Phosphorylated Upstream Frameshift 1–dependent Nonsense-mediated μ-Opioid Receptor mRNA Decay in the Spinal Cord Contributes to the Development of Neuropathic Allodynia–like Behavior in Rats

Adult male Sprague-Dawley rats (200 to 250 g; mean age, 7 weeks) were used for this study. In response to peer review, female Sprague-Dawley rats (200 to 250 g; mean age, 7 weeks) experiments were added to this study. All animals were maintained the under a standard 12-h photoperiod and a normal chow diet. Surgical and experimental protocols of all the animal studies were approved by Mackay Medical College, New Taipei, Taiwan (A1090028), and the Animal Research: Reporting In Vivo Experiments (ARRIVE) guidelines were followed. The sample sizes were determined on the similar published studies using the same approaches.²²  No a priori statistical power calculation was conducted. Seven rats displayed postoperative neurologic dysfunctions and were excluded from statistical analysis (two animals were removed after spinal nerve ligation, and five animals were removed after intrathecal catheter implantation).

The protocol for spinal nerve ligation operation was as described in our previous study.²³  The surgical plane of anesthesia was verified by areflexia, toe pinch, and normal undisturbed respiratory rate and pattern. The body temperature of the rats was maintained by heating pad. We did not measure the blood pressure and blood gas of rats, and this is one of the limitations of the study.

Mechanical allodynia-like behaviors were assayed by Von Frey test (Dixon up-and-down method). The motor function and coordination of animals were assayed by rotarod apparatus. Detailed Von Frey test and motor function assay were as previously reported.²⁴ 

In response to peer review, burrowing assay was added for evaluation of painlike behaviors. Before the burrowing assay, rats were acclimatized to the specific burrowing tube (transparent, 23.5 cm long × 10 cm diameter) once on 3 consecutive days (for 1 h). The burrowing tube’s open end is raised 60 mm higher than the closed end. The test tube was filled with 1 kg pea shingle gravel, 2 to 4 mm in diameter, and placed in a test cage with slightly sprinkled fresh bedding at the start of each test. The amount of gravel left in the burrow at the end of each test session was weighed and recorded.²⁵ 

Rats were deeply isoflurane anesthetized (cold, cyanotic, unmoving) in the anesthetic or euthanasia chamber,²⁶  and samples (L4 to 5) were rapidly isolated. Western blot analysis of dorsal horn was performed as described in our previous study with modifications.²³  The membranes were incubated overnight with primary antibodies, including total UPF1 (mouse, 1:500, Santa Cruz Biotechnology, USA), phosphorylated UPF1 (rabbit, 1:1000, Millipore, USA), μ-opioid receptor (rabbit, 1:500, Abcam, USA), anti-SMG7 (rabbit, 1:1,000, Genetex, USA), SMG1 (rabbit, 1:500, Cell Signaling, USA), eIF4A3 (rabbit, 1:1,000, Abcam), and glyceraldehyde 3-phosphate dehydrogenase (rabbit, 1:5,000, Genetex), followed by incubation with specific secondary antibodies at dilutions of 1:5,000 to 1:10,000 (Supplementary Content 2, https://links.lww.com/ALN/D105).

Protocol of coprecipitation was as previously described.²⁷  Briefly, dorsal horn samples were incubated with a mouse antibody against total UPF1 (1:500, Santa Cruz Biotechnology) overnight at 4°C. Magnetic bead suspensions (Millipore) were added to the immune complex, and the mixtures were incubated at 4°C for 2 to 3 h. After heating the beads at 95°C for 5 min, the bound proteins were eluted for Western blot analysis.

Section preparation and immunostaining were done as previously described with modifications.²³  The sections were incubated with the primary antibodies overnight, including rabbit anti-phosphorylated UPF1 (1:200, Millipore), rabbit anti-SMG7 (1:1,000, Genetex), rabbit anti-SMG1 (1:500, Cell Signaling), and rabbit anti-eIF4A3 (1:200, Abcam), and then the Alexa Fluor secondary antibodies were applied (1:500 to 1:1,000, Invitrogen, USA). To measure immunofluorescence intensity, cell numbers in the lamina II dorsal horn were determined using a microscope with a 40× objective. Five sections were selected from each spinal sample. Each group had five animals.

Total RNA isolation and cDNA synthesis were as described in our previous study.²³  Briefly, 1 μg RNA was applied to the reaction of reverse transcription, and 2 μl of the cDNA products was used for quantitative polymerase chain reaction amplification with primers specific for Oprm1 (5’-AACTTTGGTGGCCAGAGATG-3’ and 5’-TGGGAAGAGAGAACTGGACAA-3’) and Gapdh (5’-ATGACTCTACCCACGGCAAG-3’ and 5’-GATCTCGCTCCTGGAAGATG-3’). Quantitative polymerase chain reaction was monitored on the QuantStudio 3 Real-Time polymerase chain reaction device (Applied Biosystems).

To estimate the half-life of µ-opioid receptor mRNAs, we performed intrathecal administration of actinomycin D (100 μg, 10 μl, Tocris Bioscience; United Kingdom) to rats. The RNA samples were collected at 0, 1, and 2 h and normalized to the 0 h baseline measurement. Detailed cDNA preparation and quantitative polymerase chain reaction protocol were described in our previous study.²³ 

The freshly dissected dorsal horn (L4 to 5) was harvested in ribonuclease-free ice-cold phosphate buffered saline. The tissue was teased apart in ice-cold phosphate buffered saline using a Dounce homogenizer. Resuspended equal pellet volume of complete RNA immunoprecipitation Lysis Buffer (Magna RNA immunoprecipitation RNA-Binding protein Immunoprecipitation Kit, catalog No. 17-700; MilliporeSigma, USA) supplemented with fresh protease and ribonuclease inhibitors. The lysate was incubated on ice for 5 min. Then the debris was removed by centrifugation (14,000 rpm, 10 min, 4°C), and the supernatant (RNA-binding protein or RNA complexes, RNA immunoprecipitation lysate) was used for RNA coimmunoprecipitation. From each sample, 10 μl of aliquot was transferred and saved as input samples for further analysis. Phosphorylated UPF1 antibody and IgG control were added to the magnetic beads, respectively, and incubated with rotation for 30 min at room temperature. The beads–antibody complex was resuspended in 900 μl RNA immunoprecipitation Buffer and 100 μl RNA immunoprecipitation lysate, and incubated with rotation overnight at 4°C. After extensive washing with cold RNA immunoprecipitation wash buffer, the immune complexes on beads were directly used to synthesize the first-strand cDNA by the Invitrogen kit. Then 2 μl each of the first-strand cDNA was used for a 25-μl polymerase chain reaction. The primers of μ-opioid receptor were used for reverse transcription-polymerase chain reaction. The polymerase chain reaction cycling conditions included an initial denaturation of 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 60 s. The specific primer pairs for Oprm1 region were as follows: Oprm1: 5’-AACTTTGGTGGCCAGAGATG-3’ and 5’-TGGGAAGAGAGAACTGGACAA-3’.

We conducted the implantation of intrathecal catheters according to our previous study.²³  NMDI14 (a potent nonsense-mediated mRNA decay inhibitor, NMDI14 targets a pocket in the nonsense-mediated mRNA decay factor SMG7 protein and disrupts SMG7–UPF1 interactions; 10, 30, 100 nM; 10 μl; MedChemExpress; USA) and eIF4A3-IN-1 (eIF4A3 inhibitor; 10, 30, 100 nM; 10 μl; MedChemExpress, USA) were administered intrathecally twice daily for 3 consecutive days. At 4 h after the last injection, behavioral studies and biochemistry analyses were conducted.

The small-interfering RNAs used in this study were all obtained from Dharmacon (USA). The sequences were as follows: total UPF1 small-interfering RNA: 5’-UCAUCAAGGUUCCCGAUAA-3’; SMG1 small-interfering RNA: 5’-GCUUAUUAGUUGAGAAGUU-3’; eIF4A3 small-interfering RNA: 5’-CGGGAGAUCAGGUCGAUAU-3’; control small-interfering RNA: 5’-UUCUCCGAACGUGUCACGUTT-3’. The preparation and intrathecal dosage of small-interfering RNA were conducted based on a previous study.²³ 

All data in this study were analyzed using Prism 6.0 (GraphPad, USA) and were expressed as mean ± SD. Before analysis of variances (ANOVAs), we confirmed that data are normally distributed using a Kolmogorov–Smirnov normality test. Statistical comparison was performed by one-way ANOVA with post hoc Bonferroni correction test, two-way ANOVA with post hoc Bonferroni correction test, or unpaired Student’s t test where appropriate. All statistical tests were two-tailed with significance set at P < 0.05.

The spinal nerve ligation procedure significantly decreased the withdrawal threshold of the ipsilateral hind paws of rats on Days 3, 7, 14, and 21 after the operation (fig. 1A; 10.84 ± 2.93 [Day –1], 11.17 ± 1.58 [Day 3], 10.58 ± 1.72 [Day 7], 14.19 ± 3.24 [Day 14], and 11.42 ± 2.18 [Day 21] in the sham ipsilateral group; 13.65 ± 1.90 [Day –1], 2.50 ± 0.50 [Day 3], 1.19 ± 0.31 [Day 7], 1.36 ± 0.75 [Day 14], and 1.44 ± 0.65 [Day 21] in the nerve ligation ipsilateral group; mean ± SD, data in grams; [Day 3, 7, 14, 21] in the sham ipsilateral group vs. [Day 3, 7, 14, 21] in the nerve ligation ipsilateral group, all P < 0.001; [Day –1] in the nerve ligation ipsilateral group vs. [Day 3, 7, 14, 21] in the nerve ligation ipsilateral group, all P < 0.001). We then examined the dynamics of total UPF1 or phosphorylated UPF1 expression in the dorsal horn in rats after spinal nerve ligation via Western blotting analysis. Spinal nerve ligation significantly increased the expression of phosphorylated UPF1, but not that of total UPF1, in the ipsilateral dorsal horn in rats on Days 3, 7, 14, and 21 after the operation with a time course that was consistent with that observed in spinal nerve ligation–induced allodynia-like behaviors in rats (fig. 1B; total UPF1: 0.08 ± 0.02 [Day –1], 0.08 ± 0.03 [Day 3], 0.09 ± 0.02 [Day 7], 0.07 ± 0.02 [Day 14], and 0.07 ± 0.01 [Day 21] in the sham ipsilateral group; 0.08 ± 0.01 [Day –1], 0.07 ± 0.02 [Day 3], 0.08 ± 0.01 [Day 7], 0.07 ± 0.01 [Day 14], and 0.09 ± 0.02 [Day 21] in the nerve ligation group; mean ± SD, data in arbitrary units; [Day 3, 7, 14, 21] in the sham ipsilateral group vs. [Day 3, 7, 14, 21] in the nerve ligation ipsilateral group, P = 0.912, 0.999, 0.998, and 0.668; [Day –1] in the nerve ligation ipsilateral group vs. [Day 3, 7, 14, 21] in the nerve ligation ipsilateral group, P = 0.961, 0.984, 0.986, and 0.976; phosphorylated UPF1: 0.33 ± 0.12 [Day –1], 0.38 ± 0.19 [Day 3], 0.34 ± 0.19 [Day 7], 0.30 ± 0.16 [Day 14], and 0.34 ± 0.17 [Day 21] in the sham ipsilateral group; 0.35 ± 0.13 [Day –1], 0.70 ± 0.11 [Day 3], 0.88 ± 0.15 [Day 7], 0.75 ± 0.16 [Day 14], and 0.73 ± 0.21 [Day 21] in the nerve ligation ipsilateral group; mean ± SD, data in arbitrary units; [Day 3, 7, 14, 21] in the sham ipsilateral group vs. [Day 3, 7, 14, 21] in the nerve ligation ipsilateral group, P = 0.025, P < 0.001, P < 0.001, and P = 0.004; [Day –1] in the nerve ligation ipsilateral group vs. [Day 3, 7, 14, 21] in the nerve ligation ipsilateral group, all P < 0.001). These results indicated that spinal phosphorylated UPF1 protein expression content was positively correlated with spinal nerve ligation–induced allodynia-like behaviors in rats. Further analysis revealed that the mechanical allodynia-like behaviors (stimulus-evoked nociception) and results of the burrowing assay (nonstimulus-evoked nociception) on Day 7 after spinal nerve ligation were not significantly different in rats of either sex (fig. 1C: 11.03 ± 2.47 [Day –1] and 1.35 ± 0.62 [Day 7] in the male group; 11.34 ± 1.25 [Day –1] and 1.52 ± 0.42 [Day 7] in the female group; mean ± SD, data in grams; [Day –1] in the male group vs. [Day 7] in the male group, P < 0.001; [Day –1] in the female group vs. [Day 7] in the female group, P < 0.001; [Day 7] in the male group vs. [Day 7] in the female group, P = 0.596; fig. 1D: 82.2 ± 23.0 [Day –1] and 25.0 ± 15.0 [Day 7] in the male group; 83.1 ± 20.3 [Day –1] and 21.8 ± 10.8 [Day 7] in the female group; mean ± SD, data in percentages; [Day –1] in the male group vs. [Day 7] in the male group, P < 0.001; [Day –1] in the female group vs. [Day 7] in the female group, P < 0.001; [Day 7] in the male group vs. [Day 7] in the female group, P = 0.677). Spinal nerve ligation significantly increased the expression of phosphorylated UPF1 in the ipsilateral dorsal horn on Day 7 in both sexes, and there was no significant difference in phosphorylated UPF1 protein expression on Day 7 in the two sexes, indicating that there was no sex difference in the dorsal horn after spinal nerve ligation (fig. 1E; 0.43 ± 0.08 [Day –1] and 0.87 ± 0.09 [Day 7] in the male group; 0.40 ± 0.06 [Day –1] and 0.96 ± 0.13 [Day 7] in the female group; mean ± SD, data in arbitrary units; [Day –1] in the male group vs. [Day 7] in the male group, P < 0.001; [Day –1] in the female group vs. [Day 7] in the female group, P < 0.001; [Day 7] in the male group vs. [Day 7] in the female group, P = 0.254). We next determined the cellular distribution of spinal nerve ligation–induced phosphorylated UPF1 expression in the dorsal horn in rats by staining with a specific antibody on Day 7 after the operation, the time point when the animals displayed steady mechanical allodynia-like behaviors and maximal phosphorylated UPF1 expression content in the dorsal horn in rats. On Day 7 after the operation, immunofluorescence showed that spinal nerve ligation increased the content of phosphorylated UPF1 in the ipsilateral dorsal horn in rats (fig. 1F). Double immunofluorescence labeling showed that phosphorylated UPF1 colocalized with a marker for neurons (fig. 1F). Overall, these results indicated that spinal nerve ligation induces allodynia-like behaviors in rats, accompanied by increased phosphorylated UPF1 expression content selectively in ipsilateral dorsal horn neurons.

We then examined whether total UPF1 small-interfering RNA can reverse neuropathic allodynia-like behaviors in spinal nerve ligation rats. Intrathecal administration of total UPF1 small-interfering RNA (5 μg, 10 μl) into naïve rats significantly reduced total UPF1 and phosphorylated UPF1 expression in the dorsal horn on Day 4, indicating the efficacy of our total UPF1 small-interfering RNA protocols in rats (fig. 2A; total UPF1: 0.06 ± 0.02 in the naïve group vs. 0.01 ± 0.01 in the total UPF1 small-interfering RNA group, mean ± SD, P < 0.001, data in arbitrary units; phosphorylated UPF1: 0.40 ± 0.09 in the naïve group vs. 0.17 ± 0.05 in the total UPF1 small-interfering RNA group, mean ± SD, P = 0.009, data in arbitrary units). For the motor function test, naïve and total UPF1 small-interfering RNA–treated rats (5 μg, 10 μl) have no significant differences, indicating that our small-interfering RNA method and total UPF1 small-interfering RNA itself did not cause motor deficits in rats (fig. 2B; 142.96 ± 25.04 [Day 0], 159.70 ± 20.21 [Day 1], 145.06 ± 25.42 [Day 2], 149.66 ± 21.47 [Day 3], 154.33 ± 18.68 [Day 4] in the naïve group vs. 138.01 ± 18.36 [Day 0], 137.06 ± 22.76 [Day 1], 151.23 ± 20.26 [Day 2], 136.25 ± 22.98 [Day 3], 151.90 ± 16.84 [Day 4] in the total UPF1 small-interfering RNA group, mean ± SD, P = 0.916 [Day 0], 0.170 [Day 1], 0.873 [Day 2], 0.530 [Day 3], and 0.979 [Day 4], data in seconds). Notably, although total UPF1 small-interfering RNA (5 μg, 10 μl) exhibited no effects on sham-operated animals (fig. 2C; 12.83 ± 4.41 [Day –1], 10.05 ± 1.41 [Day 1], 12.77 ± 2.95 [Day 3], 14.89 ± 3.40 [Day 5] and 12.33 ± 2.55 [Day 7] in the sham group vs. 12.23 ± 3.16 [Day –1], 12.67 ± 5.51 [Day 1], 11.68 ± 3.11 [Day 3], 12.29 ± 3.05 [Day 5], and 10.63 ± 2.82 [Day 7] in the sham plus total UPF1 small-interfering RNA group, mean ± SD, P = 0.948 [Day –1], 0.371 [Day 1], 0.840 [Day 3], 0.377 [Day 5], and 0.655 [Day 7], data in grams), it significantly reversed ipsilateral allodynia-like behaviors in rats subjected to spinal nerve ligation on postoperative Days 5 and 7 (fig. 2D; 1.35 ± 0.72 [Day 5] and 1.08 ± 0.60 [Day 7] in the nerve ligation group vs. 6.18 ± 2.98 [Day 5] and 8.14 ± 2.55 [Day 7] in the nerve ligation plus total UPF1 small-interfering RNA group, mean ± SD, all P < 0.001, data in grams). Moreover, total UPF1 small-interfering RNA (5 μg, 10 μl) into spinal nerve ligation rats significantly reversed spinal nerve ligation–induced the increased phosphorylated UPF1 expression in the ipsilateral dorsal horn (fig. 2E; 0.28 ± 0.11 in the sham group on Day 7, 0.71 ± 0.18 in the nerve ligation group on Day 7, and 0.36 ± 0.10 in the nerve ligation on Day 7 plus total UPF1 small-interfering RNA group, mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group vs. nerve ligation plus total UPF1 small-interfering RNA group, P = 0.004). Total UPF1 small-interfering RNA (5 μg, 10 μl) into spinal nerve ligation rats also significantly reduced total UPF1 expression in the ipsilateral dorsal horn (fig. 2E; 0.09 ± 0.03 in the sham group on Day 7, 0.10 ± 0.01 in the nerve ligation group on Day 7, and 0.02 ± 0.01 in the nerve ligation on Day 7 plus total UPF1 small-interfering RNA group, mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.988; nerve ligation group vs. nerve ligation plus total UPF1 small-interfering RNA group, P < 0.001). These data provide further evidence showing the higher phosphorylated UPF1 expression in neuropathic allodynia–like behaviors from spinal nerve ligation rats.

UPF1 is the central effector of nonsense-mediated mRNA decay, and its phosphorylation is an essential step of nonsense-mediated mRNA decay initiation.²  Notably, a study demonstrated that nonsense-mediated mRNA decay can modulate the µ-opioid receptor gene in the mouse brain.¹⁰  Moreover, microinjection of a highly selective µ-opioid receptor agonist depressed allodynia-like behaviors in rats subjected to spinal nerve ligation, and the effects were blocked by a selective µ-opioid receptor antagonist.²⁸  Thus, we next sought to determine whether spinal phosphorylated UPF1-dependent nonsense-mediated mRNA decay contributes to the regulation of µ-opioid receptor expression in rats with neuropathic allodynia–like behaviors. As expected, spinal nerve ligation significantly decreased the mRNA and protein expression contents of the μ-opioid receptor in the ipsilateral dorsal horn, which was significantly reversed by total UPF1 small-interfering RNA (5 μg, 10 μl; fig. 3, A and B) in spinal nerve ligation rats (mRNA: 0.87 ± 0.11 in the sham group on Day 7, 0.50 ± 0.11 in the nerve ligation group on Day 7, and 0.83 ± 0.16 in the nerve ligation plus total UPF1 small-interfering RNA group on Day 7, mean ± SD, data in folds. sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.002; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus total UPF1 small-interfering RNA, P = 0.006; protein: 0.04 ± 0.01 in the sham group on Day 7, 0.01 ± 0.01 in the nerve ligation group on Day 7, and 0.04 ± 0.01 in the nerve ligation group on Day 7 plus total UPF1 small-interfering RNA; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus total UPF1 small-interfering RNA, P = 0.009). To determine whether the spinal µ-opioid receptor is a true target of phosphorylated UPF1–dependent nonsense-mediated mRNA decay during neuropathic allodynia–like behaviors in rats, we further performed experiments in rats, including estimated mRNA half-life measurements, mRNA or pre-mRNA quantifications, and RNA immunoprecipitation. First, we measured the estimated and relative half-life of µ-opioid receptor mRNA in rats. To do this, we inhibited transcription in the dorsal horn of control (missense small-interfering RNA rats, 5 μg, 10 μl) and total UPF1 small-interfering RNA rats (5 μg, 10 μl) using intrathecal administration of actinomycin D (100 μg, 10 μl). The mRNA in the dorsal horn was harvested at 0-min (baseline), 60-min, and 120-min time points and subjected to quantitative polymerase chain reaction. In UPF1 small-interfering RNA–treated rats, the estimated half-life of the µ-opioid receptor mRNA was significantly increased (fig. 3C; 0.57 ± 0.17 [60 min] in the missense small-interfering RNA group vs. 0.69 ± 0.24 [60 min] in the total UPF1 small-interfering RNA group, P = 0.398; mean ± SD, data in folds; 0.05 ± 0.01 [120 min] in the missense small-interfering RNA group vs. 0.26 ± 0.15 [120 min] in the total UPF1 small-interfering RNA group, P = 0.015; mean ± SD, data in folds). This result exemplifies that in the absence of phosphorylated UPF1-dependent nonsense-mediated mRNA decay, there is reduced degradation of µ-opioid receptor mRNA, as would be expected of a natural nonsense-mediated mRNA decay target in rats. Subsequently, we quantified µ-opioid receptor mRNAs relative to pre-mRNA contents in the ipsilateral dorsal horn after spinal nerve ligation rats. µ-Opioid receptor mRNA content was significantly decreased relative to their pre-mRNA content after spinal nerve ligation (0.26 ± 0.08-fold, P < 0.001 vs. sham Day 7 plus missense small-interfering RNA group), and the decreases were significantly blocked by total UPF1 small-interfering RNA (5 μg, 10 μl; fig. 3D; 1.37 ± 0.26 in the sham group on Day 7 plus missense small-interfering RNA, 2.17 ± 0.33 in the sham group on Day 7 plus total UPF1 small-interfering RNA, 0.26 ± 0.08 in the nerve ligation group on Day 7 plus missense small-interfering RNA, and 0.77 ± 0.14 in the nerve ligation group on Day 7 plus total UPF1 small-interfering RNA group; mean ± SD, data in folds; sham group on Day 7 plus missense small-interfering RNA vs. sham group on Day 7 plus total UPF1 small-interfering RNA group, P < 0.001; sham group on Day 7 plus missense small-interfering RNA vs. nerve ligation group on Day 7 plus missense small-interfering RNA, P < 0.001; nerve ligation group on Day 7 plus missense small-interfering RNA vs. nerve ligation group on Day 7 plus total UPF1 small-interfering RNA, P = 0.013). Finally, to confirm the targeting of µ-opioid receptor mRNAs by the phosphorylated UPF1-dependent nonsense-mediated mRNA decay machinery in rats with neuropathic allodynia–like development, we coimmunoprecipitated µ-opioid receptor mRNA with phosphorylated UPF1 in rats after spinal nerve ligation. Spinal nerve ligation significantly increased coimmunoprecipitated µ-opioid receptor mRNA with phosphorylated UPF1 in rats, and the increases were significantly blocked by total UPF1 small-interfering RNA (5 μg, 10 μl; fig. 3E; 2.5 ± 0.5 in the sham group on Day 7, 19.7 ± 1.5 in the nerve ligation group on Day 7, and 6.9 ± 0.9 in the nerve ligation group on Day 7 plus total UPF1 small-interfering RNA group; mean ± SD, data in percentages; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus missense small-interfering RNA, P < 0.001). These results revealed a direct interaction and binding between µ-opioid receptor mRNAs and the phosphorylated UPF1-dependent nonsense-mediated mRNA decay machinery in neuropathic allodynia–like development in rats. Thus, these data indicate the occurrence of spinal nerve ligation-induced degradation of µ-opioid receptor mRNA by phosphorylated UPF1-dependent nonsense-mediated mRNA decay in the dorsal horn in rats.

Nonsense-mediated mRNA decay factor SMG7 can bind to phosphorylated UPF1 to increase RNA degradation.²⁹  Therefore, we further speculated that spinal nerve ligation–activated phosphorylated UPF1 followed by the degradation of µ-opioid receptor mRNA contributes to neuropathic allodynia–like development in rats by nonsense-mediated mRNA decay factor SMG7 binding to phosphorylated UPF1 in the dorsal horn. In the ipsilateral dorsal horn in rats, spinal nerve ligation–enhanced phosphorylated UPF1-positive, SMG7-positive, and phosphorylated UPF1 or SMG7 double-labeled immunofluorescence contents (fig. 4A). Moreover, consistent with allodynia-like behaviors in rats and increased phosphorylated UPF1 expression, spinal nerve ligation significantly increased the nonsense-mediated mRNA decay factor SMG7 in the ipsilateral dorsal horn in rats on Day 7 (fig. 4B; 0.10 ± 0.04 in the sham group on Day 7 vs. 0.24 ± 0.07 in the nerve ligation group on Day 7, P = 0.006; mean ± SD, data in arbitrary units). Interestingly, intrathecal administration of NMDI14 (a potent nonsense-mediated mRNA decay inhibitor that targets a pocket in the nonsense-mediated mRNA decay factor SMG7 protein and disrupts SMG7–UPF1 interactions; 10, 30, 100 nM; 10 μl) into spinal nerve ligation rats dose-dependently restored the decreased ipsilateral paw withdrawal threshold (fig. 4C; 2.06 ± 0.66 in the nerve ligation group on Day 7 vs. 6.31 ± 2.03 [10 nM], 8.96 ± 1.40 [30 nM], and 9.94 ± 1.98 [100 nM] in the nerve ligation group on Day 7 plus NMDI14, all P < 0.001; mean ± SD, data in grams). Intrathecal administration of NMDI14 (100 nM, 10 μl) reversed the spinal nerve ligation–increased total UPF1-SMG7 interaction in the ipsilateral dorsal horn (approximately threefold of sham operation group on Day 7, approximately 0.5-fold of spinal nerve ligation group on Day 7; fig. 4D). RNA coimmunoprecipitation showed that spinal nerve ligation significantly increased the content of coimmunoprecipitated µ-opioid receptor mRNA bound to phosphorylated UPF1 in the ipsilateral dorsal horn; this effect was reversed by intrathecal administration of NMDI14 (100 nM, 10 μl) into spinal nerve ligation rats (fig. 4E; 2.1 ± 0.4 in the sham group on Day 7, 18.4 ± 3.0 in the nerve ligation group on Day 7, and 3.5 ± 1.8 in the nerve ligation group on Day 7 plus NMDI14; mean ± SD, data in percentages; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P < 0.001). Furthermore, intrathecal administration of NMDI14 (100 nM, 10 μl) into spinal nerve ligation rats significantly reversed spinal nerve ligation–associated decreases in µ-opioid receptor expression content in the ipsilateral dorsal horn (fig. 4F). However, it did not affect spinal nerve ligation–associated increases in phosphorylated UPF1 and nonsense-mediated mRNA decay factor SMG7 expression contents in the ipsilateral dorsal horn (fig. 4F; total UPF1: 0.05 ± 0.02 in the sham group on Day 7, 0.05 ± 0.01 in the nerve ligation group on Day 7, and 0.05 ± 0.01 in the nerve ligation group on Day 7 plus NMDI14; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.979; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P = 0.993; phosphorylated UPF1: 0.40 ± 0.13 in the sham group on Day 7, 0.97 ± 0.12 in the nerve ligation group on Day 7, and 0.85 ± 0.10 in the nerve ligation group on Day 7 plus NMDI14; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P = 0.350; SMG7: 0.08 ± 0.02 in the sham group on Day 7, 0.35 ± 0.15 in the nerve ligation group on Day 7, and 0.30 ± 0.10 in the nerve ligation group on Day 7 plus NMDI14; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.009; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P = 0.906; µ-opioid receptor: 0.06 ± 0.01 in the sham group on Day 7, 0.02 ± 0.01 in the nerve ligation group on Day 7, and 0.05 ± 0.01 in the nerve ligation group on Day 7 plus NMDI14; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P < 0.001). These data indicate that nonsense-mediated mRNA decay factor SMG7 binds phosphorylated UPF1 to mediate the degradation of µ-opioid mRNA in the dorsal horn, contributing to the development of spinal nerve ligation–induced neuropathic allodynia–like behaviors in rats.

SMG1 kinase–mediated UPF1 phosphorylation creates a binding platform for nonsense-mediated mRNA decay factor SMG7 for mRNA degradationa.¹⁴  Therefore, we further identified the potential role of SMG1 kinase–mediated phosphorylation of UPF1 in phosphorylated UPF1-SMG7 or µ-opioid receptor signaling in the dorsal horn during neuropathic allodynia–like development in rats. Spinal nerve ligation significantly increased SMG1 kinase expression in the ipsilateral dorsal horn on Day 7 after the operation (fig. 5A; 0.06 ± 0.02 in the sham group on Day 7 vs. 0.20 ± 0.08 in the nerve ligation group on Day 7; mean ± SD, data in arbitrary units; P = 0.005). Western blotting analysis verified the treatment efficacy of SMG1 small-interfering RNA (5 μg, 10 μl) in the dorsal horn in naïve rats (fig. 5B; 0.08 ± 0.02 in the naïve group vs. 0.03 ± 0.01 in the SMG1 small-interfering RNA group; mean ± SD, data in arbitrary units; P = 0.005). Treatment with SMG1 small-interfering RNA (5 μg, 10 μl) had no significant effects on motor performance in naïve rats (fig. 5C), suggesting that our small-interfering RNA procedure and SMG1 small-interfering RNA itself failed to affect motor function in rats (135.15 ± 18.75 [Day 0], 143.21 ± 17.85 [Day 1], 153.61 ± 19.26 [Day 2], 149.36 ± 18.59 [Day 3], 143.20 ± 14.19 [Day 4] in the naïve group vs. 141.51 ± 13.30 [Day 0], 157.23 ± 13.81 [Day 1], 148.18 ± 22.46 [Day 2], 134.61 ± 19.77 [Day 3], 146.21 ± 25.35 [Day 4] in the total UPF1 small-interfering RNA group; mean ± SD; P = 0.821 [Day 0], 0.390 [Day 1], 0.866 [Day 2], 0.353 [Day 3], and 0.957 [Day4], data in seconds). Although SMG1 small-interfering RNA exhibited no effects on sham-operated animals (5 μg, 10 μl; fig. 5D; 12.39 ± 1.94 [Day –1], 10.73 ± 3.04 [Day 1], 10.69 ± 1.62 [Day 3], 11.30 ± 2.65 [Day 5] and 10.04 ± 2.00 [Day 7] in the sham group vs. 12.46 ± 2.14 [Day –1], 11.74 ± 3.29 [Day 1], 11.92 ± 2.15 [Day 3], 11.67 ± 2.53 [Day 5], and 12.31 ± 4.04 [Day 7] in the sham plus total UPF1 small-interfering RNA group; mean ± SD; P = 0.999 [Day –1], 0.805 [Day 1], 0.724 [Day 3], 0.971 [Day 5], and 0.342 [Day 7], data in grams), it markedly reversed ipsilateral allodynia-like behaviors of spinal nerve ligation rats on Day 7 (fig. 5E; 1.26 ± 0.69 [Day 7] in the sham group vs. 6.26 ± 0.76 [Day 7] in the sham plus total UPF1 small-interfering RNA group; mean ± SD, data in grams; P < 0.001). Moreover, intrathecal administration of SMG1 small-interfering RNA (5 μg, 10 μl) into spinal nerve ligation rats significantly reversed spinal nerve ligation–associated increases in the expression contents of phosphorylated UPF1 and SMG1 kinase, but not that of nonsense-mediated mRNA decay factor SMG7, and significantly reversed the spinal nerve ligation–associated decrease in µ-opioid receptor expression in the ipsilateral dorsal horn on Day 7 (fig. 5F; SMG1: 0.05 ± 0.02 in the sham group on Day 7, 0.19 ± 0.04 in the nerve ligation group on Day 7, and 0.08 ± 0.01 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P < 0.001; total UPF1: 0.05 ± 0.01 in the sham group on Day 7, 0.05 ± 0.02 in the nerve ligation group on Day 7, and 0.06 ± 0.01 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD; data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.988; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P = 0.972; phosphorylated UPF1: 0.34 ± 0.09 in the sham group on Day 7, 0.73 ± 0.10 in the nerve ligation group on Day 7, and 0.43 ± 0.06 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P = 0.002; SMG7: 0.09 ± 0.03 in the sham group on Day 7, 0.28 ± 0.07 in the nerve ligation group on Day 7, and 0.27 ± 0.06 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.003; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P > 0.999; µ-opioid receptor: 0.05 ± 0.01 in the sham group on Day 7, 0.01 ± 0.01 in the nerve ligation group on Day 7, and 0.05 ± 0.02 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.002; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P = 0.003). Furthermore, intrathecal administration of NMDI14 (100 nM, 10 μl) or total UPF1 small-interfering RNA (5 μg, 10 μl) into spinal nerve ligation rats did not affect the spinal nerve ligation–associated increase in SMG1 kinase expression in the ipsilateral dorsal horn on Day 7 after operation (fig. 5G; NMD14: 0.07 ± 0.04 in the sham group on Day 7, 0.19 ± 0.05 in the nerve ligation group on Day 7, and 0.21 ± 0.04 in the nerve ligation group on Day 7 plus NMDI14; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.005; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P = 0.913; SMG1 small-interfering RNA: 0.08 ± 0.02 in the sham group on Day 7, 0.19 ± 0.05 in the nerve ligation group on Day 7, and 0.19 ± 0.04 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.005; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P = 0.999). Intrathecal administration of SMG1 small-interfering RNA (5 μg, 10 μl) into spinal nerve ligation rats reversed spinal nerve ligation-associated increases in total UPF1-SMG1, -phosphorylated UPF1, and -SMG7 interactions (approximately 2-fold, approximately 3-fold, and approximately 1.5-fold of sham operation group on Day 7; approximately 0.7-fold, approximately 0.5-fold, and approximately 0.7-fold of spinal nerve ligation group on Day 7) in the ipsilateral dorsal horn on Day 7 (fig. 5H). On postoperative Day 7, SMG1 small-interfering RNA administration (5 μg, 10 μl) reversed spinal nerve ligation–associated increases in the numbers of SMG1-positive, phosphorylated UPF1-positive, SMG7-positive, and SMG1 or phosphorylated UPF1 or SMG7 triple-labeled puncta in the ipsilateral dorsal horn in rats (fig. 5I). In addition, RNA coimmunoprecipitation showed that spinal nerve ligation significantly increased the content of coimmunoprecipitated µ-opioid receptor mRNA bound to phosphorylated UPF1 in the ipsilateral dorsal horn; this effect was reversed by SMG1 small-interfering RNA (5 μg, 10 μl; fig. 5J; 2.0 ± 0.7 in the sham group on Day 7, 18.0 ± 1.6 in the nerve ligation group on Day 7, and 7.6 ± 1.0 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA; mean ± SD, data in percentages; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus SMG1 small-interfering RNA, P < 0.001). Thus, these data indicate that SMG1 kinase-mediated phosphorylation of UPF1 creates a binding platform for nonsense-mediated mRNA decay factor SMG7 and triggers the degradation of µ-opioid receptor mRNA in the dorsal horn during spinal nerve ligation-induced neuropathic allodynia–like development in rats.

The factor eIF4A3 is a core component of the exon-junction complex, and it is required for SMG1 kinase–mediated UPF1 phosphorylation and nonsense-mediated mRNA decay.²  Therefore, we next determined whether eIF4A3 affects SMG1 kinase or phosphorylated UPF1-SMG7 or µ-opioid signaling in the dorsal horn in rats after spinal nerve ligation. First, images of spinal slices revealed spinal nerve ligation–associated increases in eIF4A3-positive, phosphorylated UPF1–positive, and eIF4A3 or phosphorylated UPF1 double-labeled immunofluorescence contents in the ipsilateral dorsal horn in rats (fig. 6A). Moreover, on Day 7 after the operation, spinal nerve ligation significantly increased eIF4A3 expression content in the ipsilateral dorsal horn (fig. 6B; 0.10 ± 0.02 in the sham group on Day 7 vs. 0.20 ± 0.04 in the nerve ligation group on Day 7; mean ± SD, data in arbitrary units; P = 0.001). Western blotting analysis confirmed the treatment efficacy of eIF4A3 small-interfering RNA in the dorsal horn in naïve rats (5 μg, 10 μl; fig. 6C; 0.14 ± 0.02 in the naïve group vs. 0.06 ± 0.01 in the eIF4A3 small-interfering RNA group; mean ± SD, data in arbitrary units; P < 0.001). Treatment with eIF4A3 small-interfering RNA (5 μg, 10 μl) had no significant effects on motor performance in naïve rats, suggesting that our small-interfering RNA procedure and eIF4A3 small-interfering RNA itself failed to affect motor function in rats (fig. 6D; 140.60 ± 12.55 [Day 0], 129.35 ± 18.46 [Day 1], 152.63 ± 16.66 [Day 2], 137.31 ± 14.62 [Day 3], 147.13 ± 21.83 [Day 4] in the naïve group vs. 130.91 ± 19.49 [Day 0], 146.10 ± 23.79 [Day 1], 147.66 ± 22.29 [Day 2], 139.53 ± 23.09 [Day 3], 142.36 ± 24.93 [Day 4] in the eIF4A3 small-interfering RNA group; mean ± SD; P = 0.681 [Day 0], 0.322 [Day 1], 0.904 [Day 2], 0.980 [Day 3], and 0.911 [Day 4], data in seconds). Although eIF4A3 small-interfering RNA (5 μg, 10 μl) exhibited no effects on sham-operated animals (fig. 6E; 11.18 ± 2.48 [Day –1], 10.49 ± 2.15 [Day 1], 11.69 ± 2.21 [Day 3], 11.59 ± 1.54 [Day 5] and 10.35 ± 2.11 [Day 7] in the sham group vs. 11.92 ± 1.85 [Day –1], 10.51 ± 2.82 [Day 1], 9.95 ± 0.75 [Day 3], 11.54 ± 1.82 [Day 5], and 11.66 ± 2.09 [Day 7] in the sham plus eIF4A3 small-interfering RNA group; mean ± SD; P = 0.827 [Day –1], 0.999 [Day 1], 0.358 [Day 3], 0.999 [Day 5], and 0.558 [Day 7], data in grams), it significantly reversed ipsilateral allodynia-like behaviors of spinal nerve ligation rats on postoperative Days 5 and 7 (fig. 6F; 1.09 ± 0.90 [Day 5] and 0.98 ± 0.52 [Day 7] in the sham group vs. 4.81 ± 0.77 [Day 5] and 8.39 ± 2.71 [Day 7] in the sham plus total UPF1 small-interfering RNA group; mean ± SD; P = 0.003 [Day 5] and P < 0.001 [Day 7], data in grams). Interestingly, spinal administration of eIF4A3-IN-1 (a novel inhibitor of eIF4A3; 10, 30, 100 nM; 10 μl) but not of vehicle solution dose-dependently attenuated allodynia-like behaviors in spinal nerve ligation rats, as shown by the restoration in the tactile threshold of the ipsilateral hind paw of spinal nerve ligation rats (fig. 6G; 1.11 ± 0.57 in the nerve ligation group on Day 7 vs. 4.08 ± 1.61 [10 nM], 6.53 ± 2.44 [30 nM], and 8.57 ± 2.71 [100 nM] in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD; P = 0.028 [10 nM], P < 0.001 [30 nM], and P < 0.001 [100 nM]; data in grams). Moreover, intrathecal administration of eIF4A3 small-interfering RNA (5 μg, 10 μl) or eIF4A3-IN-1 (100 nM, 10 μl) into spinal nerve ligation rats significantly reversed spinal nerve ligation–associated increases in the expression contents of eIF4A3, SMG1 kinase, and phosphorylated UPF1, but not that of nonsense-mediated mRNA decay factor SMG7, and reversed spinal nerve ligation-decreased µ-opioid receptor expression in the ipsilateral dorsal horn (fig. 6H; eIF4A3: 0.10 ± 0.04 in the sham group on Day 7, 0.25 ± 0.06 in the nerve ligation group on Day 7, 0.11 ± 0.02 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 0.11 ± 0.03 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.004; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P = 0.005; nerve ligation group on Day 7 vs. eIF4A3-IN-1, P = 0.009; SMG1: 0.04 ± 0.01 in the sham group on Day 7, 0.25 ± 0.07 in the nerve ligation group on Day 7, 0.12 ± 0.05 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 0.14 ± 0.03 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P = 0.012; nerve ligation group on Day 7 vs. eIF4A3-IN-1, P = 0.045; total UPF1: 0.07 ± 0.02 in the sham group on Day 7, 0.08 ± 0.02 in the nerve ligation group on Day 7, 0.07 ± 0.03 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 0.08 ± 0.03 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P > 0.999; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P > 0.999; nerve ligation group on Day 7 vs. eIF4A3-IN-1, P = 0.999; phosphorylated UPF1: 0.35 ± 0.07 in the sham group on Day 7, 0.75 ± 0.15 in the nerve ligation group on Day 7, 0.45 ± 0.11 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 0.44 ± 0.07 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P = 0.003; nerve ligation group on Day 7 vs. eIF4A3-IN-1, P = 0.002; SMG7: 0.10 ± 0.01 in the sham group on Day 7, 0.28 ± 0.04 in the nerve ligation group on Day 7, 0.28 ± 0.03 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 0.30 ± 0.05 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P > 0.999; nerve ligation group on Day 7 vs. eIF4A3-IN-1, P = 0.981; µ-opioid receptor: 0.05 ± 0.01 in the sham group on Day 7, 0.20 ± 0.01 in the nerve ligation group on Day 7, 0.05 ± 0.08 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 0.05 ± 0.01 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P = 0.005; nerve ligation group on Day 7 vs. eIF4A3-IN-1, P = 0.004). Intrathecal administration of total UPF1 small-interfering RNA (5 μg, 10 μl), SMG1 small-interfering RNA (5 μg, 10 μl), and NMDI14 (100 nM, 10 μl) did not affect spinal nerve ligation-induced increases in eIF4A3 expression in the ipsilateral dorsal horn in rats on postoperative Day 7. However, total UPF1 small-interfering RNA administration resulted in slightly decreased but not significantly different of eIF4A3 expression in the ipsilateral dorsal horn samples in rats on Day 7 (fig. 6I; 0.11 ± 0.02 in the sham group on Day 7, 0.27 ± 0.05 in the nerve ligation group on Day 7, 0.23 ± 0.06 in the nerve ligation group on Day 7 plus total UPF1 small-interfering RNA, 0.25 ± 0.05 in the nerve ligation group on Day 7 plus SMG1 small-interfering RNA, and 0.27 ± 0.07 in the nerve ligation group on Day 7 plus NMDI14; mean ± SD, data in arbitrary units; sham group on Day 7 vs. nerve ligation group on Day 7, P = 0.005; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus total UPF1 small-interfering RNA, P = 0.974; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus total UPF1 small-interfering RNA, P = 0.999; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus NMDI14, P > 0.999). Furthermore, intrathecal administration of eIF4A3 small-interfering RNA (5 μg, 10 μl; approximately 0.8-fold, approximately 0.6-fold, approximately 0.3-fold, and approximately 0.5-fold of spinal nerve ligation group on Day 7) or eIF4A3-IN-1 (100 nM, 10 μl; approximately 0.8-fold, approximately 0.7-fold, approximately 0.3-fold, and approximately 0.6-fold of spinal nerve ligation group on Day 7) into spinal nerve ligation rats also significantly reversed total UPF1-eIF4A3, -SMG1, -phosphorylated UPF1, and -SMG7 interactions in the ipsilateral dorsal horn on Day 7 after the operation (fig. 6J). Intrathecal administration of eIF4A3 small-interfering RNA (5 μg, 10 μl) or eIF4A3-IN-1 (100 nM, 10 μl) reversed spinal nerve ligation–associated increases in the numbers of SMG1-positive, phosphorylated UPF1–positive, SMG7-positive, and SMG1 or phosphorylated UPF1 or SMG7 triple-labeled puncta in the ipsilateral dorsal horn in rats (fig. 6K). RNA immunoprecipitation showed that spinal nerve ligation significantly increased the content of coimmunoprecipitated µ-opioid receptor mRNA bound to phosphorylated UPF1 in the ipsilateral dorsal horn; this effect was reversed by eIF4A3 small-interfering RNA (5 μg, 10 μl) or eIF4A3-IN-1 (100 nM, 10 μl; fig. 6L; 1.9 ± 0.4 in the sham group on Day 7, 19.2 ± 5.7 in the nerve ligation group on Day 7, 7.0 ± 1.0 in the nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, and 4.2 ± 1.2 in the nerve ligation group on Day 7 plus eIF4A3-IN-1; mean ± SD, data in percentages; sham group on Day 7 vs. nerve ligation group on Day 7, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3 small-interfering RNA, P < 0.001; nerve ligation group on Day 7 vs. nerve ligation group on Day 7 plus eIF4A3-IN-1, P < 0.001). Thus, eIF4A3 triggers SMG1 kinase–mediated phosphorylation of UPF1, which recruits nonsense-mediated mRNA decay factor SMG7 to bind phosphorylated UPF1, to mediate the degradation of µ-opioid receptor mRNA in the dorsal horn during spinal nerve ligation–induced neuropathic allodynia–like behaviors in rats.

Here, we identify a role for the spinal phosphorylated UPF1–dependent nonsense-mediated mRNA decay mechanism in spinal nerve ligation–induced neuropathic allodynia–like behaviors in rats. Importantly, eIF4A3 leads to increased UPF1 phosphorylation by SMG1 kinase and the interaction of phosphorylated UPF1 with nonsense-mediated mRNA decay factor SMG7, subsequently triggering the µ-opioid receptor mRNA decay in spinal nerve ligation–induced neuropathic allodynia-like behaviors in rats (Supplementary Content 1, https://links.lww.com/ALN/D104). In addition, because sex is a biologic variable, we examined the mechanical allodynia-like behaviors (stimulus-evoked nociception), burrowing behaviors (nonstimulus-evoked nociception), and spinal phosphorylated UPF1 expression in both sexes of spinal nerve ligation rats, and the results suggested that painlike behaviors and spinal phosphorylated UPF1 expression were not sex-dependent after spinal nerve ligation. Notably, burrowing is a natural rodent behavior used to assess well-being, along with the assessment of painlike behaviors.³⁰  Concerns about the clinical translatability of stimulus-evoked nociception have led in recent years to an increase in the development and implementation of nonstimulus-evoked methods such as burrowing.³¹  Therefore, our results enhance the clinical relevance of animal models of future painlike behaviors.

Nonsense-mediated mRNA decay can regulate the stability of mutated and nonmutated transcripts, and it has complex tasks in cellular activities;³²  thus, nonsense-mediated mRNA decay plays a key role in not only cell physiology but also many diseases. A study showed that the depletion of UPF1 allows nonsense-mediated mRNA decay activity inhibition without compromising cell survival, suggesting that the use of UPF1-dependent nonsense-mediated mRNA decay inhibitors could be a therapeutic strategy.³³  Our data suggest that phosphorylated UPF1–dependent nonsense-mediated mRNA decay initiates µ-opioid receptor mRNA degradation in spinal nerve ligation–induced neuropathic allodynia–like behaviors in rats. Moreover, intrathecal administration of eIF4A3 small-interfering RNA, eIF4A3-IN-1, SMG1 small-interfering RNA, NMDI14, or UPF1 small-interfering RNA all ameliorated allodynia-like behavior in rats by inhibiting this pathway. As several nonsense-mediated mRNA decay inhibitors are already available in the clinic for the treatment of hereditary diseases,³²  these UPF1-related compounds could be redirected to treat neuropathic allodynia patients, pending the comprehension of the various nonsense-mediated mRNA decay regulatory mechanisms. Furthermore, nonsense-mediated mRNA decay inhibition has been tested in preclinical studies showing favored production of neoantigens by cancer cells, which can stimulate the triggering of an antitumor immune response.³²  Thus, these compounds associated with UPF1 also represent an opportunity for new cancer pain therapy strategies while preserving nonsense-mediated mRNA decay inhibition’s intrinsic tumor suppressor functions.

The conserved proteins UPF1, UPF2, and UPF3 are the core components of the nonsense-mediated mRNA decay machinery and are required for nonsense-mediated mRNA decay in all eukaryotic organisms. Interestingly, in addition to UPF1, other components of the nonsense-mediated mRNA decay pathway have been linked with neuronal function, neural diseases, and learning or memory-related plasticity. Neuron-specific disruption of UPF2 in adulthood attenuates learning, memory, spine density, and synaptic plasticity and potentiates perseverative or repetitive behavior.³⁴  Knockdown of UPF3B in cultured hippocampal neurons affects the expression of mRNA targets of nonsense-mediated mRNA decay and alters neuronal plasticity.³⁵  Specifically, activity-regulated cytoskeleton-associated protein mRNA has been identified as an nonsense-mediated mRNA decay target important for synaptic plasticity and the neurotransmission of nociception at the spinal cord level.³⁶  Central sensitization of spinal neurons relies on molecular processes very similar to those underlying associative learning and long-term potentiation.⁵  Intriguingly, our current studies indicated that UPF-dependent nonsense-mediated mRNA decay contributes to the pathologic progression of neuropathic allodynia–like behaviors in rats. These data imply that the spinal nonsense-mediated mRNA decay (UPF1, UPF2, UPF3)–activity-regulated cytoskeleton-associated protein–neuropathic allodynia–like behavior pathway in rats needs to be investigated further.

SMG1 kinase can phosphorylate UPF1 at the N- and C-terminal tails, in turn allowing the recruitment of the nonsense-mediated mRNA decay factors SMG5, SMG6, and SMG7. These proteins share a phosphoserine-binding domain similar to that found in 14–3–3 proteins.³⁷  SMG5 and SMG7 form a stable heterodimer, and SMG5–SMG7 complex binds phosphorylated residues in the C terminus of UPF1.³⁸  SMG6 is not part of the SMG5–SMG7 complex and has been shown to bind phosphorylated residues in the N terminus of UPF1.¹⁴  A study found that SMG6 14–3–3-like domain’s phosphoserine-binding site is similar to that of SMG5 and can mediate a weak phosphorylation-dependent interaction with UPF1 in crystal structure analysis.³⁹  Moreover, tethering assays have shown that the N terminus of UPF1 is dispensable.³⁸  Therefore, we chose nonsense-mediated mRNA decay factor SMG7 as a target in this study. Our findings are consistent with mounting evidence showing that spinal nerve ligation increases nonsense-mediated mRNA decay factor SMG7 expression and the phosphorylated UPF1-SMG7 interaction in the dorsal horn on Day 7 after the operation. Moreover, we intrathecally administered NMDI14 (which disrupts SMG7–UPF1 interactions) to spinal nerve ligation rats, which ameliorated spinal nerve ligation–induced allodynia-like behaviors in rats, reversed spinal nerve ligation–associated increases in phosphorylated UPF1-SMG7 interactions, and reversed spinal nerve ligation–associated decreases in µ-opioid receptor expression in the dorsal horn. However, NMDI14 did not affect spinal nerve ligation–associated increases in phosphorylated UPF1 and nonsense-mediated mRNA decay factor SMG7 contents in the dorsal horn. Furthermore, a study showed that SMG5 and SMG7 augment phosphorylated UPF1 binding to nonsense-mediated mRNA decay targets; conversely, SMG5- and SMG7-targeted small-interfering RNA reduces UPF1 binding to an nonsense-mediated mRNA decay target.⁴⁰  Our findings further supported that NMDI14 reversed spinal nerve ligation–increased content of coimmunoprecipitated µ-opioid receptor mRNA bound with phosphorylated UPF1, suggesting that it is the heterodimer that associates with and enhances the binding of phosphorylated UPF1 to µ-opioid receptor mRNAs. Interestingly, SMG5 and SMG7 also mediate UPF1 dephosphorylation.⁴¹  Overexpression of the SMG5 mutant retains its interaction with phosphorylated UPF1 but cannot induce its dephosphorylation and impair nonsense-mediated mRNA decay, suggesting that nonsense-mediated mRNA decay requires phosphorylated UPF1 dephosphorylation.⁴¹  Thus, the roles of spinal SMG5, SMG7, or phosphorylated UPF1 dephosphorylation in spinal nerve ligation–induced neuropathic allodynia–like behaviors in rats need to be investigated further.

UPF1 has a central role in the nonsense-mediated mRNA decay pathway.²  UPF1 also functions in other pathways. A study reported strong evidence for an exciting new function of UPF1, i.e., enzymatic involvement of UPF1 in ubiquitination, which promotes protein decay.⁴²  The authors provided a biologic link for this new function by showing that UPF1 acts as an E3 ligase via its Really Interesting New Gene domain to promote the ubiquitination and degradation of the MYOD protein.⁴²  Downregulation of MYOD-1 gene expression is associated with muscle pain and damage.⁴³  Considering that our previous study demonstrated that E3 ligase–mediated ubiquitination mechanisms coordinately regulate the contents of synaptic proteins in neuropathic allodynia–like behaviors in rats, it is possible that spinal UPF1 provides a mechanistic link between the RNA and protein decay machineries during neuropathic pain.

Mounting evidence supports an increased risk of harm associated with long-term conventional opioid agonist therapy due to tolerance that leads to overdosing, dependence, and a variety of other side effects, such as constipation and respiratory depression, that arise from prolonged use.⁴⁴  In the clinic, strategies for mitigating opioid side effects include reducing analgesics dose and the duration of treatment by interrupting infusions of sedative or analgesic agents daily or modulating infusions on the basis of analgesic assessment and sedation scores, using multimodal analgesic agents, and rotating analgesic agents sequentially.⁴⁵  However, these strategies still cannot effectively reduce opioid side effects in the majority of patients. For these reasons, one of the most promising strategies in the search for new analgesics is the development of compounds that activate the opioid system without tolerance or other side effects. Considering that the µ-opioid receptor is the primary and most widely studied mediator of opioid activity, including that of morphine, our investigations focused on this receptor. Interestingly, µ-opioid receptor mRNA has been identified as a nonsense-mediated mRNA decay target during the study of neurologic diseases.¹⁰  Our current study demonstrated that the spinal phosphorylated UPF1–dependent nonsense-mediated mRNA decay mechanism promotes the degradation of µ-opioid receptor mRNA in the dorsal horn during spinal nerve ligation–induced neuropathic allodynia–like behaviors in rats. Moreover, eIF4 contributes to neuropathic allodynia-like behaviors in rats by participating in the downregulation of µ-opioid receptor in injured dorsal root ganglion.⁴⁶  The µ-opioid receptor activates signaling pathways and is implicated in protecting neurons from apoptosis by inhibiting eIF4 interaction.⁴⁷  Our current study demonstrated that spinal administration of eIF4A3-IN-1 or eIF4A3 small-interfering RNA attenuated allodynia-like behaviors in rats subjected to spinal nerve ligation and reversed SMG1 or phosphorylated UPF1-SMG7 or µ-opioid receptor signaling in the dorsal horn. Thus, the established functional interactions between µ-opioid receptor and the spinal eIF4A3 or phosphorylated UPF1–dependent nonsense-mediated mRNA decay mechanism offer an opportunity to uncover new targets to develop treatments for neuropathic allodynia–like behaviors without tolerance or other side effects. Endogenous opioid receptors consist of μ-opioid receptors, δ-opioid receptors, κ-opioid receptors, and opioid receptor–like 1 receptors. These receptor mRNAs are widely expressed throughout the central nervous system, where they contribute to nociception.⁴⁸  Increasing evidence implicates posttranscriptional events in the regulation of opioid receptor expression or activities in the central nervous system, including alternative splicing, mRNA stability, translation, RNA polyadenylation, RNA transport, and covalent modification of the receptors, i.e., the μ-opioid receptor, δ-opioid receptor, and κ-opioid receptor genes.⁴⁹  Moreover, blockade of κ-opioid receptor reduces mechanical hyperalgesia and anxiety-like behavior in a rat model of trigeminal neuropathic allodynia–like behavior.⁵⁰  Thus, other opioid receptors for UPF1-dependent nonsense-mediated mRNA decay in neuropathic allodynia–like behaviors in rats need to be investigated further.

This research was supported by the Ministry of Science and Technology, Taipei, Taiwan: MOST 108-2320-B-715-002-MY3, 108-2314-B-715-004-MY3, 108-2811-B-715-506, 109-2811-B-715-503, 110-2813-C-715-001-B, and 111-2320-B-715-002 to Dr. Peng, and MOST 109-2320-B-715-005-, MOST 110-2320-B-715-001-, and MOST 111-2320-B-715-003-MY3 to Dr. Hsieh; by the Mackay Memorial Hospital (Taipei, Taiwan): MMH-MM-10705, MMH-MM-10803, MMH-MM-10902, MMH-MM-11009, and MMH-MM-111-05 to Dr. Peng, MMH 108-59, MMH-109-47, MMH-108-087, MMH-MM-10814, and MMH-MM-10913 to Dr. Yang, and MMH-MM-11006 and MMH-MM-111-12 to Dr. Hsieh; by the Department of Medicine, Mackay Medical College (New Taipei, Taiwan): 1071B16, 1081B03, 1091B01, MMC-RD-110-1B-P029, MMC-RD-110-1H-P008, and MMC-RD-111-1B-P009 to Dr. Peng, and 1091A01 to Dr. Hsieh; and by the Department of Health, Taichung Hospital, Executive Yuan, Taichung, Taiwan: 10914 and 11124 to Dr. Yeh.

The authors declare no competing interests.

Supplemental Content 1: Spinal phosphorylated upstream frameshift 1 (UPF1)–dependent nonsense-mediated mRNA decay mechanism in nociception rat, https://links.lww.com/ALN/D104

Supplemental Content 2: Full-length Western blot images, https://links.lww.com/ALN/D105