Selective role for TRPV4 ion channels in visceral sensory pathways

Abstract

Background & aims: Although there are many candidates as molecular mechanotransducers, so far there has been no evidence for molecular specialization of visceral afferents. Here, we show that colonic afferents express a specific molecular transducer that underlies their specialized mechanosensory function: the transient receptor potential channel, vanilloid 4 (TRPV4).

Methods: We found TRPV4 mRNA is highly enriched in colonic sensory neurons compared with other visceral and somatic sensory neurons. TRPV4 protein was found in colonic nerve fibers from patients with inflammatory bowel disease, and it colocalized in a subset of fibers with the sensory neuropeptide CGRP in mice. We characterized the responses of 8 subtypes of vagal, splanchnic, and pelvic mechanoreceptors.

Results: Mechanosensory responses of colonic serosal and mesenteric afferents were enhanced by a TRPV4 agonist and dramatically reduced by targeted deletion of TRPV4 or by a TRP antagonist. Other subtypes of vagal and pelvic afferents, by contrast, were unaffected by these interventions. The behavioral responses to noxious colonic distention were also substantially reduced in mice lacking TRPV4.

Conclusions: These data indicate that TRPV4 contributes to mechanically evoked visceral pain, with relevance to human disease. In view of its distribution in favor of specific populations of visceral afferents, we propose that TRPV4 may present a selective novel target for the reduction of visceral pain, which is an important opportunity in the absence of current treatments.

Figure 1. Expression of TRPV4 in visceral afferents

A–C. Fluorescence in situ hybridization of TRPV4 mRNA expression (red) in nodose ganglia and dorsal root ganglia (DRG), combined with retrograde labelling of visceral sensory neurons from stomach (A), and colon (B and C) with CTB-FITC (green). Arrows indicate neurons showing retrograde labeling from the gut: blue arrows indicate those positive for TRPV4, and yellow arrows denote TRPV4-negative neurons. Graphs show proportion of neurons expressing TRPV4 in either whole DRG or in retrogradely labelled cells (p<0.001 vs general population, t-test). D. In situ hybridization showed the majority TRPV4 positive neurons were 15–20 μm in size (i), similar to the distribution of retrogradely labeled visceral afferents (CTB-FITC) in each ganglion (ii). These data indicate that differences in mRNA levels are not related to cell sizes. Most of these are in the C-fiber range, which was confirmed by conduction velocity (Fig 4). E. (i) Agarose gel electrophoresis of amplified RT-PCR products from three different ganglia from wild-type (+/+) and TRPV4 null-mutant (−/−) mice using primers specific for TRPV4 and β-actin; no = no RNA template added. (ii) Relative mRNA levels in each sensory ganglion following analysis of real-time PCR cycle thresholds. Thoracolumbar (TL) DRG had lowest levels of TRPV4 expression, which was significant compared with nodose ganglia and lumbosacral DRG (*p<0.05, **p<0.01 Mann-Whitney U test). (iii) In contrast colonic neurons within TL DRG showed the highest levels of relative expression – approximately 20-fold that in the whole TL DRG (p<0.001).

Figure 2. Colocalization of CGRP and TRPV4 in peripheral endings in mouse colon

A. i & ii: Fibers colocalizing CGRP and TRPV4 are restricted to the outer layers (arrows, ser=serosa), with the mucosa (muc) showing little selective label. v & vi: Section showing mesenteric blood vessels (bv) are flanked by fibers colocalizing CGRP and TRPV4 (arrows). iii & iv, vii & viii: Wholemounts of mesenteric blood vessels also showing colocalization of CGRP and TRPV4 in nerve fibers (arrows). Scale bars on sections=25μm, on wholemounts=250μm. B. i & ii: CGRP was present in dense networks of fibers within mucosal villi in a colonic section. TRPV4 does not co-localize with CGRP in mucosa. v & vi: CGRP is abundant in submucosa and external muscle layers of mouse colon, whereas TRPV4 is absent. iii & iv: CGRP+ fibers around myenteric ganglia in mouse colon wholemount do not colocalize TRPV4. Scale bars in B=250μm.

Figure 3. Selective deficits in mechanosensory function in TRPV4 null

A. Most splanchnic colonic afferents belonged to two classes based on location of receptive field - mesenteric and serosal afferents. Upper graphs: Both showed dramatically reduced stimulus-response functions to a static von Frey hair (vfh) in TRPV4−/− mice compared with +/+ (p<0.0001, 2-way ANOVA). Significance at individual stimulus intensities is shown by asterisks (* p<0.05, ** p<0.01, ***p<0.001, Bonferroni test). Lower graphs: activation thresholds to von Frey hairs were significantly increased for a greater percentage of both afferent subtypes. B. Four major classes of colonic pelvic afferents were recorded-serosal, mucosal, muscular and muscular/mucosal based on the location and responsiveness of their receptive fields. Pelvic serosal afferents showed similar deficits in TRPV4−/− to their splanchnic equivalents (i), whereas all other pelvic afferent subtypes were unchanged regardless of their adequate stimulus (ii–v). C. Two classes of gastroesophageal vagal afferents were observed-mucosal and tension receptors. Neither showed any significant change (NS=not significant) in mechanosensitivity in TRPV4−/− mice.

Figure 4. TRPV4 null-mutants have normal afferent biophysical properties

A. Splanchnic colonic mesenteric (i) and serosal (ii) afferents show similar conduction velocities. B. Both populations also had similar activation thresholds to electrical stimuli delivered by a concentric electrode placed on their receptive fields.

Figure 5. Original records showing effects of TRPV4 ligands on colonic afferent mechanosensitivity

Upper traces show instantaneous frequency plots, and lower traces show original recordings of neural activity before (i) and after (ii) incubation with ligands. A. The TRPV4 agonist 5,6-EET increased mechanical sensitivity of a splanchnic mesenteric afferent fiber to a 2g probe. B. The non-selective TRP channel antagonist Ruthenium red caused inhibition of mechanosensitivity to this stimulus in a different fiber.

Figure 6. TRPV4 ligands selectively modulate mechanosensory function

A. Splanchnic colonic mesenteric afferents show increased responses to static application of a 2g von Frey hair in the presence of the TRPV4 agonist 5,6-EET (i), which is lost in the TRPV4 null-mutant (ii). Conversely, the non-selective TRP blocker ruthenium red concentration-dependently inhibits mechanosensory responses (iii)– an effect that is also lost in the TRPV4 null-mutants (iv). Data not shown indicate similar effects on mesenteric endings. B. Pelvic colonic serosal afferents show a similar pattern of potentiation by 5,6-EET and inhibition by ruthenium red, which was also lost in the TRPV4 null-mutant. C. No significant effect of TRPV4 ligands was seen in vagal tension receptors. Data not shown indicate similar lack of effect of these compounds on pelvic mucosal, muscular, and muscular/mucosal endings. *P<0.01, **P<0.001 vs control, repeated measures ANOVA with Dunnett’s post-hoc test.

Figure 7. Translational studies of TRPV4 function and expression

A. Abdominal EMG responses of conscious mice to colorectal balloon distensions (5x80mmHg) were reduced by 55% in TRPV4 null-mutants (N=7, p<0.05). B. 14μm section of colon from a patient with Crohn’s disease (CD) showing serosal blood vessel encircled by TRPV4-immunopositive fiber. Punctate TRPV4 labelling is also evident in other areas of the serosa. C. Varicose TRPV4-immunopostive ending in longitudinal muscle within 100μm of serosal border in Crohn’s disease. This was seen in only one of 5 CD patients. D. Section of serosa/mesentery of a CD patient approximately 2mm from colon, showing 2 blood vessels surrounded by TRPV4-positive fibers. E. Circular and longitudinal muscle layers were TRPV4-negative. ser=serosa, lm=longitudinal muscle, cm=circular muscle, bv=serosal blood vessel. Scale bars B-D=40μm, E=500μm.