Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jun 2.
Published in final edited form as: Cell. 2015 Dec 17;163(7):1783–1795. doi: 10.1016/j.cell.2015.11.060

Genetic identification of an expansive mechanoreceptor sensitive to skin stroking

Ling Bai 1,2,#, Brendan P Lehnert 1,#, Junwei Liu 3, Nicole L Neubarth 1, Travis L Dickendesher 1, Pann H Nwe 1, Colleen Cassidy 4, C Jeffery Woodbury 5, David D Ginty 1
PMCID: PMC4890169  NIHMSID: NIHMS787587  PMID: 26687362

Summary

Touch perception begins with activation of low-threshold mechanoreceptors (LTMRs) in the periphery. LTMR terminals exhibit tremendous morphological heterogeneity that specifies their mechanical receptivity. In a survey of mammalian skin, we found a preponderance of neurofilament-heavy chain+ circumferential endings associated with hair follicles, prompting us to develop a genetic strategy to interrogate these neurons. Targeted in vivo recordings revealed them to be Aβ Field-LTMRs, identified 50 years ago but largely elusive thereafter. Remarkably, while Aβ Field-LTMRs are highly sensitive to gentle stroking of the skin, they are unresponsive to hair deflection, and they encode skin indentation in the noxious range across large, spotty receptive fields. Individual Aβ Field-LTMRs form up to 180 circumferential endings, making them the most anatomically expansive LTMR identified to date. Thus, Aβ Field-LTMRs are a major mammalian LTMR subtype that forms circumferential endings in hairy skin, and their sensitivity to gentle skin stroking arises through integration across many low-sensitivity circumferential endings.

Introduction

Sensitivity to mechanical stimuli is crucial for communication and survival of all organisms. In the case of mammalian skin, innocuous mechanical stimuli are detected by a diverse group of cutaneous low-threshold mechanoreceptor (LTMR) subtypes with distinct morphologies, physiological properties and functions. LTMRs are pseudo-unipolar sensory neurons whose cell bodies reside within dorsal root ganglia (DRG) or trigeminal ganglia. Their touch sensitivity derives from mechanosensitive peripheral axonal terminals that reside in the skin, where they associate with end organs that determine the geometry of their axonal terminals and force sensitivity (Loewenstein, 1969; Zimmerman et al., 2014). Touch stimuli are encoded and further transferred via axons to the central nervous system (CNS), where LTMR inputs carrying distinct aspects of touch information converge and integrate (Abraira and Ginty, 2013; Lechner and Lewin, 2013).

In mammalian hairy skin, at least seven LTMR subtypes tuned to distinct but overlapping features of a tactile stimulus have been described (Abraira and Ginty, 2013; Horch et al., 1977). Each subtype forms specialized axonal terminals associated with end organs. Aβ RA-LTMRs, Aδ-LTMRs and C-LTMRs form lanceolate endings, which extend along the long axis of hair follicles and are rapidly or intermediately adapting to skin indentation (Li et al., 2011). Aβ SA1-LTMRs form disc-like axonal terminals that associate with Merkel cells, which are located in the basal epidermis around guard hair follicles and are themselves mechanically sensitive, endowing Aβ SA1-LTMRs with their characteristic slowly adapting response property (Ikeda et al., 2014; Maksimovic et al., 2014; Woo et al., 2014). In contrast, Aβ RA2-LTMRs terminate deep in the dermis where they are wrapped by layers of cushion-like lamellar cells within Pacinian corpuscles; these serve as high pass mechanical filters that underlie the unique Aβ RA2-LTMR high frequency vibration tuning property. Aβ SA2-LTMRs also terminate in the dermis and may form Ruffini endings (Chambers et al., 1972). The least understood LTMR subtype, the Aβ Field-LTMR (formerly called field receptor), first described nearly 50 years ago in the cat by Burgess and colleagues, is rapidly or intermediately adapting to skin indentation (Burgess et al., 1968). These neurons are sensitive to gentle stroking of the skin but rarely respond to deflection of individual hairs (Horch et al., 1977). The terminal structure and skin arborization patterns of Aβ Field-LTMRs have not been determined, leaving open the question of how their distinct physiological properties arise.

All LTMRs have a central axonal projection that innervates the spinal cord dorsal horn. Undirected recordings indicate that five Aβ-LTMR subtypes, including the Aβ Field-LTMR, also have an axonal branch that ascends via the dorsal column (DC) to the dorsal column nuclei (DCN) in the brainstem (Horch et al., 1976), and these projections are considered important for the perception of discriminative touch (Ballermann et al., 2001; Dobry and Casey, 1972). Here we have taken an unbiased retrograde labeling approach to visualize the cutaneous endings of Aβ-LTMRs with axons that ascend the DC to the DCN of mice. These experiments guided molecular-genetic strategies to label the Aβ-LTMRs for which genetic tools do not currently exist, specifically Aβ SA1-LTMRs as well as large myelinated cutaneous sensory neurons that give rise to circumferential endings associated with hair follicles. Importantly, physiological recordings from genetically labeled circumferential ending neurons reveal them to display key hallmarks of the elusive Aβ Field-LTMRs. We found that in comparison to other hairy skin LTMRs, this LTMR subtype exhibits striking peripheral receptive field properties, unique ultrastructural features, characteristic LTMR central innervation patterns, and highly distinct tuning properties. Our findings establish the morphological and physiological properties of the Aβ Field-LTMR and support a model in which the weak mechanosensitivity and expansive organization of its circumferential endings explain this neuron’s sensitivity to gentle stroking of the skin.

Results

Identification of cutaneous sensory neurons that innervate the DCN

To visualize the cutaneous terminal morphologies of uncharacterized DCN-projecting Aβ-LTMR subtypes, we designed an unbiased strategy to selectively label neurons exhibiting direct DCN projections. AAV2/1-Cre virus was injected bilaterally into the rostral DC at the first cervical level of R26LSL-tdTomato mice to retrogradely infect DRG neurons and visualize their cutaneous axonal projections (Figure 1A). TdTomato+ DRG neurons have large cell bodies and are neurofilament-heavy chain+ (NFH+) (Figure S1A), but they do not bind to the lectin IB4 and very few express CGRP (Figures S1C and S1D). These findings indicate that our retrograde labeling strategy is specific because small or medium sized DRG sensory neurons, including C-LTMRs, Aδ-LTMRs and most nociceptors, do not have an axonal branch that projects directly to the brainstem.

Figure 1. Genetic labeling of sensory neurons that innervate the dorsal column nuclei.

Figure 1

Open in a new tab

(A-D) The C1 DC of R26LSL-tdTomato mice was injected with AAV2/1-Cre virus to retrogradely label neurons that project to the DCN. (B-D) In hairy skin, at least 4 types of terminals are labeled by tdTomato: an unknown ending type innervating S100+ terminal organs that resemble Pacinian corpuscles (B), Merkel endings from Aβ SA1-LTMRs that associate with Troma1+ Merkel cells (C), lanceolate endings that are NFH+ (97.8 ± 1.2%, 156 terminals, n = 3) and derived from Aβ RA-LTMRs (D, asterisk), as well as circumferential endings that are also NFH+ (99.5% ± 0.5%, 465 terminals, n = 3) (D, arrowhead). (E) NFH+ circumferential endings, NFH+ lanceolate endings, and Merkel endings innervate 94 ± 3%, 8 ± 2%, and 0.8 ± 0.2% of hair follicles, respectively (3964 hair follicles, n = 3). (F-I) Immunostaining of NFH and Tuj1 on hairy skin sections reveals that NFH+ circumferential endings (arrowhead) can be found in mouse (F), cat (G), dog (H), and macaue(I). (J) Diagram of the TrkC and Ret intersectional genetic labeling strategy. (K) Double immunostaining of hairy skin sections from TrkCCreER; Retf(CFP) mice treated with 0.1 tamoxifen at P5 reveals that CFP specifically labels the majority of NFH+ circumferential endings (79.6 ± 3.3%, 1142 terminals, n = 4) and all of CFP+ circumferential endings are NFH+ (99.2 ± 0.2%, 904 terminals, n = 4). (L) Whole-mount immunostaining of hairy skin from TrkCCreER; Retf(CFP) mice treated with 3mg tamoxifen at E12.5 reveals that CFP specifically labels Aβ SA1-LTMRs innervating Troma1+ Merkel cells (40 ± 4%, 78 terminals, n = 3). (M) Quantification of the percentage of sensory endings labeled by CFP in K-L. Scale bar: 50 μm (B-D, F-I, K-L). See also Figure S1, S2, and S3.

We next visualized the cutaneous terminal morphologies of tdTomato+ neurons. In hairy skin, tdTomato+ neurons form four types of endings. Close to the epidermis and between hair follicles, tdTomato+ axonal terminals were observed wrapped by S100+ Schwann cells (Figure 1B) in structures that resemble Pacinian corpuscles (Luo et al., 2009); however, unlike Pacinian corpuscles, these tdTomato+ endings do not reside within the deep dermis. This is a curious neuronal population with unknown function. The other three types of tdTomato+ axons observed form terminals associated with hair follicles. At the neck of guard hair follicles, tdTomato+ axonal endings associate with Troma1+ Merkel cells (Figure 1C); these tdTomato+ neurons are Aβ SA1-LTMRs. A third population forms NFH+ lanceolate endings (Figure 1D); these neurons are Aβ RA-LTMRs. A fourth class of tdTomato+ DRG neurons displays circumferential endings that also encircle hair follicles below the sebaceous gland (Figure 1D). Similar to Aβ SA1-LTMR and Aβ RA-LTMR endings, the majority of tdTomato+ circumferential endings are NFH+ (Figure 1C-1D), suggesting these neurons have myelinated axons with conduction velocities (CVs) in the Aβ range. Whole-mount staining of back hairy skin revealed that NFH+ circumferential endings innervate 94% of hair follicles, whereas Aβ RA-LTMRs and Aβ SA1-LTMRs innervate only 8% and 0.8% of hair follicles, respectively (Figure 1E). Thus, we have identified a DCN-projecting DRG neuronal population with a striking preponderance of NFH+ circumferential endings in hairy skin of mice.

Circumferential endings were first described in rodents and cats in the 1980s (Millard and Woolf, 1988; Rice and Munger, 1986). Just as there are multiple sensory neuron subtypes with lanceolate endings, experiments examining expression patterns of molecular markers showed there are multiple neuronal types with circumferential endings (Fünfschilling et al., 2004). Indeed, we found that in mouse hairy skin, there are only two molecularly distinct types of circumferential endings surrounding hair follicles, one expressing CGRP and the other expressing NFH (Figure 1E, S1E-S1H). Furthermore, all circumferential endings labeled by DCN retrograde labeling are NFH+ and CGRP- (Figure 1D and S1I); thus, only neurons with NFH+ circumferential endings project directly to the DCN. In a multi-species survey of cutaneous sensory neurons, we found that NFH+ circumferential endings are also present in cats, dogs, and macaques (Figure 1F-I), indicating that these neurons are prevalent amongst mammals.

Intersectional genetic strategies to selectively label neurons with NFH+ circumferential endings and Aβ SA1-LTMRs

The anatomical, electrophysiological, and functional properties of DRG neurons with circumferential endings surrounding hair follicles are unknown. We therefore sought to generate molecular genetic tools that enable investigation of these neurons. NT3-TrkC signaling mediates development of NFH+ circumferential endings (Albers et al., 1996) and half of DCN-projecting DRG neurons are TrkC+ (Figure S1B). Moreover, TrkC+ DRG neurons of adult mice can be subdivided into three mutually exclusive populations: Ret+ neurons; Parvalbumin+ (PV+) proprioceptors; and CGRP+ peptidergic nociceptors (Figure S3A and S3B). Thus, we tested the hypothesis that the TrkC+/Ret+ population includes neurons with NFH+ circumferential endings.

We first generated a TrkCtdTomato knock-in mouse line (Figure S2A and S2B), in which the pattern of tdTomato expression recapitulates that of TrkC (Figure S2C). In hairy skin, tdTomato is found in all NFH+ circumferential endings (Figure S2D), all Aβ SA1-LTMR endings associated with Merkel cells, and a subset of free nerve endings (Figure S2E). Importantly, the NFH+ circumferential endings and Merkel endings labeled by TrkCtdTomato are also RetCFP+ (Figure S2F, S3C, and S3D).

To selectively label the TrkC+/Ret+ DRG neuronal populations, we designed an intersectional genetic labeling strategy. For this, we generated a TrkCCreER knock-in mouse line (Figure S2A and S2B). When TrkCCreER mice were crossed with Retf(CFP) mice (Uesaka et al., 2008), which express CFP under the control of the endogenous Ret locus following Cre-mediated recombination (Figure 1J), tamoxifen treatment promoted CFP expression exclusively in TrkC+ cells (Figure S3F and S3I). Interestingly, we found that treatment of TrkCCreER; Retf(CFP) mice with 0.1 mg of tamoxifen at P5 led to expression of CFP in the majority of NFH+ circumferential endings (Figure 1K and 1M), but not in other populations, including neurons with Merkel endings, lanceolate endings, free nerve endings, or expressing CGRP+ or PV+ (Figure 1K, 1M, S3G, and S3H). Based on the efficiency of terminal labeling and the percentage of labeled NeuN+ DRG neurons, we estimate that neurons with NFH+ circumferential endings represent 4.3 ± 0.3% of all thoracic DRG neurons (3853 neurons counted in 4 mice), which is comparable to the density of Aβ RA-LTMRs and Aδ-LTMRs (Li et al., 2011).

In contrast, treatment of TrkCCreER; Retf(CFP) mice with 3 mg of tamoxifen at E12.5 led to expression of CFP exclusively in Merkel endings that are derived from Aβ SA1-LTMRs (Figure 1L and 1M), but not in other populations (Figure 1L, 1M, S3J, and S3K). Thus, the tamoxifen-inducible TrkCCreER; Retf(CFP) intersectional genetic strategy enables selective labeling of either neurons with NFH+ circumferential endings or Aβ SA1-LTMRs, depending on the time and dosage of tamoxifen delivery.

Neurons with NFH+ circumferential endings project to the deep spinal cord dorsal horn and the DCN

We next used the TrkCCreER; Retf(CFP) intersectional genetic labeling strategy to visualize the central projections of neurons with NFH+ circumferential endings. In the spinal cord, these CFP+ neurons project within lamina IIiv to lamina IV, partially overlapping with the PKCγ+ interneurons of lamina IIiv and III but not with IB4+ terminals in lamina IIid (Figure 2A-2B). These spinal cord terminations overlap with, but are slightly more dorsal than those of other Aβ-LTMR subtypes, including Aβ RA-LTMRs (Luo et al., 2009) and Aβ SA1-LTMRs that innervate laminar III to V (Figure 2C-2D), and they resemble the spinal cord innervation patterns of a subset of myelinated nociceptors (Boada and Woodbury, 2008). Furthermore, CFP+ circumferential ending neurons have central projections that terminate in the gracile and the dorsal cuneate nuclei of the DCN (Figure 2F-2G), but are excluded from the ventral cuneate and external cuneate nuclei (Figure 2E-2G), similar to Aβ RA-LTMRs (Luo et al., 2009).

Figure 2. Neurons with NFH+ circumferential endings innervate the deep spinal cord dorsal horn as well as the DCN.

Figure 2

Open in a new tab

(A-D) Spinal cord sections from TrkCCreER; Retf(CFP) mice treated with 0.1 mg tamoxifen at P5 or 3mg tamoxifen at E12.5. The central projections of neurons with NFH+ circumferential endings labeled by CFP are located in lamina IIiv through IV (n = 9), ventral to lamina IIid labeled by IB4 (A) but partially overlapping with lamina IIiv labeled by PKCγ (B). The central projections of Aβ SA1-LTMRs labeled by CFP are located in lamina III through V (n = 4), ventral to lamina IIid and lamina IIiv labeled by IB4 (C) and PKCγ (D). (E-G) Transverse brainstem sections from TrkCCreER; Retf(CFP) mice treated with 0.1mg tamoxifen at P5 reveals the innervation pattern in the DCN, which is marked by vGlut1. (H-M) Whole-mount AP staining of the skin (H and J), brainstem (I and K), and spinal cord (L and M) from TrkCCreER; Brn3af(AP) mice treated with 0.001mg tamoxifen at P8 reveals the peripheral terminals and central projections of neurons with circumferential endings (H, I and L) or Merkel endings (J, K, and M). Dorsal view of the left DCN and left thoracic spinal cord reveals that axons from a single labeled circumferential ending neuron and Aβ SA1-LTMR project into the spinal cord and form collaterals in both the rostral and caudal directions (L and M). The rostral end of this axon extends to the DCN, where it forms collaterals (I and K). Scale bar: 100 μm (A-G) and 200 μm (H-M). See also Figure S4.

In order to visualize the central projections of individual NFH+ circumferential ending neurons, we next used a sparse genetic labeling strategy, employing TrkCCreER in combination with Brn3af(AP), a Cre-dependent alkaline phosphatase (AP) reporter that is expressed in most NFH+ DRG neurons but not in the spinal cord (Badea et al., 2012). The central projections of individual NFH+ circumferential ending neurons bifurcate upon innervating the spinal cord, then project in both the rostral and caudal directions and sprout multiple collaterals that span 3.7 ± 0.3 spinal segments (Figure 2L and S4A). Moreover, the rostral axonal branch of each NFH+ circumferential ending neuron projects to the DCN where it forms collaterals (Figure 2I and S4A). These central projection patterns resemble hairy skin Aβ-LTMR subtypes including Aβ RA-LTMRs (Niu et al., 2013) and Aβ SA1-LTMRs, which form collaterals spanning 3.9 ± 0.2 spinal segments (Figure 2M and S4B) and innervate the DCN (Figure 2K) although, interestingly, the majority of Aβ SA1-LTMRs in DRGs below T10 do not project to the DCN (Figure S4B). These anatomical findings implicate DRG neurons with NFH+ circumferential endings in tactile perception.

Neurons with NFH+ circumferential endings are Aβ Field-LTMRs sensitive to gentle stroking of the skin but are relatively insensitive to skin indentation

In order to evaluate the physiological properties of NFH+ circumferential ending neurons, we developed a targeted in vivo electrophysiological recording paradigm (Figure 3A). We crossed TrkCCreER; Retf(CFP) with a dual recombinase dependent reporter line R26LSL-FSF-tdTomato (Ai65) and subcutaneously injected an AAV2/1-Flpo virus to label TrkC+ neurons innervating the dorsal surface of the thigh. This allowed us to confine our recordings to NFH+ circumferential neurons innervating a small region of hairy skin amenable to mechanical stimulation. Recently, NFH+ circumferential ending neurons were suggested to function as Aδ-mechanonociceptors on the basis of ex vivo recordings from a mixed population of DRG neurons labeled by DORGFP+ (Bardoni et al., 2014). Our intersectional genetic strategy allowed us to unambiguously and quantitatively explore the link between the structure and function of NFH+ circumferential ending neurons.

Figure 3. Neurons with NFH+ circumferential endings are Aβ Field-LTMRs.

Figure 3

Open in a new tab

(A) The targeted in vivo DRG recording preparation. (B) Representative responses of Aβ SA-LTMR, Aβ RA-LTMR, and NFH+ circumferential ending neurons to skin stroke delivered by a force controlled brush translated by an ultrasonic piezoelectric stage. The recorded holding force (black, ~5 mN) is shown at top. The stroke speed is 10 mm/sec. (C) Quantification of stroke-evoked action potential firing rates in Aβ SA1-LTMR (n = 4), Aβ RA-LTMR (n = 5), and NFH+ circumferential neurons (n = 10). Responses were computed as the average of the three highest firing rates observed in a 100 msec window which corresponds to the approximate time required for the brush to transit the full arbor of an Aβ-LTMR. NFH+ circumferential ending neurons responded to innocuous stroke with firing rates indistinguishable from those of other hairy skin LTMRs (p = 0.27 Welch’s t-test). (D) The conduction velocities of NFH+ circumferential ending neurons are indistinguishable from those of other hairy skin Aβ RA- and SA-LTMRs recorded (p = 0.871, Welch’s t-test). (E) Von Frey threshold measurements reveal that the force thresholds of NFH+ circumferential ending neurons are higher than those of other Aβ subtypes (90% bootstrap CI [21.2, 52.3] fold higher). (F) Temporal patterns of responses to indentation. Representative peristimulus time histograms (PSTHs) show that NFH+ circumferential neurons adapt more rapidly than Aβ SA1-LTMRs but lack the pronounced off-step responses of the Aβ RA-LTMRs. The indentation force recording (top) is aligned to the PSTH showing the mean firing rate computed across 15 stimulus presentations (7 msec bin widths). (G) Representative spiking responses of NFH+ circumferential ending neurons and Aβ RA-LTMR to deflection of guard hairs. (H) Group data showing that all Aβ RA-LTMRs respond similarly to deflection of guard hairs and skin stroke (p = 0.85, paired t-test), while NFH+ circumferential ending neurons are insensitive to guard hair deflection (p = 0.017, paired t-test). As in (C), evoked firing rates are computed over a 100 msec window. See also Figure S5.

We first assessed responsiveness of NFH+ circumferential ending neurons to skin stroke. Upon gentle stroking of the skin in the rostral to caudal direction with a fine brush (average holding force of 5 mN, 1-10 mm/sec sweep velocity), NFH+ circumferential ending neurons showed robust responses with an action potential firing frequency that equaled or exceeded those of Aβ RA-LTMRs and Aβ SA1-LTMRs labeled using Npy2r-GFP mice and TrkCCreER; Retf(CFP) mice treated with tamoxifen at E12.5, respectively (Figure 3B-3C). Moreover, electrical stimulation of the skin demonstrated that NFH+ circumferential ending neurons have a CV of 17.0 ± 2.0 m/sec, which is indistinguishable from the CVs of Aβ RA-LTMRs and Aβ SA1-LTMRs, respectively (Figure 3D). Thus, NFH+ circumferential ending neurons are Aβ-LTMRs highly sensitive to gentle skin stroke.

Interestingly, NFH+ circumferential ending neurons exhibit a much higher force threshold than Aβ RA-LTMRs and SA1-LTMRs when the skin is indented with von Frey filaments (Figure 3E). These responses exhibited considerable trial-to-trial variability; therefore, we next used a force-controlled stimulator which could be precisely and consistently positioned over the skin. We found that each NFH+ circumferential ending neuron’s receptive field displayed “hotspots”, which we subsequently targeted to assess indentation responses. NFH+ circumferential ending neurons exhibited distinct adaptation properties compared to Aβ SA1-LTMRs and Aβ RA-LTMRs. At forces near threshold their responses adapted rapidly over tens of milliseconds, while at higher forces adaptation responses to indentation steps became intermediate between Aβ SA1-LTMRs and Aβ RA-LTMRs (Figure 3F). NFH+ circumferential ending neurons also show few or no off responses, which are a hallmark of Aβ RA-LTMRs (Figure 3F), further demonstrating this is a distinct LTMR population.

We also made intracellular recordings from NFH+ circumferential neurons using an ex vivo preparation. These recordings revealed inflected somal action potentials that were broader than the uninflected somal action potentials of Aβ RA-LTMRs (Figure S5A). Consistent with the in vivo recordings, individual receptive fields were composed of multiple hotspots. Application of controlled forces using a blunt probe (1 mm diameter) centered over a hotspot revealed thatNFH+ circumferential neurons have an indentation threshold about 5 mN (Figure S5C), which is considerably higher than that of both Aβ RA-LTMRs and Aβ SA1-LTMRs (McIlwrath et al., 2007). Moreover, indentation responses became more slowly adapting and firing rates increased as forces increased into the noxious range (Figure S5C). These neurons responded far more vigorously when the same hotspots were indented with sharp probes (Figure S5D), and even more robustly when coaxial forces were applied to individual hairs (hair pull) in the hotspot (Figure S5B and S5E). Thus, while NFH+ circumferential neurons are highly responsive to gentle skin stroking, and are therefore an LTMR, these neurons also exhibit hallmarks of myelinated nociceptors (Burgess and Perl, 1967; Djouhri and Lawson, 2004; Koerber et al., 1988).

Both Aβ RA-LTMRs and Aβ Field-LTMRs are hairy skin innervating, predominantly rapidly adapting LTMR subtypes that are sensitive to skin stroking. Unlike Aβ RA-LTMRs, Aβ Field-LTMRs do not respond to deflection of individual guard hairs (Horch et al., 1977). Therefore, we next performed in vivo recordings to ask whether NFH+ circumferential ending neurons are sensitive to guard hair deflection by moving a brush across the tips of guard hairs, which extend beyond awl/auchene and zigzag hairs. In all cases tested, this stimulus elicited bursts of spikes in Aβ RA-LTMRs but failed to excite NFH+ circumferential ending neurons (Figure 3G and 3H). Taken together, the NFH+ circumferential ending neuron is an Aβ-LTMR that is exquisitely sensitive to gentle stroking of the skin, relatively insensitive to skin indentation, with a rapidly adapting discharge at low to moderate indentation forces, and insensitive to deflection of guard hairs. Thus, the properties of NFH+ circumferential ending neurons match those of both classically defined Aβ Field-LTMRs (Horch et al., 1977) as well as a subset of myelinated nociceptors (Burgess and Perl, 1967), and are hereafter referred to simply as Aβ Field-LTMRs.

Aβ Field-LTMRs associate with hair follicles yet are insensitive to hair deflection

Aβ Field-LTMRs reportedly respond to deflection of multiple hairs spanning a large stimulus “field”, which may be explained by simultaneous deflection of many hairs or the inevitable stimulation of skin that occurs when stroking a large group of hairs (Horch et al., 1977). To distinguish between these possibilities, we measured the extent of skin and hair displacement orthogonal to the skin surface and found that air puff stimuli could be titrated to elicit varying amounts of hair deflection without significant skin displacement, while indentation with a servo-actuated von Frey filament produced displacement of both hairs and skin that were similar in magnitude (Figure 4A and 4B).

Figure 4. Aβ Field-LTMRs are insensitive to hair deflection yet sensitive to direct stimulation of skin.

Figure 4

Open in a new tab

(A) Laser Doppler Vibrometric measurements of skin and hair movement in response to air puff (left, 15m/sec at the source) and indentation (right, 2g von Frey filament). Shown are representative displacements of the base of a hair shaft and a nearby patch of skin. (B) Group data showing the power spectral density averaged across stimulus presentations and measurement locations (n = 5 hairs, n = 5 skin locations). For air puff, the motion recorded at the base of hairs exceeds that of the surrounding skin by three orders of magnitude. (C) Representative in vivo recordings of Aβ Field-LTMR (blue) and Aβ RA-LTMR (black) responses to air puff and stroke, before and after hair removal. The recording was maintained continuously in all cases. (D) Air puff evoked spiking responses from Aβ Field-LTMRs (n = 6), Aβ SA1-LTMR (n = 3), Aβ RA-LTMRs (n = 4) and Aβ RA-LTMR after hair removal (n = 4). Firing rates computed over the entire 100 msec air puff are plotted against the speed of the air at the stimulator nozzle, which was placed 12mm above the skin. All Aβ RA-LTMR retained sensitivity to stroke after hair removal (C, data not shown), demonstrating that they remained mechanically sensitive despite loss of air puff responses. Traces are offset for clarity.

Next, we measured the response properties of Aβ Field-LTMRs to air puff-induced hair deflection. As expected, both air puff and skin stroking elicited robust spiking in Aβ RA-LTMRs (Figure 4C). Moreover, Aβ RA-LTMR spike frequency increased with the speed of the air puff, and reached a plateau at an air puff speed of 30 m/sec (Figure 4D). Remarkably, the response of Aβ RA-LTMRs to air puff was entirely dependent on the presence of hairs; experiments in which hairs were removed during the recording period showed that the response to air puff, but not to skin stroking, was abolished following hair removal (Figure 4C and 4D), further confirming that air puff elicits hair deflection with minimal skin indentation. In contrast, neither Aβ Field-LTMRs nor Aβ SA1-LTMRs responded to air puff at any stimulus intensity, including those that saturate the Aβ RA-LTMRs (Figure 4C and 4D). Moreover, Aβ Field-LTMR responses to gentle stroking of the skin were similar before and after removing hairs, demonstrating that the stroke sensitivity of these neurons does not rely on hair shafts. Thus, Aβ Field-LTMRs are categorically insensitive to hair deflection despite their association with hair follicles. Moreover, Aβ Field-LTMRs, Aβ RA-LTMRs and Aβ SA1-LTMRs have distinct response properties to mechanical stimulation; while all three Aβ-LTMR subtypes respond to gentle stroking of the skin, Aβ RA-LTMRs and Aβ SA1-LTMRs but not Aβ Field-LTMRs respond to gentle skin indentation, and only Aβ RA-LTMRs respond to hair deflection.

Micromechanical determinants of Aβ-LTMR sensitivity

Aβ Field-LTMRs and Aβ RA-LTMRs both form endings around hair follicles and yet they exhibit different sensitivities to hair deflection and skin indentation. We hypothesized that the ultrastructural architecture of their cutaneous endings and the mechanics of the surrounding tissues may explain this differential sensitivity. We first used transmission electron microscopy (TEM) to define the ultrastructural features of circumferential and lanceolate terminals. Interestingly, a cross sectional analysis of individual hair follicles clearly shows that while both lanceolate (Aβ RA-LTMR) and circumferential (Aβ Field-LTMR) endings are associated with hair follicles, they reside within distinct layers surrounding follicles. Lanceolate endings are located within an inner region where they closely abut the basement membrane of hair follicle epithelial cells with an average distance ~100 nm (Figure 5A and 5C), as previously described (Li and Ginty, 2014). In contrast, circumferential endings reside within an outer region of the follicle surrounding the lanceolate endings, and are thus considerably more distant from the follicle (Figure 5A-5C; average distance 4 μm). Strikingly, both types of sensory endings are aligned with collagen fibers that are organized in parallel to the orientation of the axon terminals (Figure 5A-5C).

Figure 5. Ultrastructural features of lanceolate and circumferential endings that support the distinct response properties of Aβ RA-LTMRs and Aβ Field-LTMRs.

Figure 5

Open in a new tab

(A-C) TEM images of cross-sections through a lanceolate and circumferential ending complex at a hair follicle. (B-C) Magnified view of the region in the red box of (A). The lanceolate ending (white asterisks) and terminal Schwann cell complexes are embedded in longitudinal oriented collagen fibers (L.C.) in close proximity to hair follicle epithelial cells. Circumferential endings (arrowheads) are embedded in circumferentially oriented collagen fibers (C.C.) in the outer layer. The lanceolate endings, circumferential endings, terminal Schwann cells, and hair follicle epithelia cells are pseudo colored in green, blue, and pink, and yellow, respectively. (D-I) The FEM simulations of the strain acting on lanceolate or circumferential endings in response to mechanical stimuli. (D-F) Top: Contour plots of strain distribution along the lanceolate and circumferential endings in response to 100 μm hair deflection (D), 0.8 mN skin indentation (E), and 10 mN skin indentation (F). Bottom: schematic diagram of the stimulations. (G-H) Plot of the average strain along different positions of lanceolate or circumferential terminals. The zero location is marked by the red asterisk in Figure 6D. (I) The ratio of accumulated strain in the absence and the presence of collagen layers along lanceolate or circumferential endings under either 100 μm hair deflections or 0.8mN skin indentations. Scale bar: 1 μm (A-C). See also Supplemental Table 1.

To identify an ultrastructural basis for the unique tuning properties of Aβ RA-LTMRs and Aβ Field-LTMRs, we estimated the extent to which hairy skin mechanical stimulation is transformed into strain, a normalized local deformation, acting upon lanceolate and circumferential endings using a finite element model (FEM) encompassing the mechanical properties of the skin, hair, hair follicle and the two layers of collagen networks. We simulated activation of axon terminals by delivering three stimuli that mimic those used for electrophysiological recordings (see Figures 3 and 4): 100 μm hair deflection, and 0.8 mN or 10 mN indentation applied to a skin region adjacent to the hair follicle (Figure 5D-5F, bottom). FEM simulations suggest that considerably more strain acts upon lanceolate endings than on circumferential endings in response to both hair deflection and skin indentation (Figure 5D-5H). The maximum strain on lanceolate endings in response to 100 μm hair deflection or 0.8 mN skin indentation is comparable (~2e−3), and is much higher than that acting on circumferential endings (~5e−4) (Figure 5D-5E and 5G-5H). Simulation using 10 mN skin indentation, which is close to the von Frey threshold of Aβ Field-LTMRs, predicted the maximum strain acting on circumferential endings to be ~6e−3 (Figure 5F). This is greater than the strain following 100 μm hair deflection (18 fold) and 0.8 mN skin indentation (12 fold) but is only slightly higher than the strain on lanceolate endings in response to 100 μm hair deflection and 0.8 mN skin indentation (3 fold). These simulated strain measurements are consistent with our physiological measurements of the relative sensitivities of the two neuronal types.

We next performed similar FEM simulations in the absence of the two layers of collagen observed by TEM. The strain on lanceolate endings was similar in the presence and absence of the two layers of collagen. In contrast, for circumferential endings, the strain in response to indentation in the absence of two collagen layers was approximately 50 times greater than that when both collagen layers were present (Figure 5I). Taken together, these FEM simulations suggest that the strain acting on Aβ RA-LTMR lanceolate endings is greater than the strain acting on the Aβ Field-LTMR circumferential endings following both hair deflection and skin indentation, and that the outer collagen layer may serve to dampen responses of Aβ Field-LTMR circumferential endings to skin indentation.

Aβ Field-LTMRs have unusually large receptive fields with many weakly mechanosensitive foci, and long distances between their axon terminals and spike initiation segments

Although Aβ Field-LTMRs are relatively insensitive to focal indentation, they are highly sensitive to gentle stroking of skin. How do Aβ Field-LTMRs acquire their sensitivity to gentle stroking across large areas of skin? One potential mechanism is that many weakly mechanosensitive endings of individual Aβ Field-LTMRs are elaborated over a large area of skin, and thus stimuli that sweep across the entirety of an Aβ Field-LTMR’s endings generate multiple receptor potentials that sum throughout the arbor and induce greater responses. To test this hypothesis, we first generated fine scale receptive field maps of Aβ Field-LTMRs using in vivo electrophysiological recordings. Stimuli were delivered using a sharp-tipped (20 μm in diameter) force controlled indenter, which was systematically translated over the skin of the dorsal thigh at 50 μm intervals. This distance was chosen to match the spatial scale of hair follicles and the spacing between them (Figure S6A and S6B). The resulting spatial receptive field maps are composed of the spiking responses to indentations delivered to 3200 separate positions.

Strikingly, Aβ Field-LTMR responses to 27 mN step indentations showed that the punctuate “hotspots”, described above in both in vivo and ex vivo recordings, are distributed over 3-4 mm2 of the skin. Responses to stimulation of hotspots (half-width ~ 60 μm, n = 2 with full receptive fields) exceeded 100 Hz, and hotspots were separated by insensitive stretches where skin stimulation evoked 0-2 spikes (Figure 6A and 6B). Thus, individual Aβ Field-LTMRs exhibit unusually large receptive fields composed of many weakly mechanosensitive hotspots distributed over a 3-4 mm2 area of skin.

Figure 6. Aβ Field-LTMRs have large receptive fields comprising many weak mechanosensitive endings.

Figure 6

Open in a new tab

(A) In vivo loose patch electrophysiological recordings from an Aβ Field-LTMR in response to 1 mN (left) and 27 mN (right) spatially patterned skin indentations with a 20 μm tipped tungsten probe. The indenter was positioned over the skin in 50 μm steps, such that adjacent steps correspond to adjacent rows in the raster plot. (B) Receptive field of an Aβ Field-LTMR mapped with a 27 mN indentation. Grid indentations of Aβ Field-LTMRs reveal punctuated “hotspots” that extend for 50-100 μm and are separated by insensitive patches (n = 2). (C) Whole-mount AP staining of hairy skin reveals peripheral terminals from individual Aβ Field-LTMRs. Scale bar: 500 μm (B-C). See also Figure S6.

To compare the physiological receptive fields to the peripheral anatomy of individual Aβ Field-LTMRs, we next performed sparse labeling using TrkCCreER; Brn3af(AP) mice. Individual Aβ Field-LTMR axons branch in the skin to give rise to circumferential endings associated with a remarkably large number of hair follicles (Figure 6C, S6C and S6E). Despite a large range in the number of associated hair follicle endings (20-180 hair follicles per neuron), most Aβ Field-LTMRs innervate a large area of skin (3.1 ± 0.1 mm2, Figure 6C, S6D and S6F). Importantly, these Aβ Field-LTMR morphologies bear striking resemblance to their receptive fields (Figure 6B), suggesting that individual or small groups of circumferential endings are the fundamental mechanosensitive units of Aβ Field-LTMRs.

How unique is the anatomy of the Aβ Field-LTMR? We next compared the morphological receptive fields of Aβ Field-LTMRs to all other hairy skin LTMRs, including Aβ SA1-LTMRs, Aβ RA-LTMRs, Aδ-LTMRs and C-LTMRs, using sparse genetic labeling strategies specific for each LTMR subtype. Indeed, individual Aβ RA-LTMRs (Figure 7B) associate with many fewer hair follicles (Figure 7F, S6C and S6E) and most of these neurons innervate a small skin area (Figure 7F, S6D and S6F) compared to Aβ Field-LTMRs (Figure 7A and 7F). Individual Aβ SA1-LTMRs typically innervate only one or two touch domes (74% or 26%, respectively) (Figure 7C and 7F), consistent with other measurements (Lesniak et al., 2014; Wu et al., 2012). The remaining hairy skin LTMRs, including Aδ-LTMRs and C-LTMRs, have both smaller morphological receptive fields and innervate fewer hair follicles (Figure 7D-7F, and S6C-S6F). Thus, Aβ Field-LTMRs have unusually expansive receptive fields containing the largest number of specialized mechanosensory terminals of any known mammalian LTMR type.

Figure 7. Aβ Field-LTMRs have large terminal fields and a long distance between their axon terminals and SISs.

Figure 7

Open in a new tab

(A-E) Whole-mount AP staining of hairy skin reveals peripheral terminals from individual LTMR subtypes sparsely labeled by a Brn3af(AP) reporter line. (F) Quantification of the number of innervated hair follicles and terminal field area from different LTMR subtypes including Aβ Field-LTMRs (n = 125, 8 mice), Aβ RA-LTMRs (n = 127, 10 mice), Aβ SA1-LTMRs (n = 31, 16 mice), Aδ-LTMRs (n = 55, 4 mice), as well as C-LTMRs (n = 41, 3 mice). Each dot represents a single neuron. (G-I) Whole-mount immunostaining of hairy skin reveals the relationship between MBP, the SIS as inferred from βIV-Spectrin and the terminals of three Aβ-LTMRs subtypes. Arrows or arrowheads point out the SISs or nodes of Ranvier, respectively. Note that the red βIV-Spectrin puncta in G (arrowhead) is associated with a different myelinated axon that is not belong to the genetically labeled Aβ Field-LTMRs. (J) Quantification of the distance of non-myelinated axons reveals that Aβ Field-LTMRs have longer non-myelinated axons (153 ± 50, n = 5) compared to Aβ RA-LTMRs (6.0 ± 1.8, n = 12) (p = 0.03, student t-test) and Aβ SA1-LTMRs (18.8 ± 0.7, n = 8) (p = 0.04, student t-test). Scale bar: 500 μm (A-E). 20 μm (G-I). See also Figure S6 and S7.

Given the Aβ Field-LTMR’s sensitivity to gentle skin stroke, we next asked whether the expansive terminal fields of Aβ Field-LTMRs exhibit morphological features consistent with subthreshold integration across multiple circumferential endings. We therefore determined the length of unmyelinated axon segments spanning between the Aβ Field-LTMR spike initiation site (SIS) and its circumferential terminals, and the findings were compared to those of Aβ SA1-LTMRs and Aβ RA-LTMRs. SISs were visualized using myelin basic protein (MBP) in combination of βIV-Spectrin immunostaining (Lesniak et al., 2014; Yang et al., 2007). The average distance between sensory terminals and the SIS was calculated for individual Aβ-LTMRs that were either genetically labeled (Aβ RA-LTMRs or Aβ Field-LTMRs) or immuno-labeled (Aβ SA1-LTMRs). Our analysis showed that SISs and initial myelination segments are uniformly localized within very close proximity to the sensory terminals of both Aβ SA1-LTMRs and Aβ RA-LTMRs (Figure 7H-7J). In contrast, Aβ Field-LTMR SISs are located at variable distances (Figure 7G and 7J) from their circumferential terminals and their unmyelinated axon segments are significantly longer than those of Aβ SA1-LTMRs and Aβ RA-LTMRs (Figure 7J and S7C-S7E). In some cases, Aβ Field-LTMR SISs were located too far from the hair follicle to identify (Figure S7D, data not shown). Similar findings were obtained in experiments in which the three types of Aβ-LTMRs were immuno-labeled using NFH (Figure S7A-S7B). Thus, Aβ Field-LTMRs are unique among hairy skin Aβ-LTMR subtypes in that their initial sites of myelination and SISs are often localized far from the circumferential endings around hair follicles. Taken together, Aβ Field-LTMRs have large, spotty receptive fields containing many weakly mechanosensitive circumferential endings distributed across a large area of skin and SISs located at variable and often considerable distances from the hair follicles with which they associate. We propose that these unique physiological and morphological properties underlie the sensitivity of Aβ Field-LTMRs to gentle stroking across large fields of skin.

Discussion

Sensory neurons that innervate the skin exhibit tremendous anatomical diversity. For the LTMRs, which mediate our sense of touch, the specialized anatomy of their endings is intimately related to their functions as mechanotransducers. Here we report the generation of murine genetic tools that label Aβ Field-LTMRs, a major population comprising more than 4% of DRG neurons in the mouse. Aβ Field-LTMRs are prevalent across mammalian species, and we find that in mouse each innervates up to 180 hair follicles with specialized NFH+ circumferential endings. Electrophysiological recordings demonstrate that they are highly sensitive to gentle stroking of the skin but do not respond to hair deflection. Moreover, Aβ Field-LTMRs are increasingly responsive to skin indentation in the noxious range and have large, spotty indentation receptive fields; these properties are unique amongst the Aβ-LTMR subtypes and have been ascribed to Aβ nociceptors (Djouhri and Lawson, 2004). Therefore, Aβ Field-LTMRs may have a nociceptor function, and a test of this possibility will require activation, silencing or ablation approaches and behavioral measures.

The Aβ Field-LTMR is highly sensitive to gentle skin stroking while, on the other hand it is remarkably insensitive to innocuous skin indentation and completely unresponsive to hair deflection. Our in vivo recordings, FEM, and morphological analyses support a model in which Aβ Field-LTMRs attain this unique sensitivity through integration across a large number of weakly mechanosensitive hotspots distributed across skin. Although the precise number of hotspots is difficult to ascertain, their approximate number and distribution is strikingly similar to those of circumferential endings from individual Aβ Field-LTMRs, visualized by sparse genetic labeling. The simplest interpretation of this observation is that each Aβ Field-LTMR circumferential ending is a weakly mechanosensitive unit.

Why are individual circumferential endings of Aβ Field-LTMRs weakly mechanosensitive compared to Aβ RA-LTMR lanceolate endings? One important clue stems from the observation that Aβ RA-LTMR lanceolate endings are closely associated with hair follicle epithelial cells, whereas Aβ Field-LTMR circumferential endings are not. In fact, the naked axonal membrane of individual lanceolate projections resides within 100 nm of the hair follicle epithelial cell membrane (Li and Ginty, 2014) (Figure 5A and 5C). This intimate physical apposition strongly implicates the lanceolate membrane immediately adjacent to the hair follicle epithelial cell as the site where hair deflection-induced mechanical forces are transduced into Aβ RA-LTMR axonal membrane depolarization and excitation. In contrast, Aβ Field-LTMR circumferential endings are located about 4 μm from hair follicle epithelial cells where they are enveloped by terminal Schwann cell processes. Another distinguishing feature is that Aβ RA-LTMR lanceolate endings are arranged within an inner follicle region containing longitudinally oriented collagen fibrils, whereas Aβ Field-LTMR circumferential endings are located within an outer follicle region comprised of circumferentially oriented collagen fibrils. Our FEM simulations suggest that these features render Aβ Field-LTMRs less susceptible to strain following both hair deflection and skin indentation.

Our analysis of MBP and βIV-Spectrin localization patterns suggests an additional defining feature of Aβ Field-LTMRs. We found that Aβ Field-LTMR SISs are often distant from their circumferential endings, whereas, in all cases analyzed, Aβ RA-LTMR and Aβ SA1-LTMR SISs are located immediately adjacent to their sensory terminals. Together, these findings support a model in which three distinguishing features of Aβ Field-LTMRs render these neurons selective for stroke among innocuous touch stimuli. These features are: 1) Aβ Field-LTMR circumferential endings are located in an outer layer of circumferentially oriented collagen fibers and thus, compared to Aβ RA-LTMR lanceolate endings, circumferential endings are relatively insensitive to hair deflection or innocuous skin indentation; 2) The terminal fields of Aβ Field-LTMRs are extremely large with individual neurons innervating up to 180 hair follicles in areas that encompass ~3 mm2, enabling convergence and summation of receptor potentials emanating from multiple circumferential endings; and 3) Aβ Field-LTMR SISs are typically located more than 100 μm away from the circumferential endings, suggesting a lower transformation efficiency of receptor potentials to action potentials as well as integration across multiple terminals in the receptive field. We propose that when a “field” of weakly mechanosensitive circumferential endings is activated, receptor potentials emanating from individual circumferential endings converge or summate at distally located SISs to initiate Aβ Field-LTMR spiking. In contrast, the extremely close apposition of hair follicle epithelial cells and Aβ RA-LTMR lanceolate endings and thus the high sensitivity to shear on these endings during follicle movement, as well as the immediate adjacency of these lanceolate endings to SISs enable Aβ RA-LTMRs to spike following deflection of a single hair. Thus, the unique physiological response properties of Aβ Field-LTMRs are a consequence of their terminal morphology, mechanical linkage to the follicle and surrounding skin, and the location of their SISs. It is also likely that LTMRs differ in expression of mechanically-gated ion channels and transduction machinery, and such molecular differences may also contribute to their unique physiological response properties.

We propose that Aβ Field-LTMRs, with their characteristic expansive receptive fields comprised of large numbers of weakly mechanosensitive circumferential endings, contribute to LTMR activity ensembles that underlie percepts associated with gentle stroking across large fields of hairy skin. Future work will define the contributions of Aβ Field-LTMRs to tactile perception and sensory-motor reflexes, and the postsynaptic partners and circuits in the dorsal horn and brainstem that receive, integrate, and process Aβ Field-LTMR and other LTMR subtype activities.

Experimental Procedures

Procedures are summarized below; see Supplemental Experimental Procedures for more detailed descriptions of each section.

Mouse Lines

TrkCtdTomato and TrkCCreER mice were generated by introducing a myristoylation signal tagged tdTomato- or CreER-Frt-Neomycine-Frt-loxP cassette into the first coding ATG in exon 1 of the TrkC gene (Figure S2A-S2B). Other mouse lines, including the Brn3af(AP), CGRPα-GFP and Npy2r-GFP BAC transgenic, RetCreER, RetCFP, Retf(CFP), R26LSL-YFP, R26LSL-tdTomato (Ai9), R26FSF-LSL-tdTomato (Ai65), THiCreER, TrkBCreER mouse lines, have been described previously (Badea et al., 2009; Gong et al., 2003; Li et al., 2011; Luo et al., 2009; Madisen et al., 2015; Rotolo et al., 2008; Rutlin et al., 2014; Uesaka et al., 2008).

Histological Analyses

Immunohistochemistry of tissue sections, whole mount immunohistochemistry, and whole mount AP histochemistry were performed using standard procedures (Li et al., 2011).

Electrophysiological Recordings

In vivo recordings were performed from genetically identified L4 DRG neurons in isoflurane anesthetized mice. To measure stimulus-evoked action potentials, a borosilicate extracellular electrode was visually guided under reflective optics to fluorescently labeled neurons and a loose seal formed. Stimuli were delivered to the hairy skin of the dorsal hindlimb. Intracellular recordings using an ex vivo skin-nerve preparation of 2-3 months old animals were performed with dorsal back skin, dorsal cutaneous nerve, thoracic DRGs and spinal cord intact (Woodbury et al., 2001).

Electron Microscopy

Fixed hairy skin was stained with 1% osmium tetroxide/1.5% potassium ferrocyanide followed by 1% uranyl acetate, gradually dehydrated with ethanol, and embedded in Epon for ultrathin sectioning and TEM imaging.

FEM simulations

FEM simulations of the hair follicle ultrastructure were built in Abaqus Standard (ver. 6.13). The model includes five parts: the skin, the hair core, the surrounding hair follicle, and two layers containing collagen fibers with co-axial and circumferential orientations.

Supplementary Material

Supplementa

Acknowledgements

We thank Michael Rutlin for help with targeting vector design, Lishi Li and Steven Hsiao for the macaque skin images, Hideki Enomoto for Retf(CFP) and RetCFP mice, Jeremy Nathans for Brn3af(AP) mice, Matthew Rasband for the βIV-Spectrin antibody, Tom Ferstl for tissue samples, and Emily Kuehn for AAV virus characterization. We are grateful to Robert LaMotte for assistance with the in vivo physiological recording preparation, Rachel Wilson for use of the Laser Doppler Vibrometer, Maria Erissson (Harvard EM core facility), Ciara Bolger and Wei-Chung Lee for assistance with the EM analysis, and Matthew Pecot, Zhang Yilei and members of the Ginty laboratory for valuable comments on the manuscript. This work was supported by NIH grants T32 NS007484-14 (BPL), F32 NS095631-01 (BPL), NS44049 (CJW), DE022750 and NS34814 (DDG). DDG is an investigator of the Howard Hughes Medical Institute.

Footnotes

Author Contributions

L.B. developed genetic labeling strategies and performed the anatomical characterization with assistance from T.L.D. and P.N.; B.P.L. performed electrophysiological and mechanical measurements with assistance from N.L.N.; J.L. performed FEM simulations; C.J.W. and C.C. performed the ex vivo recording experiments. L.B., B.P.L., and D.D.G. wrote the paper with input from all authors.

References

  1. Abraira VE, Ginty DD. The sensory neurons of touch. Neuron. 2013;79:618–639. doi: 10.1016/j.neuron.2013.07.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Albers KM, Perrone TN, Goodness TP, Jones ME, Green MA. Cutaneous Overexpression of NT-3 Increases Sensory and Sympathetic Neuron Number and Enhances Touch Dome and Hair Follicle Innervation. J. Cell Biol. 1996;134:487–497. doi: 10.1083/jcb.134.2.487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Badea TC, Cahill H, Ecker J, Hattar S, Nathans J. Distinct roles of transcription factors brn3a and brn3b in controlling the development, morphology, and function of retinal ganglion cells. Neuron. 2009;61:852–864. doi: 10.1016/j.neuron.2009.01.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Badea TC, Williams J, Smallwood P, Shi M, Motajo O, Nathans J. Combinatorial expression of Brn3 transcription factors in somatosensory neurons: genetic and morphologic analysis. J. Neurosci. 2012;32:995–1007. doi: 10.1523/JNEUROSCI.4755-11.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Ballermann M, McKenna J, Whishaw IQ. A grasp-related deficit in tactile discrimination following dorsal column lesion in the rat. Brain Res. Bull. 2001;54:237–242. doi: 10.1016/s0361-9230(01)00431-2. [DOI] [PubMed] [Google Scholar]
  6. Bardoni R, Tawfik VL, Wang D, François A, Solorzano C, Shuster S. a, Choudhury P, Betelli C, Cassidy C, Smith K, et al. Delta opioid receptors presynaptically regulate cutaneous mechanosensory neuron input to the spinal cord dorsal horn. Neuron. 2014;81:1312–1327. doi: 10.1016/j.neuron.2014.01.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Boada MD, Woodbury CJ. Myelinated skin sensory neurons project extensively throughout adult mouse substantia gelatinosa. J. Neurosci. 2008;28:2006–2014. doi: 10.1523/JNEUROSCI.5609-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Burgess BYPR, Perl ER. Myelinated afferent fibers responding specifically to noxious stimulation of the skin. J. Physiol. 1967;190:541–562. doi: 10.1113/jphysiol.1967.sp008227. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Burgess PR, Petit D, Warren RM. Receptor Types in Cat Hairy Supplied Skin by Myelinated Fibers. J. Neurophysiol. 1968;31:833–848. doi: 10.1152/jn.1968.31.6.833. [DOI] [PubMed] [Google Scholar]
  10. Chambers MR, Andres KH, Duering MV, Iggo A. The structure and function of the slowly adapting type II mechanoreceptor in hairy skin. Quarlerly J. Exp. Physiol. 1972;57:417–445. doi: 10.1113/expphysiol.1972.sp002177. [DOI] [PubMed] [Google Scholar]
  11. Djouhri L, Lawson SN. Abeta-fiber nociceptive primary afferent neurons: a review of incidence and properties in relation to other afferent A-fiber neurons in mammals. Brain Res. Brain Res. Rev. 2004;46:131–145. doi: 10.1016/j.brainresrev.2004.07.015. [DOI] [PubMed] [Google Scholar]
  12. Dobry PJK, Casey KL. Roughness discrimination in cats with dorsal column lesions. Brain Res. 1972;44:385–397. doi: 10.1016/0006-8993(72)90310-1. [DOI] [PubMed] [Google Scholar]
  13. Fünfschilling U, Ng Y-G, Zang K, Miyazaki J-I, Reichardt LF, Rice FL. TrkC kinase expression in distinct subsets of cutaneous trigeminal innervation and nonneuronal cells. J. Comp. Neurol. 2004;480:392–414. doi: 10.1002/cne.20359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gong S, Zheng C, Doughty ML, Losos K, Didkovsky N, Schambra UB, Nowak NJ, Joyner A, Leblanc G, Hatten ME, et al. A gene expression atlas of the central nervous system based on bacterial artificial chromosomes. Nature. 2003;425:917–925. doi: 10.1038/nature02033. [DOI] [PubMed] [Google Scholar]
  15. Horch KW, Burgess PR, Whitehorn D. Ascending collaterals of cutaneous neurons in the fasciculus gracilis of the cat. Brain Res. 1976;117:1–17. doi: 10.1016/0006-8993(76)90552-7. [DOI] [PubMed] [Google Scholar]
  16. Horch KW, Tuckett RP, Burgess PR. A key to the classification of cutaneous mechanoreceptors. J. Invest. Dermatol. 1977;69:75–82. doi: 10.1111/1523-1747.ep12497887. [DOI] [PubMed] [Google Scholar]
  17. Ikeda R, Cha M, Ling J, Jia Z, Coyle D, Gu JG. Merkel cells transduce and encode tactile stimuli to drive Aβ-afferent impulses. Cell. 2014;157:664–675. doi: 10.1016/j.cell.2014.02.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Koerber HR, Druzinsky RE, Mendell LM. Properties of somata of spinal dorsal root ganglion cells differ according to peripheral receptor innervated. J. Neurophysiol. 1988;60:1584–1596. doi: 10.1152/jn.1988.60.5.1584. [DOI] [PubMed] [Google Scholar]
  19. Lechner SG, Lewin GR. Hairy sensation. Physiology (Bethesda) 2013;28:142–150. doi: 10.1152/physiol.00059.2012. [DOI] [PubMed] [Google Scholar]
  20. Lesniak DR, Marshall KL, Wellnitz S. a, Jenkins B. a, Baba Y, Rasband MN, Gerling GJ, Lumpkin E. a. Computation identifies structural features that govern neuronal firing properties in slowly adapting touch receptors. Elife. 2014;3:e01488. doi: 10.7554/eLife.01488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Li L, Ginty DD. The structure and organization of lanceolate mechanosensory complexes at mouse hair follicles. Elife. 2014;3:e01901. doi: 10.7554/eLife.01901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Li L, Rutlin M, Abraira VE, Cassidy C, Kus L, Gong S, Jankowski MP, Luo W, Heintz N, Koerber HR, et al. The functional organization of cutaneous low-threshold mechanosensory neurons. Cell. 2011;147:1615–1627. doi: 10.1016/j.cell.2011.11.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Loewenstein WR. Rate sensitivity of a biological transducer. Ann. N. Y. Acad. Sci. 1969;156:892–900. doi: 10.1111/j.1749-6632.1969.tb14021.x. [DOI] [PubMed] [Google Scholar]
  24. Luo W, Enomoto H, Rice FL, Milbrandt J, Ginty DD. Molecular identification of rapidly adapting mechanoreceptors and their developmental dependence on ret signaling. Neuron. 2009;64:841–856. doi: 10.1016/j.neuron.2009.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Madisen L, Garner AR, Shimaoka D, Chuong AS, Klapoetke NC, Li L, van der Bourg A, Niino Y, Egolf L, Monetti C, et al. Transgenic Mice for Intersectional Targeting of Neural Sensors and Effectors with High Specificity and Performance. Neuron. 2015;85:942–958. doi: 10.1016/j.neuron.2015.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Maksimovic S, Nakatani M, Baba Y, Nelson AM, Marshall KL, Wellnitz S. a, Firozi P, Woo S-H, Ranade S, Patapoutian A, et al. Epidermal Merkel cells are mechanosensory cells that tune mammalian touch receptors. Nature. 2014;509:617–621. doi: 10.1038/nature13250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. McIlwrath SL, Lawson JJ, Anderson CE, Albers KM, Koerber HR. Overexpression of neurotrophin-3 enhances the mechanical response properties of slowly adapting type 1 afferents and myelinated nociceptors. Eur. J. Neurosci. 2007;26:1801–1812. doi: 10.1111/j.1460-9568.2007.05821.x. [DOI] [PubMed] [Google Scholar]
  28. Millard CL, Woolf CJ. Sensory innervation of the hairs of the rat hindlimb: a light microscopic analysis. J. Comp. Neurol. 1988;277:183–194. doi: 10.1002/cne.902770203. [DOI] [PubMed] [Google Scholar]
  29. Niu J, Ding L, Li JJ, Kim H, Liu J, Li H, Moberly A, Badea TC, Duncan ID, Son Y-J, et al. Modality-based organization of ascending somatosensory axons in the direct dorsal column pathway. J. Neurosci. 2013;33:17691–17709. doi: 10.1523/JNEUROSCI.3429-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Rice FL, Munger BL. A Comparative Light Microscopic Analysis of the Sensory Innervation of the Mystacial Pad. II. The Common Fur Between the Vibrissae. J. Comp. Neurol. 1986;252:186–205. doi: 10.1002/cne.902520205. [DOI] [PubMed] [Google Scholar]
  31. Rotolo T, Smallwood PM, Williams J, Nathans J. Genetically-Directed, Cell Type-Specific Sparse Labeling for the Analysis of Neuronal Morphology. PLoS One. 2008;3:e4099. doi: 10.1371/journal.pone.0004099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rutlin M, Ho C-Y, Abraira VE, Cassidy C, Bai L, Woodbury CJ, Ginty DD. The Cellular and Molecular Basis of Direction Selectivity of Aδ-LTMRs. Cell. 2014;159:1640–1651. doi: 10.1016/j.cell.2014.11.038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Uesaka T, Nagashimada M, Yonemura S, Enomoto H. Diminished Ret expression compromises neuronal survival in the colon and causes intestinal aganglionosis in mice. J. Clin. Invest. 2008;118:1890–1898. doi: 10.1172/JCI34425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Woo S-H, Ranade S, Weyer AD, Dubin AE, Baba Y, Qiu Z, Petrus M, Miyamoto T, Reddy K, Lumpkin EA, et al. Piezo2 is required for Merkel-cell mechanotransduction. Nature. 2014;509:622–626. doi: 10.1038/nature13251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Woodbury CJ, Ritter AM, Koerber HR. Central anatomy of individual rapidly adapting low-threshold mechanoreceptors innervating the hairy skin of newborn mice: Early maturation of hair follicle afferents. J. Comp. Neurol. 2001;436:304–323. [PubMed] [Google Scholar]
  36. Wu H, Williams J, Nathans J. Morphologic diversity of cutaneous sensory afferents revealed by genetically directed sparse labeling. Elife. 2012;1:e00181. doi: 10.7554/eLife.00181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yang Y, Ogawa Y, Hedstrom KL, Rasband M. βIV spectrin is recruited to axon initial segments and nodes of Ranvier by ankyrinG. J. Cell Biol. 2007;176:509–519. doi: 10.1083/jcb.200610128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Zimmerman A, Bai L, Ginty DD. The gentle touch receptor of mammalian skin. Science. 2014;346:950–954. doi: 10.1126/science.1254229. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementa
NIHMS787587-supplement-Supplementa.pdf (980.5KB, pdf)

RESOURCES