Activation of the bile acid receptors TGR5 and FXR in the spinal dorsal horn alleviates neuropathic pain

Abstract Aims Beyond digestion, bile acids have been recognized as signaling molecules with broad paracrine and endocrine functions by activating plasma membrane receptor (Takeda G protein‐coupled receptor 5, TGR5) and the nuclear farnesoid X receptor (FXR). The present study investigated the role of bile acids in alleviating neuropathic pain by activating TGR5 and FXR. Method Neuropathic pain was induced by spared nerve injury (SNI) of the sciatic nerve. TGR5 or FXR agonist was injected intrathecally. Pain hypersensitivity was measured with Von Frey test. The amount of bile acids was detected using a bile acid assay kit. Western blotting and immunohistochemistry were used to assess molecular changes. Results We found that bile acids were downregulated, whereas the expression of cytochrome P450 cholesterol 7ahydroxylase (CYP7A1), a rate‐limiting enzyme for bile acid synthesis, was upregulated exclusively in microglia in the spinal dorsal horn after SNI. Furthermore, the expression of the bile acid receptors TGR5 and FXR was increased in glial cells and GABAergic neurons in the spinal dorsal horn on day 7 after SNI. Intrathecal injection of either TGR5 or FXR agonist on day 7 after SNI alleviated the established mechanical allodynia in mice, and the effects were blocked by TGR5 or FXR antagonist. Bile acid receptor agonists inhibited the activation of glial cells and ERK pathway in the spinal dorsal horn. All of the above effects of TGR5 or FXR agonists on mechanical allodynia, on the activation of glial cells, and on ERK pathway were abolished by intrathecal injection of the GABAA receptor antagonist bicuculline. Conclusion These results suggest that activation of TGR5 or FXR counteracts mechanical allodynia. The effect was mediated by potentiating function of GABAA receptors, which then inhibited the activation of glial cells and neuronal sensitization in the spinal dorsal horn.


| INTRODUC TI ON
Neuropathic pain following peripheral nerve injury is considered a disease of the nervous system. 1 There are currently no good therapies to treat neuropathic pain, highlighting the lack of understanding of detailed mechanisms and the urgency of discovering new drug targets.
After peripheral nerve injury, spinal microglia are activated, possibly by macrophage colony-stimulating factor-1 (CSF-1) and adenosine triphosphate (ATP), which are released from primary afferents, and fractalkine is released from spinal projecting neurons. 2 Activated microglia can release inflammatory mediators, such as tumor necrosis factor α (TNFα), interleukin 1β (IL-1β), and interleukin-6 (IL-6), to activate and alter the function of astrocytes. 3  produces the remaining 10% of BA and is primarily extrahepatic, including the brain. 9,10 Bile acids can exert hormone-like effects by integrating with their receptors, including the G-protein-coupled membrane receptor (TGR5) and the nuclear farnesoid X receptor (FXR), with both receptors being expressed in multiple organs, such as the liver, renal system, brain, heart, and retinas. 9 In addition to the regulation of metabolic homeostasis, bile acids and their receptors have been implicated in the regulation of the inflammatory response. [11][12][13] It has been reported that TGR5 prevents TNFα-induced adhesion molecule expression. 14 Moreover, bile acids also activate ion channels, including the bile acid-sensitive ion channel and epithelial Na + channel. Whether bile acids modulate neuropathic pain by inhibiting neuroinflammation in the spinal dorsal horn is still unknown.
Imbalance of the excitatory-inhibitory transmission contributes to neuropathic pain pathogenesis. 15 Peripheral nerve injury promotes a loss of inhibitory transmission in the spinal cord dorsal horn, and GABA A receptor-mediated phasic or tonic currents were shown to be decreased in a neuropathic pain model, chronic constriction of the sciatic nerve (CCI). 16 The subunit composition of GABA A receptors at synapses differs from that of extrasynaptic receptors, and these two types of GABA A receptors mediate two different forms of inhibition: synaptic and tonic inhibition. 17 Activation of the GABA A receptors located perisynaptically or extrasynaptically by ambient GABA is constantly present in the extracellular space, generating a constant current flow through the membrane that underlies tonic inhibition. 18 GABA A receptors that contain the δ subunit are restricted to extrasynaptic locations, 19 making them likely mediators of tonic inhibition. 20,21 The tonic inhibitory currents mediated by extrasynaptic GABA A receptors modulate neuronal excitability and synaptic transmission.
Neurosteroids potentiate GABA A receptors' function. It has been suggested that TGR5 acts as a neurosteroid receptor. 22 Bile acids, which are natural products derived from cholesterol, can modulate the function of neurotransmitter receptors, such as the muscarinic acetylcholine receptors and GABA A receptors, NMDA receptors and P2X2 receptor, 23,24 revealing their neuroactive potential activity. Therefore, we hypothesized that bile acids play a critical role in inhibiting neuropathic pain by modulating the functional properties of GABA A receptors following peripheral nerve injury.
Therefore, we hypothesized that the activation of bile acid receptors could inhibit neuroinflammatory responses in the spinal cord by modulating GABA A receptor and thereby alleviate neuropathic pain.
Here, we studied the changes in total bile acid levels, the alterations in bile acid rate-limiting synthesis enzymes, and bile acid receptor expression in the lumbar spinal dorsal horn after peripheral nerve injury. The effects and mechanisms by which exogenous bile acid receptor agonists inhibit neuropathic pain were also investigated.

Conclusion:
These results suggest that activation of TGR5 or FXR counteracts mechanical allodynia. The effect was mediated by potentiating function of GABA A receptors, which then inhibited the activation of glial cells and neuronal sensitization in the spinal dorsal horn.

K E Y W O R D S
astrocyte, bile acid, mechanical allodynia, nuclear farnesoid X receptor, spinal cord, Takeda G protein-coupled receptor 5

| MATERIAL S AND ME THODS
This study was not preregistered.

| Animals
Male C57BL/6 mice weighing 25-31 g were obtained from the Animal Experimental Center, Sun Yat-sen University, China. Mice were housed with free access to sterile water and standard laboratory food in a temperature-controlled room maintained at 24°C ± 1°C and 50%-60% humidity. All experimental procedures were approved by the Animal Care Committee of Sun Yat-sen University (No. SYXK (yue) 2017-0081) and were carried out following the Regulations for the Administration of Affairs Concerning Experimental Animals (China) and the ethical guidelines for the investigation of experimental pain in conscious animals (Zimmermann 1983). All of the experiments were conducted during the light phase in a double-blinded manner. No statistical methods were used to predetermine the sample size but were based on our previous experience. 25,26 In total, 177 mice were used in the present study; among them, two mice died after surgery and one mouse was euthanized due to excessive suffering from pain. Eight mice were excluded because no mechanical allodynia was induced after SNI surgery or because intrathecal catheterization failed. The experimental design and animal group classfication are shown in Figure 1.

| Assessment of mechanical sensitivity
The mechanical sensitivity of mice was assessed with the up-down method following a previous study 25,27 using a series of filaments with logarithmically incremental stiffness (0.4, 0.6, 1, 1.4, 2, 4, 6, and 8) (Stoelting Co). After habituation on a Plexiglass box with a wire grid floor, filaments were applied in either ascending or descending strengths to determine the filament strength closest to the hind paw withdrawal threshold. Each filament was applied for a maximum of 6 s in each trial. Quick withdrawal or licking the paw in response to the stimulus was considered a positive response. The 50% paw withdrawal threshold was then calculated according to a previous study. 27

| Immunofluorescence staining
Mice were anesthetized with pentobarbital and transcardially perfused with 0.9% saline followed by 4% paraformaldehyde in 0. For double staining, secondary primary antibodies were added after incubation with the primary antibodies for the same procedure as described above.
For quantification of immunofluorescence staining, the optical density of positive signals per section was measured by ImageJ Version: k 1.45. Every fifth section was picked from a series of sections, 4-6 sections for each animal were selected randomly, and the average was calculated.

| Intrathecal injection
For single intrathecal injection, direct transcutaneous intrathecal injection was performed following previous work. 28

| Bile acid quantification
The lumbar spinal cord tissue was dissected under 0.5% pentobarbital sodium (50 mg/kg) anesthesia, after which approximately 10-25 mg of the tissue was homogenized in 1 mL of a dry-ice chilled solution containing 80% methanol in H 2 O. The homogenates were then centrifuged at 13,000 g for 15 min to remove the insoluble fraction. The supernatants that contained soluble metabolites were then lyophilized and subsequently resuspended in 50 μL of sterile H 2 O.
The amount of bile acids was detected using a bile acid assay kit (Sigma-Aldrich), which provided a convenient fluorimetric means to measure total bile acid, according to the manufacturer's guidelines.
3α-Hydroxysteroid dehydrogenase reacts with bile acid, converting NAD to NADH and reducing a probe to a highly fluorescent product.
The profiles of bile acids were determined by a fluorescence microplate reader (λex = 530 nm/λem = 585 nm). The results were then normalized by the tissue weight.

| Statistical analysis
Data are expressed as the mean ± SEM and were analyzed with GraphPad Prism 8.0 Software. Shapiro-Wilk tests for normality were used to assess data distribution. Data were considered F I G U R E 1 Experimental design and animal group classification. In experiment 1, after the baseline 50% paw withdrawal threshold (50% PWT) detection on day −3, mice received SNI surgery on day 0, after which the changes in bile acids and expression of CYP7A1, and TGR5/ FXR in the spinal dorsal horn were investigated by western blotting and immunofluorescence on days 7 and 14 after the surgery. In experiment 2, an intrathecal catheter was implanted, and SNI surgery was performed on day 0. After recovery for 7 days, TGR5 agonist INT-777 or FXR agonist INT-747 was injected into the subarachnoid space via the catheter, the 50% PWT was measured by behavioral test at 1-9 h and 1-6 days after the drug treatment. SBI-115, Z-guggulsterone or bicuculline was injected intrathecally to block TGR5, FXR, or GABA A receptor, respectively. In experiment 3, the molecular changes including p-ERK, Iba1, and GFAP in the spinal cord at 2 h after the intrathecal injection of INT-777 or INT-747 were investigated by western blotting and immunofluorescence, whether the changes were abolished by GABA A antagonist was also investigated. BL, baseline; IF, immunofluorescence; SNI, spared nerve injury; WB, western blotting. non-normally distributed if p < 0.05. Data that do not exhibit a normal distribution were analyzed via non-parametric tests, including the Friedman ANOVA for repeated measurements, followed by the Wilcoxon matched-pairs test or the Mann-Whitney U test.
The normally distributed data were analyzed by one-way ANOVA followed by Tukey's post hoc test or two-tailed unpaired Student's t-test. p < 0.05 was considered statistically significant. When nonparametric tests are used, they are specified in the figure legend.
No statistical methods were used to predetermine the sample size, which was based on our previous experience. No tests for outliers were conducted on the data.

| SNI induced mechanical allodynia, decreased the level of bile acids and upregulated CYP7A1 in microglia in the lumbar spinal dorsal horn
Consistent with previous reports, SNI decreased the 50% paw withdrawal threshold on day 4 and day 7, which lasted until day 14 ( Figure 2A). Furthermore, we found that the level of total bile acids in the lumbar spinal dorsal horn was significantly decreased on days 7 and 14 after SNI ( Figure 2B). To test whether SNI changed the expression of cholesterol 7 α-hydroxylase (CYP7A1), a ratelimiting enzyme for bile acid synthesis, in the spinal dorsal horn, we performed western blotting and immunofluorescence staining.  Figure 2E) that expressed CYP7A1-IR increased, while the percentage of astrocytes ( Figure 2G) and neurons ( Figure 2I) that expressed CYP7A1-IR did not change significantly on day 7 after SNI.
These results indicated that the synthesis of bile acids in microglia of the spinal dorsal horn was increased after peripheral nerve injury.

| TGR5 was upregulated in the spinal dorsal horn by SNI
We next tested whether the expression of the bile acid membrane receptor TGR5 in the spinal dorsal horn was changed after SNI. The results showed that 7 days after SNI, the protein expression of TGR5 was significantly upregulated and returned to the baseline level on day 14 ( Figure 3A). Immunofluorescence experiments further confirmed the western blotting data ( Figure 3B-I), showing that on day 7 after SNI, TGR5-IR was located in the bilateral spinal dorsal horn ( Figure 3B) although SNI surgery was performed unilaterally at the left hind paw. Compared with the sham group, the increase in TGR5-IR was mainly located in laminae I-II, to a lesser extent in laminae III-IV, and much less in laminae V on day 7 after SNI ( Figure 3C).
Double immunofluorescence staining further showed that on day 7, TGR5 was located in Iba1-positive microglia ( Figure 4A

| Activation of TGR5 alleviated the established neuropathic pain induced by SNI, and the effect was blocked by SBI-115 or bicuculline
Having demonstrated that the concentration of bile acids and the expression of its receptor TGR5 in the spinal dorsal horn were changed after peripheral nerve injury. We next tested whether activation of TGR5 could affect neuropathic pain behaviors. To To determine whether SNI attenuated GABA A receptor δ subunit expression, one main subunit that constitutes extrasynaptic GABA A receptors, 19 we conducted western blotting experiments, which confirmed that there was a significant decrease in GABA A receptor δ subunit expression in SNI mice on day 7 compared with sham | 1987 WU et al. Figure S1). These results suggest that GABA A receptormediated tonic inhibition declined after SNI.

controls (
Bile acids can modulate the function of neurotransmitter receptors, including GABA A receptors. 23,24 Therefore, we next tested whether bile acids played a critical role in neuropathic pain by modulating the functional properties of GABA A receptor. The results demonstrated that intrathecal injection of the GABA A receptor antagonist bicuculline (125 μg, 5 μL), which was applied 30 min F I G U R E 2 SNI decreased the 50% paw withdrawal threshold, decreased the level of total bile acids, and upregulated CYP7A1 in microglia in the lumbar spinal dorsal horn. (A) SNI decreased the 50% paw withdrawal threshold on the ipsilateral hind paw on days 4, 7, and 14. **p < 0.01 compared with the sham group (Mann-Whitney U test, n = 6/group). (B) The level of total bile acids in the lumbar spinal dorsal horn on day 7 and day 14 after SNI. **p < 0.01 compared with the sham group. (C) Western blotting shows the expression of CYP7A1 in the ipsilateral lumbar spinal dorsal horn 7 and 14 days after SNI or sham operation. The quantification of CYP7A1 normalized to β-actin is shown under the representative bands (n = 5/group). **p < 0.01 compared with the sham group. (D, F, and H) In the ipsilateral lumbar spinal dorsal horn, CYP7A1 was located mainly in Iba1-labeled microglia (D) and to a much lesser extent in GFAP-labeled astrocytes (F) but not in NeuN-marked neurons (H) 7 days after either sham or SNI operation. (F, G and I), The percentage of microglia, astrocytes and neurons that expressed CYP7A1 in sham and SNI mice (day 7). Bar = 200 μm. ****p < 0.0001 compared with the sham group.

| Activation of TGR5 by INT-777 inhibited the expression of p-ERK and the activation of glial cells in the spinal dorsal horn, and the effect was blocked by bicuculline
It has been well established that neuroinflammation following glial cell activation in the spinal cord is important for the maintenance of neuropathic pain. 30

| FXR was upregulated in the spinal dorsal horn by SNI
We next tested whether the expression of the bile acid nuclear receptor FXR in the spinal dorsal horn was changed after SNI. The results showed that 7 days after SNI, the protein expression of FXR was significantly increased, which was further increased on day 14 ( Figure 7A

| DISCUSS ION
Treatment of neuropathic pain remains a challenging problem.

F I G U R E 8
The cell types that expressed FXR in the spinal dorsal horn from sham and SNI mice. (A, C, and E) In the ipsilateral lumbar spinal dorsal horn, FXR was located in GFAP-labeled astrocytes (C) and GAD67-GFP-positive neurons (E) but not in Iba1-labeled microglia (A) 7 days after sham and SNI operations. (B, D, and F) The percentages of microglia, astrocytes, and GAD67-GFPpositive neurons that expressed FXR in sham and SNI mice on day 7 are shown. **p < 0.01, ****p < 0.0001 compared with the sham group. Bar = 200 μm.

| Activation of bile acid receptors protects against neuropathic pain after peripheral nerve injury
Our previous studies have shown that the expression of liver X receptor (LXR) and translocator protein (18 kDa) (TSPO) was increased in the spinal dorsal horn after peripheral nerve injury and that either LXR agonists or TSPO agonists could alleviate neuropathic pain behaviors. 25,26 Similarly, here we showed that TGR5 and FXR were upregulated in the bilateral spinal dorsal horn after  We show herein that the GABA A receptor δ subunit was downregulated on day 7 after SNI, suggesting that GABA A receptorsmediated tonic inhibition in the spinal dorsal horn was reduced. It has been shown that inflammation changes the GABAergic neurotransmitter system, and vice versa, GABAergic signaling can curb the inflammatory response. 49 It is very likely that there is a positive loop between the neuroinflammatory response and reduced GABA A receptors inhibition after peripheral nerve injury.
Therefore, the suppression of GABA synaptic transmission after peripheral nerve injury might further potentiate the neuroinflammatory response and aggravate pain perception ( Figure 11). Our present study further showed that GABA A receptors antagonist bicuculline blocked the inhibitory effects of TGR5 or FXR agonists on neuropathic pain behaviors and on the activation of glial cells, which has been reported to be critical to the development and maintenance of neuropathic pain by initiating the neuroinflammatory response after peripheral nerve injury. These results suggest that bile acids exert their neuroprotective effect by modulating the function of glial cells by strengthening the inhibitory effects of extrasynaptic GABA A receptors. 37 In general, neurosteroids are potent modulators of the GABA A receptors potentiating GABAmediated responses at low concentrations and thereby enhancing increased inhibitory tone. 50,51 TGR5, currently known as a bile acid receptor, may also act as a neurosteroid receptor in response to allopregnanolone and other neurosteroids, 22

ACK N OWLED G M ENTS
We thank Bin Jiang (Sun Yat-sen University) for providing the GAD67-GFP reporter mice.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no competing financial interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data supporting the findings of this study are provided within the study and its supplementary information. Additional information will be provided upon reasonable request.