Liver X Receptor α Is Involved in Counteracting Mechanical Allodynia by Inhibiting Neuroinflammation in the Spinal Dorsal Horn

NEUROPATHIC pain is a debilitating pain state, often caused by lesion or disease of the peripheral or central nervous system.¹  The management of neuropathic pain is still challenging, because it does not respond to most currently used analgesic drugs.

Liver X receptors (LXRs) are considered as intracellular cholesterol sensors, for which the activation leads to decreased plasma cholesterol.²  Two known different isoforms, LXRα and LXRβ, have been recognized. LXRα is mainly expressed in the liver, whereas LXRβ is broadly expressed.^3,4  LXRs ligands, including oxysterols GW3965 and T0901317, decrease lipopolysaccharide-induced expression of proinflammatory genes in activated macrophages.⁵  T0901317 has recently been used to reduce neuroinflammation after spinal cord injury.⁶  However, whether LXRα plays a role in neuropathic pain is still unknown.

In response to nervous tissue damage, spinal microglia and astrocytes were activated, resulting in overproduction of inflammatory cytokines, such as tumor necrosis factor α (TNF-α), interleukin 1β (IL-1β), interleukin 6, and thus contributing to neuropathic pain.^7–11  In contrast to proinflammatory cytokines, the antiinflammatory cytokine interleukin 10 (IL-10) has been shown to be gradually reduced as neuropathic pain develops.¹²  Intrathecal injection of plasmid DNA encoding IL-10 or recombinant rat IL-10 could suppress neuropathic pain symptoms.^13,14 

Therefore, we hypothesized that activation of LXRα could inhibit the neuroinflammatory responses in the spinal cord and thereby alleviate neuropathic pain. Here we first studied the time course of LXRα expression in the lumbar spinal dorsal horn after spared nerve injury and then investigated the effects of LXR agonists on the development and maintenance of mechanical allodynia. The mechanisms by which LXR agonists inhibit mechanical allodynia were also explored.

Male Sprague–Dawley rats weighing 180 to 200 g and LXRα knockout (LXRα [–/–]) C57 and wild-type male mice weighing 25 to 31g were used in this study. LXRα (–/–) mice were purchased from Jackson Laboratory (USA). Animals were group housed, with free access to standard rodent chow and water. The animals were exposed to a 12-h light/dark cycle. All of the studies were approved by the Sun Yat-Sen University Animal Care and Use Committee (Guangzhou, Guangdong, People’s Republic of China) and were carried out in accordance with the guidelines of the National Institutes of Health on animal care and the ethical guidelines for investigation of experimental pain in conscious animals.¹⁵  All of the animals were randomly assigned to different treatment groups.

In total, 424 rats and 171 mice were used in the present study; among them, three rats and one mouse died after surgery, two rats were euthanized due to excessive suffering from pain. Thirty-five rats and 17 mice were excluded because no mechanical allodynia was induced after spared nerve injury surgery or because catheterization was unsuccessful.

Two synthetic LXR agonists, T0901317 and GW3965, were used in the current study. The EC₅₀ of T0901317, which activates both LXRα and LXRβ, is approximately 20 nM.¹⁶  GW3965 has a greater affinity toward LXRβ (EC₅₀ = 30 nM) than LXRα (EC₅₀ = 190 nM).¹⁷  T0901317 was purchased from Cayman Chemical (USA). GW3965 and aminoglutethimide (R(+)-p-aminoglutethimide) were purchased from Sigma-Aldrich (USA). All of the drugs were dissolved in dimethyl sulfoxide. T0901317 and GW3965 were injected intrathecally with polyethylene 10, for which the tip was put on the spinal lumbar enlargement level 1 week before. Whether polyethylene 10 was correctly placed in intervertebral space was confirmed by paralysis in the bilateral hind limb after 2% lidocaine injection (7 µl for rats and 3 µl for mice) through the tube within 30 s. Drugs or vehicle was administered slowly in volumes of 10 µl (for rats) or 3 µl (for mice).

Spared nerve injury was carried out following the previous method.¹⁸  Briefly, rats and mice were anesthetized with isoflurane (1.5 to 2.5%) and a mixture of 30% N₂O and 70% O₂. The sciatic nerve of the left leg was exposed into common peroneal, tibial nerve, and sural nerve, and then the common peroneal and tibial nerves were ligated and cut, whereas the sural nerve was left intact. For the animals in the sham group, the sciatic nerve was only exposed but not ligated or cut.

Mechanical sensitivity of the animals before and after spared nerve injury was assessed with the up–down method following the previously described method.¹⁹  Briefly, after habituation for 10 to 15 min (rats) or 45 min to 1 h (mice), a series of filaments with varying forces (0.4, 0.6, 1, 1.4, 2, 4, 6, 8, and 15 g for rats and 0.04, 0.07, 0.16, 0.4, 0.6, 1, 1.4, and 2 g for mice) were applied to the plantar surface of the hind paw. Each stimulus consisted of a maximum 6-s application of the filament, and quick withdrawal in response to the stimulus was considered a positive response. The 50% paw withdrawal thresholds were then calculated.

To test whether T0901317 could affect the mechanical sensitivity in the sham-operated rats, the incidence of paw withdrawal to four different forces of filaments (2, 6, 15, and 26 g) was measured according to previous work.²⁰  Each filament was applied once per second to the plantar surface eight times. Ten trials were performed on each hind paw. The percentage of positive trials was recorded. The behavioral test was performed in a blinded fashion.

Animals were anesthetized with sodium pentobarbital (50 mg/kg) after defined survival time. If the animals received T0901317 or GW3965 treatment, the tissues were collected 2 days after injection. The ipsilateral spinal dorsal horn was dissected and homogenized in 15 mM Tris buffer (pH 7.6; 250 mM sucrose, 1 mM magnesium chloride, 1 mM dithiothreitol, 2.5 mM EDTA, 1 mM EGTA, 50 mM sodium fluoride, 10 μg/ml leupeptin, 1.25 μg/ml pepstatin, 2.5 μg/ml aprotin, 2 mM sodium pyrophosphate, 0.1 mM sodium orthovanadate, 0.5 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture; Roche Molecular Biochemicals, Switzerland). The samples were sonicated and then centrifuged at 13,000g for 15 min. The isolated proteins were separated by sodium dodecyl sulphate–polyacrylamide gel electrophoresis and transferred onto a polyvinylidene fluoride membrane (Bio-Rad Laboratories, Inc., USA). After incubation with blocking buffer for 1 h at room temperature, the blots were incubated with primary antibody against IL-1β (1:200, Abcam; United Kingdom), IL-10 antibody (1:500, Abcam), or TNFα (1:200; Santa Cruz Biotechnology, Inc., USA) overnight at 4°C, followed by incubation with secondary antibody horseradish peroxidase–conjugated rabbit antigoat, goat antimouse, or goat antirabbit immunoglobulin G (1:8000, 1:5000, or 1:8000; Kirkegaard & Perry Laboratories, Inc., USA) for 2 h at room temperature. The blots were developed with enhanced chemiluminescence (Clarity Western electrochemiluminescence substrate, Bio-Rad Laboratories, Inc.) and detected by a Tanon 5200 imager (Tanon Science & Technology Co., Ltd., China). The intensities of the blots were quantified by Tanon MP (Tanon Science & Technology Co., Ltd., China) software and normalized against a loading control (β actin).

Immunohistochemistry was performed following our previous study.²¹  Briefly, rats or mice were randomly chosen from a cohort used for behavioral studies. If the animals received T0901317 or GW3965 treatment, the tissues were collected 2 days after injection. After defined survival times, the animals were anesthetized and perfused with saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer. The lumbar spinal cord segments were removed and postfixed in the 4% paraformaldehyde for 3 h and then replaced with 30% sucrose overnight. Cryostat sections (25 μm) were cut in a cryostat (Leica CM1900, Leica Biosystems, Germany) and processed for immunohistochemistry. For double immunofluorescence staining, primary antibody for LXRα (1:200; Abcam) was incubated together with antibody for mouse monoclonal neuronal-specific nuclear protein (neuronal marker, 1:200; Millipore Bioscience Research Reagents, USA), glial fibrillary acidic protein ([GFAP] astrocyte marker, 1:1000; Cell Signaling Technology, USA), or goat polyclonal antiionized calcium-binding adaptor molecule 1 ([Iba1] microglia marker, 1:800; Abcam). After incubation for more than two nights at 4°C, the sections were incubated with cy3-conjugated (1:500; Jackson Immuno-Research, USA) and fluorescein isothiocyanate–conjugated secondary antibodies (1:400; Jackson ImmunoResearch) for 1 h at room temperature. For negative control sections, the above procedures were followed, except the primary antibody was omitted. The stained sections were then examined with a Leica fluorescence microscope (Leica Biosystems), and images were captured with a Leica DFC350 FX camera (Leica Camera AG). For quantification of immunofluorescence staining, the area of LXRα- immunoreactivity per section was measured in the spinal dorsal horn (laminae I to V) using a Leica Qwin V3 digital-image processing system (Leica Camera AG).

Data were expressed as mean ± SD and analyzed with SPSS 15.0 (SPSS Inc., USA). The data of behavioral tests were analyzed using repeated-measures two-way ANOVA with the Tukey post hoc test. Immunohistochemistry and Western blot data were analyzed by one-way ANOVA followed by the Tukey post hoc test. P < 0.05 was considered statistically significant.

Compared with the sham group (fig. 1A, reference value of 100), the expression of LXRα in the ipsilateral (fig. 1A) and contralateral (fig. 1B) lumbar spinal dorsal horns increased significantly from day 1 to day 14 after the surgery. Immunohistochemistry showed that, compared with the sham group (fig. 1C), the expression of LXRα was increased 1 (fig. 1D) and 7 days (fig. 1E) after spared nerve injury. LXRα was increased in the bilateral spinal dorsal horn 7 days after spared nerve injury (fig. 1F). In the negative control group (fig. 1G), no signals were detected. Compared with the sham group, the areas of LXRα immunoreactivity were all increased in laminae I to II, III to IV, and V in the ipsilateral spinal dorsal horn (fig. 1H).

In the gray matter, LXRα was located mainly in neurons (fig. 2A) but not in astrocytes (fig. 2B) or microglia (fig. 2C) 7 days after sham operation, and 7 days after spared nerve injury, LXRα was still located mainly in neurons (fig. 2D) but not in astrocytes (fig. 2E) or microglia (fig. 2F). In the white matter, LXRα was located mainly in neurite outgrowth inhibitor A–labeled oligodendrocytes (fig. 2G) but not in astrocytes (fig. 2H) and microglia (fig. 2I) in sham-operated rats, whereas 7 days after spared nerve injury, LXRα was located mainly in oligodendrocytes (fig. 2J) and to a lesser extent in astrocytes (fig. 2K) and microglia (fig. 2L).

We found that LXRα was also up-regulated in lumbar 5 dorsal root ganglia after spared nerve injury (Supplemental Digital Content 1A, http://links.lww.com/ALN/B469). LXRα was expressed in neurofilament protein 200 and isolectin B₄–labeled neurons but not in GFAP-labeled satellite cells (Supplemental Digital Content 1B, http://links.lww.com/ALN/B469).

Additional experiments were performed to test whether the activation of LXRs could affect the development and maintenance of mechanical allodynia. We found that intrathecal injection of T0901317, started 30 min before spared nerve injury, and thereafter once daily for 2 days, blocked mechanical allodynia dose dependently. T0901317 at 19.2 µg completely blocked the decrease in the bilateral 50% paw withdrawal threshold on the ipsilateral hind paw, and the effect persisted until the end of the experiment. T0901317 at 4.8 µg produced a weaker antiallodynia effect (fig. 3A). In contrast, in the dimethyl sulfoxide–treated rats, 50% paw withdrawal threshold decreased in the bilateral hind paw 4 days after spared nerve injury, which persisted for at least 21 days (fig. 3, A and B). A single intrathecal injection of T0901317 (19.2 µg) 7 days after spared nerve injury reversed the decrease of the 50% paw withdrawal threshold in the bilateral side (fig. 3C). T0901317 could induce expression not only of LXR target genes but also of the nuclear xenobiotic receptor pregnane X receptor target genes, a characteristic that more specific LXR agonists such as GW3965 do not have.^17,22  In the present study, to test whether the effect of T0901317 was specific, GW3965 was also injected intrathecally into spared nerve injury rats. As shown in figure 3D, GW3965 (6.2 µg) could also attenuate the mechanical allodynia when applied 7 days after spared nerve injury. A 50% paw withdrawal threshold was also increased for 1 day when 19.2 µg T0901317 was administrated 21 days after spared nerve injury (fig. 3E). For the paw withdrawal incidence, no significant difference between the dimethyl sulfoxide- and T0901317-treated groups (19.2 µg) was found (fig. 3F).

Intrathecal injection of T0901317 (4.4 µg) significantly increased the bilateral 50% paw withdrawal threshold in wild-type mice 1 to 3 days after the injection, whereas the same dose of T0901317 had no effect on the decrease of 50% paw withdrawal threshold in these LXRα (–/–) mice at the same time points (fig. 4, A and B). The analgesic effect of GW3965 (5.6 µg) was also completely abolished in LXRα (–/–) mice (fig. 4, C and D). Activation of astrocyte and microglia in the spinal cord was enhanced in LXRα (–/–) mice after spared nerve injury, and GW3965 inhibited activation of glial cells, decreased the expression of IL-1β and TNF-α, and increased the expression of IL-10 in the spinal dorsal horn induced by spared nerve injury in wild-type but not in LXRα (–/–) mice.

We further investigated whether LXRα modulated astrocytes and microglial activation (fig. 5). We found that, compared with the wild-type mice (fig. 5, Aa and Ag), 9 days after spared nerve injury, astrocytes (fig. 5Ac) and microglia (fig. 5Ai) in the ipsilateral lumbar spinal dorsal horn were highly activated, because the expression of GFAP and Iba1 was increased and the size of the cells was enlarged. Knockout of LXRα alone did not induce activation of astrocyte (fig. 5Ab) and microglia (fig. 5Ah); however, knockout of LXRα activated more astrocytes in the spinal dorsal horn 9 days after spared nerve injury, as shown in figure 5Ad and figure 5Aj. GW3965 could only inhibit the activation of astrocytes (fig. 5Ae) and microglia (fig. 5Ak) in the wild-type mice but not in the LXRα (–/–) mice, because we can see that the shapes and numbers of astrocytes and microglia in the GW3965-treated (LXRα [–/–]) group were comparable with those in the dimethyl sulfoxide–treated LXRα (–/–) group (fig. 5, Af and Al). Furthermore, treatment of GW3965 could decrease the expression of IL-1β (fig. 5D; 158 ± 74) and TNF-α (fig. 5E; 119 ± 57) and increase the expression of IL-10 (fig. 5F) in the wild-type mice that received spared nerve injury (215 ± 69 for IL-1β; 196 ± 48 for TNF-α; and 63 ± 22 for IL-10) but not in the LXRα (–/–) mice that received spared nerve injury, 9 days after spared nerve injury, the expression of IL-1β, but not IL-10 and TNF-α was increased in LXRα (–/–) mice, compared to the wild-type mice.

Intrathecal injection of T0901317 twenty-one days after spared nerve injury did not affect the activation of microglia and astrocytes (Supplemental Digital Content 2, http://links.lww.com/ALN/B470). Compared with the vehicle-treated group (Supplemental Digital Content 2A, http://links.lww.com/ALN/B470), the positive area of Iba1 (Supplemental Digital Content 2B, http://links.lww.com/ALN/B470) and GFAP (Supplemental Digital Content 2E, http://links.lww.com/ALN/B470) 1 h after T0901317 treatment was not changed. The same results were observed 2 h after T0901317 treatment (Supplemental Digital Content 2C and 2F, http://links.lww.com/ALN/B470).

It has been shown that LXRs function via the promotion of neurosteroidogenesis.²³  Here first we tested whether endogenous neurosteroid played a role in pain hypersensitivity and neuroinflammation after spared nerve injury. As shown in Supplemental Digital Content 3 (http://links.lww.com/ALN/B471), intraperitoneal injection of neurosteroid synthesis inhibitor aminoglutethimide (10 min after spared nerve injury and once every third day) could facilitate the development of mechanical allodynia but did not affect the maintenance of mechanical allodynia. In aminoglutethimide-treated spared nerve injury rats, mechanical allodynia in the ipsilateral hind paw developed 1 day after spared nerve injury, whereas in the vehicle-treatment group, mechanical allodynia did not develop until 2 days after spared nerve injury.

We next tested whether neurosteroidogenesis was involved in the inhibitory effect of LXR agonists on mechanical allodynia. As shown in figure 6A, expression of cytochrome P450 side chain cleavage ([P450scc] the enzyme responsible of the conversion of cholesterol into pregnenolone) was significantly increased 7 days after spared nerve injury and remained at a significantly high level at day 14, whereas steroidogenic acute regulatory protein ([StAR] molecules involved in cholesterol shuttling into the mitochondria) was increased at 4 days after spared nerve injury. The expression of P450scc was not changed, whereas the expression of StAR was up-regulated (fig. 6B) after T0901317 treatment. The expression of StAR and P450scc could only be up-regulated 7 days after spared nerve injury in the wild-type mice but not in LXRα (–/–) mice (fig. 6C). Pretreatment with aminoglutethimide (30 min before T0901317) could completely block the effect of T0901317 on the decrease of the 50% paw withdrawal threshold induced by spared nerve injury (fig. 6D). After T0901317 treatment, the expression of StAR was located mainly in neurons (fig. 6E) and microglia (fig. 6F) in the spinal gray matter, but not in astrocytes (fig. 6G), whereas in the white matter, StAR was located mainly in oligodendrocytes (fig. 6H) and microglia (fig. 6I) but not in astrocytes (fig. 6J). A similar expression pattern of StAR was observed from rats that received spared nerve injury surgery 9 days before (data not shown).

T0901317 decreased the upregulation of LXRα induced by spared nerve injury (fig. 7). Again, the effect was prevented by aminoglutethimide (fig. 7). T0901317 alone had no effect on the expression of LXRα in sham rats.

Expression of both GFAP and Iba1 in the ipsilateral lumbar spinal dorsal horn (fig. 8A) was significantly higher at day 9 in dimethyl sulfoxide–treated spared nerve injury rats than that in dimethyl sulfoxide–treated sham rats. In addition, the expression of GFAP and Iba1 in aminoglutethimide- and T0901317-treated spared nerve injury rats was higher than that of T0901317-treated spared nerve injury rats, and there was no difference from the dimethyl sulfoxide–treated spared nerve injury group. T0901317 alone had no effect on the expression of GFAP and Iba1 in sham rats.

In dimethyl sulfoxide–treated spared nerve injury rats (9 days after spared nerve injury), the expression of IL-1β was increased, whereas IL-10 was decreased in the ipsilateral spinal dorsal horn. The increased IL-1β and decreased IL-10 produced by spared nerve injury were suppressed by T0901317. The effect of T0901317 on IL-1β (fig. 8B) and IL-10 (fig. 8B) was completely blocked by aminoglutethimide. After T0901317 treatment, the expression of IL-1β was located mainly in neurons and astrocytes (fig. 8, C and D) but not in microglia (fig. 8E) in spinal gray matter. In the white matter, IL-1β was located mainly in oligodendrocytes (fig. 8F), with a lesser extent in astrocytes (fig. 8G) and microglia (fig. 8H). Only neurons (fig. 8I), but not astrocytes (fig. 8J) or microglia (fig. 8K), express IL-10 in the gray matter, whereas in the white matter, IL-10 was expressed mainly in oligodendrocytes (fig. 8L) but not in astrocytes (fig. 8M) or in microglia (fig. 8N). The expression pattern of IL-1β and IL-10 was consistent with that obtained from rats that received spared nerve injury surgery 9 days before (data not shown).

In the present study, we showed that LXRα was up-regulated in neurons and oligodendrocytes in the spinal dorsal horn after spared nerve injury, and intrathecal injection of LXR agonists GW3965 or T0901317 could alleviate mechanical allodynia. GW3965 could inhibit the activation of microglia and astrocytes and rebalance the expression of proinflammatory and antiinflammatory cytokines. All of these above effects of LXR agonists were abolished by LXRα mutation. Furthermore, more glial cells were activated and more IL-1β was produced in the spinal dorsal horn after spared nerve injury in LXRα (–/–) mice. Aminoglutethimide, a neurosteroid synthesis inhibitor, could completely block the effects of T0901317 on the expression of LXRα, the activation of microglia and astrocytes, the up-regulation of IL-1β, and the down-regulation of IL-10 induced by spared nerve injury. Moreover, after T0901317 treatment, the expressions of StAR, IL-1β, and IL-10 were all located mainly in neurons and oligodendrocytes, indicating that LXR agonists can exert a stimulating effect by activating LXRα located in both neurons and oligodendrocytes and thus promoting the synthesis of neurosteroids to counteract neuropathic pain.

Our previous work had shown that translocator protein (18kDa) was up-regulated in spinal dorsal horn after lumbar 5 spinal nerve ligation and that a translocator protein agonist could reverse neuropathic pain behaviors, including mechanical allodynia and thermal hyperalgesia.²⁴  Similarly, here we showed that LXRα was up-regulated in the bilateral spinal dorsal horn after spared nerve injury and that a single dose of LXR agonists (either T0901317 or GW3965) inhibited mechanical allodynia. The inhibitory effect of LXR agonists was abolished by LXRα mutation. Moreover, T0901317 could decrease the up-regulated LXRα while inhibiting mechanical allodynia, and the effect of T0901317 on LXRα was again abrogated by aminoglutethimide. These results suggest that the function of LXRα is in trying to inhibit neuropathic pain. Once pain was alleviated by LXRα agonists, the expression of LXRα would decrease. After aminoglutethimide treatment, because the animals kept experiencing pain, LXRα overexpression was kept there to counteract the pain state. This also suggests that, similar to translocator protein, LXRα acts as a self-protecting mechanism in animals experiencing pain. Previous studies have demonstrated the potent neuroprotective effect of LXRs in many disease models. For example, it has been reported that LXRs protect against a loss of dopaminergic neurons and of dopaminergic fibers projecting to the striatum,²⁵  that LXRα reduces cardiac hypertrophy by improving glucose uptake and use,²⁶  and that activation of LXRs protects N-methyl-d-aspartate–induced inner retinal damage.²⁷  It was also reported that LXR and its target gene adenosine triphosphate–binding cassette transporters could modulate diabetic retinopathy outcome in streptozotocin-induced diabetes mellitus²⁸  and in experimental autoimmune uveitis.²⁹  As a key sensor of cholesterol, LXRs react to the increased cellular level of cholesterol, leading to transcription of a series of genes and thus protection for the cells. Therefore, it is possible that cholesterol overload is the common reason for all the above diseases, and LXRs are the common target to ameliorate these diseases. The long-lasting effect of a single injection of an LXR agonist suggests that promoting the function of LXRα is an effective way to counteract neuropathic pain. Additional experiments should be performed to test whether LXRβ plays a role in neuropathic pain.

In the present study, we found that in the lumbar spinal cords of sham-operated animals from the wild-type group, the expression of IL-10, IL-1β, and TNF-α was at the same level as from LXRα (–/–) mice, and these animals are not in pain (data not shown). These results, together with the result showing that T0901317 had no effect on the paw withdrawal incidence in the sham-operated rats, suggest that LXRαs do not exert effects under normal conditions. Moreover, the expression of IL-1β was increased after spared nerve injury in LXRα (–/–) mice compared with the wild-type mice, whereas the expression of IL-10 and TNF-α was not changed. These results suggest that, under neuropathic pain conditions, LXRα mutation only affects expression of part of the cytokines.

Activation of the LXRs may benefit several neurologic disorders by modulating functions of neurons directly or indirectly. LXR activation improves axonal regeneration in patients with stroke,³⁰  protects dopaminergic neurons in Parkinson disease,²⁵  and reduces the loss of cholinergic neurons in Alzheimer disease.³¹  The LXR ligand could reduce markers of neuroinflammation, such as nitric oxide synthase, and reverse microglia from activated states to resting states.³²  In accordance with these studies, we found that T0901317 inhibited the activation of glial cells in the spinal dorsal horn. In addition, the expression of TNF-α and IL-1β was down-regulated and IL-10 was up-regulated by T0901317. By rebalancing the expression of proinflammatory and antiinflammatory cytokines, T0901317 may optimize the environment in which the neurons live. More importantly, the inhibitory effect of the LXR agonist could be abolished in LXRα (–/–) mice, suggesting that LXRα plays an important role in mediating the inhibitory effect of GW3965 and T0901317. Two days after LXR treatment, LXRα, StAR, IL-1β, and IL-10 were all located in neurons and oligodendrocytes, indicating that the LXR agonist can exert a stimulating effect by activation of LXRα located in both neurons and oligodendrocytes. Because the effects of T0901317 could be blocked by neuroactive steroid synthesis inhibitor aminoglutethimide, LXR agonists may first increase the synthesis of neurosteroids and then decrease the synthesis of IL-1β, hence inhibiting the glial cell activation and at last the excitability of neurons. IL-10, an antiinflammatory cytokine, could inhibit the production of many proinflammatory cytokines, including IL-1β, TNF-α, and interleukin 6.^33,34  It is probable that LXR agonists first increase the synthesis of IL-10 and then decrease the production of proinflammatory cytokines. These results also indicate that, after nerve injury, neurons and oligodendrocytes can combine hands to counteract the abnormal state. In the present study, LXR agonists were injected via the intrathecal routine, which may affect both lumbar dorsal root ganglias and spinal cord. Because LXRα is also expressed in dorsal root ganglia neurons, LXRα here may also be involved in mediating the inhibitory effect of LXR agonists.

Previous data indicate that neuroactive steroids, including pregnenolone, progesterone, and allopregnanolone, are synthesized and actively metabolized in the central and peripheral nervous system and exert a neuroprotective effect in states of neuropathic pain induced by lumbar 5 spinal nerve ligation,³⁵  chronic constriction injury,³⁶  or spinal cord injury.³⁷  We have shown previously that a single application of Ro5-4864, a specific translocator protein agonist,²⁴  and pregnenolone, a neurosteroid precursor,³⁸  inhibits the established neuropathic pain behaviors induced by lumbar 5 spinal nerve ligation. In the current study, we found that endogenous neurosteroids could postpone the development of mechanical allodynia. A line of study has reported that a LXR agonist increases levels of neuroactive steroid and prevents diabetes-induced peripheral neuropathy.²³  Recently, it has been reported that LXRs directly modulate StAR expression in the adrenal gland,³⁹  a transfer protein regulating cholesterol shuttling into mitochondria, which is a key step in the initiation of steroid hormone synthesis. Consistently, in the present study, we found that T0901317 up-regulated expression of StAR. These results suggest that activation of LXRs stimulates the synthesis of neuroactive steroids. In addition, we found that the effects of T0901317 on mechanical allodynia, activation of glial cells, and expression of cytokines could be prevented by a neurosteroid synthesis inhibitor. Therefore, T0901317 may function via neurosteroids. Improving the local neurosteroid synthesis by activation of LXRs may be a valid way to treat neuropathic pain.

In conclusion, the present study demonstrated that a novel role of spinal LXRα in neuropathic pain, that is, activation of LXRα at the early time after nerve injury, can inhibit mechanical allodynia via modulating the neuroinflammatory responses.

Supported by grants from the National Natural Science Foundation (Beijing, People’s Republic of China, Nos. U1201223, 81200856, and 81471250); Fundamental Research Funds for the Central Universities (Beijing, People’s Republic of China, No. 13ykpy05); Nature Science Foundation of Guangdong Province of China (Guangzhou, People’s Republic of China, No. 2014A030313029); Scientific Research Foundation of Guangdong Province of China (Guangzhou, People’s Republic of China, No. 2016A020215035); and by scholarships from the Chinese Scholarship Council (Beijing, People’s Republic of China). All grants were awarded to Dr. Wei except No. U1201223 which was awarded to Dr. Liu.

The authors declare no competing interests.