Dynorphin and Enkephalin Opioid Peptides and Transcripts in Spinal Cord and Dorsal Root Ganglion During Peripheral Inflammatory Hyperalgesia and Allodynia

Abstract

Understanding molecular alterations associated with peripheral inflammation is a critical factor in selectively controlling acute and persistent pain. The present report employs in situ hybridization of the 2 opioid precursor mRNAs coupled with quantitative measurements of 2 peptides derived from the prodynorphin and proenkephalin precursor proteins: dynorphin A 1-8 and [Met⁵]-enkephalin-Arg⁶-Gly⁷-Leu⁸. In dorsal spinal cord ipsilateral to the inflammation, dynorphin A 1-8 was elevated after inflammation, and persisted as long as the inflammation was sustained. Qualitative identification by high performance liquid chromatography and gel permeation chromatography revealed the major immunoreactive species in control and inflamed extracts to be dynorphin A 1-8. In situ hybridization in spinal cord after administration of the inflammatory agent, carrageenan, showed increased expression of prodynorphin (Pdyn) mRNA somatotopically in medial superficial dorsal horn neurons. The fold increase in preproenkephalin mRNA (Penk) was comparatively lower, although the basal expression is substantially higher than Pdyn. While Pdyn is not expressed in the dorsal root ganglion (DRG) in basal conditions, it can be induced by nerve injury, but not by inflammation alone. A bioinformatic meta-analysis of multiple nerve injury datasets confirmed Pdyn upregulation in DRG across different nerve injury models. These data support the idea that activation of endogenous opioids, notably dynorphin, is a dynamic indicator of persistent pain states in spinal cord and of nerve injury in DRG. PERSPECTIVE: This is a systematic, quantitative assessment of dynorphin and enkephalin peptides and mRNA in dorsal spinal cord and DRG neurons in response to peripheral inflammation and axotomy. These studies form the foundational framework for understanding how endogenous spinal opioid peptides are involved in nociceptive circuit modulation.

Keywords: Opioid; allodynia; chronic pain; gene expression; inflammation; morphine; neuropathic pain.

Figure 1.

Behavioral assessment of hyperalgesia after peripheral inflammation with complete Freund’s adjuvant (CFA). Rats were unilaterally injected with 150 μL of 1:1 saline:CFA emulsion and behavioral characterization of several sensory modalities was performed. (A) Inflammation caused significant edema lasting at least 28 days; **, P < .01; 2-way ANOVA followed by Holm-šidák. (B) Thermal sensitivity was evaluated using radiant thermal stimuli. Significantly reduced withdrawal latencies occurred for the inflamed paw which resolved completely by 14 days. (C) Mechanical sensitivity was recorded as the threshold response to von Frey filament testing. Inflammation reduced mechanical force necessary to induce withdrawal and this effect largely resolved after 28 days. (D) Evoked guarding behavior was stimulated by plantar pinprick. No guarding occurred at baseline, but in inflamed animals prolonged guarding lasted at least 21 days. (E) Spontaneous guarding behavior and weight bearing was assessed by observation of the animals while standing on a wire mesh grid. Behavioral scores were assigned in 6 consecutive 10-second epochs. 0, normal weight bearing; 1, partial weight bearing; 2, no weight bearing; 3, spontaneous nocifensive behavior (licking, quivering, or excessive guarding). Animals showed significantly more spontaneous guarding on the inflamed paw for at least 2 weeks. Error bars show standard error of the mean (A, B) or standard deviation (C, D). Significance testing was performed using 2-way ANOVA (A, B), multiple Mann-Whitney U-tests (C, E), or exact binomial test (D). For all results, Holm-šidák corrections were made to account for multiple comparisons; *, P < .05; **, P < .01.

Figure 2.

Time course of dynorphin A 1-8 and MERGL peptide content in dorsal and ventral spinal cord after inflammation. Peptide measurements were performed for 2 opioid peptides that act in the dorsal spinal cord. (A) Dynorphin A 1-8 (YGGFLRRI) is processed from the preprodynorphin precursor, where it is one of several neuroactive products. A simplified diagram, adapted from, shows some of the major products of this precursor protein. Met- and Leu-enkephalin-containing sequences are shown in gray shading. (B) A similar model is presented for preproenkephalin processing showing the location of the MERGL (YGGFMRGL) sequence. (C-D) Animals were injected with 200 μL 1:1 CFA saline emulsion, and reinjected with the same adjuvant at the end of day 8. Measurements were taken at day 0, 3, 8. A final sample was taken at day 14 after the initial inflammation, with reinjection performed on day 8. (C) Dynorphin A 1-8 was significantly elevated at day 3, day 8, and day 14 in the dorsal spinal cord in the lumbar dorsal spinal cord (ANOVA with post hoc Scheffé test; N = 5), but was not significantly altered in the ventral lumbar spinal cord. (D) Levels of MERGL were not altered in either dorsal or ventral lumbar spinal cord.

Figure 3.

Time course segmental specificity and temporal model. To investigate the time course of a single CFA injection, 1 hind paw was injected at day 0 with 200μL of a 1:1 emulsion of complete Freund’s adjuvant and sterile saline. (A) Time course. The dynorphin content in spinal cord was significantly elevated at each time point relative to the contralateral side of the same animals using a paired t-test at each time point (N = 4; *, P < .02; **, P < .001). An ANOVA was performed with 1 repeated measure, using the Scheffé post hoc test, showing no difference in the levels of dynorphin A 1-8 on the contralateral side over time. The same test also revealed that the day 5 levels of dynorphin on the inflamed side were higher than at other time points (⁺, P < .05). (B) Segmental specificity. Segmental specificity of the increase in dynorphin content was assessed by comparing lumbar spinal cord, which receives innervation from the inflamed hind limb with cervical spinal cord, which does not. Only the ipsilateral inflamed dorsal spinal cord showed significant elevation of dynorphin peptide (1-way ANOVA with post hoc Scheffé test; **, P < .001). Error bars represent the standard error of the mean. (C) Multimodal temporal model. Based on the data in the present report, and previous literature, we modeled the time course of behavioral, inflammatory edema, and molecular events in response to peripheral inflammation. At onset phase, Fos mRNA spikes briefly with immediate edema and thermal hypersensitivity. Subsequently, Pdyn mRNA rises, which is then followed by Dyn A1-8 peptide content accumulation. Thermal sensitivity resolves faster than edema, while mechanical sensitivity resolves similarly to edema (as shown in Fig 1). We hypothesize that period up to ~day 4 represents a process of active dynorphin peptide release which keeps tissue levels from accumulating. Subsequently, as hyperalgesia resolves there is a overshoot in tissue peptide content.

Figure 4.

(A) Gel chromatographic and (B) high-performance liquid chromatographic analyses of dynorphin A 1-8 immunoreactivity. (A) Acidic extracts of lumbar dorsal cord from inflamed and control animals were adsorbed on a C-18 Sep-Pak, which was washed with water and then eluted with 60% acetonitrile, .1% TFA. The eluate was dried in a vacuum centrifuge. The sample was resuspended in 500 μL of 1 M acetic acid, which also was the mobile phase. One mL fractions were collected, lyophilized and analyzed for dynorphin A 1-8 immunoreactivity. One major peak of immunoreactivity was detected which eluted in the position of synthetic dynorphin A 1-8 standard. Note the greater amount of immunoreactivity in the extract from inflamed tissue. (B) Dorsal cord extracts were prepared as described for gel filtration, resuspended in 250 μL .1% TFA and injected into an Altex 5 μ octadecylsilane reversed phase column. Solvent delivery was 1 mL/min. The column was eluted with a linear gradient of 20 to 60% acetonitrile in .1% TFA over 40 minutes (dashed line). One mL fractions were collected and assayed for dynorphin A 1-8 after drying under reduced pressure. The elution positions of synthetic dynorphin A 1-8 and 1-17 are shown (arrows). One major peak eluting in the position of dynorphin A 1-8 was observed. As with size exclusion chromatography, a larger dynorphin A 1-8 peak was observed in extracts from inflamed spinal cord relative to control.

Figure 5.

Neuropeptide content of 5 discrete brain regions after peripheral inflammation. Neuropeptide content (pmoles/mg protein) was measured in periaqueductal gray (PAG), hypothalamus, septum, caudate, and substantia nigra of 4 rats. The Freund’s adjuvant-saline emulsion was injected into both hind paws, and rats were euthanized after 6 days. Bilateral brain regions were pooled. Since sulfated cholecystokinin-8 (CCK-8) immunoreactivity is not extracted efficiently with acidic conditions, tissues were first extracted with 90% methanol and then re-extracted with acid mixture. The 2 supernatants were combined and aliquots were measured for CCK-8-like immunoreactivity, dynorphin A 1-8, and MERGL. Values represent mean ± SEM. Comparisons were made with Student’s t-test; *, P < .02.

Figure 6.

Fluorescent in situ hybridization of prodynorphin and proenkephalin mRNA 2 days after peripheral inflammation. Animals were inflamed by intraplantar injection of 6 mg of carrageenan in 150μL sterile saline and tissue was harvested 2 days later. Coronal sections of lumbar spinal cords from 4 unilaterally inflamed rats were stained using multiplex ISH using probes against Penk (green) and Pdyn (red) transcripts with DAPI as a nuclear marker. (A) All 4 stained images were scanned and spatially registered to show aggregate staining across all sections analyzed using a spinal laminar overlay template adapted from. (B) The upregulation of Pdyn in the ipsilateral dorsal horn is evident in the ipsilateral superficial laminae, where the Pdyn+ cells are concentrated. (C, D) Representative images of the medial aspect of the superficial laminae are shown for the (C) ipsilateral and (D) contralateral sides, red fluorescent Pdyn+ cells indicated by arrows. The signal is comparatively weak on the noninflamed control side. (E) Graph depicting the overall distribution of staining showed Penk in multiple laminae, with a concentration in lamina I/IIo, whereas Pdyn was mostly observed in lamina I/IIo. (F) Intensity and area of Pdyn staining was examined per pixel subsequent to thresholding. The average intensity of staining was significantly higher in the ipsilateral dorsal spinal cord compared to the contralateral side. The area (in pixels) of detectable Pdyn signal also was strongly increased on the ipsilateral side. (G) For Penk, the intensity measurement was increased significantly but by a much smaller magnitude, whereas the area was not different. (H) A small number (approximately 7) cells showed colocalization of Penk and Pdyn. (I-K) A representative image of a Pdyn+/Penk+ co-positive cell is shown (arrow). Statistics were performed using a 2-tailed Mann-Whitney U-test; **, P < .01. Error bars represent standard error of the mean. (Color version of figure is available online.)

Figure 7.

Modulation of prodynorphin (Pdyn), proenkephalin (Penk), and galanin (Gal) mRNA expression in the DRG after sciatic nerve transection or spinal nerve transection or spinal nerve ligation. (A-C) Sciatic transcetion. The sciatic nerve was transected and L4 and L5 DRGs were collected at 0d, 1d, 3d, 10d, 30d, and 90d after the axotomy. Both Pdyn and Gal were significantly elevated in the dorsal root ganglia, whereas Penk was significantly decreased. Statistics were performed in MAGIC with experimental time points compared to the naÿve control. The contralateral side is shown, but statistics were not performed for ipsilateral versus contralateral at each time point, because the contralateral sample for some time points were single replicates (10d, 30d, 90d). In the ipsilateral time points, some error bars are smaller than the marker size (not shown). (D-F) Molecular meta-analysis of spinal nerve injury studies. The experimental axotomy findings in A-C were also investigated by realigning and reanalyzing 4 RNA-Seq datasets from the SRA database from previously published reports. Also included is an examination the transcriptional effects of carrageenan inflammation on DRG gene expression. Statistics were performed to examine the overall statistical effect between all axotomy samples and all controls in aggregate (see methods). All 3 genes were significant when considering the data from all 5 datasets in aggregate. In contrast, carrageenan did not alter DRG expression of Pdyn, Penk, or Gal. Error bars show the standard error of the mean.