FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7

Article

FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7 Highlights

Authors

d

Loss of FGF13 in mechanoheat nociceptors selectively abolishes heat nociception

Liu Yang, Fei Dong, Qing Yang, ..., Limin Chen, Lan Bao, Xu Zhang

d

Heat-evoked firing of action potentials is impaired in FGF13deficient nociceptors

Correspondence

d

Heat-facilitated FGF13/Nav1.7 interaction maintains Nav1.7 levels in cell membranes

In Brief

d

Disrupting FGF13/Nav1.7 interaction reduces heat-evoked neuron firing and behaviors

Yang et al., 2017, Neuron 93, 1–16 February 22, 2017 ª 2017 Elsevier Inc. http://dx.doi.org/10.1016/j.neuron.2017.01.009

[email protected]

Yang et al. reveal that non-secretory FGF13 in mechanoheat nociceptors selectively regulates heat nociception by interacting with Nav1.7. The heatfacilitated FGF13/Nav1.7 interaction promotes the current density and cell membrane localization of Nav1.7, enabling the heat-evoked neuronal firing and nociceptive behaviors.

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Neuron

Article FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7 Liu Yang,1,8 Fei Dong,1,8 Qing Yang,3 Pai-Feng Yang,4 Ruiqi Wu,5 Qing-Feng Wu,1 Dan Wu,1 Chang-Lin Li,1 Yan-Qing Zhong,1 Ying-Jin Lu,1 Xiaoyang Cheng,7 Fu-Qiang Xu,5 Limin Chen,3,4 Lan Bao,2,6 and Xu Zhang1,6,9,* 1Institute of Neuroscience and State Key Laboratory of Neuroscience, CAS Center for Excellence in Brain Science and Intelligence Technology 2State Key Laboratory of Cell Biology, CAS Center for Excellence in Cell Biology, Institute of Biochemistry and Cell Biology Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China 3Shanghai Clinical Research Center, Chinese Academy of Sciences/XuHui Central Hospital, Shanghai 200031, China 4Department of Radiology and Radiological Science, Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN 37232, USA 5Key Laboratory of Magnetic Resonance in Biological Systems and State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan 430071, China 6School of Life Science and Technology, ShanghaiTec University, Shanghai 201210, China 7Discipline of Neuroscience and Department of Anatomy, Histology and Embryology, Collaborative Innovation Center for Brain Science, Shanghai Jiaotong University School of Medicine, Shanghai 200025, China 8Co-first author 9Lead Contact *Correspondence: [email protected] http://dx.doi.org/10.1016/j.neuron.2017.01.009

SUMMARY

The current knowledge about heat nociception is mainly confined to the thermosensors, including the transient receptor potential cation channel V1 expressed in the nociceptive neurons of dorsal root ganglion (DRG). However, the loss of thermosensors only partially impairs heat nociception, suggesting the existence of undiscovered mechanisms. We found that the loss of an intracellular fibroblast growth factor (FGF), FGF13, in the mouse DRG neurons selectively abolished heat nociception. The noxious heat stimuli could not evoke the sustained action potential firing in FGF13-deficient DRG neurons. Furthermore, FGF13 interacted with the sodium channel Nav1.7 in a heat-facilitated manner. FGF13 increased Nav1.7 sodium currents and maintained the membrane localization of Nav1.7 during noxious heat stimulation, enabling the sustained firing of action potentials. Disrupting the FGF13/Nav1.7 interaction reduced the heat-evoked action potential firing and nociceptive behavior. Thus, beyond the thermosensors, the FGF13/Nav1.7 complex is essential for sustaining the transmission of noxious heat signals. INTRODUCTION Fibroblast growth factors (FGFs) are classified into secretory (FGF1–10 and FGF15–23) or intracellular, non-secretory forms (FGF11–14) (Guillemot and Zimmer, 2011; Zhang et al., 2012). Secretory FGFs are known for their functions in neural develop-

ment, but knowledge about intracellular FGFs remains limited (Guillemot and Zimmer, 2011; Zhang et al., 2012). FGF13 is an intracellular FGF expressed in the nervous system (Hartung et al., 1997). It is transiently expressed in cortical neurons during brain development (Wu et al., 2012) but maintains high-level expression from development to adulthood in the dorsal root ganglion (DRG) (Hartung et al., 1997; Li et al., 2002). FGF13 has two major alternative splicing isoforms, i.e., FGF13A and FGF13B. FGF13A contains the nucleus localization signal, while FGF13B is distributed in the cytoplasm and contributes to dynamic signaling processes (Wu et al., 2012). Noxious stimuli (e.g., thermal, mechanical, and chemical) cause pain by activating cutaneous Ad and C nociceptors, the peripheral terminals of small-diameter DRG neurons. Our recent neuron-typing study analyzed the transcriptome and functions of DRG neurons, and at least six types of mechanoheat nociceptors (MHNs) were identified (Li et al., 2016). The transduction of nociceptive signals involves special membrane proteins sensitive to a particular stimulus. For instance, the transient receptor potential cation channel V1 (TRPV1) expressed in small DRG neurons serves as a thermosensor (>42
C). However, the TRPV1 gene knockout mice only showed a partial reduction of the nociceptive behaviors induced by intense heat stimuli (>52
C) and did not exhibit apparent reduction in response to noxious heat stimuli below 48
C (Caterina et al., 2000; Davis et al., 2000), suggesting the existence of undiscovered mechanisms for heat nociception. The transduction of nociceptive signals also requires voltagegated sodium (Nav) channels, which are critical for the generation of action potentials (APs). Of the nine a subunits of Nav channels (Catterall et al., 2005), Nav1.6, Nav1.7, Nav1.8, and Nav1.9 are expressed in DRG neurons and contribute to somatosensory signal transmission (Dib-Hajj et al., 2010). Particularly, the tetrodotoxinsensitive (TTX-S) Nav1.7 is critical to nociception (Dib-Hajj et al., Neuron 93, 1–16, February 22, 2017 ª 2017 Elsevier Inc. 1

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Figure 1. FGF13 in DRG Neurons Is Required for Heat Nociception (A) Immunostaining shows that FGF13 is present in DRG neurons, afferent fibers in the laminae I-II of spinal cord (SC), and efferent fibers in the plantar glabrous skin of Fgf13F/Y (F/Y) mice. FGF13 staining is abolished in SNS-Cre/Fgf13F/Y (/Y) mice. Scale bars, 50 mm for DRG, 100 mm for SC and skin. (B) Immunoblotting shows that the FGF13 level decreases in both the DRG and SC of /Y mice. (C–F) /Y mice do not respond to noxious heat. The tail-flick (C), Hargreaves (D), and hot plate (E) tests show that F/Y mice display nociceptive responses to noxious heat stimulation and that the response latency decreases as the stimuli temperature increases. However, /Y mice do not respond to noxious heat (legend continued on next page)

2 Neuron 93, 1–16, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

2013; Lee et al., 2014). In humans, gain-of-function mutations of Nav1.7 lead to severe thermal hyperalgesia (Dib-Hajj et al., 2005; Fertleman et al., 2006; Yang et al., 2004), while loss-of-function mutations of Nav1.7 result in complete insensitivity to noxious stimuli (Cox et al., 2006; Emery et al., 2015). A recent study on Nav1.7 gene knockout mice confirmed the requirement of Nav1.7 in nociception (Gingras et al., 2014). Furthermore, knockdown of Nav1.7 attenuates complete Freund’s adjuvant (CFA)induced thermal hyperalgesia (Yeomans et al., 2005). Deletion of Nav1.7 in mouse DRG neurons impairs acute nociception and inflammation- and burn injury-induced thermal hyperalgesia (Minett et al., 2012; Nassar et al., 2004; Shields et al., 2012). Intracellular FGFs may interact with Nav channels and modulate the gating properties and current densities of sodium channels. FGF12 interacts with Nav1.5 and Nav1.9 (Liu et al., 2001, 2003). In hippocampal and cerebellar neurons, FGF14 is localized at the initial segment of axons and interacts with Nav1.1, Nav1.2, and Nav1.6 (Laezza et al., 2007, 2009; Lou et al., 2005; Xiao et al., 2013). The genetic ablation of Fgf12 and Fgf14 in mice results in ataxia and paroxysmal dystonia (Goldfarb et al., 2007; Wang et al., 2002) and impairs neurotransmission in the hippocampus (Xiao et al., 2007). FGF13 co-localizes with Nav1.6 at the Ranvier nodes along the myelinated nerve fibers, and FGF13 overexpression increases the sodium current density in ND7/23 cells (Wittmack et al., 2004). We have shown that FGF13 is expressed in small DRG neurons (Li et al., 2002). However, its function in somatosensation remains unknown. Here, we reveal a role of FGF13 in noxious heat sensation in mice. The FGF13 deletion in DRG neurons selectively abolished heat nociception and reduced the heat-evoked neuronal firing. FGF13 interacted with Nav1.7, and this heat-facilitated interaction maintained the cell membrane localization of Nav1.7, enabling the heat-evoked AP firing. Disrupting the FGF13/Nav1.7 interaction reduced both heat-induced AP firing and nociceptive behavior. Thus, our study provides a molecular mechanism essential for heat nociception beyond thermosensors. RESULTS FGF13 Is Expressed in Mechanoheat Nociceptive Neurons We first examined the FGF13 expression profile in the nervous systems of adult mice by immunoblotting and found that

FGF13 was abundant in the DRG (Figure S1A). Immunostaining showed that FGF13 was present in 83.4% ± 4.3% (n = 3 mice) of small DRG neurons (Figure 1A; Figure S1B). FGF13 was present in both the dorsal root and the sciatic nerve (Figure S1C), suggesting that FGF13 synthesized in somata of DRG neurons is transported to their central and peripheral terminals. Double immunostaining showed that 60.6% ± 6.6% of calcitonin generelated peptide (CGRP)-positive neurons (n = 3 mice, 100–200 neurons per mouse) and 86.1% ± 3.2% of isolectin B4 (IB4)-positive neurons (n = 3 mice, 100–200 neurons per mouse) expressed FGF13 (Figure S1D). FGF13 was also detected in the DRG afferent fibers in the laminae I-II of spinal cord (SC), where C and Ad fibers terminate, as well as efferent fibers in the glabrous skin (Figure 1A; Figure S1E). We used single-cell RNA sequencing and electrophysiological recording to categorize DRG neuron types (Li et al., 2016) and found that FGF13 was highly expressed in five types of MHNs (C1, C2, C4, C5, and C6) (Figure S1F) (Li et al., 2016). Nav1.7, Nav1.8, and Nav1.9 were also expressed in these neuron types, while TRPV1 was mainly expressed in C1 and C2 (Figure S1F). The co-expression of Nav1.7 and Nav1.8 in DRG neurons was confirmed with dual fluorescent in situ hybridization (81.8% ± 0.5% of Nav1.7 mRNA-expressing neurons contained Nav1.8 mRNA and 95.0% ± 0.3% of Nav1.8 mRNA-expressing neurons contained Nav1.7 mRNA, n = 3 mice, 100–200 neurons per mouse) (Figure S1G). Loss of FGF13 Selectively Abolishes Noxious Heat Sensation To determine the FGF13 function, we deleted FGF13 gene (Fgf13) in small DRG neurons by crossing homozygous Fgf13flox mice with a bacterial artificial chromosome (BAC) transgenic mouse line expressing Cre recombinase controlled by promoter elements of the Nav1.8 gene (SNS-Cre) (Agarwal et al., 2004) (Figure S2A). Then, behavioral tests were performed with the control Fgf13-flox mice (Fgf13F/Y) and the Fgf13 knockout mice (SNS-Cre/Fgf13F/Y; Fgf13/Y). In Fgf13/Y mice, FGF13 expression was reduced in DRG neurons, afferent fibers in the dorsal spinal cord and efferent fibers in the skin (Figures 1A and 1B). There were no apparent changes in the total number and morphology of DRG neurons (Figures S2B and S2C), the structure of their fibers and synapses in the dorsal spinal cord, or their nerve innervation in the skin (Figures S2D–S2G). The synaptic transmission from the DRG to dorsal spinal cord neurons was

stimulation (n = 14 in the tail-flick test; n = 10 in the Hargreaves test; n = 13 for F/Y and n = 14 for /Y in the hot plate test). In the CFA-induced inflammatory pain model, F/Y mice show hyperalgesic responses to radiant noxious heat applied to the plantar skin. /Y mice do not show hyperalgesia (n = 12 for F/Y and n = 14 for /Y) (F). All the tests were stopped at the cutoff (C.O.) time if the animal did not respond. (G) The two-temperature choice test shows that /Y mice exhibit normal sensitivity to innocuous temperatures (n = 12–20 for F/Y and n = 8–15 for /Y) (left). /Y mice are less sensitive to noxious heat conditions than F/Y mice, but /Y mice exhibit normal sensitivity to noxious cold (n = 9 for F/Y and n = 7 for /Y) (right). (H) The fMRI activation maps describe brain activity in response to noxious heat (47.5
C) stimulation applied to the left hindpaw of mice. The top and middle rows show fMRI activation maps of F/Y (n = 16) and /Y (n = 15) mice, respectively. The bottom row shows the statistical difference map overlaid with a mouse atlas for the identification of brain regions with altered noxious heat responsiveness. Scale bar, t value range. L, left; R, right; d, dorsal; v, ventral; Amyg, amygdala; CPu, caudate nucleus and putamen (striatum); HPC, hippocampus; Ins, insular cortex; MD, mediodorsal nuclei of thalamus; PAG, periaqueductal gray; PF, parafascicular nuclei of thalamus; RSG, retrosplenial granular cortex; S1 and S2, the primary and secondary somatosensory cortices. (I and J) Comparison of changes in the BOLD signal of fMRI between F/Y and /Y mice in ten brain regions (I and J) and between the left (ipsilateral) and right (contralateral to the site of heat stimulation) hemispheres in six brain regions (I). (K) Amplitude differences in percentage of fMRI signal changes between F/Y and /Y mice in ten brain regions. The data are shown as mean ± SEM; **p < 0.01 and ***p < 0.001 versus the indicated group. See also Figures S1–S4.

Neuron 93, 1–16, February 22, 2017 3

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

also normal (Figure S2H). Thus, the tissue structure and synaptic transmission in Fgf13/Y mice are largely unaltered. Interestingly, the noxious heat-induced behavioral responses were abolished in Fgf13/Y mice, while the innocuous temperature detection and the responses to noxious mechanical stimuli were not affected. It is generally accepted that heat nociceptors are activated by noxious heat above 43
C. In the tail-flick test at 48
C, 50
C, or 52
C, Fgf13/Y mice did not respond to the heat stimulation even when the cutoff time of 20 s, 15 s, or 10 s was reached, respectively (Figure 1C). Similarly, Fgf13/Y mice did not exhibit avoidance behavior during the Hargreaves test (radiant heat test) (Figure 1D). In the hot plate test, at 50
C, 52
C, or 55
C (with cutoff time of 70 s, 50 s, or 30 s, respectively), the mutant mice did not respond to the heat stimulation (Figure 1E). Furthermore, in the CFA inflammatory pain model, Fgf13/Y mice were insensitive to the noxious heat stimulation, while Fgf13F/Y mice showed obvious hyperalgesic responses (Figure 1F). The mechanical hypersensitivity induced by CFA or the spared nerve injury (SNI), a model of neuropathic pain, was not altered by FGF13 loss (Figures S3A and S3B). These results indicate that noxious heat sensation is absent in Fgf13/Y mice. In contrast, Fgf13/Y mice had normal responses to evaporative cooling (Figure S3C). In the two-temperature choice test, Fgf13/Y mice tended to choose the innocuous warm temperature (30
C) and avoid noxious cold (5
C and 15
C) or noxious heat (48
C and 55
C), suggesting that they were able to distinguish different temperatures; yet, their tolerance to noxious heat was increased (Figure 1G). Fgf13/Y mice exhibited normal mechanical nociceptive responses in both the von Frey and Randall-Selitto (tail clip) tests (Figure S3D). We used formalin test to examine the chemical nociception and found that Fgf13/Y mice, like wild-type mice, exhibited two-phase nociceptive responses, but the phase II was delayed without changes in its peak level (Figure S3E), suggesting that chemical nociception was mildly changed. Moreover, FGF13-deficient mice showed normal locomotor functions in rotarod and open field tests (Figures S3F and S3G). Furthermore, microarray analysis showed no prominent change (>2-fold change, p < 0.05, n = 3) in mRNA levels of other FGFs, sodium channels, thermosensors—such as TRP channels (Caterina et al., 2000; Davis et al., 2000)—and anoctamin 1 (Ano1) (Cho et al., 2012) in the DRG of Fgf13/Y mice (Figure S3H). This is consistent with the result of semiquantitative reverse transcription PCR (RT-PCR) (Figure S3I). However, the protein levels of Nav1.8 and TRPV1 were elevated in the DRG of Fgf13/Y mice (Figure S4A), suggesting a compensatory reaction in the mutant mice. Whole-cell patch-clamp recording showed that both the percentage of small DRG neurons sensitive to TRPV1 agonist capsaicin and the amplitude of capsaicin-induced current were increased (Figure S4B). Therefore, Fgf13/Y mice exhibit a selective loss of heat nociception despite the compensatory increase of TRPV1 function. Noxious Heat-Induced Brain fMRI Activities Are Eliminated in Fgf13–/Y Mice In Fgf13F/Y mice, noxious heat stimuli (47.5
C) applied to the surface of left hindpaw elicited robust fMRI activation across the brain, but these stimuli elicited limited activation and more nega-

4 Neuron 93, 1–16, February 22, 2017

tive signals in Fgf13/Y mice (Figure 1H). To determine the mostaffected brain regions, we compared the activation maps between the Fgf13F/Y and the Fgf13/Y groups and described the response differences in a statistical difference map (Figure 1H). The FGF13 deficiency eliminated the majority of noxious heat-evoked responses. The primary (S1) and secondary (S2) somatosensory, retrosplenial granular (RSG), and insular (Ins) cortices, as well as the amygdala (Amyg), the periaqueductal gray (PAG), the striatum (caudate nucleus and putamen, CPu), the CA1 region of hippocampus (HPC), and the mediodorsal (MD) and parafascicular (PF) nuclei of thalamus showed different responses between two groups. Many of these brain regions are known to be involved in nociception (Treede et al., 1999). To analyze how the FGF13 loss differentially affected brain regions, we further compared the magnitudes (i.e., signal change percentages) of noxious heat responses in ten brain regions between Fgf13F/Y and Fgf13/Y mice (Figures 1I–1K). The fMRI signal magnitudes in Fgf13F/Y mice were robust (0.5%–1.5%) and over 2-fold greater than those in Fgf13/Y mice (0.2%– 0.5%). Some areas showed apparent lateralized responses. For example, the contralateral S1 (F/Y–R) of Fgf13F/Y mice responded strongly to heat stimuli, whereas the ipsilateral S1 (F/Y–L) showed very weak signal change (Figure 1I). Response magnitudes of the contra- and ipsilateral S1 in Fgf13/Y mice were equally weak, comparable to that of the ipsilateral S1 in Fgf13F/Y mice. Other brain regions of Fgf13F/Y mice, including the Amyg, S2, CPu, HPC, and Ins, were equally activated in both hemispheres. The noxious heat-induced activation signals in these regions were eliminated in Fgf13/Y mice. The responses also reduced in the RSG, two thalamic nuclei (MD and PF), and the PAG (Figure 1J). FGF13 deficiency affected the Amyg to the greatest degree, followed by the CPu, Ins, S2, and RSG. The S1 cortex exhibited the smallest difference in the magnitude (Figure 1K). Thus, FGF13 deficiency in DRG neurons reduces noxious heat-evoked responses in the brain. Noxious Heat-Evoked AP Firing Is Reduced in FGF13Deficient DRG Neurons To examine whether the heat sensation loss of Fgf13/Y mice was caused by the change of nociceptor sensitivity to heat stimuli, we applied heat ramps from room temperature (22
C–25
C) to noxious heat levels 45
C–49
C to small DRG neurons (10– 20 mm in diameter) of Fgf13F/Y and Fgf13/Y mice (Figure 2A). The percentage of noxious heat-sensitive neurons was similar in two groups (38.1% for Fgf13F/Y, n = 42; 35.9% for Fgf13/Y, n = 39) (Figure 2B). We analyzed the distribution pattern of APs through the percentage of AP numbers at each degree. In small DRG neurons of Fgf13F/Y mice, the AP firing rate started to accelerate at 42
C, reached maximum at 43
C, and remained at that level to 45
C (Figures 2A, 2C, and 2D). In contrast, the AP firing in small DRG neurons of Fgf13/Y mice gradually decreased in response to temperature elevation and tended to cease at 45
C (Figures 2A, 2C, and 2D). Moreover, the AP amplitude was relatively uniform following the all-or-none rule in Fgf13F/Y neurons, whereas the AP amplitude exhibited a rapid decline in Fgf13/Y neurons as temperature increased (Figures 2A, 2E, and 2F). The fitting curves of the AP amplitudes from 38
C to 45
C showed that the AP amplitude reduced more sharply in

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Figure 2. Heat-Evoked APs Are Reduced in FGF13-Deficient Small DRG Neurons, and Functional Deficits Are Rescued by GST-FGF13B-TAT (A) AP firing was recorded in small DRG neurons perfused with extracellular solution ranging from room temperature to 45
C–49
C. DRG neurons of F/Y mice fire APs with accelerating frequency and relatively uniform amplitude, but the DRG neurons of /Y mice show rapidly reduced AP firing. The AP firing deficits are rescued by the cell-permeable protein GST-FGF13B-TAT. (B) The number of DRG neurons responding to heat stimulation is not changed in /Y mice, while the number of heat-sensitive neurons is decreased in Trpv1/ mice. (legend continued on next page)

Neuron 93, 1–16, February 22, 2017 5

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Fgf13/Y neurons (Figure 2E). When normalized to the amplitude of the first AP, AP amplitude at 45
C was 66.2% ± 0.06% in Fgf13F/Y neurons (n = 16) and 9.3% ± 0.05% in Fgf13/Y neurons (n = 14) (Figure 2F). The heat-induced attenuation of AP firing in Fgf13/Y neurons was not caused by impaired neuronal responsiveness because APs were efficiently triggered by current injection after the heat stimulation (Figure S5A). Therefore, FGF13-deficient small DRG neurons exhibit attenuated AP firing during noxious heat stimulation, although they are capable of detecting thermal changes below 42
C. To further evaluate the distinct roles of FGF13 and TRPV1 in noxious heat sensation, we examined heat-induced AP firing in TRPV1-deficient DRG neurons for comparison. The percentage of heat-sensitive small DRG neurons (35.2%, n = 71 wild-type neurons) was decreased in Trpv1 knockout (Trpv1/) mice (16.1%, n = 78) (Figure 2B). However, the remaining heat-sensitive neurons responded to heat stimuli with an accelerating frequency and an evenly distributed amplitude in a manner that was similar to that of wild-type neurons (Figures 2C–2F). The AP amplitude at 45
C was reduced to 62.5% ± 0.12% in TRPV1-deficient neurons (n = 11), similar to that in wild-type neurons (p = 0.77). Thus, although more than half of the DRG neurons lost the responsiveness to heat stimuli in Trpv1/ mice, a considerable number of neurons remained intact in heat sensitivity. These results differ from that in Fgf13/Y mice, in which the number of noxious heat-sensitive DRG neurons was not reduced, but the AP amplitude decreased while temperature increased into noxious heat range (Figures 2A–2F), suggesting that FGF13 and TRPV1 participate in heat sensation through different mechanisms. Administration of Exogenous FGF13 Rescues Heat Nociception in Fgf13–/Y Mice To confirm the role of FGF13 in heat nociception, we further investigated whether the loss of heat sensation in Fgf13/Y mice could be reversed by acute application of exogenous FGF13 protein. GST-tagged FGF13B was fused with a TAT peptide, which is a membrane-penetrating vector developed from the human immunodeficiency virus (Schwarze et al., 1999) (GST-FGF13B-TAT). GST-FGF13B-TAT incubation (1 mM for 120 min) reversed the reduced AP firing during noxious heat stimulation in Fgf13/Y DRG neurons; the AP amplitude at 45
C returned to 64.5% ± 0.06% of the initial AP amplitude (n = 13) (Figures 2A and 2C–2F). In contrast, the control protein GST-TAT did not have the same rescue effect (Figures 2C–2F).

The proportions of neurons responding to noxious heat were similar among all experimental groups (Figure S5B). These results indicate that FGF13 is essential for the persistent AP firing of small DRG neurons during noxious heat stimulation. To study the in vivo effects of the exogenous FGF13, we first applied GST-FGF13B-TAT intrathecally (i.t.) to C57BL/6 mice and tested its metabolic kinetics. The peak level of GSTFGF13B-TAT in both the DRG and dorsal spinal cord was detected 20–60 min after injection (Figure 2G). Immunoblotting showed that the amount of the exogenous FGF13 with GST (45 kDa) was much less than that of the endogenous FGF13 (20 kDa), so GST-FGF13B-TAT effects would not be caused by an overdose (Figure S5C). Then, we applied GST-FGF13BTAT (i.t.) to Fgf13/Y mice. GST-FGF13B-TAT, but not GSTTAT, rescued the heat sensitivity of Fgf13/Y mice within an effective period after injection (Figure 2H). Immunoblotting showed that GST-FGF13B-TAT (i.t.) could be detected in both the DRG and the plantar skin of hindpaw (Figure S5D). Thus, both the reduced AP firing in small DRG neurons and the lost heat sensation in Fgf13/Y mice were reversed by exogenous FGF13 supplementation, indicating that FGF13 is necessary for heat nociception. On the other hand, the deficits of Fgf13/Y mice are not caused by developmental defects or non-specificity introduced by knockout procedures, because Fgf13/Y mice did not have morphological abnormalities and the defects were reversed by an acute supplementation of FGF13. FGF13 Directly Interacts with Nav1.7 Dual fluorescent in situ hybridization and single-cell RNA sequencing showed the co-expression of FGF13 and Nav1.7 (Figure 3A; Figure S1F). Namely, 71.8% ± 4.1% of FGF13-positive neurons (n = 3 mice, 100–200 neurons per mouse) contained Nav1.7, while 80.1% ± 1.6% of Nav1.7-positive neurons (n = 3 mice, 100–200 neurons per mouse) expressed FGF13. The majority of FGF13-containing neurons (78.1% ± 4.4%, n = 3 mice, 100–200 neurons per mouse) also expressed Nav1.8 (Figure S6A). To identify the underlying mechanism of the change in AP firing in FGF13-deficient DRG neurons, we examined the interactions between FGF13 and Nav channels in the DRG (i.e., Nav1.6, Nav1.7, Nav1.8, and Nav1.9) using co-immunoprecipitation (IP) and GST pull-down methods. In co-IP experiments with mouse DRG protein extract, FGF13 showed a potent interaction with Nav1.7 and a relatively weak interaction with Nav1.6, while FGF13 barely interacted with Nav1.8 or Nav1.9 (Figure 3B; Figures S6B–S6D). GST pull-down experiments confirmed that

(C–F) The AP firing frequency (C and D) and amplitude (E and F) is quantified. The AP firing distribution is shown by the percentage of AP spike numbers distributed at each degree from 38
C to 45
C (C). The percentage of APs firing at 45
C reveals the remaining APs at high temperature (D). The fitting curve for AP amplitudes from 38
C to 45
C in each group shows the change in amplitude with temperature (E). The ratio of the AP amplitude at 45
C to the first AP reveals the change induced by heat stimulation (F). As the temperature becomes higher than 42
C, AP frequency is accelerated and AP amplitude decreases to 66% in F/Y neurons, while AP tends to cease firing and AP amplitude drops rapidly to 9% in /Y neurons. GST-FGF13B-TAT, but not GST-TAT, rescues the AP firing of /Y neurons. Trpv1 knockout does not alter the AP firing of small DRG neurons during heat stimulation. In (C) and (E), n = 17 for F/Y, n = 13 for /Y, n = 13 for /Y + GST-FGF13B-TAT, n = 10 for /Y + GST-TAT and n = 11 for Trpv1/; in (D) and (F), n = 16 for F/Y, n = 14 for /Y, n = 13 for /Y + GST-FGF13B-TAT, n = 10 for /Y + GST-TAT, n = 11 for Trpv1/. (G) GST-FGF13B-TAT (0.5 mg/kg, i.t.) was injected into mice. Immunoblotting shows that GST-FGF13B-TAT infiltrates into the DRG and SC and interacts with Nav1.7. The interaction remains more than 1 hr after injection (n = 3). The arrows point to the band of the exogenous FGF13 with a GST tag. (H) The administration of GST-FGF13B-TAT (0.5 mg/kg or 1 mg/kg, i.t.) rescues the heat nociception of /Y mice in the tail-flick test at 52
C. The best rescue effect appears at 1 hr after injection (n = 5 for a dose of 0.5 mg/kg and n = 4 for a dose of 1 mg/kg). The test was stopped at the cutoff (C.O.) time. The data are shown as mean ± SEM; *p < 0.05, **p < 0.01, and ***p < 0.001 versus the indicated group. See also Figure S5.

6 Neuron 93, 1–16, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

(legend on next page)

Neuron 93, 1–16, February 22, 2017 7

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

FGF13 exhibited strong interaction with Nav1.7 but much weaker interaction with Nav1.8 (Figure 3C; Figure S6E). Interaction between the exogenous FGF13 (i.t.) and the endogenous Nav1.7 in the DRG and spinal dorsal horn was also detected (Figure 2G). Notably, the most potent interaction appeared 20–60 min after GST-FGF13B-TAT injection, consistent with the period of its best rescue effects on Fgf13/Y mice (Figures 2G and 2H). Moreover, we did not detect any apparent interaction between FGF13 and TRPV1 or TRPV2, two TRP channels that are activated by temperatures above 42
C or 52
C, respectively (Figures S6F and S6G). Thus, FGF13 interacts with Nav1.7 in DRG neurons, while its interactions with TRPV1 or TRPV2 are not detectable. Furthermore, we identified an FGF13-binding region in the C terminus of Nav1.7 (Nav1.7CT). Intracellular FGFs have been predicted to interact with the C-terminal region of Nav channels (Goetz et al., 2009). We constructed a plasmid encoding a fusion protein myc-CD8a-Nav1.7CT, in which Nav1.7CT was anchored to the cell membrane via myc-CD8a; myc-CD8a-Nav1.7CT, but not myc-CD8a, was co-immunoprecipitated with FGF13 (Figure 3D). To determine the precise binding region for FGF13, we made a series of truncated Nav1.7CT proteins fused with GST (Figure 3E). A GST pull-down assay showed a direct interaction between the purified FGF13 and Nav1.7CT, and the 1,876–1,896 aa in Nav1.7CT was the major binding site for FGF13 (Figure 3F). On the other hand, it seemed that FGF13 did not contain a specific region mediating its interaction with Nav1.7CT because only full-length FGF13 bound to Nav1.7CT (Figure S6H). We assume that the tertiary structure of the full-length FGF13 is critical for its interaction with Nav1.7CT. Thus, FGF13 directly interacts with the C terminus of Nav1.7. FGF13 Increases Nav1.7 Current Density Because FGF13 interacted with Nav1.7, FGF13 might regulate the function of Nav1.7. We recorded the sodium currents evoked by a step depolarization (90 mV to 50 mV in 5 mV increments) in HEK293 cells co-transfected with Nav1.7 and FGF13B-IRES-GFP. We analyzed the current density and the voltage dependence of Nav1.7 activation. In comparison with the control IRES-GFP vector, FGF13B-IRES-GFP expression increased Nav1.7 current densities without an apparent shift in the voltage dependence of Nav1.7 activation (Nav1.7 and vector: V50 = 17.5 ± 2.1 mV, n = 16; Nav1.7 and FGF13B: V50 = 22.2 ± 2.2 mV, n = 17; p = 0.14) (Figures 4A–4C). We also analyzed the fast inactivation of Nav1.7 by measuring peak sodium currents induced at 10 mV after a series of 500 ms prepulses stepping from 130 mV to 10 mV in 5 mV increments. The voltage

dependence of Nav1.7 inactivation was not changed (Nav1.7 and vector: V50 = 75.2 ± 1.2 mV, n = 15; Nav1.7 and FGF13B: V50 = 72.8 ± 1.1 mV, n = 17; p = 0.15) (Figure 4C). Thus, FGF13 increases the Nav1.7 current density without altering its activation/inactivation properties. Then, we examined the sodium currents in FGF13-deficient small DRG neurons by a step depolarization (90 mV to 50 mV in 5 mV increments). FGF13 deficiency reduced the TTX-S sodium currents without affecting the voltage dependence of channel activation (Fgf13F/Y: V50 = 29.57 ± 1.39 mV, n = 23; Fgf13/Y: V50 = 30.30 ± 1.34 mV, n = 20; p = 0.71) or inactivation (Fgf13F/Y: V50 = 75.77 ± 1.67 mV, n = 16; Fgf13/Y: V50 = 76.68 ± 1.04 mV, n = 17; p = 0.64) (Figures 4D–4F). Because temperature affects sodium currents (Arrigoni et al., 2016; Han et al., 2007), we then recorded the sodium currents of small DRG neurons at innocuous (30
C) and noxious (45
C) temperatures (Figure 4G). The sodium currents were lower in Fgf13/Y neurons than in Fgf13F/Y neurons at both temperatures. Notably, the sodium currents of Fgf13/Y neurons were even lower at 45
C than at 30
C, indicating a correlation between the decreased sodium current and the attenuated AP firing in FGF13-deficient neurons during noxious heat stimulation. In contrast, the sodium currents of Fgf13F/Y neurons were not changed by increased temperature (Figure 4G). Taken together, these findings show that FGF13 increases the current density of Nav1.7 and therefore maintains the excitability of small DRG neurons, especially in noxious heat conditions. FGF13/Nav1.7 Interaction Is Facilitated by Heat Stimulation and Maintains Nav1.7 Level in Cell Membrane Given that FGF13 was critical for the maintenance of sodium currents and neuronal excitability in noxious heat condition, we further investigated the underlying mechanism. We assumed that the availability of sodium channels might decrease during heat stimulation. The cultured DRG neurons of Fgf13F/Y and Fgf13/Y mice were treated with a water bath at 43
C for 30 s or 1 min and then immediately cooled down on ice, such that Nav1.7 trafficking was restrained, and the instantaneous localization of Nav1.7 was maintained. Then, the membrane proteins were examined by biotinylation and immunoblotting (Figure 5A). We did not detect heat-induced changes in the Nav1.7 level in the plasma membrane of Fgf13F/Y neurons. However, the membrane Nav1.7 level decreased in Fgf13/Y neurons after heat treatment (Figure 5B). The membrane Nav1.7 level was not affected by the Fgf13 knockout without heat stimuli (Figure 5B).

Figure 3. FGF13 Interacts with Nav1.7 (A) Dual fluorescent in situ hybridization shows the coexistence of FGF13 and Nav1.7 mRNA in DRG neurons (arrows). Scale bar, 50 mm. (B) Co-IP shows that FGF13 interacts with Nav1.7 in the DRG and SC. The FGF13/Nav1.7 interaction level is higher than the interaction levels between FGF13 and other sodium or TRP channels (n = 3–6). (C) A GST pull-down assay shows that GST-FGF13B protein interacts with Nav1.7. The interaction between GST-FGF13B and Nav1.8 is much weaker (n = 3). (D) The C terminus of Nav1.7 (Nav1.7CT) was fused to myc-CD8a, which anchors Nav1.7CT to the cell membrane. Co-IP shows that FGF13 interacts with Nav1.7CT fused to myc-CD8a, but not with myc-CD8a. (E) A schematic illustration of GST-flag-Nav1.7CT truncations. Nav1.7CT was truncated into nine fragments to search for the FGF13-binding site. The position of the major binding site f8 is marked in red. (F) A GST pull-down assay with two purified proteins, GST-flag-Nav1.7CT and FGF13B, shows the direct interaction between FGF13 and Nav1.7. The major interaction site is f8 (1,876–1,896 aa) of Nav1.7CT (n = 4). The data are shown as mean ± SEM; *p < 0.05, **p < 0.01, and ***p < 0.001 versus the indicated group. See also Figure S6.

8 Neuron 93, 1–16, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Figure 4. FGF13 Increases Sodium Currents (A) FGF13B expression increases Nav1.7 current in HEK293 cells transfected with the plasmid expressing Nav1.7. (B) The I-V curve shows the Nav1.7 current density in HEK293 cells is enhanced by FGF13B (n = 16 for Nav1.7 + vector and n = 17 for Nav1.7 + FGF13B). (C) The voltage dependence of Nav1.7 activation and inactivation in HEK293 cells is not changed by FGF13B (activation: n = 16 for Nav1.7 + vector and n = 17 for Nav1.7 + FGF13B; inactivation: n = 15 for Nav1.7 + vector and n = 17 for Nav1.7 + FGF13B). (D) The TTX-S sodium current is reduced in small DRG neurons of /Y mice. (E) The I-V curve shows that the TTX-S sodium current density is reduced in small DRG neurons of /Y mice (n = 23 for F/Y and n = 20 for /Y). (F) The activation and inactivation voltage dependence of the TTX-S sodium channel is not altered in small DRG neurons of /Y mice (activation: n = 23 for F/Y and n = 20 for /Y; inactivation: n = 16 for F/Y and n = 17 for /Y). (G) The sodium current of small DRG neurons was recorded at 30
C and 45
C. The I-V curve shows that the sodium current is smaller in small DRG neurons of /Y mice than in those of F/Y mice in both temperature conditions; in addition, treatment at 45
C reduces the sodium current of /Y neurons, but not that of F/Y neurons (n = 17 for F/Y and n = 15 for /Y). The data are shown as mean ± SEM; **p < 0.01 and ***p < 0.001 versus the indicated group.

Thus, in the absence of FGF13, the amount of Nav1.7 in the plasma membrane of DRG neurons decreases after noxious heat stimulation, indicating a requirement of FGF13 for maintaining Nav1.7 density in the cell membrane. Next, we wondered whether the FGF13/Nav1.7 interaction was dynamic in noxious heat conditions. We collected HEK293 cells transfected with the plasmids expressing FGF13B and Nav1.7 and added the membrane-permeant reagent disuccinimidyl suberate (DSS) to chemically crosslink FGF13 with the Nav1.7 present in the membrane and the cytosol (Figure 5C). Using immunoblotting, we identified an extra FGF13-containing band with a large molecular weight (>250 kDa, approximately the sum of the molecular weights of FGF13 and Nav1.7) after

crosslinking (Figure 5D), suggesting the formation of the FGF13/Nav1.7 complex. Then, we used surface biotinylation and immunoblotting to examine how the FGF13/Nav1.7 complexes associated with the cell membrane were affected by heat treatment. To estimate the formation of FGF13/Nav1.7 complex during heat stimulation, we performed the DSS crosslinking before and after the heat treatment (Figure 5C). We found that the FGF13/Nav1.7 complex could be detected in both the plasma membrane and the whole-cell lysate and that the heat stimulation facilitated the complex formation (Figure 5D). The heat-induced formation of the FGF13/Nav1.7 complex was also detected in DRG neurons (Figure 5E). Furthermore, we enriched the membrane component of HEK293 cells expressing

Neuron 93, 1–16, February 22, 2017 9

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

(legend on next page)

10 Neuron 93, 1–16, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

FGF13 and Nav1.7 (as functional FGF13/Nav1.7 complex was supposed to be in the cell membrane) after the FGF13/Nav1.7 interaction was trapped during heat stimulation with crosslinking reagent. After removing free-binding FGF13 and Nav1.7 using high detergent condition and repeated washing procedures in co-IP assay, we detected the enhanced FGF13/Nav1.7 interaction that was trapped during heat stimulation (Figure 5F; Figure S7A). The heat-facilitated FGF13/Nav1.7 interaction may stabilize Nav1.7 channels in the cell membrane during noxious heat stimulation. We have shown that FGF13 acts as a microtubule-stabilizing protein that is required for the migration of cortical neurons during brain development (Wu et al., 2012). We found that the heat-facilitated formation of the FGF13/Nav1.7 complex was not attenuated by the disruption of actin filaments or microtubules with cytochalasin D or nocodazole, respectively (Figure S7B). This suggests that the cytoskeleton-mediated transport system is not involved in the heat-facilitated FGF13/ Nav1.7 interaction. FGF13 and Nav1.7 might localize closely and rapidly form the complex required for the stabilization of Nav1.7 in the cell membrane. Disrupting FGF13/Nav1.7 Interaction Attenuates Noxious Heat Sensation To confirm the contribution of the FGF13/Nav1.7 complex to noxious heat-induced neuronal excitability, we prepared membrane-permeable GST-flag-Nav1.7CT-TAT (Nav1.7CT-TAT) as a competitive peptide to block the endogenous interaction between FGF13 and Nav1.7 (Figures 6A and 6B). In cultured DRG neurons, the co-IP signal of the FGF13/Nav1.7 interaction was decreased after a 30 min treatment with 1 mM Nav1.7CT-TAT (Figure 6B). Thus, Nav1.7CT-TAT is able to competitively block the FGF13/Nav1.7 interaction in DRG neurons. Then, we analyzed the effect of Nav1.7CT-TAT on AP firing following the heat stimulation in small DRG neurons of C57BL/6 mice. In neurons incubated with 1 mM Nav1.7CT-TAT (n = 16), the APs showed two patterns. In one pattern, the AP amplitude at 45
C decreased to 20.6% ± 0.08% of the initial AP, and the decay level was much greater than that of the control group treated with GST-flag-TAT (81.1% ± 0.06%, n = 12) (Figures 6C and 6D). In another firing pattern, few AP spikes were evoked by the heat stimulation (Figures 6C and 6D). Nav1.7CTTAT did not alter the firing of small DRG neurons from Fgf13/Y mice (Figure 6D). Therefore, the AP firing of small DRG neurons

treated with Nav1.7CT-TAT was attenuated during heat stimulation, which was similar to the change observed in FGF13-deficient small DRG neurons (Figures 2A–2F). Furthermore, the heat-induced AP firing was reduced in small DRG neurons treated with 10 nM Protoxin-II, a Nav1.7 inhibitor (Schmalhofer et al., 2008), and the decay of heat-induced AP amplitude was similar to that in neurons lacking FGF13 or treated with Nav1.7CT-TAT (Figure 6D). Thus, the FGF13/Nav1.7 interaction is essential for the maintenance of neuronal excitability in response to noxious heat stimuli. We further examined the effect of Nav1.7CT-TAT on behavioral responses to noxious heat stimuli. Nav1.7CT-TAT was injected into mice (intraperitoneally [i.p.], four injections, once per hour). Co-IP experiments showed that the applied Nav1.7CT-TAT reduced the FGF13/Nav1.7 interaction in the afferent fibers in the dorsal spinal cord, but not in the DRG (Figure 7A). It is possible that the abundant amount of FGF13 and Nav1.7 in the DRG veiled the interfering effects of Nav1.7CT-TAT. The Nav1.7CT-TAT treatment dose-dependently extended the response latency in the 52
C tail-flick tests (Figure 7B). Such an effect reached a peak level at 4–6 hr after the last injection. Thus, heat nociception is impaired by the disruption of FGF13/ Nav1.7 interaction. DISCUSSION FGF13/Nav1.7 Complex Is Essential for Heat Nociception This study showed that Fgf13/Y mice selectively lost responsiveness to noxious heat but remained responsive to other stimulus modalities such as innocuous temperatures and noxious mechanical stimuli. Moreover, FGF13-deficient small DRG neurons exhibited AP firing that rapidly stopped when the temperature was above 43
C. Therefore, FGF13 is selectively required for the transmission of noxious heat stimuli. TRPV1 and TRPV2 expressed in small DRG neurons have been considered as high-temperature sensors (>42
C and >52
C, respectively), while TRPM8 and TRPA1 as low-temperature sensors (25
C–28
C and 12
C–24
C, respectively) (Dhaka et al., 2006). Studies with TRPV1-deficient mice show that TRPV1 contributes mainly to the development of inflammationinduced thermal hyperalgesia and only partially to heat-induced nociceptive responses at high temperatures (>52
C) (Caterina et al., 2000; Davis et al., 2000). We found that heat nociception

Figure 5. FGF13 Maintains Nav1.7 Levels in Cell Membrane during Heat Stimulation (A) The diagram shows the experimental procedure. The DRG neurons were dissociated and stimulated in 43
C water bath for 30 s or 1 min. Then, the membrane proteins were extracted by surface biotinylation and processed for immunoblotting. (B) The amount of Nav1.7 in the plasma membrane at room temperature (RT) is not apparently changed by Fgf13 knockout. However, the Nav1.7 level in the membrane is decreased after stimulation at 43
C in /Y neurons, but not in F/Y neurons (n = 4). (C) The diagram shows the experimental procedure. The cells transfected with the plasmids expressing FGF13B and Nav1.7 were stimulated at 43
C. The DSS crosslinking agent was added before or after the heat stimulation. The membrane proteins were extracted by surface biotinylation and processed for immunoblotting. (D) Immunoblotting shows the FGF13/Nav1.7 complex formation after DSS crosslinking in HEK293 cells expressing FGF13B and Nav1.7. The amount of the FGF13/Nav1.7 complex is increased after stimulation at 43
C. The increased complex formation is detected in both the plasma membrane (PM) and the cell lysate (n = 4). (E) The FGF13/Nav1.7 complex and its increased formation after heat stimulation are also observed in DRG neurons (n = 4). (F) Co-IP shows the heat-enhanced FGF13/Nav1.7 interaction in the enriched membrane component of FGF13B- and Nav1.7-expressing HEK293 cells (n = 4). The data are shown as mean ± SEM; *p < 0.05, **p < 0.01, and ***p < 0.001 versus the indicated group. See also Figure S7.

Neuron 93, 1–16, February 22, 2017 11

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Figure 6. Nav1.7CT-TAT Blocks the FGF13/Nav1.7 Interaction in DRG Neurons and Impairs Heat-Evoked AP Firing (A) The cell-permeable protein GST-flag-Nav1.7CT-TAT (Nav1.7CT-TAT) was purified and used to block the FGF13/Nav1.7 interaction. Coomassie brilliant blue (CBB) staining shows the protein from the first four elution units (upper). HEK293 cells were incubated with different concentrations of Nav1.7CT-TAT for 1 hr. Nav1.7CT-TAT was detected in the cell lysate by immunoblotting (lower). (B) Co-IP shows that a 30 min treatment with Nav1.7CT-TAT (1 mM) disrupts the interaction between endogenous FGF13 and Nav1.7 in cultured DRG neurons (n = 3). (C and D) Heat-induced AP firing is impaired in small DRG neurons by 1 mM Nav1.7CT-TAT treatment for 30 min. The impaired AP firing exhibits two patterns: APs with reduced amplitude and APs firing with very few spikes (C). The attenuated AP firing mimics the defects observed in small DRG neurons of /Y mice (C). The ratio of the AP amplitude at 45
C to that of the first AP is used to quantify the change of AP firing induced by heat stimulation (D). Nav1.7CT-TAT did not alter the firing of small DRG neurons from /Y mice (D). The heat-induced AP firing was reduced in small DRG neurons treated with 10 nM Nav1.7 inhibitor Protoxin-II (D). n = 11 for elution buffer, n = 16 for GST-flag-Nav1.7CT-TAT, n = 12 for GST-flag-TAT, n = 9 for GST-flag-Nav1.7CT-TAT treating /Y neurons, and n = 11 for Protoxin-II. The data are shown as mean ± SEM; *p < 0.05, **p < 0.01, and ***p < 0.001 versus the indicated group.

was abolished in Fgf13/Y mice, even with the compensatory increase of TRPV1 in DRG neurons. In Trpv1/ mice, the number of heat-sensitive DRG neurons was reduced and the remaining heat-sensitive DRG neurons exhibited intact heat-induced AP firing. It is possible that a large population of the heat-sensitive DRG neurons could not be sufficiently depolarized in the absence of TRPV1 because of the lack of cation influxes, and thus, APs are not effectively initiated. However, in Fgf13/Y mice, the number of heat-sensitive DRG neurons did not decrease, but AP firing was not sustained as the temperature reached noxious heat levels, indicating a distinct role of FGF13 in heat nociception.

12 Neuron 93, 1–16, February 22, 2017

We propose a model for noxious heat sensation (Figure 7C), which may require two processes. First, heat stimuli activate thermosensors, such as TRPV1, to induce a cation influx, depolarize the membrane potential, and initiate APs. Second, the FGF13/Nav1.7 complex acts as a booster that keeps persistent AP firing during noxious heat stimulation (>43
C). Several thermosensors have been reported, such as TRPV1, TPRM3, Ano1, and K2P channels (Caterina et al., 2000; Cho et al., 2012; €l et al., 2009; Vriens et al., 2011). They Davis et al., 2000; Noe are expressed in different types of MHNs and are likely to be redundant in function so that a considerable number of heatsensitive DRG neurons and partial nociceptive behavior are

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Figure 7. Nav1.7CT-TAT Blocks Endogenous FGF13/Nav1.7 Interaction and Impairs Heat Nociception (A) Co-IP shows that the Nav1.7CT-TAT (60 mg/kg, i.p., 4 injections, one injection per hour) blocks the FGF13/Nav1.7 interaction in afferent fibers to the SC, but not in the DRG (n = 4). (B) The tail-flick test at 52
C shows that the Nav1.7CT-TAT (i.p.) dose-dependently impairs noxious heat sensation (n = 3). (C) A proposed model of the mechanism for noxious heat sensation. During heat stimulation, thermosensor activation in MHNs leads to cation influxes and initiates AP firing. The heat-facilitated FGF13/Nav1.7 interaction stabilizes Nav1.7 in the plasma membrane (PM) to sustain AP firing. A lack of FGF13 reduces the number of Nav1.7 in the PM and the heat-induced AP firing and impairs heat nociception. The data are shown as mean ± SEM; *p < 0.05 and ***p < 0.001 versus the indicated group.

retained in the absence of one thermosensor. After the thermosensors initiate APs, the FGF13/Nav1.7 complex sustains the AP firing during noxious heat stimulation. The unsustained AP firing and the abolished behavioral responses in Fgf13/Y mice indicate the absence of complementary pathways for this booster’s function in the nociceptors. Notably, a low dose of exogenous FGF13 was sufficient to rescue the AP firing of small DRG neurons and noxious heat sensation in Fgf13/Y mice, suggesting that this mechanism is highly efficient. The detection of intrathecally injected GST-FGF13B-TAT in the DRG and skin suggests that the exogenous FGF13 could act at the peripheral nerve ending, although the transport mechanism is unclear.

Furthermore, FGF13 was widely distributed in different types of MHNs (Li et al., 2016); thus, FGF13 could be a commonly used regulator in nociceptors. Taken together, heat nociception requires the FGF13/Nav1.7 complex through a distinct mechanism from thermosensors. Heat-Facilitated FGF13/Nav1.7 Interaction Is a Selective Mechanism of Nav1.7 for Heat Nociception Nav1.7 is important for both noxious heat and mechanical sensations. The gain-of-function mutations of Nav1.7 lead to severe thermal hyperalgesia in humans (Dib-Hajj et al., 2013), whereas the loss-of-function mutations result in a complete loss of

Neuron 93, 1–16, February 22, 2017 13

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

sensitivity to noxious stimuli (Cox et al., 2006; Dib-Hajj et al., 2013). The loss of Nav1.7 in mice attenuates acute pain and inflammation- and burn injury-induced thermal hyperalgesia (Minett et al., 2012; Nassar et al., 2004; Shields et al., 2012; Yeomans et al., 2005). Nav1.7 also contributes to mechanical nociception in mice (Nassar et al., 2004, 2005). TTX-resistant (TTX-R) Nav1.8 and Nav1.9 are expressed in nociceptive neurons (Dib-Hajj et al., 2010; Tate et al., 1998). However, Nav1.8- or Nav1.9-deficient mice show little change in nociceptive thresholds and mild deficits in the inflammatory hypersensitivity (Akopian et al., 1999; Amaya et al., 2006; Leo et al., 2010; Priest et al., 2005). Therefore, Nav1.7 plays a major role among the sodium channels in nociception. It is notable that a lack of Nav1.7 causes a partial loss of heat nociception in mice, in contrast with the complete loss in humans (Cox et al., 2006; Nassar et al., 2004). This difference may be attributed to the compensatory adaptation via other sodium channels in mice. For example, Nav1.7 and Nav1.8 might coordinate because both the Nav1.7 expression and the TTX-S sodium current density increase in Nav1.8 null mice (Akopian et al., 1999), and the noxious heat threshold is doubled in the Nav1.7 and Nav1.8 gene double-knockout mice (Nassar et al., 2005). Interestingly, the compensatory increase of Nav1.8 protein was detected in the FGF13-deficient DRGs. However, we did not find compensatory effects of Nav1.8 in Fgf13/Y mice, consistent with a report showing no TTX-R sodium current compensation for the Nav1.7 deficiency (Nassar et al., 2004). The loss of noxious heat sensation in Fgf13/Y mice was more prominent than that in both the Nav1.7 null mice and the Nav1.7 and Nav1.8 double-knockout mice, indicating that the FGF13 function cannot be compensated. The loss of heat nociception, but not noxious mechanical sensation, in Fgf13/Y mice suggests that the heat-facilitated FGF13/Nav1.7 interaction could be a selective mechanism in MHNs for noxious heat sensation, although Nav1.7 itself does not have the specificity and selectivity of sensory modalities. It is possible that noxious heat condition leads to conformational changes of Nav1.7 and therefore instable membrane localization. The facilitated interaction with FGF13 may stabilize the conformation and membrane-localization of Nav1.7 during the heat stimulation. Taken together, the interaction with Nav1.7 acts as a major mechanism for the FGF13 function in heat nociception, although other mechanisms might be involved in this process. Brain Regions Specifically Activated by Noxious Heat Stimuli Are Identified As pain is a complex experience that involves both sensory and emotional changes (Zhuo, 2016), we further explored how the peripheral loss of noxious heat sensation affected the brain activities. The present study demonstrates that the loss of noxious heat-induced behavior in Fgf13/Y mice directly correlates with the loss of noxious heat-evoked fMRI signals in the brain regions. Some of the regions are known to be involved in pain perception (Becerra et al., 2011), including the regions mediating sensorydiscriminative and affective-motivational aspects of nociceptive information, such as the MD and PF of thalamus and the S1, S2, and insular cortices. Moreover, the reduced fMRI signals of the Amyg, HPC, CPu, and RSG in Fgf13/Y mice provide direct

14 Neuron 93, 1–16, February 22, 2017

evidence that the limbic system and regions involved in memory and emotional reaction are also related to noxious heat information processing. Notably, the reduction in noxious heat responses varied among different brain regions, suggesting that the different ascending nociceptive pathways are distinctively affected by the lack of FGF13 in DRG neurons. In addition, we detected the reduced response to heat stimuli in the PAG, which is a key region for descending modulation of pain. Thus, the present study has identified the brain regions that process and integrate the peripheral noxious heat inputs. Potential Roles of FGF13 in Pathological Sensory Disorders We show that the thermal hyperalgesia induced by peripheral tissue inflammation was also abolished in Fgf13/Y mice. The loss of heat nociception in inflammatory conditions may be related to a reduction in cell-membrane-localized Nav1.7, which contributes to inflammation-induced thermal hyperalgesia (Minett et al., 2012; Nassar et al., 2004; Yeomans et al., 2005). Because the expression of FGF13 and Nav1.7 in DRG neurons is decreased after peripheral nerve injury (Kim et al., 2002; Li et al., 2002; Xiao et al., 2002), FGF13 could be mainly involved in inflammatory pain, but not in neuropathic pain. Furthermore, FGF13 regulates brain development, and its gene mutations in humans could cause X-linked intelligence disability (ID) (Wu et al., 2012) and genetic epilepsy and febrile seizures plus (GEFS+) (Puranam et al., 2015). Patients with Fgf13 mutations may also suffer from sensory disorders, e.g., presumably a lack of sensitivity to noxious heat stimuli, if the mutations would cause a marked reduction of FGF13 in sensory neurons. Therefore, somatosensory tests in these patients would extend our understanding of these diseases and their pathological mechanisms and help us to protect these patients from potential harms. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d d

KEY RESOURCES TABLE CONTACT FOR REAGENT AND RESOURCE SHARING EXPERIMENTAL MODEL AND SUBJECT DETAILS B Animals METHOD DETAILS B Plasmid Construction B Cell Culture and Transfection B Morphology Analysis B Immunoblotting B Preparation of the GST-Fused Proteins B Immunoprecipitation and GST Pull-Down B Cell Surface Biotinylation B Genotyping B Behavioral Tests B Infusion of Nav1.7CT-TAT and FGF13B-TAT B Preparation of an FGF13 Antibody B Electrophysiological Recording B Dual Fluorescent In Situ Hybridization

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

B

fMRI Experiment Single-Cell RNA Sequencing B Semiquantitative PCR B Microarray QUANTIFICATION AND STATISTICAL ANALYSIS DATA AND SOFTWARE AVAILABILITY B

d d

SUPPLEMENTAL INFORMATION Supplemental Information includes seven figures and two tables and can be found with this article online at http://dx.doi.org/10.1016/j.neuron.2017. 01.009. AUTHOR CONTRIBUTIONS L.Y. did major experiments. F.D. did electrophysiological recording. Q.Y., R.W., and P.-F.Y. did fMRI analysis. Q.-F.W. contributed to molecular cloning. D.W. contributed to in situ hybridization. C.-L.L. and Y.-Q.Z. did single-cell RNA-seq. Y.-J.L. contributed to immunoblotting. X.C. contributed to patchclamp recording. L.C. and F.-Q.X. designed fMRI tests and interpreted the data. X.Z., L.B., and L.Y. designed the project and wrote the paper. ACKNOWLEDGMENTS We thank Dr. F. Hofmann and Dr. S. Liang for providing the Nav1.7 plasmid and Dr. R. Kuner for SNS-Cre mice. This work was supported by NNSFC (31130066 and 31330046), SPRP (B) of CAS (XDB02010000), and STCSM (16JC1420500). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors have declared that no competing interests exist. Received: August 3, 2016 Revised: December 15, 2016 Accepted: January 4, 2017 Published: February 2, 2017 REFERENCES Agarwal, N., Offermanns, S., and Kuner, R. (2004). Conditional gene deletion in primary nociceptive neurons of trigeminal ganglia and dorsal root ganglia. Genesis 38, 122–129. Akopian, A.N., Souslova, V., England, S., Okuse, K., Ogata, N., Ure, J., Smith, A., Kerr, B.J., McMahon, S.B., Boyce, S., et al. (1999). The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways. Nat. Neurosci. 2, 541–548. Amaya, F., Wang, H., Costigan, M., Allchorne, A.J., Hatcher, J.P., Egerton, J., Stean, T., Morisset, V., Grose, D., Gunthorpe, M.J., et al. (2006). The voltagegated sodium channel Nav1.9 is an effector of peripheral inflammatory pain hypersensitivity. J. Neurosci. 26, 12852–12860. Arrigoni, C., Rohaim, A., Shaya, D., Findeisen, F., Stein, R.A., Nurva, S.R., Mishra, S., Mchaourab, H.S., and Minor, D.L., Jr. (2016). Unfolding of a temperature-sensitive domain controls voltage-gated channel activation. Cell 164, 922–936. Becerra, L., Chang, P.C., Bishop, J., and Borsook, D. (2011). CNS activation maps in awake rats exposed to thermal stimuli to the dorsum of the hindpaw. Neuroimage 54, 1355–1366. Caterina, M.J., Leffler, A., Malmberg, A.B., Martin, W.J., Trafton, J., PetersenZeitz, K.R., Koltzenburg, M., Basbaum, A.I., and Julius, D. (2000). Impaired nociception and pain sensation in mice lacking the capsaicin receptor. Science 288, 306–313. Catterall, W.A., Goldin, A.L., and Waxman, S.G. (2005). International Union of Pharmacology. XLVII. Nomenclature and structure-function relationships of voltage-gated sodium channels. Pharmacol. Rev. 57, 397–409.

Cho, H., Yang, Y.D., Lee, J., Lee, B., Kim, T., Jang, Y., Back, S.K., Na, H.S., Harfe, B.D., Wang, F., et al. (2012). The calcium-activated chloride channel anoctamin 1 acts as a heat sensor in nociceptive neurons. Nat. Neurosci. 15, 1015–1021. Cox, J.J., Reimann, F., Nicholas, A.K., Thornton, G., Roberts, E., Springell, K., Karbani, G., Jafri, H., Mannan, J., Raashid, Y., et al. (2006). An SCN9A channelopathy causes congenital inability to experience pain. Nature 444, 894–898. Davis, J.B., Gray, J., Gunthorpe, M.J., Hatcher, J.P., Davey, P.T., Overend, P., Harries, M.H., Latcham, J., Clapham, C., Atkinson, K., et al. (2000). Vanilloid receptor-1 is essential for inflammatory thermal hyperalgesia. Nature 405, 183–187. Dhaka, A., Viswanath, V., and Patapoutian, A. (2006). Trp ion channels and temperature sensation. Annu. Rev. Neurosci. 29, 135–161. Dib-Hajj, S.D., Rush, A.M., Cummins, T.R., Hisama, F.M., Novella, S., Tyrrell, L., Marshall, L., and Waxman, S.G. (2005). Gain-of-function mutation in Nav1.7 in familial erythromelalgia induces bursting of sensory neurons. Brain 128, 1847–1854. Dib-Hajj, S.D., Cummins, T.R., Black, J.A., and Waxman, S.G. (2010). Sodium channels in normal and pathological pain. Annu. Rev. Neurosci. 33, 325–347. Dib-Hajj, S.D., Yang, Y., Black, J.A., and Waxman, S.G. (2013). The NaV1.7 sodium channel: from molecule to man. Nat. Rev. Neurosci. 14, 49–62. Emery, E.C., Habib, A.M., Cox, J.J., Nicholas, A.K., Gribble, F.M., Woods, C.G., and Reimann, F. (2015). Novel SCN9A mutations underlying extreme pain phenotypes: unexpected electrophysiological and clinical phenotype correlations. J. Neurosci. 35, 7674–7681. Fertleman, C.R., Baker, M.D., Parker, K.A., Moffatt, S., Elmslie, F.V., Abrahamsen, B., Ostman, J., Klugbauer, N., Wood, J.N., Gardiner, R.M., and Rees, M. (2006). SCN9A mutations in paroxysmal extreme pain disorder: allelic variants underlie distinct channel defects and phenotypes. Neuron 52, 767–774. Gingras, J., Smith, S., Matson, D.J., Johnson, D., Nye, K., Couture, L., Feric, E., Yin, R., Moyer, B.D., Peterson, M.L., et al. (2014). Global Nav1.7 knockout mice recapitulate the phenotype of human congenital indifference to pain. PLoS ONE 9, e105895. Goetz, R., Dover, K., Laezza, F., Shtraizent, N., Huang, X., Tchetchik, D., Eliseenkova, A.V., Xu, C.F., Neubert, T.A., Ornitz, D.M., et al. (2009). Crystal structure of a fibroblast growth factor homologous factor (FHF) defines a conserved surface on FHFs for binding and modulation of voltage-gated sodium channels. J. Biol. Chem. 284, 17883–17896. Goldfarb, M., Schoorlemmer, J., Williams, A., Diwakar, S., Wang, Q., Huang, X., Giza, J., Tchetchik, D., Kelley, K., Vega, A., et al. (2007). Fibroblast growth factor homologous factors control neuronal excitability through modulation of voltage-gated sodium channels. Neuron 55, 449–463. Guillemot, F., and Zimmer, C. (2011). From cradle to grave: the multiple roles of fibroblast growth factors in neural development. Neuron 71, 574–588. Han, C., Lampert, A., Rush, A.M., Dib-Hajj, S.D., Wang, X., Yang, Y., and Waxman, S.G. (2007). Temperature dependence of erythromelalgia mutation L858F in sodium channel Nav1.7. Mol. Pain 3, 3. Hartung, H., Feldman, B., Lovec, H., Coulier, F., Birnbaum, D., and Goldfarb, M. (1997). Murine FGF-12 and FGF-13: expression in embryonic nervous system, connective tissue and heart. Mech. Dev. 64, 31–39. Kim, C.H., Oh, Y., Chung, J.M., and Chung, K. (2002). Changes in three subtypes of tetrodotoxin sensitive sodium channel expression in the axotomized dorsal root ganglion in the rat. Neurosci. Lett. 323, 125–128. Klugbauer, N., Lacinova, L., Flockerzi, V., and Hofmann, F. (1995). Structure and functional expression of a new member of the tetrodotoxin-sensitive voltage-activated sodium channel family from human neuroendocrine cells. EMBO J. 14, 1084–1090. Laezza, F., Gerber, B.R., Lou, J.Y., Kozel, M.A., Hartman, H., Craig, A.M., Ornitz, D.M., and Nerbonne, J.M. (2007). The FGF14(F145S) mutation disrupts the interaction of FGF14 with voltage-gated Na+ channels and impairs neuronal excitability. J. Neurosci. 27, 12033–12044.

Neuron 93, 1–16, February 22, 2017 15

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Laezza, F., Lampert, A., Kozel, M.A., Gerber, B.R., Rush, A.M., Nerbonne, J.M., Waxman, S.G., Dib-Hajj, S.D., and Ornitz, D.M. (2009). FGF14 N-terminal splice variants differentially modulate Nav1.2 and Nav1.6-encoded sodium channels. Mol. Cell. Neurosci. 42, 90–101. Lee, J.H., Park, C.K., Chen, G., Han, Q., Xie, R.G., Liu, T., Ji, R.R., and Lee, S.Y. (2014). A monoclonal antibody that targets a NaV1.7 channel voltage sensor for pain and itch relief. Cell 157, 1393–1404. Leo, S., D’Hooge, R., and Meert, T. (2010). Exploring the role of nociceptorspecific sodium channels in pain transmission using Nav1.8 and Nav1.9 knockout mice. Behav. Brain Res. 208, 149–157. Li, G.D., Wo, Y., Zhong, M.F., Zhang, F.X., Bao, L., Lu, Y.J., Huang, Y.D., Xiao, H.S., and Zhang, X. (2002). Expression of fibroblast growth factors in rat dorsal root ganglion neurons and regulation after peripheral nerve injury. Neuroreport 13, 1903–1907. Li, C.L., Li, K.C., Wu, D., Chen, Y., Luo, H., Zhao, J.R., Wang, S.S., Sun, M.M., Lu, Y.J., Zhong, Y.Q., et al. (2016). Somatosensory neuron types identified by high-coverage single-cell RNA-sequencing and functional heterogeneity. Cell Res. 26, 83–102. Liu, C.J., Dib-Hajj, S.D., and Waxman, S.G. (2001). Fibroblast growth factor homologous factor 1B binds to the C terminus of the tetrodotoxin-resistant sodium channel rNav1.9a (NaN). J. Biol. Chem. 276, 18925–18933. Liu, C.J., Dib-Hajj, S.D., Renganathan, M., Cummins, T.R., and Waxman, S.G. (2003). Modulation of the cardiac sodium channel Nav1.5 by fibroblast growth factor homologous factor 1B. J. Biol. Chem. 278, 1029–1036. Lou, J.Y., Laezza, F., Gerber, B.R., Xiao, M., Yamada, K.A., Hartmann, H., Craig, A.M., Nerbonne, J.M., and Ornitz, D.M. (2005). Fibroblast growth factor 14 is an intracellular modulator of voltage-gated sodium channels. J. Physiol. 569, 179–193. Minett, M.S., Nassar, M.A., Clark, A.K., Passmore, G., Dickenson, A.H., Wang, F., Malcangio, M., and Wood, J.N. (2012). Distinct Nav1.7-dependent pain sensations require different sets of sensory and sympathetic neurons. Nat. Commun. 3, 791. Nassar, M.A., Stirling, L.C., Forlani, G., Baker, M.D., Matthews, E.A., Dickenson, A.H., and Wood, J.N. (2004). Nociceptor-specific gene deletion reveals a major role for Nav1.7 (PN1) in acute and inflammatory pain. Proc. Natl. Acad. Sci. USA 101, 12706–12711. Nassar, M.A., Levato, A., Stirling, L.C., and Wood, J.N. (2005). Neuropathic pain develops normally in mice lacking both Nav1.7 and Nav1.8. Mol. Pain 1, 24. €l, J., Zimmermann, K., Busserolles, J., Deval, E., Alloui, A., Diochot, S., Noe Guy, N., Borsotto, M., Reeh, P., Eschalier, A., and Lazdunski, M. (2009). The mechano-activated K+ channels TRAAK and TREK-1 control both warm and cold perception. EMBO J. 28, 1308–1318. Priest, B.T., Murphy, B.A., Lindia, J.A., Diaz, C., Abbadie, C., Ritter, A.M., Liberator, P., Iyer, L.M., Kash, S.F., Kohler, M.G., et al. (2005). Contribution of the tetrodotoxin-resistant voltage-gated sodium channel NaV1.9 to sensory transmission and nociceptive behavior. Proc. Natl. Acad. Sci. USA 102, 9382–9387. Puranam, R.S., He, X.P., Yao, L., Le, T., Jang, W., Rehder, C.W., Lewis, D.V., and McNamara, J.O. (2015). Disruption of Fgf13 causes synaptic excitatoryinhibitory imbalance and genetic epilepsy and febrile seizures plus. J. Neurosci. 35, 8866–8881. Schmalhofer, W.A., Calhoun, J., Burrows, R., Bailey, T., Kohler, M.G., Weinglass, A.B., Kaczorowski, G.J., Garcia, M.L., Koltzenburg, M., and Priest, B.T. (2008). ProTx-II, a selective inhibitor of NaV1.7 sodium channels,

16 Neuron 93, 1–16, February 22, 2017

blocks action potential propagation in nociceptors. Mol. Pharmacol. 74, 1476–1484. Schwarze, S.R., Ho, A., Vocero-Akbani, A., and Dowdy, S.F. (1999). In vivo protein transduction: delivery of a biologically active protein into the mouse. Science 285, 1569–1572. Shields, S.D., Cheng, X., Uc¸eyler, N., Sommer, C., Dib-Hajj, S.D., and Waxman, S.G. (2012). Sodium channel Nav1.7 is essential for lowering heat pain threshold after burn injury. J. Neurosci. 32, 10819–10832. Tate, S., Benn, S., Hick, C., Trezise, D., John, V., Mannion, R.J., Costigan, M., Plumpton, C., Grose, D., Gladwell, Z., et al. (1998). Two sodium channels contribute to the TTX-R sodium current in primary sensory neurons. Nat. Neurosci. 1, 653–655. Treede, R.D., Kenshalo, D.R., Gracely, R.H., and Jones, A.K. (1999). The cortical representation of pain. Pain 79, 105–111. Vriens, J., Owsianik, G., Hofmann, T., Philipp, S.E., Stab, J., Chen, X., Benoit, M., Xue, F., Janssens, A., Kerselaers, S., et al. (2011). TRPM3 is a nociceptor channel involved in the detection of noxious heat. Neuron 70, 482–494. Wang, Q., Bardgett, M.E., Wong, M., Wozniak, D.F., Lou, J., McNeil, B.D., Chen, C., Nardi, A., Reid, D.C., Yamada, K., and Ornitz, D.M. (2002). Ataxia and paroxysmal dyskinesia in mice lacking axonally transported FGF14. Neuron 35, 25–38. Wittmack, E.K., Rush, A.M., Craner, M.J., Goldfarb, M., Waxman, S.G., and Dib-Hajj, S.D. (2004). Fibroblast growth factor homologous factor 2B: association with Nav1.6 and selective colocalization at nodes of Ranvier of dorsal root axons. J. Neurosci. 24, 6765–6775. Wu, Q.F., Yang, L., Li, S., Wang, Q., Yuan, X.B., Gao, X., Bao, L., and Zhang, X. (2012). Fibroblast growth factor 13 is a microtubule-stabilizing protein regulating neuronal polarization and migration. Cell 149, 1549–1564. Xiao, H.S., Huang, Q.H., Zhang, F.X., Bao, L., Lu, Y.J., Guo, C., Yang, L., Huang, W.J., Fu, G., Xu, S.H., et al. (2002). Identification of gene expression profile of dorsal root ganglion in the rat peripheral axotomy model of neuropathic pain. Proc. Natl. Acad. Sci. USA 99, 8360–8365. Xiao, M., Xu, L., Laezza, F., Yamada, K., Feng, S., and Ornitz, D.M. (2007). Impaired hippocampal synaptic transmission and plasticity in mice lacking fibroblast growth factor 14. Mol. Cell. Neurosci. 34, 366–377. Xiao, M., Bosch, M.K., Nerbonne, J.M., and Ornitz, D.M. (2013). FGF14 localization and organization of the axon initial segment. Mol. Cell. Neurosci. 56, 393–403. Yang, Y., Wang, Y., Li, S., Xu, Z., Li, H., Ma, L., Fan, J., Bu, D., Liu, B., Fan, Z., et al. (2004). Mutations in SCN9A, encoding a sodium channel alpha subunit, in patients with primary erythermalgia. J. Med. Genet. 41, 171–174. Yeomans, D.C., Levinson, S.R., Peters, M.C., Koszowski, A.G., Tzabazis, A.Z., Gilly, W.F., and Wilson, S.P. (2005). Decrease in inflammatory hyperalgesia by herpes vector-mediated knockdown of Nav1.7 sodium channels in primary afferents. Hum. Gene Ther. 16, 271–277. Zhang, Z.N., Li, Q., Liu, C., Wang, H.B., Wang, Q., and Bao, L. (2008). The voltage-gated Na+ channel Nav1.8 contains an ER-retention/retrieval signal antagonized by the beta3 subunit. J. Cell Sci. 121, 3243–3252. Zhang, X., Bao, L., Yang, L., Wu, Q., and Li, S. (2012). Roles of intracellular fibroblast growth factors in neural development and functions. Sci. China Life Sci. 55, 1038–1044. Zhuo, M. (2016). Neural mechanisms underlying anxiety-chronic pain interactions. Trends Neurosci. 39, 136–145.

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

STAR+METHODS KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies IHC/WB/IP: anti-FGF13

Santa Cruz Biotechnology

Cat# sc-16811 RRID: AB_2104044

IHC: anti-CGRP

DiaSorin

Cat# 24112 RRID: AB_572217

IHC: anti-PGP9.5

Ultraclone

Cat# RA95101 RRID: AB_2313685

IHC: anti-tuj1

Abcam

Cat# ab107216 RRID: AB_10899689

IHC/WB: anti-Nav1.8

Alomone Labs

Cat# ASC-016 RRID: AB_2040188

IHC: Fluorescein-labeled IB4

Vector Laboratories

Cat# FL-1201 RRID: AB_2314663

IHC: FITC donkey anti-goat

Jackson ImmunoResearch

Cat# 705-095-147 RRID: AB_2340401

IHC: FITC donkey anti-rabbit

Jackson ImmunoResearch

Cat# 711-095-152 RRID: AB_2340596

IHC: Cy3 donkey anti-goat

Jackson ImmunoResearch

Cat# 705-165-147 RRID: AB_2307351

IHC: Cy3 donkey anti-rabbit

Jackson ImmunoResearch

Cat# 711-165-152 RRID: AB_2307443

IHC: Cy5 donkey anti-goat

Jackson ImmunoResearch

Cat# 705-175-147 RRID: AB_2340415

IHC: Cy5 donkey anti-rabbit

Jackson ImmunoResearch

Cat# 711-175-152 RRID: AB_2340607

IHC: RRX donkey anti-chicken

Jackson ImmunoResearch

Cat# 703-295-155 RRID: AB_2340371

WB/IP: anti-FGF13

This paper

immunization with FGF13B 182-192 aa

WB: anti-Nav1.7

Millipore

Cat# MABN41 RRID: AB_10808664

WB: anti-Nav1.6

Millipore

Cat# AB5580 RRID: AB_2314858

WB: anti-Nav1.9

Alomone Labs

Cat# ASC-017 RRID: AB_2040200

WB: anti-TRPV1

Santa Cruz Biotechnology

Cat# sc-12498 RRID: AB_2241046

WB: anti-actin

Chemicon

Cat# MAB1501 RRID: AB_2223041

WB: anti-transferrin receptor

Invitrogen

Cat# 13-6890 RRID: AB_2533030

WB: anti-flag

Abmart

Cat# M20008

WB: anti-myc

Developmental Studies Hybridoma Bank

Cat# 9E 10 RRID: AB_2266850

WB: anti-TRPV2

Sigma-Aldrich

Cat# SAB2501068 RRID: AB_10602720

WB: anti-GST

Proteintech Group

Cat# 10000-0-AP RRID: AB_11042316

WB: anti-pan-cadherin

Cell Signaling Technology

Cat# 4068 RRID: AB_2158565

WB: anti-HSP90

Cell Signaling Technology

Cat# 4874 RRID: AB_2121214

IP: normal Goat IgG

Santa Cruz Biotechnology

Cat# sc-2028 RRID: AB_737167

IP: normal Rabbit IgG

Santa Cruz Biotechnology

Cat# sc-2027 RRID: AB_737197

IP: anti-myc

Sigma-Aldrich

Cat# C3956 RRID: AB_439680

IP: anti-GST

Amersham Pharmacia Biotech

Cat# 27-4577-01 RRID: AB_771432

Hoechst 33258

Sigma-Aldrich

Cat# 94403

Coomassie brilliant blue R-250

Sangon Biotech

Cat# A602151

Chemicals, Peptides, and Recombinant Proteins

Ponceau S

Sangon Biotech

Cat# A100860

Paraformaldehyde

Sigma-Aldrich

Cat# P6148

Formaldehyde

Sigma-Aldrich

Cat# F1635

Saturated picric acid

Sigma-Aldrich

Cat# P6744

Benzyl Alcohol

Sigma-Aldrich

Cat# 402834

Benzyl Benzoate

Sigma-Aldrich

Cat# B6630

Thrombin

Sigma-Aldrich

Cat# T6884

DSS

Thermo Fisher Scientific

Cat# 21655

Sulfo-NHS-LC-Biotin

Pierce Protein Biology

Cat# 21335 (Continued on next page)

Neuron 93, 1–16.e1–e9, February 22, 2017 e1

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

Streptavidin-agarose beads

Pierce Protein Biology

Cat# 20353

Protoxin-II

Smartox Biotechnology

Cat# 07PTX002

Capsaicin

Tocris

Cat# 0462

KOD-plus Mutagenesis Kit

Toyobo

Cat# SMK-101

Glutathione-Sepharose beads

GE Healthcare

Cat# 17075601

Protein G beads

Roche

Cat# 1243233

MEM-Per Plus Kit

Thermo Fisher Scientific

Cat# 89842

SuperScript II reverse transcriptase

Thermo Fisher Scientific

Cat# 18064014

Li et al., 2016

GEO: GSE63576

Cell Bank of the Chinese Academy of Sciences (Shanghai, China)

Cat# GNHu18

Mouse: C57BL/6J

Shanghai Laboratory Animal Center, Chinese Academy of Sciences (Shanghai, China)

N/A

Mouse: SNS-Cre

Agarwal et al., 2004

N/A

Mouse: Fgf13F/Y or Fgf13F/F

This paper

N/A

Mouse: TRPV1 KO

the Jackson Laboratory

Cat# 003770 RRID: IMSR_JAX:003770

Bacteria: Escherichia coli BL21

TIANGEN BIOTECH

Cat# CB105

Critical Commercial Assays

Deposited Data Single-RNA sequencing data Experimental Models: Cell Lines HEK293

Experimental Models: Organisms/Strains

Recombinant DNA FGF13B (with a separate EGFP)

This paper

N/A

FGF13B (full-length)-myc

This paper

N/A

GST-FGF13B

This paper

N/A

FGF13BD1-37-myc

This paper

N/A

FGF13BD38-75-myc

This paper

N/A

FGF13BD76-99-myc

This paper

N/A

FGF13BD100-145-myc

This paper

N/A

FGF13BD146-192-myc

This paper

N/A

Nav1.7-flag

This paper

N/A

GST-flag-Nav1.7CT

This paper

N/A

GST-flag

This paper

N/A

GST-flag-Nav1.7CT-f1

This paper

N/A

GST-flag-Nav1.7CT-f2

This paper

N/A

GST-flag-Nav1.7CT-f3

This paper

N/A

GST-flag-Nav1.7CT-f4

This paper

N/A

GST-flag-Nav1.7CT-f5

This paper

N/A

GST-flag-Nav1.7CT-f6

This paper

N/A

GST-flag-Nav1.7CT-f7

This paper

N/A

GST-flag-Nav1.7CT-f8

This paper

N/A

GST-flag-Nav1.7CT-f9

This paper

N/A

GST-flag-TAT

This paper

N/A

myc-CD8a

Zhang et al., 2008

N/A

myc-CD8a-Nav1.7CT

This paper

N/A (Continued on next page)

e2 Neuron 93, 1–16.e1–e9, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

GST-flag-Nav1.7CT-TAT

This paper

N/A

GST-FGF13B-TAT

This paper

N/A

Nav1.7

Klugbauer et al., 1995

N/A

Nav1.8-myc

Zhang et al., 2008

N/A

See Table S1 and S2

N/A

Clampfit 10.3

Axon Instruments

http://mdc.custhelp.com/app/answers/ detail/a_id/18779/kw/clampfit%2010

Origin pro 8.6.0

Microcal

http://www.originlab.com/Origin

GraphPad Prism

GraphPad Software

http://www.graphpad.com/scientificsoftware/prism/

Imaris

Bitplane

http://www.bitplane.com/imaris/imaris

ImageJ

National Institutes of Health

https://imagej.nih.gov/ij/

Sequence-Based Reagents Primers for plasmid, oligomer and PCR Software and Algorithms

CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents may be directed to and will be fulfilled by the Lead Contact Xu Zhang ([email protected]). EXPERIMENTAL MODEL AND SUBJECT DETAILS Animals Experiments were performed according to the guidelines of the Committee for Research and Ethical Issues of the International Association for the Study of Pain and were approved by the Committee of Use of Laboratory Animals and Common facility, Institute of Neuroscience, the Chinese Academy of Sciences. Mice were raised together with littermates in pathogen-free environment and their health status was routinely checked. No more than 5 mice were housed in one cage. Food and water were provided ad libitum. Mice were maintained in a 12-hr light/dark cycle at 22
C–26
C. Experiments were conducted during the light phase of the cycle. 2- to 4-month-old male mice were used for all in vivo and in vitro experiments. We generated the conditional knockout mice lacking FGF13 specifically in small DRG neurons. The FGF13 gene is located in the X chromosome. We made the Fgf13-loxP mice by flanking its exons 2 and 3 which encode the FGF13 core region with loxP sequences. The FGF13 gene was deleted selectively in DRG neurons by crossing Fgf13-loxP mice with BAC transgenic mice expressing Cre recombinase controlled by promoter elements of the Nav1.8 gene, which is mainly expressed in small DRG neurons (SNS-Cre). This gene deletion mediated by SNS-Cre started at the perinatal stage, thereby minimizing the risk of developmental defects. The Fgf13 conditional knockout mice were viable and fertile and did not exhibit visible abnormalities. Then, experiments were performed with the Fgf13 knockout mice (Fgf13/Y) and the control Fgf13-loxP mice (Fgf13F/Y). Trpv1/ mice were purchased from the Jackson Laboratory. C57BL/6J mice purchased from Shanghai Laboratory Animal Center, Chinese Academy of Sciences (Shanghai, China). METHOD DETAILS Plasmid Construction All primers and oligonucleotides used for the construction of plasmids expressing FGF13B and FGF13B truncations, GST-FGF13B, Nav1.7, Nav1.8, CD8a-Nav1.7CT, GST-Nav1.7CT fragments, GST-flag-Nav1.7CT-TAT and GST-FGF13B-TAT are listed in Table S1. The cDNA of mouse FGF13B was cloned into pCAG-IRES-EGFP, pcDNA3.1/myc-his(-) and pGEX vectors for the expression of FGF13B (with a separately expressed EGFP), FGF13B (full-length)-myc and GST-FGF13B, respectively. The plasmids expressing FGF13BD1-37-myc, FGF13BD38-75-myc, FGF13BD76-99-myc, FGF13BD100-145-myc and FGF13BD146-192-myc were modified from the FGF13B (full-length)-myc plasmid using the KOD-plus Mutagenesis Kit (Toyobo). The cDNA of human Nav1.7 was constructed in a pcDNA3.1-mod vector (Klugbauer et al., 1995), and a flag tag was inserted into the C terminus of Nav1.7 using the KOD-plus Mutagenesis Kit (Toyobo). Full length Nav1.8 was cloned into pEGFP-N3 vector and the GFP tag of pEGFP-N3 vector was replaced by a myc tag (Zhang et al., 2008). The C terminus (5248-5931 bp) of human Nav1.7 (Nav1.7CT) following a flag tag was cloned into a pGEX vector for the construction of the GST-flag-Nav1.7CT plasmid. The plasmids expressing GST-flag and the GST-flag-Nav1.7CT fragments (f1-f9) were modified from the GST-flag-Nav1.7CT plasmid by deletion using the KOD-plus Mutagenesis Kit (Toyobo). Glycine10-TAT was made by oligonucleotides denaturing and annealing. The Glycine10-TAT sequence was inserted into the

Neuron 93, 1–16.e1–e9, February 22, 2017 e3

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

C terminus of GST-flag-Nav1.7CT and GST-FGF13B for the construction of the GST-flag-Nav1.7CT-TAT and GST-FGF13B-TAT plasmids, respectively. The GST-flag-TAT plasmid was made from the GST-flag-Nav1.7CT-TAT plasmid by deletion using the KOD-plus Mutagenesis Kit (Toyobo). The Nav1.7CT sequence was cloned into a pMyc-CD8a construct (Zhang et al., 2008). The rat CD8a sequence on the N terminus was used to anchor Nav1.7CT to the cell membrane. Cell Culture and Transfection HEK293 cells were cultured in MEM with 10% fetal bovine serum. The cells were transfected with 1–2 mg plasmid per 35-mm dish or 2-3 mg plasmid per 60-mm dish using Lipofectamine 2000 reagent (Invitrogen). The cells were used for the following experiments 24–48 hr after transfection. DRGs from 2- to 4-month-old mice were carefully isolated and digested in oxygenated DMEM containing collagenase (0.4 mg/mL), trypsin (1 mg/mL) and DNase (0.1 mg/mL) for 30 min at 37
C. After gentle trituration, dissociated cells were plated onto poly-D-lysinecoated glass coverslips in DMEM/F12 (1:1) media supplemented with 100 U/mL penicillin, 0.1 mg/mL streptomycin and 10% fetal bovine serum. Morphology Analysis 2- to 4-month-old mice were perfused with 4% paraformaldehyde (PFA). The lumber (L) 4–5 DRGs and spinal cord were dissected and post-fixed in 4% PFA for 2 hr. Then the tissues were dehydrated and embedded for cryostat sectioning. The cryostat sections were incubated in primary antibodies at 4
C overnight, and then incubated in secondary antibodies at 37
C for 45 min. Skin immunostaining was performed with non-perfused mice. The glabrous skin of mouse hindpaw was dissected and fixed in Zamboni’s fixative [2% formaldehyde, 15% saturated picric acid, 0.1 M phosphate buffer (pH 7.3)] for 18 hr at 4
C. The cryostat sections at 30 mm were placed in phosphate-buffered saline (PBS) with 0.1% Triton X-100 and stained as free floating sections. The primary antibodies against FGF13 (1:3000, Go; Santa Cruz Biotechnology), CGRP (1:1000, Rb; DiaSorin), PGP9.5 (1:1000, Rb; Ultraclone), tuj1 (1:1000, Chicken; Abcam) and Nav1.8 (1:500, Rb; Alomone Labs) were used. The secondary antibodies were FITC, Cy3 or Cy5-conjugated donkey against goat, rabbit or chicken antibodies (1:100; Jackson ImmunoResearch). Fluorescein-labeled IB4 (1:200; Vector Laboratories) stained the sections at 37
C for 45 min. Free floating sections were counterstained with Hoechst 33258 (1:5000; Sigma-Aldrich) at room temperature for 15 min. For whole-mount DRG immunostaining, L5 DRGs were dissected and fixed in 4% PFA at 4
C overnight. The DRGs were washed and penetrated with 0.3% Triton X-100 in PBS every one hour for 5 hr. Then the tissues were incubated with primary antibodies in blocking solution (75% 0.3% PBST, 20% DMSO and 5% Donkey Serum) for 4 d and then were washed with 0.3% Triton X-100 in PBS every one hour for 5 hr. The tissues were incubated with secondary antibodies in blocking solution for 2 d and then washed with 0.3% Triton X-100 in PBS every one hour for 5 hr. All the washing and incubation procedures were performed on a rocking platform at room temperature. The tissues were dehydrated in methanol for 1 hr and cleared in Benzyl Alcohol: Benzyl Benzoate 1:2. The tissues were mounted in Benzyl Alcohol/Benzyl Benzoate and the images were acquired using Leica SP8 confocal microscope. The total cell number of DRGs was manually counted with ImageJ cell counter tool. Nissl staining was performed with 0.1% cresyl violet acetate solution. The sections were stained in cresyl violet acetate solution for 5–10 min at 37
C, rinsed in water, dehydrated in 70, 95 and 100% ethanol, cleared in xylene and finally mounted for observation. To examine the ultrastructure of the spinal cord, we fixed the mice with 4% PFA and 0.05% glutaraldehyde, and dissected the lumbar spinal cord. The vibratome sections of the lumbar spinal cord were post-fixed in 0.5% osmium tetroxide and embedded in Epon 812. The ultrathin sections of the spinal cord dorsal horn were examined with an electron microscope. Immunoblotting The cerebral cortex, cerebellum, thalamus, hippocampus, DRG, spinal cord, dorsal root, sciatic nerve and plantar skin tissues from adult mice were dissected and homogenized. The tissue homogenates and cultured cells were lysed with RIPA buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.1% Triton X-100, 10% glycerol, 0.05% bovine serum albumin, 1 mM PMSF, 10 mg/mL aprotinin, 1 mg/mL pepstatin and 1 mg/mL leupeptin] or [10 mM HEPES (pH 7.5), 10 mM NaF, 120 mM NaCl, 5 mM EDTA, 1% Triton X-100, 1 mM PMSF, 10 mg/mL aprotinin, 1 mg/mL pepstatin and 1 mg/mL leupeptin]. The denatured samples were loaded for SDS-PAGE, transferred, probed with antibodies and visualized with enhanced chemiluminescence. The antibodies against FGF13 (1:500, Go), FGF13 (1:5000, Rb; made by immunization with the C terminus of FGF13B 182-192 aa), Nav1.7 (1:500, Mo; Millipore), Nav1.8 (1:500), Nav1.6 (1:500, Rb; Millipore), Nav1.9 (1:500, Rb; Alomone Labs), TRPV1 (1:500, Go, Santa Cruz Biotechnology), actin (1:100000, Mo; Chemicon), transferrin receptor (1:500, Mo; Invitrogen), flag (1:2000, Mo; Abmart), myc (1:1000, Mo; Developmental Studies Hybridoma Bank), TRPV2 (1:500, Go; Sigma-Aldrich), GST (1:500, Go; Amersham Pharmacia Biotech), pan-cadherin (1:2000, Rb, Cell Signaling Technology), HSP90 (1:2000, Rb, Cell Signaling Technology) were used. We also stained the SDS-PAGE gels and immunoblotting membranes with Coomassie brilliant blue R-250 (Sangon Biotech) (0.25% solution in methanol: water: acetic acid 5:4:1) and Ponceau S (Sangon Biotech) (0.5% solution in 1% acetic acid), respectively, to show the GST-fused proteins. Preparation of the GST-Fused Proteins The GST-fused proteins were expressed in Escherichia coli BL21. The bacteria were grown in 2 3 YTA media and the protein expression was induced by 1 mM Isopropyl b-D-thiogalactopyranoside (IPTG). Then the bacteria were centrifuged, resuspended and e4 Neuron 93, 1–16.e1–e9, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

sonicated for the protein release. The proteins were purified using Glutathione-Sepharose beads (GE Healthcare), concentrated and quantified before use. Immunoprecipitation and GST Pull-Down The cultured cells or dissected tissues were lysed in ice cold RIPA buffer. The lysate was centrifuged and 5% of the supernatant was taken for a whole-cell lysate sample. The remaining supernatant was precipitated with 1–5 mg of antibody at 4
C overnight and afterward protein G beads (Roche) at 4
C for 2 hr. The immunoprecipitated sample was denatured and prepared for immunoblotting. The immunoprecipitation was performed with antibodies against FGF13 (Go), FGF13 (Rb), GST (Go, Amersham Pharmacia Biotech) and myc (Rb, Sigma-Aldrich). The GST-fused protein (5 mg) was incubated with the cell lysate at 4
C overnight. Then, the GST-fused protein was precipitated with 10 mL of Glutathione-Sepharose beads at 4
C for 2 hr. The precipitant was washed, denatured and prepared for immunoblotting. To detect the direct interaction between FGF13 and Nav1.7CT, we prepared purified FGF13B by removing GST from GSTFGF13B using thrombin (Sigma-Aldrich). The GST-FGF13B slurry (100 mL, 1 mg/mL) was digested with 0.02 units of thrombin at room temperature for 12 hr. The GST and undigested GST-FGF13B proteins were attached with the beads and were removed by centrifugation. Then, the purified FGF13B protein was collected from the supernatant. We incubated 10 mL of supernatant (diluted in 400 mL of RIPA buffer) with 5 mg of GST-flag-Nav1.7CT or the GST-flag-Nav1.7CT fragments for the detection of their direct interaction. To test the effect of heat stimulation on the FGF13/Nav1.7 interaction, we collected HEK293 cells co-transfected with FGF13B and Nav1.7, and added DSS (Thermo Fisher Scientific) to the cell suspension shortly before or after the heat stimulation (43
C for 1 min) to chemically crosslink the FGF13 and Nav1.7 that interacted during the heat stimulation. The reaction was stopped after 45 min incubation at 4
C. The cells were collected and mildly lysed with Permeabilization Buffer (MEM-Per Plus Kit, Thermo Fisher Scientific). The cytosol component was separated from the supernatant after centrifugation. The pellet (enriched membrane component) was further lysed with crosslinking RIPA buffer [50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM PMSF, 10 mg/mL aprotinin, 1 mg/mL pepstatin and 1 mg/mL leupeptin]. The high detergent condition in crosslinking RIPA buffer effectively prevented interaction between free FGF13 and Nav1.7 so that only the FGF13 and Nav1.7 crosslinked with covalent bond was detected. After the incubation with FGF13 antibody and protein G beads, the proteins attached to the protein G beads were washed repeatedly to further remove the freely bound Nav1.7. In this way, we uncovered the FGF13/Nav1.7 interaction regulated by the heat stimulation. Cell Surface Biotinylation Cultured cells were dissociated using trypsin and incubated in Ca2+/Mg2+ PBS. The cells were transferred into a small volume of 100 mL per tube for rapid heating. The cells were treated with or without noxious heat stimulation and then were cooled on ice immediately. Sulfo-NHS-LC-Biotin (Pierce Protein Biology) was added into each cell suspension at a working concentration of 0.25 mg/mL. The biotinylation reaction was stopped after 45 min of incubation at 4
C. Then, the cells were lysed with RIPA buffer, and 5% of the lysate was used as a whole-cell lysate sample and the remaining lysate was precipitated with streptavidin-agarose beads (Pierce Protein Biology) overnight at 4
C. The precipitated sample was denatured and prepared for immunoblotting. In the crosslinking assay, DSS (Thermo Fisher Scientific) was added to the cell suspension shortly before or after heat stimulation. The cell suspension was crosslinked at 4
C for 10 min. Then, the cells were biotinylated, lysed, precipitated and processed for immunoblotting. Genotyping The Fgf13 conditional knockout mice were identified by genotyping. A small piece of mouse tail or toe was digested in Proteinase K buffer. The genomic DNA was extracted with phenol/chloroform, precipitated with isopropanol, washed with 75% ethanol and dissolved in water. The mouse genotype was then identified by PCR. The primers for the identification of Fgf13 flox were 50 -CTA GAGGTCCTTTGATGGGA-30 and 50 -GTGGAAGAGATTCAGGATTC-30 . The primers for the identification of SNS-Cre were 50 ATTTGCCTGCATTACCGGTC-30 and 50 -GCATCAACGTTTTCTTTTCGG-30 . The FGF13 gene is located in the X chromosome. Male mice carrying Fgf13 flox and SNS-Cre (Fgf13-/Y) were used as the experimental group, while male mice carrying Fgf13 flox without SNS-Cre (Fgf13F/Y) were used as the control group. Behavioral Tests Behavioral tests were performed with 2- to 4-month-old male mice. In the tail-flick, Hargreaves and hot plate tests, the response latency to nociceptive stimuli was measured. In von Frey and Randall-Selitto test, the threshold of nociceptive responses was measured. The tests above were stopped at a cutoff time or force for animal protection. In evaporative cooling and formalin test, the duration of nociceptive responses was measured. The Tail-Flick Test The tail of mouse was immersed in a water bath at 48
C, 50
C or 52
C. The cutoff time was 20 s, 15 s, or 10 s, respectively. The Hargreaves Test Mice were habituated in plastic chambers, and radiant light was applied to one of their hindpaws. The radiant light was applied when the mice were resting quietly and was stopped immediately after the movement of the hindpaw. In the inflammatory pain model, 20 mL of CFA was intraplantarly injected into one hindpaw of each mouse, and we tested the noxious heat responses using Hargreaves test when applied radiant heat to the hindpaw with CFA injection.

Neuron 93, 1–16.e1–e9, February 22, 2017 e5

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

The Hot Plate Test Mice were put on a hot plate at a temperature of 50
C, 52
C or 55
C, and the cutoff time was 70 s, 50 s or 30 s, respectively. The von Frey Test The mice were habituated in plastic chambers on a mesh floor. The von Frey filaments with forces of increasing grades were applied to the mice hind paw. One filament was applied 5 times in a round of testing. The filament force evoking paw withdrawals for more than 3 times in a round of testing was defined as the mechanical threshold. The cutoff force was 4 g. To prepare the neuropathic pain model induced by spared nerve injury (SNI), we transected the common peroneal and tibial branches of the right sciatic nerve with 1 mm of nerve removed and left the sural nerve intact. The von Frey test was performed on the lateral part of the right plantar surface where the sural nerve innervates. The Randall-Selitto Test A gradually increasing pressure was applied to the mouse tail. The pressure force causing a tail withdrawal was defined as the mechanical threshold. The cutoff force was 200 g. The Evaporative Cooling Test The mice were habituated in a plastic chamber on a mesh floor. A drop (10–20 mL) of acetone was applied to the mice hind paw with 1 mL syringe. The duration of the flinching or licking behavior within 1 min was measured. The Rotarod Test The mice were tested on a rotarod with the velocity increasing from 4 rpm to 40 rpm within 5 min. The mice were pre-trained for 2 d for adaptation. Then the duration time on the rotarod before the mice fell off was recorded. The Open Field Test The mice were put in an open field (45 3 45 cm) and their locomotor activities within 30 min were recorded and analyzed. The total moving distance and the moving velocity were measured for the evaluation of the locomotor activity. The Two-Temperature Choice Test In the two-temperature choice test, two cold/hot plates (20 3 20 cm each) were joined together. The plates were divided into two 40 3 10 cm lanes in which the mice were free to move. The behavior of mice was recorded for 10 min in each round of test, and the time they spent on each plate was analyzed. The Formalin Test The 2% formalin diluted in saline was injected into the plantar skin of left hindpaw of the mouse. The spontaneous nociceptive responses were recorded and analyzed. The duration of licking or biting behavior was counted for 1 hr in 5 min interval. The first phase of nociceptive responses was analyzed during 0–10 min, and the second phase of nociceptive responses was analyzed during 10–60 min. Infusion of Nav1.7CT-TAT and FGF13B-TAT GST-flag-Nav1.7CT-TAT (Nav1.7CT-TAT) was used for disruption of the FGF13/Nav1.7 interaction. Nav1.7CT-TAT was dissolved in elution buffer (50 mM Tris-HCl, pH 8.0) and was injected intraperitoneally for 4 times (one injection an hour) before the behavioral tests. Co-IP was used to test whether Nav1.7CT-TAT was able to disrupt the FGF13/Nav1.7 interaction. The tissues were collected 4–5 hr after the injections when Nav1.7CT-TAT was most effective in the behavioral tests. To test the rescue effects of GST-FGF13BTAT on the behavioral deficits, we dissolved GST-FGF13B-TAT in the elution buffer and injected it intrathecally (i.t.) into Fgf13/Y mice. Then the behavioral tests were performed at the set time points. We performed the co-IP experiments with mouse DRG or spinal cord tissues collected at 20 min, 1 hr and 2.5 hr after GST-FGF13B-TAT injection (i.t.) to detect whether the rescue effects in the behavioral tests were correlated with the interaction of exogenous FGF13B and Nav1.7. These time points (20 min, 1 hr and 2.5 h) corresponded to the most effective stage and the decaying stage of the GST-FGF13B-TAT effects in the behavioral tests. The vehicle (elution buffer) did not cause apparent changes in nociceptive behaviors when injected either intraperitoneally or intrathecally. Preparation of an FGF13 Antibody A peptide mapping at the C terminus of FGF13B (182-192 aa) was prepared and chemically conjugated with the immunogenic carrier protein keyhole limpet hemocyanin (KLH). The conjugated peptide was mixed with Freund’s complete adjuvant and used to immunize New Zealand rabbits. The immunizations were repeated 3 weeks and 5 weeks later. The sera were taken and stored at 70
C. The antibody efficiency was validated in HEK293 cells transfected with FGF13 plasmids and the DRG tissue samples. The antibody specificity was verified by a pre-absorption assay. Electrophysiological Recording All chemicals used in the electrophysiological experiments were purchased from Sigma-Aldrich. The voltage and current electrophysiological data were analyzed using Clampfit 10.3 (Axon instruments), Origin pro 8.6.0 (Microcal) and PRISM (GraphPad Software). Voltage-Clamp Recording For sodium currents recording, small DRG neurons cultured for 12–24 hr or HEK293 cells transfected with the indicated plasmids were used for voltage-clamp recordings. Patch pipettes with 1–2 MU resistance were prepared using a puller (P-97 Sutter Instrument). The pipette solution contained (in mM): 140 CsF, 10 NaCl, 1 EGTA, 10 dextrose and 10 HEPES, pH 7.3 (adjusted with

e6 Neuron 93, 1–16.e1–e9, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

CsOH). The bath solution for DRG neurons was (in mM): 30 NaCl, 110 choline chloride, 3 KCl, 1 MgCl2, 1 CaCl2, 20 TEACl, 5 CsCl, 0.1 CdCl2 and 10 HEPES, pH 7.3 (adjusted with NaOH). The bath solution for HEK293 cells was (in mM): 140 NaCl, 3 KCl, 1 MgCl2, 1 CaCl2 and 10 HEPES, pH 7.3 (adjusted with NaOH). The whole-cell configuration was obtained in voltage-clamp mode using an Axopatch 700B amplifier (Molecular Devices). Data were acquired via a Digidata 1440 A/D converter (Molecular Devices) at 50 kHz and filtered at 10 kHz. Voltage errors were minimized with 80%–90% series resistance compensation, and linear leak currents and capacitance artifacts were subtracted out using the P/6 method. In both HEK293 cells transfected with Nav1.7 and DRG neurons, sodium currents were evoked by 100-ms pulses from a holding potential of 100 mV to a range of depolarizing steps (90 to +50 mV in 5-mV increments). For DRG neurons, we used 1 mM TTX in the bath solution to isolate the TTX-S currents. The TTX-S currents were obtained by the subtraction of the sodium currents obtained before and after TTX treatment. DRG neurons were held at 100 mV to recover the majority of TTX-S channels from inactivation. DRG neurons with Nav1.9 currents were excluded from the data analysis. The peak current induced by each depolarizing potential was plotted into I-V curves. Activation curves were obtained by calculating conductance (G) at each voltage (V) using the equation G = I/(V – Vrev) with Vrev being the calculated reversal potential. Activation curves were fitted with the Boltzmann distribution equation as follows: G = ðGmax =1 + eðV1=2;act V=kÞ Þ, where Gmax is the maximal sodium conductance, V1/2,act is the potential of half-maximal activation, V is the test potential and k is the slope factor. Steady-state fast inactivation was assessed with a series of 500-ms prepulses that started from the holding potential of 100 mV, ranged from 130 mV to 10 mV in 5-mV increments, and were followed by a 100-ms depolarization to 10 mV. The peak inward currents were normalized to the maximal peak current (Imax). The normalized curves were fitted with a Boltzmann distribution equation: I=Imax = ð1=1 + eðVm V1=2;inact =kÞ Þ, where Vm is the preconditioning pulse potential and V1/2, inact is the potential of half-maximal inactivation. The sodium currents of DRG neurons in innocuous temperature and noxious heat condition were recorded. The sodium currents were evoked by 100-ms pulses from a holding potential of 100 mV to a range of depolarizing steps (90 to +50 mV in 10-mV increments). The recording started when the bath solution had been adjusted to the set temperatures. For the recording of capsaicin-induced currents, dissociated small DRG neurons were held at 70 mV. Pipettes solution contained (in mM): 135 K gluconate, 0.5 CaCl2, 2 MgCl2, 5 KCl, 5 EGTA and 5 HEPES (pH 7.4). The bath solution contained (in mM): 150 NaCl, 5 KCl, 1 MgCl2, 2.5 CaCl2 and 10 HEPES, 10 D-glucose (pH 7.3). Inward currents induced by 1 mM capsaicin for 30 s were recorded. Current-Clamp Recording Pipettes (2–3 MU) were filled with solution contained (in mM) 140 KCl, 0.5 EGTA, 5 HEPES and 3 Mg-ATP, adjusted to pH 7.3 with KOH. The bath solution contained (in mM): 140 NaCl, 3 KCl, 2 MgCl2, 2 CaCl2 and 10 HEPES, pH 7.3, adjusted with NaOH. A wholecell configuration was obtained at room temperature in voltage-clamp mode with a holding potential of –70 mV and then the recording was performed after switching to current-clamp mode. Only the cells with stable resting potentials below 40 mV were used in this study. The data were filtered at 5 kHz and digitized at 20 kHz. APs induced by heat stimulation were recorded in current-clamp mode at temperatures ranging from room temperature (22
C– 25
C) to noxious heat levels (45
C–49
C). If the increasing temperature evoked AP firing in a small DRG neuron, the neuron was considered a heat-sensitive neuron. As duration of heat exposure might also affect AP firing, we tried to differentiate whether the decreased AP amplitude was caused by noxious heat stimulation or the duration of heat exposure. We held the temperature between 42
C and 43
C, but not above 43
C. We found the AP amplitude barely decreased after 10 s heat (42
C–43
C) exposure (Fgf13F/Y 97.5% ± 0.9%, n = 23; Fgf13/Y 94.7% ± 1.7%, n = 12) and slightly decreased after 20 s heat (42
C–43
C) exposure (Fgf13F/Y 87.5% ± 2.7%, n = 23; Fgf13/Y 84.8% ± 4.6%, n = 12). In comparison, when treated with noxious heat (> 43
C), most Fgf13/Y neurons exhibited dramatic AP amplitude decrease within 10 s treatment. So the decreased AP amplitude was mainly caused by noxious heat stimulation but not extending heat exposure. In order to test the rescue effects of GST-FGF13B-TAT, we treated DRG neurons from Fgf13/Y mice with 1 mM GST-FGF13B-TAT or GST-TAT for 2 hr. To disrupt the endogenous FGF13/Nav1.7 interaction, we incubated DRG neurons from C57BL/6J mice with 1 mM GST-flag-Nav1.7CT-TAT or GST-flag-TAT for 30 min. For each heat-sensitive DRG neuron, we counted the number of AP spikes within one degree as the temperature increased. The analysis of AP spike numbers revealed the AP distribution within the temperature ranging from 38
C to 45
C. We depicted the change of AP amplitudes using a fitting curve of the AP amplitudes from 38
C to 45
C. The change of AP amplitudes in noxious heat stimulation was also quantified by the ratio of the AP amplitude at 45
C to that of the first AP. We also recorded the APs induced by 50-ms current injections starting from 50 pA in 5-pA increments before and after heat stimulation to make sure that the neurons were healthy and did not lose responsiveness. Temperature Control The temperature of perfusion solution was controlled by a TC-344B (Warner Instruments). The temperature was increased gradually from room temperature to the maximal temperature in 80–120 s. Recoding of Spinal Cord sEPSC We recorded the spontaneous excitatory postsynaptic currents (sEPSCs) in spinal cord laminae II. Vibratome sections of the mice lumbar spinal cord were perfused with oxygenated Krebs’ solution (in mM: 117 NaCl, 3.6 KCl, 2.5CaCl2, 1.2 MgCl2, 1.2 NaH2PO4, 25 NaHCO3 and 11 D-glucose). The pipette solution contained (in mM) 135 K-gluconate, 0.5 CaCl2, 2 MgCl2, 5 KCl, 5 EGTA, 5 HEPES and 5 D-glucose. Each neuron was recorded by a whole-cell patch-clamp for 15 min. The frequency and amplitude of sEPSC were analyzed by Axograph (Molecular Devices).

Neuron 93, 1–16.e1–e9, February 22, 2017 e7

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Dual Fluorescent In Situ Hybridization The probes for the detection of mouse FGF13 and Nav1.7 mRNAs were amplified from the mouse cDNA by PCR. The primers producing mouse Nav1.7 probe were 50 - GAGGACCTGGACCCATAC-30 and 50 -TGAGTCATTAGCCGAAAC-30 . The primers producing mouse FGF13 probe were 50 -TCGCTCATCCGGCAAAAGAG-30 and 50 -GGTTCTGTTATAGAGCCCTC-30 . The primers producing mouse Nav1.8 probe were 50 -AATCTGGGGAGTTGGATTCTCT-30 and 50 -ATGTGTCTCTGGAAGTCCTGGT-30 . For the dual labeling of FGF13 and Nav1.7, the probe for FGF13 mRNA was labeled with digoxigenin (DIG) and the probe for Nav1.7 mRNA was labeled with fluorescein (FITC). For the dual labeling of Nav1.7 and Nav1.8, the probe for Nav1.7 mRNA was labeled with digoxigenin (DIG) and the probe for Nav1.8 mRNA was labeled with fluorescein (FITC). The cryostat sections of the fresh DRG tissue were fixed with 4% PFA in Diethy Pyrocarbonate (DEPC)-PBS for 20 min, then acetylated and prehybridized in hybridization buffer at 67
C for 3 hr. Then, the sections were incubated in the hybridization buffer containing the antisense probes (1 mg/mL of each probe) at 67
C for 16 hr. After hybridization, the sections were washed, blocked and incubated with anti-FITC- horseradish peroxidase (HRP) (1:4000, Roche) at 4
C overnight and then treated with tyramide signal amplification (TSA)-plus 2,4-dinitrophenyl (DNP) reagents (1:100, PerkinElmer) for 10 min. Then the sections were incubated with anti-digoxigenin alkaline phosphatase (anti-DIG-AP) (1:1000, Roche) and anti-dinitrophenyl Alexa488 (anti-DNP Alexa488) (1:500, Molecular Probes) in the blocking buffer at 4
C overnight. The sections were washed and then developed with 2-hydroxy-3-naphtoic acid-20 -phenylanilide phosphate/Fast Red (HNPP/FR) (1:100, Roche). fMRI Experiment For fMRI scans, mice were sedated and mechanically ventilated using light isoflurane anesthesia (0.7%–1%) delivered with O2:N2 70:30. Each anesthetized mouse was placed in a custom-designed magnetic resonance cradle with the head secured by ear and bite bars. The cradle contained a circulating warm system for the maintenance of the core body temperature at 37.5
C–38.5
C. A pressure sensor (SA Instruments Inc) was positioned under the abdomen of the mice to monitor respiration (maintained at 80–100 breaths/min). An advanced thermal stimulator (ATS) thermode (Medoc, ramp rate 8
C/s) was positioned on the left hind paw to deliver noxious heat stimuli. To map the neural responses, we ran nine alternations of baseline temperature (32
C, 30 s in duration) and noxious heat temperature (47.5
C, 21 s in duration) in each run. Typically, 2–3 runs were collected and each run consisted of 228 image volumes. The thermode remained in contact with the skin during the entire imaging session. All fMRI scans were performed in a 7-Tesla small animal scanner with a 30-cm diameter bore (Bruker Biospec 7030 USR). A 4-element phased array coil was used to receive the radio frequency signal and a linear volume coil was used to transmit the radio frequency pulses. The T2-weighted (T2W) structural images were obtained using a rapid acquisition with refocused echoes (RARE) sequence [TR = 3000 ms, TE = 8.5 ms, matrix size = 192 3 128 and field of view (FOV) = 24 mm 3 16 mm]. Prior to functional image scans, a global shimming calculation of the entire brain was applied to improve the magnetic field homogeneity (MAPSHIM, Bruker). Twelve coronal functional echo planar imaging (EPI) images (in plane resolution of 0.25 3 0.25 mm2 and 0.5 mm slice thickness) were acquired using gradient-echo echo-planar imaging (GE-EPI) sequences (FOV = 24 3 16 mm, matrix size = 96 3 64, TR = 2500 ms, TE = 13.5 ms and bandwidth 200 kHz). Fifteen sets of T2W structural images were randomly selected from the experimental groups, realigned with each other and then averaged to generate a set of reference images. For each animal, the T2W images were co-registered to the reference images using a linear image registration tool (FLIRT, FSL) for group analysis. Using the hemodynamic response function (HRF) convolved with the boxcar function of the stimulation protocol (3dDeconvolve, AFNI) as a predictor, a voxel-wise correlation analysis of fMRI signals was performed for each group (Fgf13F/Y and Fgf13/Y). Group activation maps were displayed as statistical t-value maps with warm- and cold-color indicating the increase and decrease of blood oxygen level dependent (BOLD) signals, respectively. Student’s t test (3dttest++, AFNI) was used to quantify the statistical differences between the group activation maps. EPI voxels (with a minimum cluster size of 5 contiguous voxels) that exhibited significant signal differences between the two groups (p < 0.01, with a corrected false discovery rate) were included in the difference maps. Thresholded group activation maps and the difference maps were superimposed on T2-weighted structural images for display. To quantify the fMRI BOLD response magnitude to noxious heat stimuli in each group (Fgf13F/Y and Fgf13/Y mice), we extracted BOLD signal time courses from the voxels in each region of interest (ROI). A mouse atlas and the statistical activation difference map were used to identify the ROIs. The BOLD signal changes (in percentage) in each ROI before and after the stimulation were calculated and averaged across the mice. The changes were compared between the two groups of mice (Fgf13F/Y and Fgf13/Y mice). Single-Cell RNA Sequencing The detailed procedures were described previously (Li et al., 2016). DRG neurons were digested, triturated and plated on coverslips. Then the neurons were randomly selected, picked by a glass pipette and lysed. The RNA of a single neuron was extracted for reverse transcription. The cDNA of each neuron was amplified for the construction of sequencing libraries. The cDNA libraries were used for high-coverage single-cell RNA sequencing. DRG neurons were classified into ten clusters by unsupervised hierarchical clustering analysis. The gene expression pattern in each cluster was analyzed on basis of the sequencing data.

e8 Neuron 93, 1–16.e1–e9, February 22, 2017

Please cite this article in press as: Yang et al., FGF13 Selectively Regulates Heat Nociception by Interacting with Nav1.7, Neuron (2017), http:// dx.doi.org/10.1016/j.neuron.2017.01.009

Semiquantitative PCR The primer sequences are provided in Table S2. The mRNAs were reverse-transcribed with SuperScript II reverse transcriptase (Thermo Fisher Scientific). PCR was performed according to standard methods. As too many PCR cycles will cover mild expression changes, we used less cycle numbers which were adjusted according to the abundance of each gene. Microarray We used L4 and L5 DRGs of F/Y and /Y mice (3 mice each group) in the microarray assay to investigate their gene expression. Total RNA of DRG was extracted using TRIZOL Reagent (Life technologies) and purified by RNeasy mini kit (QIAGEN) and RNase-Free DNase Set (QIAGEN). Total RNA was amplified and labeled by Low Input Amp Labeling Kit, One-color (Agilent technologies). Agilent SurePrint G3 Mouse Gene Expression Microarray 8 3 60K was used for hybridization. Each slide was hybridized with 600 ng Cy3labeled cRNA using Gene Expression Hybridization Kit (Agilent technologies) in Hybridization Oven. After 17-hr hybridization, slides were washed in staining dishes (Thermo Shandon) with Gene Expression Wash Buffer Kit (Agilent technologies). Slides were scanned by Agilent Microarray Scanner (Agilent technologies) with default settings, Dye channel: Green, Scan resolution = 3 mm, PMT 100%, 20 bit. Data were extracted with Feature Extraction software 10.7 (Agilent technologies). Raw data were normalized by Quantile algorithm, limma packages in R. QUANTIFICATION AND STATISTICAL ANALYSIS The data are presented as mean ± SEM. Sample number (n) values are indicated in figures, figure legends or results section. Two groups were compared by a two-tailed, unpaired Student’s t test. Comparisons between two groups with multiple time or voltage points were performed by a two-way ANOVA with Bonferroni’s post hoc test. In electrophysiological experiments, multiple groups were compared by a one-way ANOVA with Tukey’s post hoc test. In fMRI experiments, differences in the peak BOLD signal amplitude between hemispheres or groups were examined using an unpaired t test. Statistical analysis was performed using PRISM (GraphPad Software). Differences were considered significant at p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001 and N.S. not significant). DATA AND SOFTWARE AVAILABILITY Single RNA-sequencing data can be acquired from Gene Expression Omnibus, GEO: GSE63576.

Neuron 93, 1–16.e1–e9, February 22, 2017 e9