Research ArticlePain

Inhibition of the kinase WNK1/HSN2 ameliorates neuropathic pain by restoring GABA inhibition

See allHide authors and affiliations

Science Signaling  29 Mar 2016:
Vol. 9, Issue 421, pp. ra32
DOI: 10.1126/scisignal.aad0163

“WNK”ing out pain

Mutations in the HSN2 exon present in the nervous system–specific isoform of the kinase WNK1 cause an ulcerating neuropathy disorder called hereditary sensory and autonomic neuropathy type IIA (HSANII). HSANII affects the peripheral and spinal nerves and results in loss of touch, temperature, and pain sensation. Kahle et al. generated transgenic mice specifically lacking this alternatively spliced variant of WNK1, which is present in the spinal cord’s dorsal horn, the gateway for pain processing from the periphery to the brainstem. These mice exhibited no gross neurological defects and did not exhibit symptoms of HSANII, likely because mutations in HSANII patients generate a truncated form of the kinase that has an intact catalytic domain. The HSN2-deficient mice were protected from pain hypersensitivity in a model of neuropathic pain resulting from peripheral nerve injury, but not in an inflammatory pain model. Mechanistically, HSN2-deficient mice had less phosphorylation of the K+-Cl cotransporter KCC2 in the nerves, which resulted in an increase in KCC2 activity, a decrease in the amount of Cl in the nerves, and restoration of the inhibitory response to GABA. Thus, by alleviating GABA “disinhibition,” a known major contributor to neuropathic pain, drugs that inhibit HSN2 might reduce injury-induced neuropathic pain.


HSN2 is a nervous system predominant exon of the gene encoding the kinase WNK1 and is mutated in an autosomal recessive, inherited form of congenital pain insensitivity. The HSN2-containing splice variant is referred to as WNK1/HSN2. We created a knockout mouse specifically lacking the Hsn2 exon of Wnk1. Although these mice had normal spinal neuron and peripheral sensory neuron morphology and distribution, the mice were less susceptible to hypersensitivity to cold and mechanical stimuli after peripheral nerve injury. In contrast, thermal and mechanical nociceptive responses were similar to control mice in an inflammation-induced pain model. In the nerve injury model of neuropathic pain, WNK1/HSN2 contributed to a maladaptive decrease in the activity of the K+-Cl cotransporter KCC2 by increasing its inhibitory phosphorylation at Thr906 and Thr1007, resulting in an associated loss of GABA (γ-aminobutyric acid)–mediated inhibition of spinal pain-transmitting nerves. Electrophysiological analysis showed that WNK1/HSN2 shifted the concentration of Cl such that GABA signaling resulted in a less hyperpolarized state (increased neuronal activity) rather than a more hyperpolarized state (decreased neuronal activity) in mouse spinal nerves. Pharmacologically antagonizing WNK activity reduced cold allodynia and mechanical hyperalgesia, decreased KCC2 Thr906 and Thr1007 phosphorylation, and restored GABA-mediated inhibition (hyperpolarization) of injured spinal cord lamina II neurons. These data provide mechanistic insight into, and a compelling therapeutic target for treating, neuropathic pain after nerve injury.


Chronic neuropathic pain after peripheral nerve injury is common, debilitating, often resistant to treatment, and an economic burden on society (1). Novel therapeutic strategies and druggable targets based on disease pathogenesis are needed. One contributor to neuropathic pain after nerve injury is the central disinhibition of GABAA (γ-aminobutyric acid type A) receptor–dependent spinal nociceptive pathways (2) that can result from loss of function of the K+-Cl cotransporter KCC2 (encoded by SLC12A5), which leads to Cl accumulation in dorsal horn (DH) postsynaptic neurons (3). A compelling potential analgesic strategy is to restore ionotropic GABA-mediated inhibition by enhancing KCC2 activity in DH neurons (4). However, the specific mechanisms responsible for the injury-induced impairment of KCC2-dependent Cl extrusion are poorly understood, and pharmacologic KCC2 activators are only beginning to be discovered (4).

The study of rare monogenic diseases can identify key regulatory genes in complex pathways, and these genes may also be relevant to more common forms of diseases with similar symptoms. Hereditary sensory and autonomic neuropathy type IIA [HSANII; OMIM (Online Mendelian Inheritance in Man) #201300] is an autosomal recessive ulcero-mutilating neuropathy characterized by the progressive reduction of pain, temperature, and touch sensation, and neurodegeneration in DH and dorsal root ganglia (DRG) neurons of the spinal cord and neurodegeneration of peripheral nerves (5). PRKWNK1, encoding the kinase WNK1, is alternatively spliced with one isoform containing a nervous system–specific exon called HSN2 (6). Mutations in the HSN2 exon cause HSANII (6). Intronic deletions in PRKWNK1, which cause overexpression of a WNK1 isoform lacking the HSN2-specific exon in the kidney, result in pseudohypoaldosteronism type 2C (PHA2C; OMIM #614492) (7), an autosomal dominant form of Cl-sensitive hypertension resulting from WNK1-dependent constitutive phosphorylation and activation of the Na+-Cl cotransporter NCC, a renal-specific cation-Cl cotransporter (CCC) related to KCC2 (8).

Transcripts encoding the HSN2 variant of WNK1 (referred to hereafter as WNK1/HSN2) are highly expressed in the DH, DRG, and peripheral nerves (9), but the normal function, downstream targets, and pathogenic mechanism by which mutations in WNK1/HSN2 cause disease are unknown. WNK1-dependent inhibitory phosphorylation of KCC2 maintains the depolarizing action of GABA in the developing mouse brain (10). In immature neurons, WNK1 inhibition triggers a hyperpolarizing shift in GABA activity by reducing KCC2 Thr906 and Thr1007 phosphorylation and enhancing KCC2-mediated Cl extrusion. However, whether WNK1/HSN2 regulates KCC2 in the spinal cord is unknown. To begin to elucidate the role of WNK1/HSN2 in vivo, we generated a Wnk1/Hsn2 knockout mouse and investigated the development of neuropathic pain in the spared nerve injury (SNI) pain model and the complete Freund’s adjuvant (CFA) model of inflammatory pain.


Wnk1ΔHsn2Hsn2 mice have normal spinal and sensory nerve morphology and no gross neurological deficits

We used Cre recombinase technology to generate a knockout mouse model of the Wnk1/Hsn2 isoform by specifically targeting the Hsn2 exon (fig. S1A). Homozygous animals with the Wnk1/Hsn2-floxed allele (Wnk1flox/flox) were crossed to the pCX-NLS-Cre general deletor strain, and the resulting heterozygote animals (Wnk1ΔHsn2/+) were crossed together to generate both mutant (Wnk1ΔHsn2Hsn2) and wild-type (Wnk1+/+) animals. This resulted in specific depletion of the Wnk1/Hsn2 isoform, as revealed by the lack of Wnk1/Hsn2 transcripts in Wnk1ΔHsn2Hsn2 cerebrum, cerebellum, spinal cord, and liver (Fig. 1A). Wnk1ΔHsn2Hsn2 mice exhibited no gross anatomical abnormalities, including ulcerative mutilations in either upper or lower limbs, after up to 80 weeks of observation. Histological examination of small and large nerve fibers in lumbar 4 (L4) dorsal and ventral spinal roots and of sural sensory nerves revealed normal axonal distribution and morphology in Wnk1ΔHsn2Hsn2 mice (Fig. 1B).

Fig. 1 Wnk1/Hsn2 knockout in mice results in a mild sex-dependent loss of distal thermal sensitivity and reduces chronic pain hypersensitivity after peripheral nerve injury.

(A) Genomic analysis confirming that HSN2 is absent. Genomic analysis was performed with samples from 3-month-old mice by reverse transcription polymerase chain reaction (RT-PCR) with primers specific for between exon 8 and Hsn2. Br, brain; Cer, cerebellum; SC, spinal cord; Li, liver. (B) Axonal distribution and morphology of nerves in Wnk1ΔHsn2Hsn2 mice. (Left) Histological transverse sections of ventral and dorsal spinal roots of L4 as well as sural nerves of 11-month-old Wnk1+/+ and Wnk1ΔHsn2Hsn2 mice are shown (scale, 10 mm). (Right) Quantification of nerve diameter and distribution of nerve fibers in Wnk1ΔHsn2Hsn2 mice and Wnk1+/+ mice are shown. The Kruskal-Wallis H test was performed on three to six mice of each genotype. Error bars represent SEM. See table S1 for statistical analysis. No statistically significant differences were detected. (C) Response of Wnk1ΔHsn2Hsn2 males and females to heat. Latency of females and males for tail withdrawal from either 47° or 49°C water is shown. Tail-withdrawal test was performed on 8 Wnk1+/+ and 7 Wnk1ΔHsn2Hsn2 males and 8 Wnk1+/+ and 10 Wnk1ΔHsn2Hsn2 females. Error bars represent mean ± SEM, **P < 0.05, *P < 0.01, two-way analysis of variance (ANOVA) with Bonferroni post hoc testing. (D) Response of Wnk1ΔHsn2Hsn2 mice to cold and mechanical sensation. (Left) Response to acetone evaporation as a measure of cold allodynia. Duration of paw flicking response to evaporation of acetone was tested on seven to eight mice per group (Wnk1ΔHsn2Hsn2, Wnk1+/+). (Right) Response to von Frey stimulation as a measure of mechanical hypersensitivity. Withdrawal threshold to von Frey hairs of varying diameters that apply a range of forces to the paw (measured in grams) was tested on seven to eight mice per group (Wnk1ΔHsn2Hsn2, Wnk1+/+). Mixed-model ANOVA with Bonferroni post hoc testing was performed on test results with between-factor variable being genotype and within-factor variable being time (days). Error bars represent mean ± SEM. *P < 0.01. BL, baseline.

Wnk1ΔHsn2Hsn2 mice performed similarly to Wnk1+/+ mice in the SHIRPA protocol (a standardized, three-stage battery of tests of general behavioral and neurological function) (11), the elevated plus maze and light-dark box test (assessing fear of new places and anxiety), and the hole-board test (assessing spontaneous motor activity and exploration) (Table 1). To monitor nociceptive responses, we measured sensitivity to hot and cold (using multiple assays), tactile sensitivity to pressure, and sensitivity to chemical irritants or inflammatory stimuli. Wnk1ΔHsn2Hsn2 male mice exhibited a significantly increased latency to tail withdrawal at both 47° and 49°C (Fig. 1C), but otherwise, Wnk1ΔHsn2Hsn2 mice exhibited similar response thresholds compared to Wnk1+/+ mice in tests assessing noxious thermal and mechanical sensitivity (Table 2). Global deletion of Wnk1/Hsn2 in mice, therefore, did not phenocopy the sensory neuropathy seen in HSANII patients, which is due to truncating mutations that leave the N-terminal kinase domain intact (6). Indeed, Wnk1/Hsn2 knockout produced no significant neurological effects in these assays other of a mild, male-specific reduction in distal thermal sensitivity.

Table 1 Performance of Wnk1+/+ and Wnk1ΔHsn2Hsn2 mice in assays testing general central nervous system function.

Four of each genotype were assessed in each test.

View this table:
Table 2 Nociception test summary for Wnk1+/+ and Wnk1ΔHsn2Hsn2 mice.

Statistical results of nociception tests performed on Wnk1ΔHsn2Hsn2 and Wnk1+/+ animals for both naïve mice as well as CFA and SNI pain models are presented. CFA nociception tests were performed 24 hours after CFA injection into the hind paw. SNI nociception data from 21 days after SNI are shown. Where appropriate, the degrees of freedom of the error in the ANOVA are indicated as F(n, n).

View this table:

Wnk1/Hsn2 knockout reduces pain hypersensitivity associated with nerve injury

Given the HSANII phenotype and localization of WNK1/HSN2 (6, 9), we hypothesized that WNK1/HSN2 had a role in chronic pain hypersensitivity after nerve injury or in response to inflammation. We compared the neurobehavioral responses of Wnk1ΔHsn2Hsn2 and Wnk1+/+ mice in two different pain models. We used the SNI model of chronic neuropathic pain by ligating the tibial and common peroneal nerves (12), and the CFA model of inflammatory pain by injecting CFA into the hind paw to trigger inflammation. To assess sensitivity to cold, we measured the duration of hind paw flicking in response to the sensation induced by the application and evaporation of acetone to the foot pad; to assess sensitivity to heat, we measured the time to move the paw away from a high-energy beam of light (the Hargreaves assay); and to assess sensitivity to mechanical stimuli, we used the von Frey assay of poking the paw with a filament and recording the force required to elicit a withdrawal response. Because SNI animals do not show heat allodynia (12), we compared the sensitivity to cold in the SNI pain model, sensitivity to heat in the CFA pain model, and mechanical allodynia in both models. In the SNI pain model, Wnk1ΔHsn2Hsn2 mice at both early and delayed time points (up to 21 days) exhibited significantly attenuated responses to the evaporation of acetone and pressure measured with von Frey filaments, indicating decreased cold hyperalgesia and mechanical allodynia, respectively (Fig. 1D and Table 2). In contrast, in the CFA pain model, Wnk1ΔHsn2Hsn2 mice performed similarly to Wnk1+/+ mice in tests measuring mechanical (von Frey) and thermal (Hargreaves) nociception (Table 2). Therefore, Wnk1/Hsn2 knockout reduced pain hypersensitivity associated with nerve injury but not hypersensitivity resulting from inflammation.

Knockout of Wnk/Hsn2 prevents SNI-induced inhibitory phosphorylation of KCC2

The WNK family of kinases stimulate the downstream SPS1-related proline/alanine-rich kinase (SPAK) or the SPAK homolog oxidative stress responsive kinase-1 (OSR1), to regulate the CCCs (13), stimulating the Na+-K+-Cl cotransporter NKCC1 and inhibiting KCC2, through phosphorylation at multiple residues in a homologous motif in each transporter (NKCC1Thr202/Thr207/Thr212; KCC2Thr906/Thr1007) (14, 15). Dephosphorylation of KCC2 at these sites significantly stimulates KCC2 activity (14, 15), such that activation of the GABAA Cl-conducting channel elicits hyperpolarization even in neurons with extremely negative resting membrane potentials (16). To determine whether lack of WNK1-dependent phosphorylation of CCCs contributed to the decreased pain hypersensitivity in the Wnk1ΔHsn2Hsn2 mice, we examined the phosphorylation status of NKCC1Thr202/Thr207/Thr212 and KCC2Thr906/Thr1007 in the spinal cords from naïve and SNI-subjected mice (Fig. 2, A and B). We observed a low amount of phosphorylation of KCC2 at Thr906 and a higher amount of phosphorylation at Thr1007 in the Wnk1+/+ animals under control conditions, and the phosphorylation at these two sites in the naïve Wnk1ΔHsn2Hsn2 mice was significantly less than that of the naïve Wnk1+/+ mice (P < 0.001). SNI significantly increased phosphorylation at both sites in KCC2 in mice of both genotypes; however, the increase in the phosphorylation status in the Wnk1ΔHsn2Hsn2 mice resulted in an amount of phosphorylation that was similar to those seen in naïve Wnk1+/+ mice.

Fig. 2 Antagonizing spinal Wnk1/Hsn2 kinase signaling decreases maladaptive KCC2 inhibitory phosphorylation and normalizes depolarizing GABA-evoked responses after nerve injury.

(A) Spinal cord homogenates were subjected to Western blot analysis of either immunoprecipitates or lysates with the indicated antibodies. Proteins or phosphorylated proteins recognized by the antibodies that were used for immunoprecipitation are indicated with IP; proteins recognized by antibodies that were used for Western blotting are indicated with IB. The WNK1/HSN2 antibody is specific for this isoform. Western blots from one of two independent experiments are shown. (B) Quantification of indicated Western blot data from results shown in (A). The quantification (ratio calculation) is based on (phospho KCC2)/(total KCC2), or HSN2/ERK1 (extracellular signal–regulated kinase 1). Statistical analysis was performed on each Western blot separately and shown are the results of unpaired t test for the three samples shown in (A). Values represent mean ± SEM, ***P < 0.001, **P < 0.01, *P < 0.05. (C) Representative currents recorded in spinal cord lamina II neurons from naive and SNI-subjected (1 week) adult mice in response to 30-ms puffs of GABA (1 mM) in the presence of a Cl load (29 mM), introduced through the patch pipette. Slices were incubated for 30 to 45 min with vehicle control or STOCK 1S-50699 (10 μM) before recording. (D) Current-voltage (I-V) relationship for GABA normalized currents (I/−I80mV) obtained from naïve mice under control conditions (n = 13) or after SNI treatment (1 week) in the presence of vehicle control (n = 9) or STOCK1S-50699 (n = 14). Conditions as described in (C). (E) Apparent EGABA measured in lamina II spinal cord neurons from naïve mice under control conditions (n = 13) or after treatment with STOCK1S-50699 (n = 6). Also shown are EGABA measurements in sham-operated mice (1 week) under control conditions (n = 6) or after treatment with STOCK1S-50699 (n = 5), and in SNI (1 week) under control conditions (n = 9) or after STOCK1S-50699 treatment (n = 14). Data are presented as means ± SEM. Two-way ANOVA with factors surgery and treatment and Bonferroni post hoc testing (**P < 0.05, *P < 0.01). (F) Apparent EGABA measured in lamina II neurons from Wnk1+/+ and Wnk1ΔHsn2Hsn2 mice under control or SNI (1 week) conditions. Data are presented as means ± SEM, unpaired t test, **P < 0.05.

In the Wnk1+/+ mice, the amount of WNK1/HSN2 was also increased by SNI. In contrast, phosphorylation of NKCC1Thr202/Thr207/Thr212 and the abundance of SPAK and the related kinase OSR1 were unchanged by WNK1/HSN2 knockout in either naïve mice or after SNI (Fig. 2, A and B). We did not detect differences under any condition in the phosphorylation of SPAK at Ser373, a WNK phosphorylation target in the noncatalytic C-terminal domain with an unclear role in SPAK regulation (17), indicating that the degree of SPAK Ser373 phosphorylation is independent of WNK1/HSN2 abundance. We quantified the amount of the kinase ERK1 as a negative control because this kinase has not been implicated as a regulator of KCC2 or as a component of the WNK pathway.

Genetic knockout or pharmacologic inhibition of Wnk1/Hsn2 restores SNI-induced GABA-mediated depolarization in lamina II neurons

Impaired KCC2-dependent Cl extrusion results in depolarizing GABA-evoked responses in rat spinal nociceptive pathways after SNI (3, 4). We performed electrophysiology of lamina II neurons in spinal cord slices derived from naïve mice, sham-operated controls, or those subjected to the SNI pain model in the presence of vehicle or STOCK1S-50699 (18), a specific inhibitor of WNK signaling that activates the KCCs by reducing KCC inhibitory phosphorylation by SPAK or OSR1 (15) (Fig. 2, C to E). To detect the activity of KCC2, we delivered Cl through the patch pipette and measured the response of the neurons to GABA. If KCC2 is active, the excess Cl is extruded and GABA promotes hyperpolarization; if KCC2 activity is inhibited, the intracellular Cl remains high and the apparent reversal potential for the GABA-induced current (EGABA) shifts to a less negative value. Thus, by measuring the apparent EGABA under these Cl challenge conditions, we can indirectly monitor KCC2 activity (19). We found that EGABA in lamina II neurons from SNI mice was significantly less negative than that in lamina II neurons from naïve mice or sham-operated controls. In mice subjected to SNI, EGABA was restored to more negative values by slice incubation with STOCK1S-50699 for 30 to 45 min before recording, reflecting an increase in neuronal Cl extrusion capacity. Similarly, in lamina II neurons derived from Wnk1ΔHsn2Hsn2 mice, EGABA was not altered after SNI (Fig. 2F). STOCK1S-50699 had no significant effect on EGABA in naïve animals (Fig. 2E).


Here, we report the generation of a Wnk1/Hsn2 knockout mouse inspired by the mutation in the HSN2 exon that causes HSANII. Global deletion of Wnk1/Hsn2 in mice did not produce the phenotype of HSANII patients. Whereas patients with HSANII exhibit a progressive neuropathy characterized by the progressive reduction of pain, temperature, and touch sensation, and neuronal degeneration (6), Wnk1ΔHsn2Hsn2 mice lacked any neurological defects except for a mild male-specific reduced distal thermal sensitivity. The difference between these phenotypes may be due to the nature of the HSN2 mutation in either case. The mutations that cause HSANII result in a truncated protein that retains the N-terminal kinase domain (6), which may produce a kinase lacking critical regulatory regions and with altered substrate specificity. In contrast, Wnk1ΔHsn2Hsn2 does not produce the nervous system–predominant WNK1/HSN2 isoform but otherwise produces fully functional WNK1 lacking this HSN2 exon in other tissues. Thus, this knockout model is useful for exploring the specific functions of this isoform without compromising other WNK1 functions.

Our results suggested that WNK1/HSN2 contributes to a maladaptive decrease in KCC2 activity and loss of GABA-mediated inhibition of neuronal activity that is associated with chronic pain hypersensitivity after nerve injury. Furthermore, our data indicated that the reduction in KCC2 activity in response to nerve injury resulted from an increase in WNK1/HSN2-dependent inhibitory phosphorylation of KCC2 at Thr906 and Thr1007. Kinases are highly druggable molecules that have been successfully targeted in cancer but have been infrequently targeted for neurological diseases (20). The nervous system–specific expression of WNK1/HSN2, its importance for human physiology as evidenced by the diseases associated with its mutation and the beneficial impact of its genetic inhibition on neuropathic pain behavior without producing other major neurological phenotypes in mice, suggests that WNK1/HSN2 is a compelling therapeutic target worthy of future investigation.


Generation of floxed Hsn2 and knockout mice

The Hsn2-containing genomic regions and its neighboring exons (8b, 11) were obtained by PCR amplification of three separate fragments from total genomic DNA of 129S1/SvImJ origin (table S2). The three separate fragments were assembled using the SLIC (sequence- and ligation-independent cloning) method and cloned in a modified pDELBOY-3X vector (Addgene, plasmid 13440) (2124). The construct included 5′ and 3′ homology arms, allowing homologous recombination within the Wnk1 gene, as well as the Hsn2 exon flanked with two loxP sites. The construct also contained a phosphoglycerate kinase (PGK) promoter–driven neomycin resistance cassette flanked by two frt sites within the 5′ homology arm that was used as a positive selection marker and a PGK-driven diphtheria gene used as a negative selection marker that lay downstream of the 3′ homology arm (fig. S1A). All construct and DNA fragments were sequence-verified at each step and checked for correct orientation before electroporation in embryonic stem cells (ESCs).

The targeting vector was electroporated in R1 (+Kitl-SlJ) ESCs using a Gene Pulser Xcell electroporator (Bio-Rad). Briefly, 25 μg of vector DNA was linearized using the unique Not I restriction site and electroporated in 10 × 106 to 20 × 106 ESCs with the following conditions: ionic strength, low; pulse length, 960 μF; and voltage, 250 V at room temperature. The cells were grown in double selection media, and proper recombinant clones were first verified by PCR screening (table S2). Clones passing this first screening process were further confirmed by Southern blot analyses (fig. S1B). Using this procedure, one correctly targeted ESC clone was recovered from 300 candidates. The correctly targeted clone was microinjected into C57BL/6J mouse blastocysts from which eight resulting male chimeras from the same clone were obtained that showed germline transmission. Three resulting strains (351, 352, and 355) were backcrossed for two to three generations to mice on the C57BL/6J background (N2-3) and kept for histological and nociception characterization.

Mice were screened for carrying a loxP-flanked Hsn2 allele using PCR genotyping (see table S2 for primers). The Hsn2 exon was removed by crossing homozygote animals harboring the Wnk1/Hsn2-floxed allele (Wnk1flox/flox) to animals expressing the Cre recombinase under the activity of the chicken β-actin gene (general deletor, pCX-NLS-Cre animals). The resulting animals contained an intact allele of the Hsn2 exon and a recombined allele (Wnk1ΔHsn2/+). The Cre recombinase transgene was outbred, and experimental animals were generated by subsequent mating of two heterozygote animals for the Hsn2 recombined allele (X/Y, Wnk1ΔHsn2/+; X/X, Wnk1ΔHsn2/+) that yielded both knockout (Wnk1ΔHsn2Hsn2) and wild-type animals (Wnk1+/+). All mice were maintained on a standard light cycle of 12 hours on and 12 hours off and had food and water ad libitum. This study was conducted in compliance with the ethic committees at the Centre Hospitalier de l’Université de Montréal and McGill University and was approved by the Boston Children’s Hospital Animal Care and Use Committee. All experiments were performed in a blinded fashion in a quiet room (temperature, 22° ± 1°C) from 9:00 a.m. to 6:00 p.m.

RNA extractions and RT-PCR

RNA extractions from different neuronal and nonneuronal tissues (brain, cerebellum, spinal cord, liver, and kidney) of Wnk1ΔHsn2Hsn2 and Wnk1+/+ animals were performed using TRIzol Reagent (Invitrogen by Life technologies, catalog no. 15596-018). Reverse transcription was performed with 1 μg of total RNA per reaction mix using SuperScript III Reverse Transcriptase (Invitrogen, catalog no. 18080-044). PCR was performed with 5 μl of first-strand complementary DNA using primer pairs listed in table S2 and Taq DNA Polymerase (Qiagen). A thermocycler was used with the following parameters: a nucleic acid denaturation/reverse transcriptase inactivation step (94°C, 4 min) followed by 35 cycles of denaturation (94°C, 1 min) and annealing (55°C, 40 s) and primer extension (72°C, 50 s) followed by final extension incubation (72°C, 10 min). For each amplification, an internal control (RNA polymerase) was co-amplified.

Tissue preparation and histological analyses

Mice from the strain 352, between 10 and 13 months of age, were anesthetized with ketamine (100 mg/ml) and xylazine (20 mg/ml) before perfusion with saline solution and 3% glutaraldehyde. The dorsal and ventral spinal roots (n = 3 to 4 mice per group) from the L4 spinal section, as well as sural nerves (n = 5 to 6 mice per group), were dissected and fixed in 3% glutaraldehyde in phosphate buffer (78 mM Na2HPO4, 22 mM KH2PO4, pH 7.4) overnight. After several washes, the tissues were osmicated with 2% osmium tetroxide (Canemco), dehydrated, and embedded with Epon [TAAB 812 resin (45%), distilled dodecenyl succinic anhydride (DDSA) (35%), nadic methyl anhydride (18%), tri(dimethylaminomethyl)phenol (DMP30) (2%)] (Canemco). The tissues were cut (5 μm thick) and mounted on glass slides. Sections were stained with toluidine blue and observed under light microscopy. The number of axons was counted and classified into 1-μm groups for each type of nerve. Axon diameters were calculated at magnification ×320 from surface areas using the particle analysis function of Image J ( The total number of axons for each type of nerve was counted, averaging 600, 530, and 2100 axons for dorsal roots, ventral roots, and sural nerves, respectively.


The SHIRPA test was performed according to the method described by Rogers et al. (11). Briefly, SHIRPA is a three-stage protocol designed to identify neurologic and behavioral phenotypic abnormalities in mice. The first stage provides a general behavioral observation profile, during which the tester observes undisturbed mice for any abnormal behaviors or movements. The second stage includes multiple behavioral tests that generate quantitative phenotypic data, including tests of locomotor activity, food and water intake, balance and coordination, and analgesia, along with histological and biochemical analysis of tissues sampled from test mice. The third stage adds more detailed testing based on known or potential modes of neurologic disease, including tests of anxiety, learning and memory, prepulse inhibition, and others.

Elevated plus maze test

The elevated plus maze (Noldus) was constructed of black plastic with two “open” arms with no walls and two “closed” arms with walls (30 × 5 cm) extending out opposite from each other from a central platform (decision zone) to create a “plus” shape. The arms were raised 85 cm above the floor. At the start of the trial, a mouse was placed on the center platform of the maze, facing a closed arm, and allowed to explore the apparatus for 5 min. The maze was cleaned between subjects with a weak ethanol solution. A computer-assisted video tracking system (Ethovision XT 9.0, Noldus) recorded the total time spent exploring the different regions of the maze [open arms, center (decision zone), and closed arms]. The percent time spent in open arms (defined as when all four paws were placed on the open arm) was used as a surrogate measure of exploratory behavior in a novel environment; mice with lower amounts of open arm exploratory activity have greater neophobia (fear of new stimuli).

Light-dark box test

The light and dark box was a plastic box divided into two compartments and connected by an aperture in the divider separating the two compartments. One compartment was constructed of white plastic and was 66% of the box, and the other compartment was made of black plastic and formed 33% of the box. There was also a lid on top of the black compartment. A computer-assisted video tracking system (Ethovision XT 9.0, Noldus) recorded the total time spent exploring the different regions of the box. The percent time spent in the white side of the box (defined as when all four paws were placed in the white side of the box) was used as a surrogate measure of exploratory behavior in a novel environment; mice with lower amounts of white side exploratory activity have greater neophobia.

Hole board test

The hole-board test was used to assess mouse exploratory behavior in a novel environment. The apparatus was an acrylic box with a floor and four walls (40 cm × 40 cm × 30 cm). The plastic floor contained nine holes evenly spaced apart that were 2 cm deep. Mice were placed in the box for 15 min. Gross (for example, walking and running) and fine (for example, active grooming but not moving from one position) motor activity pertaining to location in the box and proximity to the nine holes was recorded through interruption of infrared beams located in the walls of the arena (for horizontal activity and fine movements) and in the holes to measure exploratory head dips.

Nociception tests

All behavioral experiments were performed with handlers blinded to groups. Wnk1ΔHsn2Hsn2 and Wnk1+/+ mice from three different strains backcrossed onto the C57BL/6J background were used for nociception analyses (strains 351, 352, and 355) in two different centers (McGill University and Boston Children’s Hospital). The animals were tested simultaneously over the same day, tested in multiple litters, and assessed in the same cohort of animal (n = 11 to 12 per genotype for each group). All mice were acclimated to their novel environment for 3 months before the study. All mice were 5 to 11 months old when the study began with the exception of mice undergoing SNI, which was performed at 3 months. All nociceptive assays executed in this study have been fully described in detail in the literature (2529). Tests measuring mechanical (von Frey) and thermal (Hargreaves) nociception stimuli were used in the CFA model of inflammatory pain as described (30). For neuropathic pain, mechanical allodynia (von Frey filaments test) (31) and cold hypersensitivity [evaporation of a drop of acetone (32)] were assessed before and up to 21 days after peripheral nerve injury (SNI).


SNI surgery was performed under 3% induction/2% maintenance with isoflurane on adult mice. The tibial and common peroneal branches of the sciatic nerve were tightly ligated with a 5.0 silk suture and transected distally, whereas the sural nerve was left intact as previously described (12). After injury, the incision was sutured and mice were allowed to recover on heated pads before being returned to their home cages.

Statistical analyses

Statistical data were analyzed using STATISTICA 10 (StatSoft Inc.) and SPSS. Data were initially examined using Shapiro-Wilk test for normal distribution. Data that did not fit a normal distribution underwent nonparametric analysis, whereas data that were normally distributed were subjected to parametric analysis. Axon diameter, Hargreaves plantar, hotplate Hargreaves, hot plate, von Frey, and formalin data were not normally distributed and, therefore, underwent nonparametric analyses (Kruskal-Wallis and ANOVA). Tail-clip data and tail-withdrawal data were analyzed using factorial ANOVA for genotype and sex effects. All chronic tests (SNI mechanical sensitivity and CFA mechanical and thermal sensitivity) were analyzed using mixed-model ANOVA. Post hoc tests with Bonferroni correction were performed for between- and within-subject comparisons when appropriate. All data are reported as means ± SEM.


Buffer A contained 50 mM tris-HCl (pH 7.5) and 0.1mM EGTA. Lysis buffer was 50 mM tris-HCl (pH 7.5), 1 mM EGTA, 1 mM EDTA, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 1% (w/v) Triton X-100, 0.27 M sucrose, 0.1% (v/v) 2-mercaptoethanol, and protease inhibitors (cOmplete protease inhibitor cocktail tablets, Roche; 1 tablet per 50 ml). Tris-buffered saline (TBS)–Tween (TTBS) contained tris-HCl (pH 7.5), 0.15 M NaCl, and 0.2% (v/v) Tween 20. SDS sample buffer was 1× NuPAGE LDS (lithium dodecyl sulfate) sample buffer (Invitrogen) containing 1% (v/v) 2-mercaptoethanol.


Antibodies used for Western blots included antibodies raised in sheep and affinity-purified on the appropriate antigen by the Division of Signal Transduction Therapy Unit at the University of Dundee: KCC2a total antibody (residues 1 to 119 of human KCC2a); KCC3a phospho-Thr991 [residues 975 to 989 of human KCC3a phosphorylated at Thr991, SAYTYER(T)LMMEQRSRR, corresponding to residues of rat KCC2 phosphorylated at Thr906, SAYTYEK(T)LMMEQRSRR]; KCC2a phospho-Thr906 [residues 975 to 989 of human KCC3a phosphorylated at Thr991, SAYTYER(T)LMMEQRSRR]; KCC3a phospho-Thr1039/1048 [residues 1032 to 1046 or 1041 to 1055 of human KCC3a phosphorylated at Thr1039/1048, CYQEKVHM(T)WTKDKYM, corresponding to residues of rat KCC2 phosphorylated at Thr1006, TDPEKVHLTW(T)KDKSV]; NKCC1 total antibody (residues 1 to 288 of human NKCC1); NKCC1 phospho-Thr203/Thr207/Thr212 antibody [residues 198 to 217 of human NKCC1 phosphorylated at Thr203, Thr207 and Thr212, HYYYD(T)HTN(T)YYLR(T)FGHNT]; SPAK total antibody (full-length glutathione S-transferase–tagged human SPAK protein); SPAK and OSR1 (S-motif) phospho-Ser373/Ser325 antibody [367 to 379 of human SPAK, RRVPGS(S)GHLHKT, which is similar to residues 319 to 331 of human OSR1 in which the sequence is RRVPGS(S)GRLHKT]; HSN2 total antibody (full-length human WNK1/HSN2 protein); and ERK1 total antibody (full-length human ERK1 protein). KCC2 total antibody (residues 932 to 1043 of rat KCC2) was purchased from NeuroMab. Secondary antibodies coupled to horseradish peroxidase (HRP) used for immunoblotting were obtained from Pierce. Immunoglobulin G used in control immunoprecipitation experiments was affinity-purified from preimmune serum using protein G–Sepharose.

Phosphorylation-specific antibody immunoprecipitations

KCCs phosphorylated at the KCC2 Thr906 and Thr1007 equivalent residue were immunoprecipitated from clarified spinal cord lysates. The phosphorylation-specific antibody was coupled to protein G–Sepharose at a ratio of 1 mg of antibody per 1 ml of beads. A total of 2 mg of clarified spinal cord lysate was incubated with 15 μg of antibody conjugated to 15 μl of protein G–Sepharose in the presence of 20 μg/ml of the corresponding dephosphorylated peptide dissolved in buffer. Incubation was carried for 2 hours at 4°C with gentle agitation, and the immunoprecipitates were washed three times with 1 ml of lysis buffer containing 0.15 M NaCl and twice with 1 ml of buffer A. Bound proteins were eluted with 1× LDS sample buffer.


Spinal cord lysates (15 μg) in SDS sample buffer were subjected to electrophoresis on polyacrylamide gels and transferred to nitrocellulose membranes. The membranes were incubated for 30 min with TTBS containing 5% (w/v) skimmed milk. The membranes were then immunoblotted in 5% (w/v) skimmed milk in TTBS with the indicated primary antibodies overnight at 4°C. Sheep antibodies were used at a concentration of 1 to 2 μg/ml. The incubation with phosphorylation-specific sheep antibodies was performed with the addition of 10 μg/ml of the dephosphorylated peptide antigen used to raise the antibody. The blots were then washed six times with TTBS and incubated for 1 hour at room temperature with secondary HRP-conjugated antibodies diluted 5000-fold in 5% (w/v) skimmed milk in TTBS. After repeating the washing steps, the signal was detected with the enhanced chemiluminescence reagent. Immunoblots were developed using a film automatic processor (SRX-101, Konica Minolta Medical), and films were scanned with a 600-dpi resolution on a scanner (PowerLook 1000, UMAX). Figures were generated using Photoshop/Illustrator (Adobe).

Electrophysiological recordings in spinal neurons

All electrophysiology experimental procedures were conducted in conformity with the recommendations of the European Union directive on animal experimentation (2012/63/EU). Whole-cell patch-clamp recordings were performed in 300-μm transverse slices of lumbar segments L4 to L6 as previously described (4, 33, 34). For experiments involving SNI mice, all recordings were performed in the ipsilateral side of the nerve injury. Recordings were performed at room temperature in standard artificial cerebrospinal fluid. Data were filtered at 10 kHz and acquired using pClamp 10 (Molecular Devices). Borosilicate patch pipettes (4 to 6 megohms) were filled with 120 mM K-methanesulfonate, 25 mM KCl, 10 mM Hepes, and 2 mM MgCl2 (pH 7.3). To measure EGABA under a Cl challenge, neurons in lamina II were recorded, whereas GABA (1 mM) was puffed locally for 30 ms. Experimental EGABA was extrapolated from GABA I-V curve relationship. For STOCK1S-50699 (InterBioScreen Ltd.) experiments, slices were incubated in STOCK1S-50699 (10 μM for 30 to 45 min) before recordings.


Fig. S1. Strategy for making and validation of the Wnk1/Hsn2 knockout mice.

Table S1. Nerve size quantification summary of Wnk1+/+ and Wnk1ΔHsn2Hsn2 mice.

Table S2. PCR primers used.


Acknowledgments: We thank M. Attiwell (Montreal General Hospital, McGill University, Montréal, Québec, Canada) for her help and expertise in sectioning Epon-embedded tissues for microscopy. The general deletor pCX-NLS-Cre animals were provided by the laboratory of A. Nagy (University of Toronto, Toronto, Ontario, Canada). We also thank G. Houle and M. Dona (Centre hospitalier de l’Université de Montréal Research Center) for their help in collecting and analyzing data. The general behavior phenotyping tests were performed in the IDDRC Neurodevelopmental Behavior Core, Boston Children’s Hospital (CHB IDDRC, P30 HD18655). Funding: This work was supported by the Canadian Institutes of Health Research (grant no. 179251) to G.A.R.; the Manton Center for Orphan Disease Research at Boston Children’s Hospital and Harvard Medical School, and a Harvard-MIT Basic Neuroscience Grant to K.T.K.; the Réseau de Médecine Génétique appliquée, Claude-Laberge postdoctoral fellowship to V.L.; the Fonds de la Recherche en Santé du Québec postdoctoral fellowship to V.L.; the Canadian Institutes of Health Research postdoctoral fellowship to J.-F.S.; and RO1DE022912 grant to A.L. Author contributions: K.T.K., J.-F.S., V.L., P.I., and G.A.R. conceived and designed the study. K.T.K., J.-F.S., V.L., A.L., J.Z., N.A., T.O., J.L., D.R., P.H., G.C., R.G., J.C.S.M., S.G.S., J.D., J.S.M., C.W., P.A.D., and P.I. performed the experiments. K.T.K., J.-F.S., V.L., A.L., J.Z., N.A., P.A.D., C.J.W., P.I., and G.A.R. analyzed the data. K.T.K., J.-F.S., V.L., J.Z., A.R.K., P.A.D., P.I., and G.A.R. prepared the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Wnk1ΔHsn2Hsn2 mice are available from G.A.R. with material transfer agreement with Montreal Neurological Institute and Hospital.
View Abstract

Stay Connected to Science Signaling

Navigate This Article