Delta opioid receptors presynaptically regulate cutaneous mechanosensory neuron input to the spinal cord dorsal horn

Abstract

Cutaneous mechanosensory neurons detect mechanical stimuli that generate touch and pain sensation. Although opioids are generally associated only with the control of pain, here we report that the opioid system in fact broadly regulates cutaneous mechanosensation, including touch. This function is predominantly subserved by the delta opioid receptor (DOR), which is expressed by myelinated mechanoreceptors that form Meissner corpuscles, Merkel cell-neurite complexes, and circumferential hair follicle endings. These afferents also include a small population of CGRP-expressing myelinated nociceptors that we now identify as the somatosensory neurons that coexpress mu and delta opioid receptors. We further demonstrate that DOR activation at the central terminals of myelinated mechanoreceptors depresses synaptic input to the spinal dorsal horn, via the inhibition of voltage-gated calcium channels. Collectively our results uncover a molecular mechanism by which opioids modulate cutaneous mechanosensation and provide a rationale for targeting DOR to alleviate injury-induced mechanical hypersensitivity.

Figure 1. DORGFP mouse faithfully reports DOR expression in CNS and DRG, where DOR is predominantly expressed by large-diameter, myelinated (NF200+) neurons

(A) Co-staining with anti-GFP and -NF200 antibodies shows that 67% of DORGFP+ DRG neurons express NF200. (B) Oprd1 mRNA is also predominantly present in NF200+ neurons in DRG from wildtype mice. (C) Control showing that in the absence of the specific probe against Oprd1 sequence there is no in situ hybridization signal in DRG neurons. (D) Substance P+ (SP+) cells almost never contain Oprd1 mRNA in DRG neurons from wildtype mice. (E) The great majority of NF200− cells that contain Oprd1 mRNA bind the isolectin B4 (IB4). (F) Quantification of A–E and of A–B from Figure S1. The left panel indicates neurons that coexpress DOR among those that express SP, or bind IB4, or express NF200. The right panel indicates DOR+ neurons that coexpress NF200, and DOR+ NF200− neurons that coexpress SP or bind IB4. (G) The distribution of fluorescence in brain slices from DORGFP mice is identical to the binding pattern of DOR radioligands [³H]-Deltorphin II and [³H]-DPDPE in wildtype mice. Radiolabeling is lost in Oprd1 null mice, demonstrating specificity. (H) In wildtype mouse spinal cord [³H]-Deltorphin II binds to DORs that are present throughout the grey matter, consistent with DORGFP fluorescence pattern. Data are represented as means ± SEM. Scale bars equal 50 μm. See also Figure S1.

Figure 2. DOR-expressing myelinated DRG neurons express the markers of Aβ low-threshold mechanoreceptors TrkC and Ret

(A) TrkC is expressed by about 62% of NF200+ DORGFP+ DRG neurons. Arrows indicate cells where costaining occurs. (B) Quantification of (A) and (C)–(I). (C) Proprioceptors, identified by the coexpression of parvalbumin and Runx3, do not express DORGFP. Note that some DORGFP+ neurons individually express either parvalbumin or Runx3. (D) Approximately 70% of DORGFP+ TrkC+ neurons coexpress Ret (arrows) indicating that these neurons are likely myelinated cutaneous mechanoreceptors, including Aβ LTMRs. (E) About 40% of large-diameter myelinated (NF200+) DORGFP+ neurons (arrows) and almost all small-diameter DORGFP+ neurons express Ret. (F) Approximately 36% of NF200+ DORGFP+ neurons express CGRP (arrows). (G) Large-diameter DORGFP+ DRG neurons do not express the noxious heat sensor TRPV1. A very small number of small-diameter neurons coexpress DOR and TRPV1 (arrows). (H) Large-diameter DORGFP+ neurons that express CGRP also immunostain for the nociceptor marker TrkA (arrows). (I) Almost all large-diameter DORGFP+ neurons express either TrkC or TrkA. (J) Molecular identity of myelinated (NF200+) DORGFP+ neuron subpopulations. Data are represented as means ± SEM. Scale bars equal 50 μm.

Figure 3. Myelinated DRG neurons that express MOR coexpress CGRP, but not markers of Aβ low-threshold mechanoreceptors

(A) Virtually all MOR+ myelinated neurons (NF200+) express CGRP (arrows indicate NF200+ MOR+ neurons), indicating that these neurons are TrkA+ nociceptors (see Figure S2C). (B) NF200+ MOR+ neurons (arrows) rarely express TrkC. (C) The great majority of NF200+ MOR+ neurons express neither Ret nor TrkC (see (B)), indicating that they are neither proprioceptors nor Aβ LTMRs. (D) A small subpopulation of NF200+ MOR+ neurons also expresses TRPV1 (arrows). (E) MOR+ neurons do not express the D-hair receptor marker TrkB. (F) Quantification of (A)–(E). (G) Molecular identity of myelinated (NF200+) MOR+ neuron subpopulations. Data are represented as means ± SEM. Scale bars equal 50 μm.

Figure 4. DOR and MOR are colocalized at the plasma membrane of CGRP+ A nociceptors

(A) Stereotaxic injection of the B fragment of cholera toxin (CTB) into the dorsal column nuclei (DCN) labels Aβ LTMR cell bodies in lumbar DRGs. Only DORGFP+ DRG neurons display CTB immunoreactivity (22.9% ± 2.9%), while MOR+ (0.0% ± 0.0%) and DORGFP+ MOR+ cells do not (n=294 CTB+ cells). (B) DORGFP and MOR are expressed by largely non-overlapping populations of DRG neurons. Some NF200+ neurons express only DOR (white arrowhead) or MOR (black arrowhead); a third population coexpresses both receptors (arrow). (C) Quantification of (B) shows that 87.8% ± 3.9% of the DRG neurons that coexpress DOR and MOR are NF200+ (bar graph) and that 6.2% ± 1.1% of all NF200+ DRG neurons coexpress both receptors (pie-chart) (n=1091). DOR predominates in NF200+ neurons (67.3% ± 4.3%), MOR in NF200− neurons (79.8% ± 1.9%) (bar graphs). (D) DORGFP and MOR are coexpressed at the plasma membrane of large-diameter CGRP+ DRG neurons, which are most often mechanonociceptors (arrow) (Lawson et al., 2008). White arrowhead points to a large-diameter DORGFP+ CGRP− MOR− DRG neuron that likely belongs to the TrkC+ population described in Figure 1. Black arrowhead illustrates that all large-diameter MOR+ DRG neurons, and the vast majority of MOR+ small-diameter cells, coexpress CGRP. (E) All DORGFP+ MOR+ large-diameter DRG neurons coexpress CGRP (100.0% ± 0.0%) and this population corresponds to 43.3% ± 7.7% of all CGRP+ large-diameter DRG neurons (n=229). Data are represented as means ± SEM. Scale bars equal 50 μm.

Figure 5. The terminals of DORGFP-expressing DRG neurons form touch-detecting mechanosensory organs in the skin

(A) Peripheral axons of DORGFP+ DRG neurons (green, arrowheads) innervate cytokeratin 8 (CK8)-immunoreactive Merkel cells grouped in touch domes associated with guard hairs (autofluorescent hair shaft, white arrow). (B) DORGFP+ terminals form circumferential, but not longitudinal, endings (arrowheads) around follicles of most pelage hairs, including guard hairs (arrow, center of the image). (C) DORGFP+ axons end as NF200+ flattened disks (arrowheads) surrounding the base of Merkel cells (arrows). (D) DORGFP+ axons that form circumferential endings around follicles are myelinated (NF200+) and are distinct from CGRP+ terminals. Innervation of a guard hair is shown. (E) DORGFP+ NF200+ axons form Meissner corpuscles in the dermal papillae of the mouse footpad. The Meissner corpuscles are localized by immunoreactivity of the glial protein S100 (arrows). DORGFP+ NF200− axons penetrate further into the epidermis (white arrowheads), consistent with their nonpeptidergic MrgprD+ identity. Black arrowhead indicates the boundary between stratum corneum (autofluorescent dead cells) and stratum granulosum. (F) Myelinated DORGFP+ afferents with epidermal free nerve endings often coexpress CGRP (arrowheads). Circumferential endings around the follicle of the small caliber hairs (arrow) express DORGFP and NF200, but not CGRP. Autofluorescence generates the fluorescence of the hair shafts (yellow in (A), green in (B) and (D)).

Figure 6. DORGFP-expressing DRG neurons are mechanosensitive and include high- and low-threshold Aδ mechanonociceptors

(A) The conduction velocity (CV, 6.6 m/s), broad action potential (bottom left), high mechanical threshold (10.5 mN, bottom right), and activity-induced hyperpolarization of the DORGFP+ neuron (green; red: neurobiotin) in (A) are characteristics of high-threshold Aδ mechanonociceptors. (B) In comparison to (A) the DORGFP+ Aδ fiber recorded in (B) exhibits a relatively narrow action potential and low mechanical threshold (0.07 mN, similar to LTMRs), indicating that this myelinated nociceptor might also participate in touch sensation. (C) Conduction velocity (7.3 m/s ± 0.3 m/s) and mechanical threshold (1.5 mN ± 0.7 mN) of 20 DORGFP+ DRG neurons show that most neurons that project to the trunk skin are low-threshold Aδ mechanonociceptors (see example in (B)). These cells (blue symbols in right panel) generally have large receptive fields composed of multiple isolated spots, while the other, comparatively high-threshold, population of DORGFP+ mechanonociceptors (red; see example in (A)) tends to have smaller receptive fields composed of a single spot. Data are represented as means ± SEM.

Figure 7. Presynaptic DORs depress Aβ fiber input to laminae III–V spinal neurons

(A) Voltage-clamp recordings of laminae III–V neurons in spinal cord slices from wildtype mice show that the DOR agonist deltorphin II (500 nM) did not significantly change miniature EPSC (mEPSC) frequency or amplitude (n=15). (B) Example of EPSCs evoked in a lamina III spinal neuron by stimulation of monosynaptically connected Aβ fibers using a paired pulse protocol. Deltorphin II (500 nM) depressed EPSC peak amplitude and altered the ratio between the amplitudes of the 2^nd and 1^st peaks. (C) Time course of the deltorphin II inhibitory effect on EPSC peak amplitudes. (D) Deltorphin II (500 nM) significantly reduced the first peak amplitude and increased the paired pulse ratio in 10 of the 17 recorded cells (responsive neurons). (E) GDPβS (1 mM in internal solution) did not block deltorphin II’s effect (8/14 recorded neurons with GDPβS responded to deltorphin II, a proportion similar to (D)), which confirms a presynaptic mechanism of action. (F) The DOR antagonist naltrindole (1 μM) prevented deltorphin II’s effect in all recorded neurons (n=9). (G) Summary showing all recorded cells. Red, deltorphin II responsive neurons; black, non-responsive neurons; blue, deltorphin II responsive neurons with GDPβS; brown, non-responsive neurons with GDPβS; green, cells recorded in the presence of naltrindole. The dotted line represents the threshold value (−10%) of percent change of 1^st peak amplitude at which cells were considered responsive. (H) Average deltorphin II effect on 1^st peak amplitude. Color code as in (G). (I) Average deltorphin II effect on paired pulse ratio. Color code as in (G). Data are represented as means ± SEM, * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 8. DOR-mediated inhibition of voltage-gated calcium channels in cultured large-diameter DRG neurons and functional organization of DOR and MOR in cutaneous somatosensory neurons

(A) Large-diameter NF200+ DORGFP+ cells express TrkC in primary DRG culture (arrows). (B) DOR-expressing large-diameter DRG neurons (brightfield image at right) are identified by direct visualization of DORGFP fluorescence (arrow, left image). (C) Deltorphin II (500 nM) reduces amplitude and increases rise time of voltage-activated calcium currents in DORGFP+ DRG neurons, but not in DORGFP- DRG neurons. (D) Average effect of deltorphin II on current amplitude in DORGFP+ (n=11) and DORGFP− large-diameter neurons (n=7). (E) Average effect of deltorphin II on current rise time in DORGFP+ (n=11) and DORGFP− large-diameter neurons (n=7). Results are expressed as mean ± SEM, *** p < 0.001. (F) Labeling of IB4+ neurons in DRG culture from a wildtype mouse. (G) In the DRG culture shown in (F) deltorphin II (500 nM) predominantly inhibited VGCCs in large-diameter neurons (> 40 pF), while DAMGO (500 nM) predominantly inhibited VGCCs in small-diameter (< 40 pF) IB4− neurons. (H) Occurrence of deltorphin II- and DAMGO-mediated inhibition of VGCCs in 44 randomly selected neurons in DRG cultures from wildtype mice. Only one cell out of 29 tested for deltorphin II- and DAMGO-responsiveness (*) showed VGCC inhibition by both drugs. (I) Schematic of the functional organization of DOR and MOR in cutaneous somatosensory neurons. DOR is expressed by populations of TrkC+ and/or Ret+ myelinated mechanoreceptors, which include SAI and RA Aβ LTMRs that form the Merkel cell-neurite complex (yellow) and Meissner corpuscles, respectively. DOR is also expressed by myelinated mechanonociceptors (AM), some of which are Aδ fibers, have low threshold, and respond to innocuous mechanical stimuli. The axons derived from DOR+ TrkC+ AM Aδ mechanonociceptors, which often coexpress Ret, may form circumferential endings around hair follicle, although this remains speculative. Additionally, DOR is expressed by Ret+ MrgprD+ C mechanonociceptors (CM) that innervate epidermal stratum granulosum. By contrast MOR is predominantly expressed by capsaicin- and heat-sensitive TRPV1+ C nociceptors (CH) (many of which are visceral, not cutaneous, afferents) and regulates heat pain perception. Colocalization of DOR and MOR occurs in myelinated TrkA- and CGRP-expressing neurons, which are presumptive high-threshold Aδ or Aβ mechanonociceptors. Most TH+ C LTMRs, TrkB+ D-hair mechanoreceptors, and Ret+ Aβ LTMRs that form longitudinal endings around hair follicles do not appear to express DOR or MOR. Data are represented as means ± SEM.