Hedgehog signaling regulates nociceptive sensitization

Abstract

Background: Nociceptive sensitization is a tissue damage response whereby sensory neurons near damaged tissue enhance their responsiveness to external stimuli. This sensitization manifests as allodynia (aversive withdrawal to previously nonnoxious stimuli) and/or hyperalgesia (exaggerated responsiveness to noxious stimuli). Although some factors mediating nociceptive sensitization are known, inadequacies of current analgesic drugs have prompted a search for additional targets.

Results: Here we use a Drosophila model of thermal nociceptive sensitization to show that Hedgehog (Hh) signaling is required for both thermal allodynia and hyperalgesia following ultraviolet irradiation (UV)-induced tissue damage. Sensitization does not appear to result from developmental changes in the differentiation or arborization of nociceptive sensory neurons. Genetic analysis shows that Hh signaling acts in parallel to tumor necrosis factor (TNF) signaling to mediate allodynia and that distinct transient receptor potential (TRP) channels mediate allodynia and hyperalgesia downstream of these pathways. We also demonstrate a role for Hh in analgesic signaling in mammals. Intrathecal or peripheral administration of cyclopamine (CP), a specific inhibitor of Sonic Hedgehog signaling, blocked the development of analgesic tolerance to morphine (MS) or morphine antinociception in standard assays of inflammatory pain in rats and synergistically augmented and sustained morphine analgesia in assays of neuropathic pain.

Conclusions: We demonstrate a novel physiological role for Hh signaling, which has not previously been implicated in nociception. Our results also identify new potential therapeutic targets for pain treatment.

Figure 1. hedgehog Mutants Fail to Develop Thermal Allodynia and Hyperalgesia

(A) Behavioral responses of control (w¹¹¹⁸) and hedgehog (hh^ts2) mutant larvae to a normally nonnoxious stimulus of 38°C. Behaviors were classified as no response (white, >20 s), slow withdrawal (gray, between 5 and 20 s), or fast withdrawal (black, <5 s). Responses were measured with and without ultraviolet irradiation (UV)-induced tissue damage, in the presence or absence of a 24 hr 29°C heat shock administered after UV treatment. n = triplicate sets of 30 larvae per condition. Brackets with asterisks represent statistically significant comparisons by Fisher’s exact test. (B) Behavioral responses to a noxious suprathreshold stimulus of 45°C. Withdrawal latency was measured with and without UV-induced tissue damage in the presence or absence of an 8 hr 29°C heat shock administered after UV treatment. n = 50 larvae per condition. Error bars represent standard error of the mean (SEM). Brackets with asterisks represent statistically significant comparisons by two-way repeated-measures analysis of variance (ANOVA). n.s. indicates not significant.

Figure 2. Hedgehog Signaling Components Are Required in Nociceptive Sensory Neurons

In both panels, ppk1.9-Gal4 drives (>) expression of the indicated upstream activating sequence (UAS) transgenes or no transgene (ppk1.9-Gal4 crossed to w¹¹¹⁸ = ppk/+) in nociceptive sensory neurons. (A) Behavioral responses of larvae of indicated genotypes to a stimulus of 38°C 24 hr after UV treatment. n = triplicate sets of 30 larvae per condition. Asterisk represents statistical significance (p < 0.05) versus ppk/+ control by Fisher’s exact test. (B) Response of larvae of indicated genotypes to a 45°C stimulus without UV and 8 hr after UV treatment. n = 50 larvae. Bracket with asterisk represents statistically significant comparison by Fisher’s exact test. Error bars represent SEM. See also Figure S1.

Figure 3. Blocking Hedgehog Signaling Components Does Not Affect Nociceptive Sensory Neuron Morphology or Baseline Nociception

(A and B) Whole mounts of dissected L3 larvae of the indicated genotypes. Sensory neuron morphology (ppk1.9Gal4, ppk-eGFP) is shown. (A) Control. (B) UAS-smo^IR. Scale bar represents 100 µm. (C) Quantification of total dendritic arbor size (see Experimental Procedures) when ppk1.9-Gal4 drives expression of the indicated UAS transgenes in nociceptive sensory neurons. n = 10 larvae per genotype. (D) Baseline nociception in response to a 45°C or 48°C stimulus in the absence of UV damage when ppk1.9-Gal4 drives expression of the indicated UAS transgenes in nociceptive sensory neurons. n = 50 larvae. Error bars represent SEM. See also Figures S2 and S3.

Figure 4. Epistasis Analysis of Tumor Necrosis Factor versus Hedgehog and Tumor Necrosis Factor or Hedgehog versus Transient Receptor Potential Channels

ppk1.9-Gal4 drives expression of the indicated UAS transgenes in nociceptive sensory neurons. (A and B) Constitutive activation of Hedgehog (Hh) signaling in the absence of UV irradiation produced thermal allodynia to a 38°C stimulus (A) and thermal hyperalgesia to a 45°C stimulus (B). (C) Constitutive activation of tumor necrosis factor (TNF) signaling causes allodynia that is reduced by knockdown of Wengen but not Smoothened. (D) Constitutive activation of Hh signaling causes allodynia that is reduced by knockdown of Smoothened but not Wengen. (E) RNAi directed against both Painless and TRPA1 reduce UV-induced thermal allodynia. (F) RNAi directed against TRPA1 prevents UV-induced thermal hyperalgesia whereas Painless RNAi-expressing larvae develop thermal hyperalgesia. Both comparisons are to the altered thresholds for baseline 45C in RNAi-expressing larvae. (G) Hh- and TNF-induced allodynia depend on Painless but not TRPA1. (H) Both Painless and TRPA1 reduce responsiveness to a 45°C stimulus in unirradiated larvae. Abbreviations are as follows: IR, inverted repeat; DN, dominant negative. n = triplicate sets of 30 larvae per condition. Error bars represent SEM.

Figure 5. Cyclopamine Modulates Morphine Analgesia in Rat Models of Inflammatory and Neuropathic Pain

(A) Inflammatory pain paradigm. Hindpaws were injected with saline (sham, n = 6), or complete Freund’s adjuvant (CFA). Thermal analgesia was assessed using paw withdrawal latency (PWL) following intrathecal administration of 10% Captisol (vehicle, n = 5), cyclopamine (CP, n = 6), morphine (MS, n = 6), or cyclopamine and morphine (MS + CP, n = 6). Abbreviations are as follows: BS, baseline before paw injection; 0, baseline after paw injection, prior to intrathecal drug administration. *p < 0.01 versus vehicle. (B) Neuropathic pain paradigm. Rats underwent sciatic nerve ligation or sham (n = 7) operation. Animals received daily intrathecal injections of vehicle (n = 3) or drugs (CP, n = 6; MS, n = 6; or MS + CP, n = 5). Mechanical allodynia was assessed using Von Frey filaments. Abbreviations are as follows: BS, baseline measurement prior to operation; 0, baseline 2 weeks after operation, prior to intrathecal drug administration. *p < 0.001 versus vehicle. All data are ±SEM. (C) Inflammatory pain paradigm with peripheral drug administration. Hindpaws were injected with CFA and cyclopamine, morphine, or a combination of both drugs. Animals received daily intraplantar injections of drugs. n = 9 animals per treatment group. Thermal analgesia was assessed using PWL. *p < 0.0001 versus vehicle (two-way repeated-measures ANOVA). All data are 6SEM.

Figure 6. Model of Hedgehog-Induced Thermal Allodynia and Hyperalgesia

Hh (from an as yet undetermined source tissue) acts through Patched (whose overexpression blocks sensitization), which subsequently inhibits Smoothened (required for sensitization) to activate signal transduction. The subsequent signaling steps likely involve the transcriptional activity of Cubitus interruptus (required for sensitization) and other components of the canonical Hh signaling pathway. For thermal allodynia, this pathway acts through the transient receptor potential (TRP) channel Painless, and for thermal hyperalgesia it requires the TRP channel dTRPA1.