Characterization of dural afferent neurons innervating cranial blood vessels within the dura in rats

Accepted Manuscript Research report Characterization of dural afferent neurons innervating cranial blood vessels within the dura in rats Michiko Nakamura, Il-Sung Jang PII: DOI: Reference:

S0006-8993(18)30335-4 https://doi.org/10.1016/j.brainres.2018.06.007 BRES 45839

To appear in:

Brain Research

Received Date: Revised Date: Accepted Date:

13 February 2018 5 June 2018 7 June 2018

Please cite this article as: M. Nakamura, I-S. Jang, Characterization of dural afferent neurons innervating cranial blood vessels within the dura in rats, Brain Research (2018), doi: https://doi.org/10.1016/j.brainres.2018.06.007

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Characterization of dural afferent neurons innervating cranial blood vessels within the dura in rats Michiko Nakamura1,2 and Il-Sung Jang1,2 1

Department of Pharmacology, School of Dentistry, Kyungpook National University, Daegu 41940, Republic of Korea

2

Brain Science & Engineering Institute, Kyungpook National University, Daegu 41940, Republic of Korea

Abbreviated title: Properties of dural afferent neurons

Number of pages: 38 pages for text, 8 main figures, 3 supplementary figures, and 1 supplementary table Number of words: 245 for Abstract, 530 for Introduction, and 1645 for Discussion

Please send all correspondence to: Il-Sung Jang, PhD, Professor Department of Pharmacology, School of Dentistry, Kyungpook National University 2177 Dalgubeol-daero, Jung-gu, Daegu 41940, Republic of Korea Tel: +82-53-660-6887, Fax: +82-53-424-5130, E-mail: [email protected]

Abbreviations; AITC; allyl isothiocyanate, αβ-me-ATP; αβ-methylene-ATP, CGRP; calcitonin gene related peptide, ICa; voltage-gated Ca2+ currents, NF; neurofilament, TG; trigeminal ganglia, TRPA1; TRP ankyrin 1, TRPM8; TRP melastatin 8, TRPV1; transient receptor potential vanilloid 1, TTX; tetrodotoxin, TTX-R; tetrodotoxin-resistant, TTX-S; tetrodotoxin-sensitive, V1; ophthalmic division, V2; maxillary division, V3; mandibular division

1

Abstract

Dural afferent neurons are implicated in primary headaches including migraine. Although a significant portion of primary afferent neurons innervating the dura are myelinated A-type neurons, previous electrophysiological studies have primarily characterized the functional properties of small-sized C-type sensory neurons. Here we show the functional characterization of dural afferent neurons identified with the fluorescent dye DiI. DiI-positive neurons were divided into three types: small-, medium-, and large-sized neurons, based on their diameter, area, and membrane capacitance. The immunoreactivity of NF200, a marker of A-type myelinated neurons, was detected in most large-sized, but it was also present in a limited number of smalland medium-sized DiI-positive neurons. Capsaicin, a transient receptor potential vanilloid 1 agonist, induced the membrane currents in most small- and medium-sized neurons, but not in large-sized DiI-positive neurons. Tetrodotoxin-resistant Na+ channels were expressed in almost all types of DiI-positive neurons. Mechanosensitive currents were detected from a majority of large-sized, and to a lesser extent, small- and medium-sized DiI-positive neurons. The results suggest that most dural afferent neurons are nociceptive, e.g., polymodal C-type for small- and medium-sized neurons, and high-threshold nociceptive A-type mechanoreceptors for large-sized neurons. We also found that DiI-positive neurons differed with respect to passive and active membrane properties, and that sumatriptan, a representative drug used for the acute treatment of migraine attack, inhibited voltage-gated Ca2+ currents in all types of DiI-positive neurons. The present results showing the nociceptive properties of dural afferent neurons would contribute to understand the pathophysiology of primary headaches.

Keywords; migraine, dural afferent neurons, ion channels, excitability, patch clamp

2

1. Introduction

Sensory information from peripheral tissues including nociception as well as tactile sensation is conducted into the central nervous system through primary afferent neurons. The soma of these neurons are located in the sensory ganglia, such as the dorsal root ganglia and trigeminal ganglia (TG), and can be generally classified into three types based on the soma diameter (Mense, 1990; Lawson, 2002; Le Pichon and Chesler, 2014). Small (somatic diameter of <30 μm)- and medium (somatic diameter of 30–40 μm)-sized sensory neurons are regarded as C-type and Aδ-type neurons, respectively. While C-type neurons transmit polymodal nociceptive information, such as thermal and mechanical nociception, Aδ-type neurons are mainly involved in mechanical pain mediated by high-threshold mechanoreceptors (Christensen and Corey, 2007; Delmas et al., 2011). In contrast, large-sized sensory neurons (somatic diameter of >40 μm) are generally regarded as Aβ-type neurons, which detect and transmit tactile sensation mediated by low-threshold mechanoreceptors in response to innocuous stimuli (Abraira and Ginty, 2013). In addition to morphological properties, several ion channels and/or receptors can be used to distinguish between non-nociceptive and nociceptive neurons. Migraine is a common recurrent neurological disorder that is typically characterized by disabling headaches and associated symptoms including photophobia, phonophobia, nausea and vomiting (Goadsby, 2007; Diener et al., 2012). Although the pathophysiology of migraine is not yet fully understood, a growing body of literatures suggests that migraine headache is peripherally triggered by the dilatation of cranial extracerebral blood vessels within the dura mater, which is induced by calcitonin gene related peptide (CGRP) released from trigeminovascular nerves (Silberstein, 2004; Benemei et al., 2009). Migraine headache is also centrally triggered by cortical spreading depression (Leao, 1944; Lauritzen, 1994), which

3

activates trigeminovascular afferents and evokes a series of CGRP-induced neurogenic inflammation at the cortical meningeal as well as brainstem regions (Bolay et al., 2002; Diener et al., 2012). Several inflammatory mediators, such as histamine, serotonin, and prostaglandins, are known to sensitize peripheral terminals of trigeminovascular neurons, also called dural afferent neurons, which innervate blood vessels within the dura mater (Strassman et al., 1996; Levy and Strassman, 2002a). Previous morphological studies have shown that the intracranial dura are densely innervated by myelinated A-fibers as well as unmyelinated C-fibers, although unmyelinated fibers comprise approximately 75–80% of the dural axons (Andres et al., 1987; Strassman et al., 2004). However, about 80% of unmyelinated fibers are autonomic axons, while essentially all myelinated fibers are sensory axons (Ruskell, 1988), suggesting that a significant portion of primary

afferent

neurons

innervating

the

dura

is

myelinated

A-type

neurons.

Electrophysiological studies analyzing the conduction velocity have also confirmed that both C-type and Aδ-type neurons are responsible for nociceptive responses to thermal, mechanical, and chemical stimuli (Bove and Moskowitz; 1997; Levy and Strassman, 2002a, 2002b). However, several studies using the whole-cell patch-clamp technique have examined the electrophysiological properties of small-sized, presumably nociceptive C-type, dural afferent neurons (Harriott and Gold, 2009; Vaughn and Gold, 2010; Yan et al., 2011), thus, the neuronal properties of other subpopulation of dural afferent neurons, e.g., medium- and large-sized neurons, still need to be elucidated. In the present study, therefore, we identified and characterized the electrophysiological properties of small-, medium-, and large-sized dural afferent neurons using the fluorescent dye DiI.

4

2. Results

2.1. General properties of dural afferent neurons Seven to 10 days after the application of DiI on the dura mater, DiI-positive neurons were mainly found in the ophthalmic division (V1) and, to a much lesser extent, the maxillary (V2) and mandibular (V3) divisions, of the TG (Supplementary Fig. S1), as shown in other studies (O'Connor and van der Kooy, 1986; Huang et al., 2012). However, DiI-positive neurons were rarely detected from the TG when DiI was applied to the cranial bone (data not shown). These DiI-positive neurons would originate from intracranial blood vessels, as shown by other studies (Harriott and Gold, 2009; Yan et al., 2011; Huang et al., 2012). In the subsequent experiments, DiI-positive neurons enzymatically isolated from the ophthalmic division of the TG were subjected to morphological, immunohistochemical, and electrophysiological experiments. We firstly examined the basic morphological properties of single dural afferent neurons. As shown in Fig. 1A and B, DiI-positive dural afferent neurons could be divided into the three populations based on the distribution of diameter (n = 1220), area (n = 1114), and membrane capacitance (n = 346), e.g., small, medium, and large groups. Both soma diameter and area of dural afferent neurons were apparently correlated to membrane capacitance (r = 0.92 and 0.93 for diameter and area, respectively, n = 346, Fig. 1C). In this study, therefore, the size of dural afferent neurons were divided into the following three groups: small (<30 μm in diameter, and <800 μm2 in area), medium (30–40 μm in diameter, and 800–1200 μm2 in area), and large (≥40 μm in diameter, and ≥1200 μm2 in area) groups. In the electrophysiological experiments, membrane capacitance was also considered as a parameter for the size distribution of dural afferent neurons (<40 pF, 40–80 pF, and ≥80 pF for small, medium, and large, respectively). (Figure 1 near here)

5

The aforementioned results suggest that small-sized, medium-sized, and large-sized DiI-positive neurons might belong to C-type, Aδ-type, and Aβ-type sensory neurons, respectively. Since NF200 is generally used to distinguish myelinated A-type and unmyelinated C-type sensory neurons (Fornaro et al., 2008; Ho and O'Leary, 2011), we next examined the expression of neurofilament (NF) 200 protein in acutely isolated DiI-positive neurons. The comparison between electrophysiological data and immunohistochemical and/or morphological data obtained from acutely isolated single neurons, rather than TG sections, would be adequate to characterize the properties of dural afferent neurons, because electrophysiological data were mainly used to characterize dural afferent neurons (see below). However, the present study cannot be excluded a possibility that the dissociation procedure differentially affects the survival of small- to large-sized DiI-positive neurons. The immunoreactivity for NF200 was detected in most large-sized DiI-positive neurons (128 of 162 neurons, 79.0%, Fig. 2A and B). However, the immunoreactivity for NF200 was not detected in most medium-sized DiI-positive neurons (29 of 143 neurons, 20.3%, Fig. 2A and B). In addition, a subset of small-sized DiI-positive neurons showed the immunoreactivity for NF200 (25 of 127 neurons, 19.7%, Fig. 2A and B), which was an unexpected result. We also examined the expression pattern of CGRP in DiI-positive neurons. Almost all of small-sized and medium-sized DiI-positive neurons expressed CGRP (94 of 99 small-sized neurons, 94.9% and 92 of 95 medium-sized neurons, 96.8%, Fig. 2C and D). The CGRP immunoreactivity was also detected in a small portion of large-sized DiI-positive neurons (23 of 98 neurons, 23.5%, Fig. 2C and D). (Figure 2 near here)

2.2. Nociceptive properties of dural afferent neurons

6

It is generally accepted that C-type and Aδ-type neurons transmit nociceptive information in response to thermal, chemical, and mechanical stimuli but Aβ-type sensory neurons transmit non-nociceptive innocuous tactile sensation. We next examined whether DiI-positive neurons express ion channels and receptors regarded as nociceptive markers using a whole-cell patch clamp technique at a VH of -80 mV (Fig. 3A). Acidic pH (pH 6.0) induced quickly desensitizing inward currents, which are mediated by acid-sensing ion channels in dural afferent neurons, but the proportion of neurons sensitive to acidic pH was quite different among DiI-positive neurons (7 of 40 small-sized neurons tested, 17.5%; 81 of 86 medium-sized neurons tested, 94.2%; 27 of 39 large-sized neurons tested, 69.2%; Fig. 3B and C). The density of pH6.0-induced currents was larger in small-sized than in medium-sized and large-sized DiI-positive neurons (Supplementary Fig. S2). The application of ATP (100 μM) also induced quickly desensitizing inward currents in most dural afferent neurons, and the proportion of neurons responding to ATP was higher in large-sized than in small-sized or medium-sized DiI-positive neurons (22 of 36 small-sized neurons tested, 61.1%, 68 of 85 medium-sized neurons tested, 80.0%, 39 of 39 large-sized neurons tested, 100%, Fig. 3B and C). The density of ATP-induced currents was larger in small-sized than in medium-sized and large-sized DiI-positive neurons (Supplementary Fig. S2). In a subset of experiments, we also examined the effect of αβ-me-ATP, a P2X1 and P2X3 receptor agonist (Coddou et al., 2011), in dural afferent neurons, because P2X3 receptors are regarded as a marker for nociceptive neurons (Bradbury et al., 1998; Dunn et al., 2001). The application of αβ-me-ATP (100 μM) induced quickly desensitizing inward currents in all small(n = 18) and medium- (n = 15) sized neurons tested (Fig. 3B), but αβ-me-ATP (100 μM) did not induce any membrane current in large-sized neurons tested (n = 10). Transient receptor potential vanilloid 1 (TRPV1) is known to be specifically expressed in C-type nociceptive neurons (Kobayashi et al., 2005; Yeo et al., 2010). The application of

7

capsaicin (0.5 μM), a TRPV1 agonist, induced inward currents in most small-sized DiI-positive neurons (37 of 40 neurons tested, 92.5%, Fig. 3B and C). However, capsaicin (0.5 μM) also induced membrane currents in most medium-sized DiI-positive neurons (80 of 83 neurons tested, 94.1%, Fig. 3B and C). In large-sized DiI-positive neurons, capsaicin induced membrane currents in only 4 of 39 neurons tested (10.2%, Fig. 3B and C). The density of capsaicin-induced currents was larger in small-sized than in medium-sized DiI-positive neurons (Supplementary Fig. S2). Menthol (300 μM), a TRP melastatin 8 (TRPM8) agonist, did not induce any membrane currents in any DiI-positive neurons tested (Fig. 3B and C). AITC (100 μM), a TRP ankyrin 1 (TRPA1) agonist, induced membrane currents in a small portion of DiI-positive neurons (0 of 21 small-sized neurons tested, 0%, 2 of 15 medium-sized neurons tested, 13.3%, 12 of 24 large-sized neurons tested, 50.0%, Fig. 3B and C). (Figure 3 near here) Among nine types of voltage-gated Na+ channels, tetrodotoxin-resistant (TTX-R) Na+ channels, such as Nav1.8 and Nav1.9, are known to be specifically expressed in small-sized and medium-sized nociceptive neurons (Chahine and O'Leary, 2014). We therefore examined the existence of TTX-R Na+ channels in dural afferent neurons. DiI-positive neurons were voltage-clamped at VHs of -80 mV or -100 mV, and voltage-gated Na+ currents were elicited by depolarizing voltage step pulses to -10 mV or -20 mV (100 ms in duration). As expected, voltage-gated Na+ currents still remained even in the presence of TTX (300 nM) in most small-sized and medium-sized DiI-positive neurons (52 of 57 small-sized neurons tested, 91.2% and 105 of 115 medium-sized neurons tested, 91.3%, Fig. 4A and B). We further examined the existence of transcript for Nav1.8 using a single-cell RT-PCR technique in DiI-positive neurons that show the voltage-gated Na+ currents in the presence of TTX. We could detect the Na v1.8 transcript in all small-sized (n = 6) and medium-sized DiI-positive neurons (n = 8) tested,

8

respectively (Fig. 4C), suggesting that TTX-R Na+ channels are expressed in small- and medium-sized DiI-positive neurons. However, it was rather unexpected that TTX-R Na+ currents could be recorded from most large-sized DiI-positive neurons (87 of 108 large-sized neurons tested, 80.6%, Fig. 4Ac and B). In all large-sized DiI-positive neurons showing TTX-R Na+ currents tested (n = 8), the Na v1.8 transcript was clearly detected (Fig. 4C). In the remaining 21 large-sized neurons (19.4%), TTX (300 nM) completely attenuated voltage-gated Na+ currents (Fig. 4Ad), and Nav1.8 transcript was not detected in these neurons (n = 4, Fig. 4C). (Figure 4 near here)

2.3. Functional properties of dural afferent neurons Next, we examined the passive and active membrane properties of dural afferent neurons in a current-clamp mode. The resting membrane potentials of DiI-positive neurons ranged from -45 mV to -80 mV, with mean values of -49.2 ± 1.5 mV (n = 46), -55.2 ± 1.3 mV (n = 74), and -58.1 ± 2.1 mV (n = 62) for small-, medium-, and large-sized (TTX-R) DiI-positive neurons, in which TTX did not block the generation of action potentials, respectively (Supplementary Table 1). In a subset of large-sized (TTX-S) DiI-positive neurons, in which TTX blocked the generation of action potentials, the resting membrane potentials were in a more hyperpolarized range (-65.9 ± 3.5 mV, n = 15, Supplementary Table 1). Next, single action potentials were elicited by brief depolarizing current injections (5 ms duration) in DiI-positive neurons (Supplementary Fig. S3). Differences in the basic properties of action potentials, such as the duration and overshoot amplitude of action potentials and afterhyperpolarization (AHP), are summarized in Supplementary Table 1 and Supplementary Fig. S3. We next examined the firing patterns of dural afferent neurons using the sustained

9

depolarizing current injection. Firstly, the rheobase current, which means the threshold amplitude of injected current needed to elicit action potentials, was determined by increasing the current amplitude injected to dural afferent neurons. The rheobase currents needed to generate action potentials were 163.5 ± 22.5 pA (n = 46), 105.7 ± 8.3 pA (n = 74), and 312.6 ± 28.5 pA (n = 62) for small-, medium-, and large-sized (TTX-R) DiI-positive neurons, respectively (Fig. 5A and Ba, Supplementary Table 1). The rheobase currents were the largest in large-sized (TTX-S) DiI-positive neurons, in which the application of 300 nM TTX abolished the generation of action potentials in a current-clamp mode and blocked voltage-gated Na+ currents (526.5 ± 57.1 pA , n = 15, Fig. 5A and Ba, Supplementary Table 1). The number of action potentials elicited by step current injections was the largest in medium-sized DiI-positive neurons (Fig. 5A and Bb). When DiI-positive neurons were stimulated with step current injections of integer of rheobase current, largely three types of firing patterns, such as phasic, tonic, and single firing types, were found in DiI-positive neurons. The phasic firing type (34 of 46 neurons tested, 73.9%) and, to a lesser extent, the single (6 of 46 neurons tested) and tonic firing types (6 of 46 neurons tested), were found in small-sized DiI-positive neurons (Fig. 5A and C). In contrast, most medium-sized DiI-positive neurons displayed the tonic firing type (64 of 74 neurons tested, 86.5%), where the firing frequency was greatly increased along with the stimulation strength (Fig. 5A and C). In large-sized DiI-positive neurons for which TTX did not abolish action potentials, both the single and tonic firing types were dominant (28 of 62 neurons tested, 45.2% for single firing pattern, 34 of 62 neurons tested, 54.8% for tonic firing pattern, respectively, Fig. 5A and C). In large-sized DiI-positive neurons for which TTX did abolish action potentials, the single firing pattern was dominant (15 of 15 neurons tested, 100%, Fig. 5A and C). (Figure 5 near here)

10

2.4. Mechanosensitive ion channels in dural afferent neurons We next examined whether dural afferent neurons express mechanosensitive ion channels using mechanical stimuli. Mechanical stimuli using a fire-polished glass pipette induced inward currents at a VH of -80 mV. As shown in Fig. 6A and B, mechanical stimuli induced inwardly directed membrane currents, and these currents were reliably reproduced during repetitive mechanical stimuli. The mechanical stimulation-induced currents were reversed at near 0 mV (Fig. 6C), which is similar to the theoretical equilibrium potential of monovalent cations in our experimental conditions. The mechanical stimulation-induced currents were greatly reduced in the Na+-free external solution (3.1 ± 0.5% of the control, n = 7, p < 0.01, Fig. 6D), and the remaining currents were completely disappeared in the both the Na+- and Ca2+-free external solution (Fig. 6Db). In addition, the amplitude mechanical stimulation-induced currents were clearly decreased by Gd3+, a non-selective cation channel blocker, in a concentration-dependent manner (Fig. 6E), suggesting that mechanosensitive ion channels expressed in dural afferent neurons are non-selective cation channels. (Figure 6 near here) The proportion of neurons responding to mechanical stimulation was higher in large-sized than in small-sized or medium-sized DiI-positive neurons (14 of 53 small-sized neurons tested, 26.4%; 17 of 42 medium-sized neurons tested, 40.5%; 32 of 41 large-sized neurons tested, 78.0%; Fig. 7A). The amplitude of mechanosensitive currents was increased with an increase in moving distance of glass pipette for mechanical stimuli (Fig. 7B), and the sensitivity to mechanical stimulation was larger in small-sized than in medium-sized or large-sized DiI-positive neurons (Fig. 67). In a current-clamp condition, mechanical stimulation elicited action potentials and the probability of action potential generation was increased with increasing

11

the stimulus intensity (Fig. 67 and E). (Figure 7 near here)

2.5. Voltage-gated Ca2+ channels in dural afferent neurons We finally examined whether sumatriptan, a selective 5-HT1B/1D receptor agonist that is widely used for the treatment of acute migraine attack (Bigal et al., 2009), can inhibit voltage-gated Ca2+ currents (ICa) in DiI-positive neurons. The ICa, which were completely blocked by 100 μM Cd2+, a general voltage-gated Ca2+ channel blocker, were recorded from three different groups of DiI-positive neurons by applying brief voltage step pulses (500 ms duration) to 0 mV at a VH of -80 mV (Fig. 8Aa). The current density of Ca2+ currents was not different among small-, medium-, and large-sized DiI-positive neurons (65.8 ± 10.5 pA/pF for small-sized neurons, n = 6, 70.2 ± 14.9 pA/pF for medium-sized neurons, n = 7, and 82.9 ± 19.5 pA/pF for large-sized neurons, n = 7, Fig. 8Ba). The application of sumatriptan (10 μM) slightly but significantly decreased the ICa in all types of DiI-positive neurons (10.4 ± 1.9% for small-sized neurons, n = 6, p < 0.01, 12.6 ± 2.0% for medium-sized neurons, n = 7, p < 0.01, and 11.9 ± 1.8% for large-sized neurons, n = 7, p < 0.01, Fig. 8Bb). The extent of sumatriptan inhibition was not different within the three groups. We also examined whether dural afferent neurons express 5-HT1B or 5-HT1D receptor subtypes using a single-cell RT-PCR technique. Transcript for both 5-HT1B and 5-HT1D receptors were simultaneously detected from in all small-sized neurons (n = 3, Fig. 8C). In the case of medium-sized and large-sized DiI-positive neurons, transcript for both 5-HT1B and 5-HT1D receptors were simultaneously detected in 3 of 4 DiI-positive neurons for each size category, and only the for 5-HT1D receptor transcript was detected in the single remaining medium-sized and large-sized DiI-positive neuron (Fig. 8C). (Figure 8 near here)

12

3. Discussion

Primary sensory neurons are heterogeneous in a number of aspects including morphological, immunohistochemical, and electrophysiological properties (Harper and Lawson, 1985; Lawson et al., 1993; Petruska et al., 2000a; 2000b; 2002; Xu et al., 2010). Previous studies have attempted to classify the subpopulations of sensory neurons based on the differences in such properties, but it is generally accepted that small-, medium-, and large-sized sensory neurons are likely to be C-, Aδ- and Aβ-type neurons, respectively, based on the size of the neuronal soma, In the present study, putative dural afferent neurons identified with the fluorescent dye DiI could be divided into three groups based on parameters regarding neuronal size, such as diameter, area, and membrane capacitance. Although a small portion of DiI-positive neurons in each group may overlap with neighboring groups, the membrane capacitance measured using an electrophysiological tool was clearly correlated to the diameter or area. On the other hand, NF200 is a representative marker for myelinated A-type sensory neurons, such as medium- and large-sized neurons (Fornaro et al., 2008; Ho and O'Leary, 2011), but this protein is also found in a subset of small-sized sensory neurons (Fornaro et al., 2008). In the present study, we found that the NF200 immunoreactivity was detected in the majority of large-sized DiI-positive neurons, but we could not detect NF200 in the majority of small- and medium-sized DiI-positive neurons. These results suggest that while most large-sized DiI-positive neurons belong to A-type sensory neurons, but most small- and medium-sized DiI-positive neurons would be C-type sensory neurons. Several ion channels and/or receptors can be used to distinguish between non-nociceptive and nociceptive neurons. For example, TRPV1, one of the nociceptive markers, is mainly expressed in small- and medium-sized sensory neurons within the DRG and TG (Holzer, 1991;

13

Caterina et al., 1997; Guo et al., 1999), and TRPV1 is exclusively found in putative C-fiber neurons (Kobayashi et al., 2005; Yeo et al., 2010). P2X3 receptors are also regarded as a marker for nociceptive neurons (Bradbury et al., 1998; Dunn et al., 2001). In the present study, we found that most small- and medium-sized DiI-positive neurons were sensitive to capsaicin and αβ-me-ATP, suggesting that small- and medium-sized DiI-positive neurons can be classified into C-type sensory neurons. In addition, TTX-R Na+ channels are mainly expressed in small- and medium-sized sensory neurons, and thus are regarded as specific markers for nociceptive C-type and Aδ-type neurons (Black et al., 1996; Chahine and O'Leary, 2014), although TTX-R Na+ channels are also expressed in large-sized sensory neurons within the sensory ganglia (Amaya et al., 2000; Dib-Hajj et al., 1998; Novakovic et al., 1998; Renganathan et al., 2000; Sangameswaran et al., 1996; Ho and O'Leary, 2011; Djouhri et al., 2003). In the present study, TTX-R Na+ currents were recorded from the majority of DiI-positive neurons regardless of neuronal size. Furthermore, a single-cell RT-PCR analysis revealed that the transcript responsible for TTX-R Na+ currents, e.g., NaV1.8, was simultaneously expressed in these neurons. A very small portion of DiI-positive neurons, which did not exhibit TTX-R Na+ currents, the NaV1.8 transcript was not detected. Interestingly, a previous study has shown that a subpopulation of large-sized (>30 μm) DRG neurons express Na V1.8 and NaV1.9 transcripts at 4–5-fold higher levels than the general population of large neurons (Ho and O'Leary, 2011). This subpopulation of large-sized sensory neurons might be similar to medium-sized (30–40 μm) DiI-positive neurons shown in the present study. Together, most small- and medium-sized DiI-positive neurons could be classified into nociceptive polymodal C-type sensory neurons. In addition, considering that NaV1.8 is present in most C- and A-fiber nociceptive neurons but not in muscle spindle or in most cutaneous Aα/β low-threshold mechanoreceptors (Djouhri et al., 2003), most large-sized DiI-positive neurons are could be classified into nociceptive A-type

14

neurons. However, further study will be necessary to elucidate whether large-sized DiI-positive neurons expressing TTX-R Na+ channels are Aδ- or Aβ-type neurons, because these neurons are defined primarily by their axon conduction velocity rather than the expression of TTX-R Na+ channels (Light and Perl, 1993), and Aβ-type nociceptive neurons have been also reported (Djouhri and Lawson, 2004). Migraine is characterized by throbbing pain, and therefore mechanosensitive or stretch-activated ion channels responding to pulsation may play pivotal roles in the pathology of migraine headache (Kaube et al., 1992; Strassman et al., 1996; Levy and Strassman, 2002b), although recent studies have shown no relationship between the throbbing pain and heart rate (Ahn, 2010; Mirza et al., 2012). A previous study has shown that putative mechanosensitive ion channels, which are activated by the application of either a hypotonic solution or a TRPV4 activator, are likely to be involved in the mechanosensation of dural afferent neurons (Wei et al., 2011). In the present study, we found that most large-sized DiI-positive neurons (78.0%), and to a lesser extent, small- and medium-sized DiI-positive neurons (26.4% and 40.5%, respectively) showed the mechanosensitive currents. These membrane currents were blocked by Gd 3+, a nonselective cation channel blocker, were reversed at near 0 mV, and were completely occluded in the absence of extracellular Na + and Ca2+, consistent with the properties of nonselective cation channels. It would be of great interest to elucidate whether such mechanosensitive ion channels expressed in dural afferent neurons play certain roles in the pathology of migraine headache. Previous studies have described the detailed electrophysiological properties of dural afferent neurons including passive and active membrane properties, but these studies have been performed on small-sized dural afferent neurons (Harriott and Gold, 2009; Vaughn and Gold, 2010; Scheff and Gold, 2011). Therefore, the electrophysiological properties of other

15

populations of dural afferent neurons are still unknown. In the present study, we found that the most electrophysiological properties of small-sized DiI-positive neurons, e.g., the characteristics of action potentials and firing pattern, were similar to those of small-sized dural afferent neurons, where these neurons had a wide action potential duration and a higher threshold for action potential generation (Harriott and Gold, 2009). These characteristics are consistent with the representative nociceptive properties of small-sized sensory neurons shown in a previous study (Xu et al., 2010). In contrast, many medium-sized DiI-positive neurons showed the very low threshold for action potential generation, a tonic firing pattern, and spontaneous and regular action potentials even at resting membrane potentials. In the case of large-sized DiI-positive neurons, the higher threshold for action potential generation and single (and to lesser extent tonic) firing patterns were detected. Given that Aδ-type nociceptive neurons transmit the sensory information for high threshold mechanosensation for mechanical pain, these large-sized DiI-positive neurons seem to be Aδ-type nociceptive neurons. Similarly, we found that the membrane currents in response to mechanical stimulation were observed in most large-sized DiI-positive neurons. It should also be noted that action potentials triggered by the depolarizing current injection were not blocked by TTX in all types of DiI-positive neurons. In addition, the peak amplitudes of action potentials were not affected by TTX in these neurons, consistent with a previous study showing that TTX-R Na+ channel underlies the action potential overshoot in the majority of nociceptive afferents (Renganathan et al., 2001). On the other hand, a subset of large-sized DiI-positive neurons without TTX-R Na+ currents, presumably Aβ-type sensory neurons, showed the much higher threshold for action potential generation, and a single firing pattern. Given that Aβ-type sensory neurons are generally regarded as low-threshold mechanoreceptors for processing innocuous tactile sensation (for review; Abraira and Ginty DD, 2013), the physiological roles of this subpopulation of large-sized DiI-positive neurons should

16

be further elucidated. CGRP is regarded as a key molecule that initiates neurogenic inflammation within the dura mater (Silberstein, 2004; Benemei et al., 2009; but see also Amin et al., 2013). The CGRP-mediated neurogenic inflammation leads to the release or production of several inflammatory mediators, e.g., prostaglandins, serotonin, and histamine, around the dura mater, and these mediators are thought to sensitize dural afferent neurons (Bolay et al., 2002; Diener et al., 2012). In this regard, C-type dural afferent neurons would play a pivotal role in the migraine pathology, as 80% of TRPV1-positive dural afferent neurons co-expresses CGRP (Shimizu et al., 2007). In the present study, we found that the CGRP immunoreactivity was found in about 80% of small- or medium-sized DiI-positive neurons examined, and that sumatriptan significantly decreased voltage-gated Ca2+ currents in these DiI-positive neurons. Although the extent of sumatriptan inhibition was small (<15%), this might be due to the expression of multiple types of Ca2+ channels are found in soma of sensory neurons (Huang et al., 1997). In fact, previous studies has shown that sumatriptan selectively inhibits N-type Ca2+ channels in axon or presynaptic terminals (Baillie et al., 2012, Choi et al., 2012). Therefore, the triptan inhibition of voltage-gated Ca2+ channels, which might be mediated by 5-HT1D receptors in C-type dural afferent neurons, and the subsequent decrease of CGRP release in the dura mater may contribute to the antimigraine action of triptans (Jennings et al., 2004). It should also be noted that most small- or medium-sized DiI-positive neurons expressed 5-HT1B in addition to 5-HT1D receptors, suggesting that 5-HT1B receptors would also be pharmacological targets of triptans in dural afferent neurons (Choi et al., 2012; but see also Jennings et al., 2004). On the other hand, we found that sumatriptan also decreased voltage-gated Ca2+ currents in large-sized DiI-positive neurons, and that both 5-HT1B and 5-HT1D receptors were found in most of large-sized neurons tested, suggesting a pharmacological role of triptans in potential high-threshold nociceptive

17

A-type mechanoreceptors. In this regard, it would be of great interest to examine whether mechanosensitive ion channels expressed in these neurons are subjected to peripheral sensitization by the CGRP-mediated neurogenic inflammation. In conclusion, we have characterized the functional properties of dural afferent neurons. Most dural afferent neurons seem to be nociceptive, e.g., polymodal C-type for small- and medium-sized neurons and high-threshold nociceptive A-type mechanoreceptors for large-sized neurons. The present results showing the nociceptive properties of dural afferent neurons would contribute to understand the pathophysiology of primary headaches.

18

4. Materials and Methods

4.1. Preparation All experiments complied with the guiding principles for the care and use of animals approved by the Council of Kyungpook National University (No. KNU 2017-63) and the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and every effort was made to minimize both the number of animals used and their suffering. Sensory neurons of the TG innervating the dura were identified after application of the retrograde tracer 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchloride (DiI) to the dura, as described previously (Harriott and Gold, 2009). Briefly, animals (Sprague Dawley rats, 120–150 g, either sex) were intraperitoneally anaesthetized with a mixture of ketamine (20 mg/kg) and xylazine (10 mg/kg). The cranial bone overlying the superior sagittal sinus was gently removed by a careful craniotomy using a dental drill, and the dura was exposed. Ten l of DiI solution (100 mg/ml in DMSO diluted in 1:10 saline) was applied to the dura. One min after the application of DiI solution, a dental resin was placed on the exposed dura to replace the removed cranial bone. The incision was closed with sutures, and penicillin G (100,000 U/kg) and naproxen (10 mg/kg) were intramuscularly injected into the rats to reduce postoperative infection and pain. Seven to 10 days after the DiI application, rats were decapitated under ketamine anesthesia (50 mg/kg, ip). The TG was dissected and a pair of ganglia (V1, ophthalmic region, Supplementary Fig. S1) was treated with a standard external solution [in mM; 150 NaCl, 3 KCl, 2 CaCl2, 1 MgCl2, 10 glucose and 10 Hepes, (pH 7.4 with Tris-base)] containing 0.3 % collagenase (type I) and 0.3 % trypsin (type I) for 40–60 min at 37°C. Thereafter, TG neurons were dissociated mechanically by triturating with fire-polished Pasteur pipettes in a culture dish (Primaria 3801, Becton Dickinson, Rutherford, NJ, USA). The isolated neurons were used for

19

immunohistochemistry or electrophysiological recordings 1–6 h after preparation. Images of TG neurons were obtained using CCD cameras (ProgRes® MF, Jenoptik, Yena, Germany or DS-Ri1, Nikon, Tokyo, Japan).

4.2. Electrical measurements All electrical measurements were performed using conventional whole-cell patch recordings and a standard patch-clamp amplifier (Axopatch 200B; Molecular Devices, Union City, CA, USA). Neurons were voltage clamped at a holding potential (V H) of -80 mV, except where indicated. Patch pipettes were made from borosilicate capillary glass (G-1.5; Narishige, Tokyo, Japan) by use of a pipette puller (P-97; Sutter Instrument Co., Novato, CA, USA). The resistance of the recording pipettes filled with the internal solution was 0.8–1.5 MΩ. Membrane potentials were corrected for the liquid junction potential and the pipette capacitance and series resistance (60–90%) were compensated for. DiI-positive or DiI-negative TG neurons were viewed under phase contrast or fluorescence on an inverted microscope (TE2000; Nikon, Tokyo, Japan). Membrane currents were filtered at 2–5 kHz, digitized at 5–20 kHz, except where indicated, and stored on a computer equipped with pCLAMP 10.3 (Molecular Devices). To isolate voltage-gated Na+ currents (TTX-R INa), the internal solution was composed of the following (in mM): 120 CsF, 20 tetraethylammonium-Cl, 10 NaCl, 2 EGTA, 2 Mg-ATP, and 10 Hepes (pH 7.2 with Tris-base). The bath solution was composed of the following (in mM): 130 NaCl, 20 TEA-Cl, 2 CaCl2, 1 MgCl2, 10 Hepes, 10 glucose, and 0.01 CdCl2 (pH 7.4 with Tris-base). To record voltage-gated Na+ currents, capacitative and leak currents were subtracted using the P/4 subtraction protocol (pCLAMP 10.3). To isolate the voltage-gated Ca2+ currents, the internal solution was composed of the following (in mM): 140 Cs-methanesulfonate, 10 CsCl, 2 EGTA, 2 Mg-ATP, and 10 Hepes (pH 7.2 with Tris-base). Bath solution was composed

20

of the following (in mM): 150 N-methyl-D-glucamine-Cl, 3 CsCl, 2 CaCl2, 1 MgCl2, 10 Hepes, and 10 glucose (pH 7.4 with Tris-base). Recordings of Ca2+ currents were generally finished within 3 min after membrane rupture, and the rundown of Ca2+ currents was negligible during this recording time. In current-clamp experiments, the internal solution was composed of the following (in mM): 140 KF, 10 KCl, 2 EGTA, 2 Mg-ATP, and 10 Hepes (pH 7.2 with Tris-base). All experiments were performed at room temperature (22–25ºC). To record mechanical stimulation-induced membrane currents, a glass pipette (G-1.5; Narishige) was placed at an angle of 45° by using a micromanipulator and was carefully touched to the surface of DiI-positive neurons. A sharp tip (diameter 2–3 μm) of glass pipette was clogged by fire polishing and the other side was connected with a microinjector (PLI-100A, Warner Instruments, Hamden, CT, USA). The pressure injection of nitrogen gas was used to facilitate the forward movement of the glass pipette. The moving distance of the tip of the glass pipette was calculated using video image, and the pressure was typically adjusted to 3–10 psi so that the 10 ms duration of the pressure injection corresponded to a movement of 2 μm, while the velocity of tip movement was 200 μm/s during the 100 ms duration of pressure injection. Mechanical stimuli were applied every 3–5 s for the full recovery of current amplitude during the recording.

4.3. Single-cell RT-PCR After whole-cell patch-clamp recording, the contents of the recorded DiI-positive neuron, including mRNA, were aspirated by applying a gentle suction through the recording pipette. Then, the harvested material in the patch pipette was expelled into a PCR tube, and reverse transcription and 1st PCR reactions were performed in the same tube using a one-step RT-PCR kit (Qiagen, Hilden, Germany). Primers used for RT-PCR were as follows: GAPDH (1st)

21

5’-catcttccaggagcgagatcc-3’/5’-cagtgagcttcccgttcagct-3’,

(2nd)

5’-tggagtctactggcgtcttcac-3’/5’-gatgcagggatgatgttctggg-3’ (product size 343 bp), Nav1.8; (1st) 5’-atcaagggtgtcagagggctc-3’/5’-cctctgtgctcggaaaggttc-3’, 5’-tgacccttacaaccagcgcag-3’/5’-cgcagacaggaagctcttctg-3’ receptor

;

(1st)

(2nd) (product

size

361

bp),

5’-gctaactacctgatcgcctcg-3’/5’-accgtggagtagaccgtgtag3’,

5-HT1B

(2nd)

5’-

tgacctgctcgtgtccatcct-3’/5’-cacaaagcagtccagcacctc -3’ (product size 322 bp), and 5-HT1D receptor;

(1st)

5’-gtcctgcaggcactcagaatc-3’/5’-gatggcccagtatctgtccag-3’,

(2nd)

5’-ccactgtcctctccaatgcct-3’/5’-gcgatgacacagagatgcagg-3’ (product size 247 bp). The nested 2nd PCR was performed using GoTaqⓇ DNA polymerase (Promega) and each first-round PCR product was used as a template (2 µl). In a subset of experiments, negative control reactions were also performed without the reverse transcription. Amplified PCR products were electrophoresed in 2% agarose gels, to which RedSafeTM Nucleic Acid Staining Solution had been added, and the gels subsequently photographed.

4.4. Data analysis The amplitudes of membrane currents were obtained by subtracting the peak value from the baseline currents using a pCLAMP 10.3 program. The continuous curve for the concentration-inhibition relationship was fitted using a least-squares fit to the following equation: I = 1 – [Cn / (Cn + IC50n)], where I is the inhibition ratio of Gd3+-induced mechanosensitive currents, C is the concentration of Gd3+, IC50 is the concentration for the half-effective response and n is the Hill coefficient. The rheobase (or threshold) currents were determined as the least amplitude of depolarizing current for the generation of action potentials. Firing patterns were determined as single, phasic, and

22

tonic patterns, based on the number and successiveness of action potentials triggered by 4-fold rheobase current injections. Numerical values are provided as the mean ± SEM using values normalized to the control. Significant differences in the mean amplitude were tested using Student’s paired two-tailed t-test, except where indicated, using absolute values rather than normalized ones. Values of p < 0.05 were considered significantly different.

4.5. Immunohistochemistry After enzymatically isolated single TG neurons had settled and adhered to a 35-mm culture dish, TG neurons were washed with phosphate-buffered saline (PBS, pH 7.4) and fixed with 4% paraformaldehyde in PBS for 30 min. After treatment with 0.1% Triton-X 100 for 15 min, neurons were incubated with PBS containing mouse anti-NF200 antibody (1:1000; Sigma, St. Louis, MO, USA) or rabbit anti-CGRP antibody (1:1000; Immunostar, Hudson, WI, USA) and 1.5% normal bovine serum for 1 h. Then neurons were incubated with Alexa Fluor 488 goat anti-mouse or anti-rabbit secondary antibody (1:2000; Thermo Fisher Scientific Korea Ltd., Seoul, Korea) for 1 h. Images of the neurons were collected with a digital camera (DS-Ri1, Nikon) and stored in a computer using the NIS-Elements program (Nikon). All procedures were performed at room temperature.

4.6. Drugs The drugs used in the present study were collagenase, trypsin, tetrodotoxin (TTX), ATP, αβ-methylene-ATP (αβ-me-ATP), capsaicin, allyl isothiocyanate (AITC), menthol, GdCl 3, ZD7288, sumatriptan, and DiI (from Sigma). All solutions containing drugs were applied using the ‘Y–tube system’ for rapid solution exchange (Murase et al., 1989).

23

5. Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (2008-0062282 and NRF-2015R1D1A1A01060873).

Conflict of interest

The authors declare no competing financial interests.

24

6. References

Abraira, V.E., Ginty, D.D., 2013. The sensory neurons of touch. Neuron 79, 618-639. Ahn, A.H., 2010. On the temporal relationship between throbbing migraine pain and arterial pulse. Headache 50, 1507-1510. Amaya, F., Decosterd, I., Samad, T.A., Plumpton, C., Tate, S., Mannion, R.J., Costigan, M., Woolf, C.J. 2000. Diversity of expression of the sensory neuron-specific TTX-resistant voltage-gated sodium ion channels SNS and SNS2. Mol. Cell. Neurosci. 15, 331-342. Amin, F.M., Asghar, M.S., Hougaard, A., Hansen, A.E., Larsen, V.A., de Koning, P.J., Larsson, H.B., Olesen, J., Ashina, M., 2013. Magnetic resonance angiography of intracranial and extracranial arteries in patients with spontaneous migraine without aura: a cross-sectional study. Lancet Neurol. 12, 454-461. Andres, K.H., von Düring, M., Muszynski, K., Schmidt, R.F., 1987. Nerve fibres and their terminals of the dura mater encephali of the rat. Anat. Embryol. (Berl.) 175, 289-301. Baillie, L.D., Ahn, A.H., Mulligan, S.J., 2012. Sumatriptan inhibition of N-type calcium channel mediated signaling in dural CGRP terminal fibres. Neuropharmacology 63, 362-367. Benemei, S., Nicoletti, P., Capone, J.G., Geppetti, P., 2009. CGRP receptors in the control of pain and inflammation. Curr. Opin. Pharmacol. 9, 9-14. Bigal, M.E., Ferrari, M., Silberstein, S.D., Lipton, R.B., Goadsby, P.J., 2009. Migraine in the triptan era: lessons from epidemiology, pathophysiology, and clinical science. Headache 49 Suppl 1, S21-S33. Black, J.A., Dib-Hajj, S., McNabola, K., Jeste, S., Rizzo, M.A., Kocsis, J.D., Waxman, S.G., 1996. Spinal sensory neurons express multiple sodium channel alpha-subunit mRNAs. Brain Res. Mol. Brain Res. 43, 117-131.

25

Bolay, H., Reuter, U., Dunn, A.K., Huang, Z., Boas, D.A., Moskowitz, M.A., 2002. Intrinsic brain activity triggers trigeminal meningeal afferents in a migraine model. Nat. Med. 8, 136-142. Bove, G.M., Moskowitz, M.A., 1996. Primary afferent neurons innervating guinea pig dura. J. Neurophysiol. 77, 299-308. Bradbury, E.J., Burnstock, G., McMahon, S.B., 1998. The expression of P2X3 purinoreceptors in sensory neurons: effects of axotomy and glial-derived neurotrophic factor. Mol. Cell. Neurosci. 12, 256-268. Caterina, M.J., Schumacher, M.A., Tominaga, M., Rosen, T.A., Levine, J.D., Julius, D., 1997. The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 389:816-824. Chahine, M., O'Leary, M.E., 2014. Regulation/modulation of sensory neuron sodium channels. Handb. Exp. Pharmacol. 221, 111-135. Choi, I.S., Cho, J.H., An, C.H., Jung, J.K., Hur, Y.K., Choi, J.K., Jang, I.S., 2012. 5-HT1B receptors inhibit glutamate release from primary afferent terminals in rat medullary dorsal horn neurons. Br. J. Pharmacol. 167, 356-367. Christensen, A.P., Corey, D.P., 2007. TRP channels in mechanosensation: direct or indirect activation? Nat. Rev. Neurosci. 8, 510-521. Coddou, C., Yan, Z., Obsil, T., Huidobro-Toro, J.P., Stojilkovic, S.S., 2011. Activation and regulation of purinergic P2X receptor channels. Pharmacol. Rev. 63, 641-683. Delmas. P., Hao, J., Rodat-Despoix, L., 2011. Molecular mechanisms of mechanotransduction in mammalian sensory neurons. Nat. Rev. Neurosci. 12, 139-153. Dib-Hajj, S.D., Tyrrell, L., Black, J.A., Waxman, S.G., 1998. NaN, a novel voltage-gated Na channel, is expressed preferentially in peripheral sensory neurons and down-regulated after

26

axotomy. Proc. Natl. Acad. Sci. U.S.A. 95, 8963-8968. Diener, H.C., Dodick, D.W., Goadsby, P.J., Lipton, R.B., Olesen, J., Silberstein, S.D., 2012. Chronic migraine--classification, characteristics and treatment. Nat. Rev. Neurol. 8, 162-171. Djouhri, L., Fang, X., Okuse, K., Wood, J.N., Berry, C.M., Lawson, S.N., 2003. The TTX-resistant sodium channel Nav1.8 (SNS/PN3): expression and correlation with membrane properties in rat nociceptive primary afferent neurons. J. Physiol. 550, 739-752. Djouhri, L., Lawson, S.N., 2004. Aβ-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. 46, 131-145. Dunn, P.M., Zhong, Y., Burnstock, G., 2001. P2X receptors in peripheral neurons. Prog. Neurobiol. 65, 107-134. Fornaro, M., Lee, J.M., Raimondo, S., Nicolino, S., Geuna, S., Giacobini-Robecchi, M., 2008. Neuronal intermediate filament expression in rat dorsal root ganglia sensory neurons: an in vivo and in vitro study. Neuroscience 153, 1153-1163. Goadsby, P.J., 2007. Recent advances in understanding migraine mechanisms, molecules and therapeutics. Trends Mol. Med. 13, 39-44. Guo, A., Vulchanova, L., Wang, J., Li, X., Elde, R., 1999. Immunocytochemical localization of the vanilloid receptor 1 (VR1): relationship to neuropeptides, the P2X3 purinoceptor and IB4 binding sites. Eur. J. Neurosci. 11, 946-958. Harper, A.A., Lawson, S.N., 1985. Electrical properties of rat dorsal root ganglion neurones with different peripheral nerve conduction velocities. J. Physiol. 359, 47-63. Harriott, A.M., Gold, M.S., 2009. Electrophysiological properties of dural afferents in the

27

absence and presence of inflammatory mediators. J. Neurophysiol. 101, 3126-3134. Ho, C., O'Leary, M.E., 2011. Single-cell analysis of sodium channel expression in dorsal root ganglion neurons. Mol. Cell. Neurosci. 46, 159-166. Holzer, P., 1991. Capsaicin: cellular targets, mechanisms of action, and selectively for thin sensory neurons. Pharmacol. Rev. 43, 143-201. Huang, C.S., Song, J.H., Nagata, K., Yeh, J.Z., Narahashi, T., 1997. Effects of the neuroprotective agent riluzole on the high voltage-activated calcium channels of rat dorsal root ganglion neurons. J. Pharmacol. Exp. Ther. 282, 1280-1290. Huang, D., Li, S., Dhaka, A., Story, G.M., Cao, Y.Q., 2012. Expression of the transient receptor potential channels TRPV1, TRPA1 and TRPM8 in mouse trigeminal primary afferent neurons innervating the dura. Mol. Pain 8, 66. Jennings, E.A., Ryan, R.M., Christie, M.J., 2004. Effects of sumatriptan on rat medullary dorsal horn neurons. Pain 111, 30-37. Kaube, H., Hoskin, K.L., Goadsby, P.J., 1992. Activation of the trigeminovascular system by mechanical distension of the superior sagittal sinus in the cat. Cephalalgia 12, 133-136. Kobayashi, K., Fukuoka, T., Obata, K., Yamanaka, H., Dai, Y., Tokunaga, A., Noguchi, K., 2005. Distinct expression of TRPM8, TRPA1, and TRPV1 mRNAs in rat primary afferent neurons with Aδ/C-fibers and colocalization with Trk receptors. J. Comp. Neurol. 493, 596-606. Lauritzen, M., 1994. Pathophysiology of the migraine aura. The spreading depression theory. Brain 117, 199-210. Lawson, S.N., 2002. Phenotype and function of somatic primary afferent nociceptive neurones with C-, Aδ- or Aα/β-fibres. Exp. Physiol. 87, 239-244. Lawson, S.N., Perry, M.J., Prabhakar, E., McCarthy, P.W., 1993. Primary sensory neurones:

28

neurofilament, neuropeptides, and conduction velocity. Brain Res. Bull. 30, 239-243. Leao, A.A.P., 1944. Spreading depression of activity in cerebral cortex. J. Neurophysiol. 7, 359-390. Le Pichon, C.E., Chesler, A.T., 2014. The functional and anatomical dissection of somatosensory subpopulations using mouse genetics. Front. Neuroanat. 8, 21. Levy, D., Strassman, A.M., 2002a. Distinct sensitizing effects of the cAMP-PKA second messenger cascade on rat dural mechanonociceptors. J. Physiol. 538, 483-493. Levy, D., Strassman, A.M., 2002b. Mechanical response properties of A and C primary afferent neurons innervating the rat intracranial dura. J. Neurophysiol. 88, 3021-3031. Light,A.R., Perl, E.R., 1993. Peripheral sensory systems. Dyck, P.J., Thomas, P.K., Griffin, J.W., Low, P.A., Poduslo J.F., (Eds.), Peripheral Neuropathy, Saunders, W.B., Philadelphia, pp. 149-165. Mense, S., 1990. Structure-function relationships in identified afferent neurones. Anat. Embryol. 181, 1-17. Mirza, A.F., Mo, J., Holt, J.L., Kairalla, J.A., Heft, M.W., Ding, M., Ahn, A.H., 2012. Is there a relationship between throbbing pain and arterial pulsations? J. Neurosci. 32, 7572-7576. Murase, K., Ryu, P.D., Randic, M., 1989. Excitatory and inhibitory amino acids and peptide-induced responses in acutely isolated rat spinal dorsal horn neurons. Neurosci. Lett. 103, 56-63. Novakovic, S.D., Tzoumaka, E., McGivern, J.G., Haraguchi, M., Sangameswaran, L., Gogas, K.R., Eglen, R.M., Hunter, J.C., 1998. Distribution of the tetrodotoxin-resistant sodium channel PN3 in rat sensory neurons in normal and neuropathic conditions. J. Neurosci. 18, 2174-2187. O'Connor, T.P., van der Kooy, D., 1986. Pattern of intracranial and extracranial projections of

29

trigeminal ganglion cells. J. Neurosci. 6, 2200-2207. Petruska, J.C., Cooper, B.Y., Gu, J.G., Rau, K.K., Johnson, R.D., 2000a. Distribution of P2X1, P2X2, and P2X3 receptor subunits in rat primary afferents: relation to population markers and specific cell types. J. Chem. Neuroanat. 20, 141-162. Petruska, J.C., Napaporn, J., Johnson, R.D., Cooper, B.Y., 2002. Chemical responsiveness and histochemical phenotype of electrophysiologically classified cells of the adult rat dorsal root ganglion. Neuroscience 115, 15-30. Petruska, J.C., Napaporn, J., Johnson, R.D., Gu, J.G., Cooper, B.Y., 2000b. Subclassified acutely dissociated cells of rat DRG: histochemistry and patterns of capsaicin-, proton-, and ATP-activated currents. J. Neurophysiol. 84, 2365-2379. Renganathan, M., Cummins, T.R., Waxman, S.G., 2001. Contribution of Nav1.8 sodium channels to action potential electrogenesis in DRG neurons. J. Neurophysiol. 86, 629-640. Ruskell, G.L., 1988. The tentorial nerve in monkeys is a branch of the cavernous plexus. J. Anat. 157, 67-77. Sangameswaran, L., Delgado, S.G., Fish, L.M., Koch, B.D., Jakeman, L.B., Stewart, G.R., Sze, P., Hunter, J.C., Eglen, R.M., Herman, R.C., 1996. Structure and function of a novel voltage-gated, tetrodotoxin-resistant sodium channel specific to sensory neurons. J. Biol. Chem. 271, 5953-5956. Scheff, N.N., Gold, M.S., 2011. Sex differences in the inflammatory mediator-induced sensitization of dural afferents. J. Neurophysiol. 106, 1662-1668. Shimizu, T., Toriumi, H., Sato, H., Shibata, M., Nagata, E., Gotoh, K., Suzuki, N., 2007. Distribution and origin of TRPV1 receptor-containing nerve fibers in the dura mater of rat. Brain Res. 1173, 84-91. Silberstein, S.D., 2004. Migraine pathophysiology and its clinical implications. Cephalalgia 24

30

(Suppl 2), 2-7. Strassman, A.M., Raymond, S.A., Burstein, R., 1996. Sensitization of meningeal sensory neurons and the origin of headaches. Nature 384, 560-564. Strassman, A.M., Weissner, W., Williams, M., Ali, S., Levy, D., 2004. Axon diameters and intradural trajectories of the dural innervation in the rat. J. Comp. Neurol. 473, 364-376. Vaughn, A.H., Gold, M.S., 2010. Ionic mechanisms underlying inflammatory mediator-induced sensitization of dural afferents. J. Neurosci. 30, 7878-7888. Xu, S., Ono, K., Inenaga, K., 2010. Electrophysiological and chemical properties in subclassified acutely dissociated cells of rat trigeminal ganglion by current signatures. J. Neurophysiol. 104, 3451-3461. Yan, J., Edelmayer, R.M., Wei, X., De Felice, M., Porreca, F., Dussor, G., 2011. Dural afferents express acid-sensing ion channels: a role for decreased meningeal pH in migraine headache. Pain 152, 106-113. Yeo, E.J., Cho, Y.S., Paik, S.K., Yoshida, A., Park, M.J., Ahn, D.K., Moon, C., Kim, Y.S., Bae, Y.C., 2010. Ultrastructural analysis of the synaptic connectivity of TRPV1-expressing primary afferent terminals in the rat trigeminal caudal nucleus. J. Comp. Neurol. 518, 4134-4146.

31

7. Figure Legends

Figure 1. Morphological properties of DiI-positive neurons. A, Typical phase contrast (Ph, left) and fluorescence (DiI, right) images of small-, medium-, and large-sized DiI-positive neurons. B, Distribution of diameter (a), area (b), and capacitance (c) of DiI-positive neurons. Bin sizes were 2 μm, 50 μm2, and 5 pF for diameter (n = 1220), area (n = 1114), and capacitance (n = 346), respectively. Continuous lines were Gaussian fitting results. Diameter was defined as the mean of the length of longest and shortest transversal lines. Area was calculated using by a CCD camera (ProgRes® MF) and its program. Membrane capacitance were measured using by the pClamp program during whole-cell patch-clamp recording. C, Scatter plots of diameter (a) or area (b) against membrane capacitance estimated from individual DiI-positive neurons. Linear lines were the least-squares linear fit.

Figure 2. Immunohistochemical properties of DiI-positive neurons. A, Typical phase contrast (Ph; left) and fluorescence (DiI; middle, NF200; right) images of small-, medium-, and large-sized DiI-positive neurons. Single TG neurons were enzymatically isolated, fixed, and then treated to anti-NF200 antibody. B, The proportion of NF200-positive and NF200-negative neurons in small- (n = 127, medium(n = 143), and large- (n = 162) sized DiI-positive neurons. The parenthesis represents the numbers NF200-positive neurons and all DiI-positive neurons. C, Typical phase contrast (Ph; left) and fluorescence (DiI; middle, CGRP; right) images of small-, medium-, and large-sized DiI-positive neurons. D, The proportion of CGRP-positive and NF200-negative neurons in small- (n = 99, medium- (n

32

= 95), and large- (n = 98) sized DiI-positive neurons. The parenthesis represents the numbers CGRP-positive neurons and all DiI-positive neurons.

Figure 3. Nociceptive properties of DiI-positive neurons. A, Typical phase contrast images of small- (a), medium- (b), and large- (c) sized DiI-positive neurons (not shown) during the whole-cell patch-clamp recording. B, Typical raw traces observed from DiI-positive neurons (small; a, medium; b, and large; c) before, during, and the application of various agonists [pH 6.0 for ASICs, ATP and αβ-me-ATP for P2X receptors, capsaicin (Caps) for TRPV1, AITC for TRPA1, and menthol (Ment) for TRPM8, respectively]. All DiI-positive neurons were clamped at -80 mV. In a, the capsaicin-induced current was truncated. C, The proportion of each agonist-positive small- (a), medium- (b), and large- (c) sized DiI-positive neurons. DiI-positive neurons that each agonist induced the membrane currents of ≥30 pA was regarded as agonist-positive neurons. The parenthesis represents the numbers of agonist-positive neurons and all neurons tested in each case.

Figure 4. TTX-R Na+ currents in DiI-positive neurons. A, Typical raw traces of voltage-gated Na+ currents observed from DiI-positive neurons (small; a, medium; b, and large; c and d) in the absence and presence of 300 nM TTX. Voltage-gated Na+ currents were elicited by brief voltage step pulses (50 ms duration, -80 mV to -10 mV in a, b, and c, and -100 mV to -20 mV in d). B, The proportion of small- (S), medium- (M), and large- (L) sized DiI-positive neurons that exhibit TTX-R Na+ currents (INa). The existence of TTX-R Na+ currents in DiI positive neurons was defined when TTX decreased voltage-gated Na+ currents by ≥90% of total Na+

33

currents. The parenthesis represents the numbers of DiI-positive neurons showing TTX-R Na+ currents and all neurons tested in each case. C, Single-cell RT-PCR analysis from small-, medium-, and large-sized DiI-positive neurons. While the transcript for NaV1.8 (361 bp) was detected from all DiI-positive neurons tested that TTX did not eliminated voltage-gated Na+ currents (n = 6 for small-sized, n = 8 for medium-sized, and n = 8 for large-sized DiI-positive neurons), it was not detected from large-sized DiI-positive neurons that TTX completely eliminated voltage-gated Na+ currents (n = 4).

Figure 5. Electrophysiological properties of DiI-positive neurons. A, Typical traces of voltage response during the depolarizing current injection in small-, medium-, and large-sized DiI-positive neurons. Large-sized DiI-positive neurons were further divided into two groups, such as TTX-R and TTX-S, according to the expression of TTX-R Na+ channels (see also Supplementary Fig. S3). The threshold currents to generate action potentials were 150 pA (small), 100 pA (medium), 280 pA (large, TTX-R), and 400 pA (large, TTX-S), and the depolarizing currents corresponding the integers of threshold (T) currents were injected to each size of DiI-positive neurons. Ba, The amplitude of threshold current (rheobase) in small-, medium-, and large- (TTX-R and TTX-S) sized DiI-positive neurons. Each column and error bar was the mean and SEM from 46 (small-sized), 74 (medium-sized), 62 (large-sized, TTX-R), and 15 (large-sized, TTX-S) experiments. b, The number of action potentials elicited by the depolarizing currents (1T to 4T) injection in small-, medium-, and large- (TTX-R and TTX-S) sized DiI-positive neurons. All points and error bars were the mean and SEM from 46 (small-sized), 74 (medium-sized), 62 (large-sized, TTX-R), and 15 (large-sized, TTX-S) experiments.

34

C, The proportion of small-, medium-, and large- (TTX-R and TTX-S) sized DiI-positive neurons that exhibit different firing patterns, such as single firing pattern, phasic firing pattern, and tonic firing pattern, in response to the depolarizing current injection (3T). The parenthesis represents the numbers of DiI-positive neurons showing different firing patterns and all neurons tested.

Figure 6. Properties of mechanosensitive currents in DiI-positive neurons. A, Typical fluorescence (DiI, left) and phase contrast (Ph; right) images of large-sized DiI-positive neurons. A fire-polished glass pipette (Stim.) was used for mechanical stimulation. For details, see Methods section. B, A typical raw trace of membrane currents during mechanical stimulation (10 μm movement). The mechanosensitive currents were elicited by the stimulating pipette in every 5 s. Inset represents mechanosensitive currents with an expanded time scale. Ca, Typical raw traces of mechanosensitive currents at various holding potentials (VH). b, Current-voltage relationship of the mechanosensitive currents. Each point and error bar was the mean and SEM from 8 experiments (2 small-sized, 2 medium-sized, and 4 large-sized neurons). Note that the reversal potential of mechanosensitive currents was near 0 mV, which was similar to the theoretical equilibrium potential of monovalent cations calculated by Nernst equation. Da, Typical raw traces of mechanosensitive currents observed before, during, and after the application of Na+-free external solution in large-sized DiI-positive neurons. Note that the mechanosensitive currents were not completely disappeared in the Na +-free external solution (arrowheads). b, Changes in the amplitude of mechanosensitive currents by Na+-free and both Na+- and Ca2+-free external solutions. **; p < 0.01.

35

Ea, Typical raw traces of mechanosensitive currents observed before, during, and after the application of 0.1 μM, 1 μM, and 10 μM Gd3+ in large-sized DiI-positive neurons. b, Concentration-inhibition relationship of Gd3+. The IC50 value calculated from curve fitting result was 0.56 μM. Each point and error bar represents the mean and SEM from 8 experiments (2 small-sized, 2 medium-sized, and 4 large-sized neurons).

Figure 7. Roles of mechanosensitive ion channels in neuronal excitability. A, The proportion of small- (S), medium- (M), and large- (L) sized DiI-positive neurons that exhibit mechanosensitive currents (Imechano). DiI-positive neurons that the mechanical stimulation of 10 μm movement induced the current amplitude of ≥30 pA was regarded as mechanosensitive neurons. The parenthesis represents the numbers of mechanosensitive neurons and all neurons tested in each size of DiI-positive neurons. B, Typical raw traces of mechanosensitive currents observed before, during, and after the various intensities of mechanical stimulation in large-sized DiI-positive neurons. C, Mechanical stimulation (2 μm increments)-response relationship for the density of mechanosensitive currents in each size of DiI-positive neurons. Each point and error bar represents the mean and SEM from 8 experiments in small-sized, medium-sized, and large-sized DiI-positive neurons, respectively. D, Typical raw traces of mechanosensitive currents (upper) and voltages (lower) observed by mechanical stimulation with various intensities in voltage-clamp (V-Clamp) and current-clamp (I-Clamp) conditions, respectively, in large-sized DiI-positive neurons. Five successive current or voltage responses were represented for each stimulation intensity in the same large-sized DiI-positive neurons. E, The probability of action potential generation by mechanical stimulation (2 μm increments)

36

in small-, medium-, and large-sized DiI-positive neurons. Ten successive mechanical stimuli were applied to DiI-positive neurons and the number of sweep that action potentials are generated by mechanical stimuli were counted. Each point and error bar was the mean and SEM from 8 experiments in small-sized, medium-sized, and large-sized DiI-positive neurons, respectively.

Figure 8. Voltage-gated Ca2+ currents in DiI-positive neurons. Aa, Typical raw traces of voltage-gated Ca2+ currents observed from DiI-positive neurons in the absence and presence of 100 μM Cd2+, a general voltage-gated Ca2+ channel blocker. Voltage-gated Ca2+ currents were elicited by brief voltage step pulses (50 ms duration, -80 mV to 0 mV. b, Typical raw traces of voltage-gated Ca2+ currents observed from DiI-positive neurons in the absence and presence of 10 μM sumatriptan, a selective 5-HT1B/1D receptor agonist. c, Time courses of the amplitude of Ca2+ currents before, during, and after application of 10 μM sumatriptan. The Ca2+ currents were elicited every 15 s. Each point and error bar represents the mean and SEM from 6 (small-sized), 7 (medium-sized), and 7 (large-sized) experiments. **; p < 0.01. Ba, The current density of voltage-gated Ca2+ currents in small- (S), medium- (M), and large(L) sized DiI-positive neurons. Each column and error bar represents the mean and SEM from 6 (small-sized), 7 (medium-sized), and 7 (large-sized) experiments. n.s; not significant (ANOVA). b, Sumatriptan-induced changes in voltage-gated Ca2+ currents in in small- (S), medium- (M), and large- (L) sized DiI-positive neurons. Each column and error bar represents the mean and SEM from 6 (small-sized), 7 (medium-sized), and 7 (large-sized) experiments. **; p < 0.01. Ca, Single-cell RT-PCR analysis from small-, medium-, and large-sized DiI-positive neurons

37

shown in A. The transcripts for both 5-HT1B (1B; 322 bp) and 5-HT1D receptors (1D; 247 bp) were simultaneously detected from most DiI-positive neurons tested (n = 4 for small-sized, n = 4 for medium-sized, and n = 4 for large-sized DiI-positive neurons). Negative; negative control. b, The proportion of small- (S), medium- (M), and large- (L) sized DiI-positive neurons that express the transcripts for 5-HT1B and/or 5-HT1D receptors. The parenthesis represents the numbers of DiI-positive neurons expressing 5-HT1B and/or 5-HT1D receptors and all neurons tested in each size of DiI-positive neurons.

38

39

40

41

42

43

44

45

46

Highlights

1. Dural afferent neurons were divided into small-, medium-, and large-sized neurons. 2. Small- and medium-sized dural afferent neurons were polymodal C-type neurons. 3. Large-sized dural afferent neurons were high-threshold Aδ-type mechanoreceptors. 4. Differential nociceptive properties would help understanding the pathology of primary headaches.

47