheres; n > 3; Fig. 5B; supplemental online Fig. 1). Thus, although the blockage of LPA1/LPA3 inhibited LPA's effect on NSC differentiation, this effect was not totally abolished, suggesting that other LPA receptors may also be involved in the inhibition of neuronal differentiation. Due to the lack of available tools to further study the receptor specificity for this biological effect, we cannot assess whether LPA2, LPA4, and LPA5 are contributing to this phenomenon. Nonetheless, our data suggest that LPA inhibits neuronal differentiation at least through the activation of LPA1 and/or LPA3. Activation of PPARγ has also been shown to modulate murine NSC proliferation and differentiation . As PPARγ is considered an intracellular LPA receptor , we tested the effect of its specific and irreversible antagonist GW9662 (1 μM ) on LPA-induced inhibition of neuronal differentiation. In its presence, the effect of LPA on neuronal differentiation was not modified, suggesting that LPA does not act through PPARγ to inhibit neuronal differentiation of hESC-derived NSC (Fig. 5C; supplemental online Fig. 1).
LPA receptors are coupled to different G proteins. When NSC were incubated with PTX (10 ng/ml), which ADP-ribosylates αi proteins, LPA's effect was maintained (33.33% ± 19.25% of neuron-forming neurospheres; n = 3), suggesting that αi proteins are not involved in the inhibition of neuronal differentiation by LPA (Fig. 5D; supplemental online Fig. 1). As LPA4 and LPA5 have been shown to activate cAMP production in some cells types [19, 20], we tested the effect of the specific protein kinase A (PKA) inhibitor H-89 (1 μM). This inhibitor had no effect by itself on neuron formation or on LPA's inhibition of neuronal differentiation (Fig. 5E; supplemental online Fig. 1), indicating that the cAMP/PKA pathway is not involved in neuronal differentiation from hESC-derived neurospheres or in LPA's inhibition of neuronal differentiation.
In many cell types, LPA activates the mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase-1 and 2 (Erk1/2) and phosphatidylinositol 3-kinase (PI3K)/Akt pathways (reviewed in ). When cells were treated with the MAPK kinase inhibitor U0126 (30 μM), neuronal differentiation was maintained in the control conditions, and LPA's effect was not blocked (Fig. 5F; n > 3; supplemental online Fig. 1), thus suggesting that LPA-induced inhibition of neuronal differentiation is Erk-independent. However, inhibition of PI3K by its specific inhibitor LY294002 (10 μM) reduced LPA's effect on neuronal differentiation (10 μM, 52.78% ± 18.47% of neuron-forming neurospheres; n > 3; p < .05), suggesting the involvement of the PI3K/Akt pathway (Fig. 5G; supplemental online Fig. 1). In addition, when the Rho pathway was inhibited by the specific p160 ROCK inhibitor Y27632 (1 μM), which alone had no marked effect, LPA's inhibition of neuronal outgrowth was partially reversed (50.00% ± 6.80% of neuron-forming neurospheres; n = 3; p < .05; Fig. 5H; supplemental online Fig. 1). The activation of the Rho pathway by LPA is generally associated with G12 proteins, suggesting that LPA could activate the Rho pathway, in a G12-dependent manner, which would lead to the inhibition of neuronal differentiation. Given that VPC32183 and Ki16425 inhibit only LPA1 and LPA3 and only partially block LPA inhibition of neuronal differentiation, LPA2, LPA4, and LPA5, which have all been described as activating G12/13 proteins [2, 21, 22], could thus inhibit neuronal differentiation through this pathway. Interestingly, coincubation of VPC32183 and Y27632 with LPA (10 μM) completely abolished the effect of LPA (100.00% ± 0.00% of neuron-forming neurospheres; n = 3; p < .001; Fig. 5I; supplemental online Fig. 1), thus suggesting that the two signaling mechanisms are additive and mediated through at least two different receptors. The additive effects of the two inhibitors on the antidifferentiation effect of LPA also suggest that one or more pathways (αi independent), activated by LPA1 and/or LPA3 other than the Rho pathway are modulating LPA's effect. The addition of LY294002 and Y27632 to neurospheres also abolished LPA's inhibition of neuronal differentiation (91.67% ± 8.33% of neuron-forming neurospheres; n = 3; p < .01; Fig. 5J; supplemental online Fig. 1). Overall, the data indicate that the combined activation of both the PI3K/Akt pathway and the Rho/ROCK pathway by LPA is responsible for the inhibition of neuronal differentiation from hESC.
Endogenous NSC are located in areas of neurogenesis within the central nervous system. In vitro, these cells are maintained as floating neurospheres, and in vivo in the adult, stem/progenitor cells can migrate to sites of neural injury . LPA receptors are expressed in neural stem and progenitor cells, and LPA has been described as able to induce morphological rearrangements, proliferation, and differentiation of neural stem and progenitor cells (reviewed in ). However, LPA receptor expression differs depending on the tissue source, and the effects of LPA on neural stem/progenitor cells seem to be dependent on the cell region of origin, the species, and the developmental stage (reviewed in ). Indeed, LPA induces proliferation of murine cortical neuroblasts in vitro , but not in vivo , whereas it has no proliferative effect on hippocampal neural progenitor cells of the embryonic rat . Furthermore, LPA induces neural differentiation of murine cortical neuroblasts, early cortical neurons, and hippocampal progenitor cells [10, 26, 27]. In cortical neuroblasts and progenitors, LPA induces the formation of retraction fibers, cell rounding, and neurite retraction [10, 24, 28, –30]. In the rat hippocampus, LPA induces neural progenitor cell aggregation, in a mechanism dependent on the activation of the Rho pathway . In vitro, its role in neurosphere formation is somewhat controversial. Indeed LPA is reported to inhibit bFGF-induced neurosphere formation by increasing neuronal differentiation of cortical neural progenitor/stem cells  on the one hand, and to stimulate neurosphere formation from neural progenitor/stem cells of forebrain origin on the other . Here we show results for hESC-derived neurospheres different from the ones observed in the murine systems. Indeed, in our conditions, LPA does not modify the differentiation of hESC into Pax-6-positive cells (which are considered NSC) but inhibits their maintenance as neurospheres (Fig. 6A). Interestingly, Svetlov et al.  observed that the clonal generation of murine NSC is mediated by LPA1, whereas the growth and proliferation of the neurospheres requires LPA1 and LPA3, again suggesting a discrepancy between species. Furthermore, our data show that in human cells, LPA specifically inhibits the differentiation of NSC toward neurons, whereas it maintains the differentiation of NSC toward astrocytes (Fig. 6A). These effects were observed on both laminin and fibronectin substrates, which normally bias the differentiation of hESC-derived NSC toward neurons and glia respectively . As LPA's inhibition of neuronal differentiation is observed under both conditions, our data suggest that LPA's effect is independent of matrix type. LPA's inhibition of neuronal differentiation is reversible, as its removal from the culture medium was followed by neuronal differentiation.
Our data also indicate that LPA's inhibition of neuronal differentiation is receptor-mediated and PI3K- and Rho-dependent (Fig. 6B). Signaling by LPA is complex because of the existence of different receptor subtypes, as well as the binding to multiple G proteins. Indeed, LPA1–3 are coupled to Gi and Gq proteins, as well as G12/13 for LPA1 and LPA2, whereas LPA4 and LPA5 are coupled to Gq and G12/13 [2, 21, 22]. Interestingly, LPA4 and LPA5 might also activate adenylate cyclase through coupling to Gs proteins [19, 20], unlike the other LPA receptors. All five LPA receptors are found to be expressed in NSC derived from hESC. The antagonist VPC32183 partially inhibits LPA's effect, suggesting the involvement of LPA1/LPA3 in LPA's inhibition of neuronal differentiation, as well as that of other LPA receptors. The lack of effect of PTX may reflect that LPA acts independently of αi, through αq, α12, and/or βγ subunits of G proteins, to inhibit neuronal differentiation. Furthermore, as LPA4 and LPA5 seem not to be coupled to Gi proteins, the lack of effect of PTX might also be reflective of their involvement in LPA's inhibition of neuronal differentiation, which is in accordance with the observed partial effect of VPC32183.
Blockage of the PI3K/Akt pathway reduced LPA's inhibition of neuronal differentiation. Furthermore, the p160 ROCK inhibitor Y27632 partially reversed LPA's effect, suggesting that the activation of Rho/ROCK pathway by LPA is involved in the inhibition of neuronal differentiation. Thus, LPA, at least through the activation of LPA1 and/or LPA3, inhibits neuronal differentiation, in a mechanism mediated by the PI3K/Akt and the Rho/ROCK pathways (Fig. 6B). To some extent, these two pathways must be independent of each other, as the coincubation of their specific inhibitors abolishes LPA's effect, a phenomenon not obtained with the sole application of one inhibitor or the other (Fig. 6B). As PTX did not modify LPA's effect on neuronal differentiation, our data suggest that in human NSC, these two pathways are αi-independent. LPA's activation of the PI3K pathway in a PTX-insensitive manner could be mediated by βγ subunits of G proteins, whereas LPA's activation of the Rho/ROCK pathway is likely to be G12/13-mediated (Fig. 6B), although it could also involve Gq (reviewed in [2, 32]).
Astrogliosis, neuronal death, and nonregeneration of neurons are essential events following injury, trauma, or hemorrhage to the nervous system. During such events, which damage the blood-brain barrier, levels of LPA within the nervous system are hypothesized to increase to 10 μM (reviewed in [33, 34]). LPA has been shown to stimulate astrocytic proliferation [7, 34], and depending on its concentration, it has also been described as promoting death of hippocampal neurons, by apoptosis (1 μM) or by necrosis (10 μM) . Moreover, LPA has recently been described as cytotoxic to the neuromicrovascular endothelium . Furthermore, our data show that a pathological concentration of LPA (10 μM) inhibits neuronal differentiation of NSC of human origin, whereas lower concentrations do not, which suggests that the presence of high levels of LPA within the central nervous system following an injury might inhibit endogenous neuronal regeneration (by inhibiting the differentiation of NSC toward neurons) and maintains gliogenesis. Our data, together with the literature, strongly support the idea that LPA is a key player on the outcome of nerve damage and/or repair following injuries. Thus, modulating LPA signaling may have a significant impact in injury within the nervous system, allowing new avenues for potential therapeutic approaches.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This study was supported by The University of Melbourne, the National Health and Medical Research Council of Australia (NHMRC 454723), and Friedreich Ataxia Research Association Australia. We are grateful to Dr. Jane Ward (Department of Pharmacology, University of Melbourne) for providing the PPARγ antagonist GW9662.