Keratinocyte-TRPV1 sensory neuron interactions in a genetically controllable mouse model of chronic neuropathic itch

Abstract

Our understanding of neural circuits that respond to skin dysfunction, triggering itch, and pathophysiological scratching remains incomplete. Here, we describe a profound chronic itch phenotype in transgenic mice expressing the tetracycline transactivator (tTA) gene within the Phox2a lineage. Phox2a; tTA mice exhibit intense, localized scratching and regional skin lesions, controllable by the tTA inhibitor, doxycycline. As gabapentin and the kappa opioid receptor agonist, nalfurafine, but not morphine, significantly reduce scratching, this phenotype has a pharmacological profile of neuropathic pruritus. Importantly, the Phox2a; tTA expression occurs in a spatially restricted population of skin keratinocytes that overlaps precisely with the skin area that is scratched. Localized G_i-DREADD-mediated inactivation of these Phox2a-keratinocytes completely reverses the skin lesions, while inducible tTA activation of keratinocytes initiates the condition. Notably, ablation of TRPV1-expressing primary afferent neurons also reduces scratching and skin lesions, but this occurs slowly, over a course of two months. In contrast denervation induced loss of all cutaneous input rapidly blocks scratching. These findings identify the cellular, molecular, and topographic basis of a robust and chronic sensory neuron-dependent and gabapentin-responsive neuropathic itch that is initiated by genetic factors within keratinocytes.

Keywords: DREADDs; TRPV1 sensory neurons; chronic itch; keratinocytes; neuropathic itch.

Fig. 1.

Constitutive tTA-GCaMP6 transgene expression in the Phox2a+ lineage provokes a neuropathic itch condition. (A) Transgenic breeding scheme to express GCaMP6 in the Phox2a lineage, using Phox2a-Cre mice and TIGRE 2.0 reporter mice. TIGRE 2.0 mice, specifically of the Ai162 line, coexpress two genes, tTA2 and GCaMP6s, after Cre-mediated recombination. (B) Analysis of spontaneous scratching in Phox2a-Cre; tTA-GCaMP6 mice. Scratching becomes obvious at 5 wk of age and increases in frequency thereafter. Phox2a-Cre+; tTA-GCaMP6 (N = 13), Cre-; tTA-GCaMP6 (N = 15). GCaMP6s and 6f data are compiled in Fig. 1B; see SI Appendix, Fig. S2A for a direct comparison of the 6s vs. 6f GCaMP variants. (C) Representative images of bilateral, shoulder-localized erosive skin lesions in a 12 wk Phox2a; tTA-GCaMP6s mouse. (D) Weekly hind paw nail trimming reduces the size of skin lesions in Phox2a; tTA-GCaMP6 mice. Trimmed (N = 5), Untrimmed (N = 4). (E) Scratching provoked by intradermal injection of chloroquine at the shoulder in Phox2a-Cre+; tTA-GCaMP6 mice compared to Cre- littermate control mice. Data points represent individual mice. For (F–I), lines connect paired data points of the same mouse 24 h apart. Intraperitoneal injection of (F) Gabapentin 30 mg/kg and (I) Nalfurafine 20 μg/kg, compared to an injection of the drug vehicle, reduced scratching bouts in Phox2a-Cre; tTA-GCaMP6 mice. However, no difference in scratching bouts after ip. injection of (G) morphine 10 mg/kg or (H) cetirizine 30 mg/kg in Cre+ mice. (J) Scratching provoked by intradermal injection of chloroquine at the cheek in Cre+; tTA-GCaMP6 mice compared to Cre- littermate control mice. Data points represent individual mice. Statistics were analyzed by two-tailed unpaired (E and J) or paired (F–I) t test. Error bars indicate the 95% CI. *P < 0.05, **P < 0.01, ns = nonsignificant.

Fig. 2.

tTA independently controls the scratching phenotype in Phox2a; tTA-GCaMP6 mice. (A) The genetic configuration of the Ai96 (GCaMP6s) line, without tTA transcriptional control, compared to two tTA-driven lines: Ai162 (GCaMP6s) and Ai148 (GCaMP6f). (B) Comparison of spontaneous scratching frequency at 11 wk of age across the indicated genotypes crossed with Phox2a-Cre or control mice without Cre. Colored dots represent individual mice. Ai9 = Cre-dependent CAG-tdTomato reporter used for fate mapping studies. (C) The genetic configuration of Ai166 is the same as Ai162 except that, instead of GCaMP6, it encodes GFP which is expressed in a random subset of Cre+ cells. (D) A maximum intensity projection image of a SHIELD-cleared, whole-mount spinal cord from a Phox2a-Cre; Ai166 mouse shows sparse, cellular GFP expression. Specifically, the cleared cervical enlargement is shown, and cervical spinal segments are annotated as (C4–C7). Inset provides a magnified image of a single GFP neuron. (E) Image of a Phox2a-Cre; Ai166 mouse showing a localized skin lesion phenotype similar in location to Phox2a; Ai162 mice. (F) Doxycycline 200 mg/kg, administered through the chow, inhibits tTA activity through direct binding, arresting its transcriptional activity. (G) Spinal cord histology confirms the downregulation of GCaMP6s in dorsal horn projection neurons of mice on Doxycycline chow. (H) Doxycycline chow provided during pregnancy and after weaning prevents the development of scratching. Dox diet “On” (N = 5), Standard diet “Off” (N = 9). (I) Doxycycline started in adulthood partially reduces spontaneous scratching. Dox On (N = 4), Dox Off (N = 5). Error bars indicate the 95% CI.

Fig. 3.

Fate mapping identifies keratinocytes within the Phox2a lineage as the topographical determinant of scratching. (A) Quantification showing comparable numbers of tTA+ and tTA− (“wild-type”) neurons in the cervical spinal cord for both Phox2a-Cre and Phox2b-Cre lines. Ai9 = Cre-dependent CAG-tdTomato reporter used to visualize neurons. (B) Phox2b-Cre; tTA-GCaMP6 mice do not exhibit the scratching phenotype observed in Phox2a-Cre; tTA-GCaMP6 mice. (C) Schematic of the Phox2b-Cre fate map (annotated in red) in the nervous system based on Cre-dependent reporter expression. Phox2b-Cre labels many of the same populations of neurons as Phox2a-Cre (compare to E). Purple color specifically annotates the XI motoneurons. (D) Whole-body cross-section histology of a 7-d-old Phox2a-Cre; Ai9 mouse reveals tdTomato reporter expression in the skin surrounding the shoulder but not in deeper soft tissues, except the sympathetic ganglia (SG). Red boxes define the boundary of the skin where reporter expression is observed, which is bilateral. The left box outlines the Left inset image. (E) The Phox2a-Cre lineage includes both neuronal populations and skin keratinocytes (approximate expression boundary in the skin is indicated by red ovals). (F) Visualized by tdTomato pseudocolored in white, a Phox2a; Ai9 embryo reveals a striking topography of reporter expression outlined in red. The red dotted line defines the boundary of Cre-mediated recombination at the neck. (G) Monochrome image showing a hairless skin lesion in an adult Phox2a-Cre; Ai9; Ai162 mouse (Left). Fluorescent imaging reveals GCaMP6s (green) and tdTomato (red) signals throughout the lesioned area but not in the surrounding skin (Right). The red dotted line indicates the neck boundary of recombination. (Scale bar, 500 μm.) Statistics were analyzed using two-tailed unpaired t test. Error bars represent 95% CI. ****P < 0.0001, ns = nonsignificant.

Fig. 4.

Skin pathology can be controlled by the selective manipulation of keratinocytes or by the ablation of TRPV1 sensory neurons. (A and B) Unilateral 4-OHT application to the skin and validation of GCaMP6 and tdTomato reporter expression at one month after treatment. A hairless area (outlined in yellow) is typical of prodromal lesion sites. (C–K) Three experiments are detailed horizontally, one experiment per row. (C, F, and I) Experimental scheme and timeline. (D, G, and J) Example images of the three differing treatments after measurement. (E, H, and K) Scratching and skin lesion analysis of each experiment. Connected lines in (E and H) are measurements from either the Left or Right side of an individual animal. As RTX-mediated ablation is irreversible, the experiment in (I–K) was compared across groups of mice. Accordingly, statistics were analyzed by two-tailed paired (E and H) or unpaired (K) t test. *P < 0.05, **P < 0.01, ***P < 0.001.