Mechanical allodynia but not thermal hyperalgesia is impaired in mice deficient for ERK2 in the central nervous system

Ò

PAIN 153 (2012) 2241–2252

www.elsevier.com/locate/pain

Mechanical allodynia but not thermal hyperalgesia is impaired in mice deficient for ERK2 in the central nervous system Yukiko Otsubo a, Yasushi Satoh a,⇑, Mitsuyoshi Kodama a, Yoshiyuki Araki a, Maiko Satomoto b, Eiji Sakamoto a, Gilles Pagès c, Jacques Pouysségur c, Shogo Endo d, Tomiei Kazama a a

Department of Anesthesiology, National Defense Medical College, Tokorozawa 359-8513, Japan Department of Anesthesiology, Tokyo Medical and Dental University Graduate School, Tokyo 113-8510, Japan Institute of Developmental Biology and Cancer Research, University of Nice Sophia-Antipolis, Centre National de la Recherche Scientifique, Unité mixte de Recherche 6543, Centre Antoine Lacassagne, Nice 06189, France d Aging Regulation Research Team, Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology, Tokyo 173-0015, Japan b c

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

a r t i c l e

i n f o

Article history: Received 1 December 2011 Received in revised form 29 June 2012 Accepted 19 July 2012

Keywords: ERK MEK inhibitor Neuropathic pain Allodynia Spinal cord Hyperalgesia

a b s t r a c t Extracellular signal-regulated kinase (ERK) plays critical roles in pain plasticity. However, the specific contribution of ERK2 isoforms to pain plasticity is not necessarily elucidated. Here we investigate the function of ERK2 in mouse pain models. We used the Cre-loxP system to cause a conditional, region-specific, genetic deletion of Erk2. To induce recombination in the central nervous system, Erk2-floxed mice were crossed with nestin promoter-driven cre transgenic mice. In the spinal cord of resultant Erk2 conditional knockout (CKO) mice, ERK2 expression was abrogated in neurons and astrocytes, but indistinguishable in microglia compared to controls. Although Erk2 CKO mice showed a normal baseline paw withdrawal threshold to mechanical stimuli, these mice had a reduced nociceptive response following a formalin injection to the hind paw. In a partial sciatic nerve ligation model, Erk2 CKO mice showed partially restored mechanical allodynia compared to control mice. Interestingly, thermal hyperalgesia was indistinguishable between Erk2 CKO and control mice in this model. In contrast to Erk2 CKO mice, mice with a targeted deletion of ERK1 did not exhibit prominent anomalies in these pain models. In Erk2 CKO mice, compensatory hyperphosphorylation of ERK1 was detected in the spinal cord. However, ERK1 did not appear to influence nociceptive processing because the additional inhibition of ERK1 phosphorylation using MEK (MAPK/ERK kinase) inhibitor SL327 did not produce additional changes in formalin-induced spontaneous behaviors in Erk2 CKO mice. Together, these results indicate that ERK2 plays a predominant and/or specific role in pain plasticity, while the contribution of ERK1 is limited. Ó 2012 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

1. Introduction Extracellular signal-regulated protein kinase (ERK) cascade involves the sequential activation of Raf, MAPK/ERK kinase (MEK), and ERK, and mediates intracellular signal transduction in response to extracellular stimuli. In neurons, ERK has an important role in plasticity and is implicated in learning and memory [15]. Recently, accumulating evidence has suggested that ERK has important roles in the modulation of nociceptive signaling. ERK is activated in the spinal cord dorsal horn (SCDH) in models of inflammatory pain [7,16]. Furthermore, inhibition of ERK phosphorylation in SCDH by MEK inhibitor PD98059 attenuates formalin-induced pain ⇑ Corresponding author. Address: Department of Anesthesiology, National Defense Medical College, 3-2 Namiki, Tokorozawa 359-8513, Japan. Tel.: +81 4 2995 1692; fax: +81 4 2992 1215. E-mail address: [email protected] (Y. Satoh).

behavior [16]. These results indicate that activation of ERK in SCDH contributes to pain hypersensitivity. In spinal nerve ligation (SNL) models, ERK phosphorylation can be detected in the SCDH with the induction of mechanical allodynia, and intrathecal injection of an MEK inhibitor, PD98059, reduces mechanical allodynia, suggesting that ERK in the spinal cord plays a critical role in the induction of mechanical allodynia following SNL [34]. ERK1 and ERK2 share 84% amino acid identity and have a very similar substrate profile [4]. Both are activated via phosphorylation by the upstream kinase MEK. Since MEK inhibitors and phosphospecific antibodies do not distinguish between ERK1 and ERK2, it is difficult to examine the specific contribution of each isoform to physiological functions. Thus, gene target method is required to study the role of specific isoform in pain sensitization. Recently, it was reported that Erk1 knockout (Erk1 KO) mice did not show significant differences from controls in pain sensitization [2].

0304-3959/$36.00 Ó 2012 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.pain.2012.07.020

2242

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

Despite this finding, the relative contribution of ERK2 to pain sensitization is not necessarily elucidated. Thus, to address the above concerns and to evaluate the specific contribution of ERK2, we chose to selectively ablate ERK2 using gene manipulation. Since Erk2-deficient mice are embryonically lethal, we employed a conditional [25], region-specific, genetic approach (Cre-loxP system) to target Erk2 using a nestin promoter-driven cre transgenic mouse line. In the spinal cord of the resulting conditional Erk2 mutant (Erk2 CKO [conditional knockout]) mice, ERK2 was abrogated in many spinal neurons and astrocytes, but not in microglia. Resultant Erk2 CKO mice showed a reduced nociceptive response after formalin injection to the hind paw, although this was not observed in Erk1 KO mice. We also found that Erk2 CKO mice exhibited reduced mechanical allodynia after a partial sciatic nerve ligation (PSNL) compared to control mice. These results suggest that ERK2 in the spinal cord plays a predominant role in pain plasticity, while ERK1 has a more limited role in these pain models. Interestingly, Erk2 CKO mice were indistinguishable from controls with respect to hyperalgesia in the PSNL model, suggesting a distinct mechanism for this type of pain plasticity compared to mechanical allodynia. 2. Methods 2.1. Mice All experiments were conducted according to the institutional ethical guidelines for animal experiments and the safety guidelines for gene manipulation experiments of the National Defense Medical College. These experiments were approved by the Committee for Animal Research at the National Defense Medical College. Erk2 CKO mice [26] and Erk1 KO mice [23] were generated as described previously. Both Erk1 and Erk2 mutant mice were backcrossed with C57BL/6 for more than 10 generations. All mice used in this study were age-matched male littermates. 2.2. Immunohistochemistry Immunohistochemical studies were performed as previously described [27]. Briefly, paraffin sections (5 lm thick) were deparaffinized and immersed in unmasking solution (Vector H3300; Vector Laboratories, Burlingame, CA, USA) for antigenic retrieval and heated in an autoclave (121 °C) for 5 min. The sections were then incubated in a nonspecific blocking reagent (Dako, Glostrup, Denmark) for 30 min to reduce background staining. Sections were then incubated with primary antibodies overnight in a humidified chamber at 4 °C. Primary antibodies used in this study were anti-ERK2 (mouse monoclonal, 1:1000; BD Biosciences, Franklin Lakes, NJ, USA), anti-phospho-ERK1/2 (rabbit polyclonal, 1:200; Cell Signaling Technology, Danvers, MA, USA), anti-ERK2 (rabbit polyclonal, 1:400; Cell Signaling Technology), anti-glial fibrillary acidic protein (GFAP) (mouse monoclonal, 1:50; Sigma-Aldrich, St. Louis, MO, USA), anti-neuronal nuclear antigen (NeuN) (mouse monoclonal, 1:100; Millipore, Billerica, MA, USA), and anti-OX42 (mouse monoclonal, 1:100; abcam, Cambridge, UK) antibodies. For brightfield dye staining, sections were incubated with peroxidase-conjugated secondary antibody (Dako EnVision+ system; Dako). Immunoreactivity was revealed using 3,3-diaminobenzine-tetrachloride (Vector Laboratories) according to the manufacturer’s instruction. Finally, the sections were counterstained with hematoxylin. For fluorescent staining, sections were incubated with Alexa-Fluor 488-conjugated goat anti-mouse immunoglobulin G (IgG) (1:200; Molecular Probes, Eugene, OR, USA) for primary antibodies derived from mouse. For primary antibody derived from

rabbit, Cy3-conjugated goat anti-rabbit IgG antibody (1:200; Jackson Immunoresearch, West Grove, PA, USA) was used. Sections were examined using a Nikon C2 confocal fluorescence microscopy system (Nikon, Tokyo, Japan). Samples from at least 4 mice per genotype were examined in each experiment. 2.3. Formalin-induced spontaneous behaviors Spontaneous behavioral response to formalin injection was assessed as previously described [2]. Briefly, mice were habituated in the plexiglass test box (17 cm in length  22 cm in width  14 cm in height) for 1 h. Then, a formalin solution (2%, 10 lL in 0.9% NaCl) was injected subcutaneously into the left hind paw and the mouse was returned to the test box immediately. The total time spent engaging in nociception behavior (licking and lifting of the injected paw) was recorded in blocks of 5 min for 1 h. 2.4. MEK inhibitor administration To inhibit the phosphorylation of ERK 1/2 in the formalin test, we used the MEK inhibitor a-[amino[(4-aminophenyl)thio]methylene]-2-(trifluoromethyl)benzeneacetonitrile (SL327) (Enzo Life Sciences, Farmingdale, NY, USA). SL327 was dissolved in dimethyl sulfoxide. SL327 crosses the blood-brain barrier [3] and effectively inhibits ERK1/2 phosphorylation in the spinal cord [2]. Mice were allowed to acclimate as described above, and then SL327 (50 mg/kg) was injected intraperitoneally. Thirty minutes later, formalin was injected subcutaneously into the left hind paw. Previous research showed that SL327 at a dose of 50 mg/kg maximally inhibited formalin-induced ERK1/2 phosphorylation in the spinal cord [2]. Some reports indicated that MEK inhibitors including U0126 and PD98059 blocked activation of another isoform, ERK5, in addition to ERK1 and 2 [17,18]. Although SL327 is a structural analog of U0126, this drug does not inhibit the phosphorylation of ERK5 in this condition (see Section 3). 2.5. Nerve injury pain model Mice were anesthetized and subjected to a partial sciatic nerve injury model as described previously [19]. Briefly, the sciatic nerve located in the right hind paw was exposed via a small incision and approximately 1/3-1/2 the diameter of the nerve was tightly ligated with an 8–0 silk suture. 2.6. Measurement of mechanical allodynia To quantify mechanical allodynia, paw withdrawal latency in response to mechanical stimulation was measured as previously described [2]. Briefly, mice were habituated for at least 1 h prior to testing in a clear acrylic cylinder (20 cm in height and 10 cm in diameter) on an elevated mesh floor. Calibrated von Frey filaments (North Coast Medical, Gilroy, CA, USA) were then applied through the mesh to the plantar surface on the hind paw. Filaments were applied for 1 s and this was repeated 5 times at a frequency of 0.5 Hz. Filaments were pressed until the filament bent. Paw movements associated with locomotion or weight shifting were not counted as responses. The lightest filament of 0.008 g bending force was initially applied, with progressively increasing force. Measurement was finished when paw withdrawals were evoked at least 3 withdrawals out of 5 applications, and the calibrated bending force of the filament was defined as the paw withdrawal threshold. Measurements were carried out before (Pre) and after surgery (Day 1, 4, 7, 14, and 21).

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

2243

2.7. Measurement of thermal (heat) hyperalgesia Responses to thermal stimulation were measured using a modified Hargreaves test [13]. Mice were placed in behavior chambers on a plantar test apparatus (Ugo Basile, Comerio, VA, Italy) with a glass plate floor. Before the test, mice were habituated for at least 1 h in a cylinder identical to that used in measurement of mechanical allodynia. A focused heat beam was applied to the plantar surface of the hind paw using a radiant heat light source (Ugo Basile). The intensity of the thermal stimulus was adjusted to achieve a baseline paw-withdrawal latency of approximately 5–7 s, on average, in naive mice. The paw withdrawal latency was measured using a timer that began automatically at the onset of heat application and stopped when the heat source was shut off upon paw withdrawal. Only quick hind paw movements (with or without licking of hind paws) away from the stimulus were considered to be part of a withdrawal response. Paw movements associated with locomotion or weight shifting were not counted as responses. The latency was determined as the average of 2 measurements per paw, alternating between the left and right paws with an interval of at least 5 min between measurements. Measurements were carried out before (Pre) and after surgery (Day 1, 4, 7, 14, and 21). 2.8. Western blot analysis Western blot analysis of proteins from spinal cord lumbar sections (L3–L6) was performed as previously described [25]. Primary antibodies used in this study were anti-ERK1/2 (rabbit polyclonal, 1:1000: Cell Signaling Technology), anti-phospho-ERK1/2 (rabbit polyclonal, 1:1000: Cell Signaling Technology), anti-ERK5 (rabbit polyclonal, 1:1000: Cell Signaling Technology), anti-phosphoERK5 (goat polyclonal, 1:100: Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-b-actin (mouse monoclonal, 1:2500: Sigma-Aldrich). Western blot images were captured on LAS3000 luminescent image analyzer (Fujifilm, Tokyo, Japan) and band levels were quantified using Multi Gauge Ver. 2.0 (Fujifilm). 2.9. Cell counting Phospho-ERK1/2-positive cells in the SCDH (lamina I, lamina II, and lamina III) were counted in sections from the L3–L6 levels of the lumber spinal cord. 2.10. Statistics Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software Inc, San Diego, CA, USA). Details regarding specific statistical tests are included in the results. Values are presented as the mean ± SEM. 3. Results 3.1. Erk2 CKO mice showed ablation of ERK2 in the spinal cord, but not in the dorsal root ganglion We targeted Erk2 using the Cre-loxP strategy. A floxed allele in which exon 2 and 3 of ERK2 were flanked by loxP sites (Fig. 1A) was constructed as reported previously [26]. A nestin promoterdriven cre transgenic mouse line, in which cre activity is confined to the central nervous system (CNS) [31], was used to drive recombination. The resulting nestin-cre+; Erk2flox/flox mice (also called as Erk2DCNS/DCNS mice or Erk2 CKO mice) were viable and fertile, with a normal appearance [26]. The littermate controls used in this study had the following genotypes: nestin-cre ; Erk2+/+, nestincre ; Erk2flox/flox, or nestin-cre+; Erk2+/+ (hereafter termed ‘‘control

Fig. 1. Generation of Erk2 conditional knockout (CKO) mice. (A) Schematic diagram of the targeted knockout of the mouse Erk2 gene. Exon 2 and 3 of extracellular signal-regulated kinase (ERK)2 were flanked by loxP sites (LoxP allele). The nestin promoter-driven cre transgenic mouse line was used to drive recombination, resulting in the excised allele. (B) A Western blot analysis of spinal cord using the anti-ERK and anti-phospho-ERK1/2 antibodies. b-actin was used as an internal control. (C) The expression levels of ERK1 were not altered in Erk2 CKO mice compared with controls. (D) The expression levels of ERK2 were significantly decreased in Erk2 CKO mice compared to controls. (E) ERK1 phosphorylation was significantly elevated in the Erk2 CKO mice compared to controls. (F) ERK2 phosphorylation was significantly decreased in the Erk2 CKO mice compared to controls. To evaluate the expression and phosphorylation, band levels were divided by their corresponding loading internal control (b-actin). ⁄⁄P < 0.01, ⁄⁄⁄P < 0.001 compared with control (t test, control: n = 6; Erk2 CKO: n = 6).

2244

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

mice’’). We did not detect any differences among control groups in this study and assumed that any phenotypic differences between them would be minimal. Western blot analysis showed that expression level of ERK2 was significantly decreased in Erk2 CKO mice compared to controls (Fig. 1B, D; t-test, t = 20.99, P < 0.001). On the other hand, no significant difference was observed in expression levels of ERK1 in the spinal cord between Erk2 CKO and control mice, indicating no compensatory changes in ERK1 expression (Fig. 1B, C; t-test, t = 0.37, P > 0.05). The density of the phospho-ERK2 band in the Erk2 CKO mice was smaller than in the controls (Fig. 1B, F; t-test, t = 11.86, P < 0.001). On the contrary, the density of the phospho-ERK1 band in the Erk2 CKO mice was greater than in the controls (Fig. 1B, E; t-test, t = 3.18, P < 0.01), suggesting the existence of compensatory upregulation in phosphorylation level. An immunohistochemical assessment showed that ERK2 was expressed throughout the spinal cord of control mice with the strongest immunoreactivity detected in the superficial dorsal horn (Fig. 2A–C). On the other hand, immunoreactivity for ERK2 was considerably abrogated in the spinal cord of Erk2 CKO mice (Fig. 2D, E), although some staining was detected in Erk2 CKO mice (Fig. 2D, F). In the dorsal root ganglion (DRG), we did not detect significant differences in ERK2 expression between Erk2 CKO (Fig. 2I, J) and control mice (Fig. 2G, H). We further analyzed the properties of ERK2 expression in the SCDH. ERK2 was expressed in the neuronal cells of control mice at 8 weeks of age, as indicated by double staining for ERK2 and the neuronal marker NeuN (Fig. 3A). On the other hand, ERK2 was considerably abrogated in neuronal cells of Erk2 CKO mice (Fig. 3B). At the level of the spinal cord, neuropathic pain has long been considered as caused by relevant changes in neurons. However, emerging lines of evidence have suggested that spinal astrocytes and microglia are activated as a key event in the pathogenesis of neuropathic pain [9]. Double staining for ERK2 and the astrocyte marker GFAP demonstrated expression of ERK2 in astrocytes of control mice (Fig. 3C). On the other hand, ERK2 was considerably abrogated in the astrocytes of Erk2 CKO mice (Fig. 3D). Double staining for ERK2 and the specific microglia marker OX-42

produced a slight expression of ERK2 in the microglia of Erk2 CKO mice (Fig. 3F) as well as in control mice (Fig. 3E). 3.2. ERK2 is required for second phase of the formalin test The formalin model [11] is frequently used to study inflammatory pain states in rodents. Subcutaneous injection of formalin into the hind paw induces a biphasic nociceptive response. The first, or acute, phase is caused by the direct chemical activation of nociceptive primary afferent fibers [24], which lasts < 10 min. The second, or tonic, phase takes place after a short recovery, which lasts approximately 40–60 min and is characterized by persistent shaking or licking of the injected paw. Accumulating evidence indicated that the second phase is related to the activity-dependent increase in the excitability of neurons within the SCDH, known as central sensitization, induced by nociceptive inputs during the first phase [6,10] or the development of inflammation [1]. Although the mechanism responsible for the central sensitization is not necessarily elucidated, the second phase could neither be induced nor maintained when N-methyl-D-aspartate receptor was blocked [32]. A previous report indicated that spinal ERK was phosphorylated rapidly within minutes after the subcutaneous injection of formalin into the hind paw, that is, during the first phase; the level of phosphorylated ERK peaked at 3 min and declined at 8 min [16]. Inhibition of spinal ERK1/2 phosphorylation by MEK inhibitors before the formalin injection reduced the nociceptive behaviors of the second phase without significant effect on the first phase [16]. These results imply that the behavioral response in the second phase is related to the phosphorylation of ERK during the first phase, although the behavioral response in the first phase is independent of ERK phosphorylation [16]. Figure 4A shows that Erk1 KO and control mice exhibited a typical biphasic nociceptive response to formalin injection consistent with a previous report [2]. There was no significant difference between genotypes (2-way repeated-measures analysis of variance [RM ANOVA], control vs Erk1 KO, F = 0.42, P > 0.05). On the other hand, we found that the licking/lifting behaviors in the second phase were significantly reduced in Erk2 CKO mice without a significant effect on the first phase (Fig. 4B). A 2-way RM ANOVA

Fig. 2. Extracellular signal-regulated kinase (ERK)2 was abrogated in the spinal cord dorsal horn (SCDH), but was normal in dorsal root ganglion (DRG) neurons. (A–C) Immunohistochemical staining for ERK2 in control mice shows that ERK2 was expressed throughout the spinal cord with the strongest immunoreactivity detected in the superficial dorsal horn (A). Higher power view of the dorsal horn (B) and the area around the central canal of the spinal cord (C). (D–F) Immunoreactivity for ERK2 was considerably abrogated in the spinal cord of Erk2 conditional knockout (CKO) mice although some staining was detected (D). Higher power view of the dorsal horn shows that strong immunoreactivity for ERK2, which was detected in controls, was abrogated (E). Some cells immunoreactive to ERK2 can be seen in the area around the central canal (F). (G–J) ERK2-immunoreactivity was seen in the DRG, both from control (G, H) and Erk2 CKO mice (I, J). Blue staining represents hematoxylin counter-staining for the cell nucleus. SC, spinal cord. Scale bars: 250 lm in A and D; 50 lm in B, C, E and F; 100 lm in G and I; 25 lm in H and J.

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

2245

phosho-ERK1 but not for phospho-ERK2 because the antibody cannot distinguish between phosho-ERK1 and phosoho-ERK2. We next analyzed the expression and phosphorylation levels of ERKs in the formalin test using a Western blot analysis (Fig. 5A–E). No significant altered expression level of ERK1 was observed between 0 and 5 min after the formalin injection both in control and Erk2 CKO mice (Fig. 5A, B). A 2-way ANOVA indicated no significant main effect of formalin injection (F = 0.21, P > 0.05) and genotype (F = 0.06, P > 0.05) (Fig. 5B). On the other hand, ERK2 expression was significantly reduced in Erk2 CKO mice compared to control (Fig. 5A, C). However, no significant alteration in ERK2 expression level was observed between 0 and 5 min after formalin injection both in control and Erk2 CKO mice (Fig. 5C). A 2-way ANOVA indicated no significant main effect of formalin injection (F = 0.50, P > 0.05), but significant main effect of genotype (F = 219.2, P < 0.0001) (Fig. 5C). ERK1 was significantly phosphorylated following a formalin injection both in control and Erk2 CKO mice (Fig. 5A, D). However, phosphorylation level of ERK1 was significantly increased in Erk2 CKO mice compared to control before formalin injection. A 2-way ANOVA indicated a significant main effect of formalin injection (F = 24.58, P < 0.001) and genotype (F = 7.22, P < 0.05) (Fig. 5D). ERK2 was significantly phosphorylated following a formalin injection in control mice (Fig. 5E) but not in Erk2 CKO mice. A 2-way ANOVA indicated a significant main effect of formalin injection (F = 23.19, P = 0.0001) and genotype (F = 106.0, P < 0.0001). 3.3. Pharmacological blockade of Erk1 in Erk2 CKO mice did not produce additional effects in the formalin test

Fig. 3. Extracellular signal-regulated kinase (ERK)2 was abrogated in neurons and astrocytes but not in microglia in the spinal cord dorsal horn of Erk2 conditional knockout (CKO) mice. (A, B) ERK2 was expressed in neuronal cells of control mice at 8 weeks of age, as indicated by double staining for ERK2 and the neuronal marker NeuN (A). ERK2 was considerably abrogated in neuronal cells of Erk2 CKO mice (B). (C, D) ERK2 was expressed in astrocytes of control mice, as indicated by double staining for ERK2 and the astrocyte marker glial fibrillary acidic protein (GFAP) (C). ERK2 was considerably abrogated in astrocytes of Erk2 CKO mice (D). (E, F) ERK2 was not abrogated in microglia in Erk2 CKO mice. Double staining for ERK2 and the microglia marker OX-42 demonstrated slight expression of ERK2 in microglia both from control (E) and Erk2 CKO mice (F). Scale bars: 50 lm.

confirmed this difference, indicating a main effect of genotype (control vs. Erk2 CKO, F = 9.71, P < 0.01). We next carried out an immunohistochemical study to investigate ERK phosphorylation in the spinal cord 5 min after the formalin injection. Staining for phospho-specific ERK1/2 antibody revealed increased phospho-ERK1/2-positive cells on the side ipsilateral to paw injection regardless of genotype, compared with the contralateral side (Fig. 4C). A 2-way ANOVA confirmed this difference, indicating a significant difference between the sides ipsilateral and contralateral to the formalin injection (Fig. 4D; ipsilateral vs. contralateral, F = 36.10, P < 0.001). However, the number of phospho-ERK1/2-positive cells on the ipsilateral side was significantly reduced in Erk2 CKO mice compared to controls (Fig. 4D; 2-way ANOVA, control vs. Erk2 CKO, F = 10.77, P < 0.01), although some positive cells were detected in Erk2 CKO mice (Fig. 4C, D). Note that these cells might include those positive for

To identify putative ERK2 functions that could be compensated for, we set out to examine the effect of the additional inhibition of ERK1 using MEK inhibitor in Erk2 CKO mice. Some reports indicated that MEK inhibitors including U0126 and PD98059 blocked activation of another isoform, ERK5, in addition to ERK1 and 2 [17,18]. In this study, we found that another MEK inhibitor, SL327 (50 mg/kg), did not block the phosphorylation of spinal ERK5 in the formalin test (Fig. 6A). Thus, we used SL327 (50 mg/ kg), which was administered intraperitoneally 30 min prior to formalin injection. Similar to the previous report [2], a Western blot analysis revealed that SL327 at this dose significantly attenuated phosphorylated ERK1 levels in the spinal cord compared to vehicle-injected mice 5 min after formalin injection, regardless of genotype (Fig. 6A, B; 2-way ANOVA, SL327 vs. vehicle, F = 173.4, P < 0.0001). SL327 also significantly reduced ERK2 phosphorylation levels in control mice compared to vehicle-treated mice (Fig. 6A, C; 2-way ANOVA, SL327 vs. vehicle, F = 328.2, P < 0.0001). There was no difference in ERK1 and ERK2 expression levels in the spinal cord of control mice between the vehicle-injection and SL327-injection conditions (Fig. 6A). SL327 at 50 mg/kg significantly attenuated behavioral responses of control mice in the second phase in the formalin test (Fig. 6D). A 2-way RM ANOVA confirmed this difference, indicating a significant main effect of SL327 treatment between SL327-injection and vehicle-injection (SL327 vs. vehicle, F = 4.52, P < 0.05). Furthermore, no significant difference in behavioral response was detected between vehicle-injected and SL327-injected Erk2 CKO mice (Fig. 6E; 2-way RM ANOVA, SL327 vs. vehicle, F = 0.10, P > 0.05), indicating that additional inhibition of ERK1 in Erk2 CKO mice did not produce additional effects on behavioral responses in the formalin test. Thus, ERK1 phosphorylation in Erk2 CKO mice has negligible effect on behavioral responses. Overall, these results indicate that the reduction of behavioral responses of control mice using SL327 was only due to reduction of ERK2 phosphorylation and no participation of ERK1 and ERK5 phosphorylation. Therefore, ERK2, but not ERK1, in the spinal cord

2246

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

Fig. 4. Abrogation of extracellular signal-regulated kinase (ERK)2 in the spinal cord resulted in the attenuation of formalin-induced spontaneous behaviors. (A, B) The amount of time spent engaging in spontaneous nociceptive behaviors (licking and lifting) after formalin (2%) injection is plotted in 5-min bins. Formalin-induced spontaneous nociceptive responses were preserved in Erk1 knockout (KO) mice (A). There were no significant differences between Erk1 KO (n = 12) and littermate controls (n = 12) detected over the course of a 60-min period. On the other hand, the second phase was significantly reduced in Erk2 conditional knockout (CKO) mice (n = 12) compared to littermate controls (n = 12) (B). We used a 2-way repeated-measures analysis of variance (ANOVA) with a Bonferroni post hoc test; ⁄⁄⁄P < 0.001. (C) The number of phosphoERK1/2-positive cells in the spinal cord dorsal horn (SCDH) 5 min after the formalin injection was reduced in Erk2 CKO mice compared with littermate controls. Scale bars: 500 lm. (D) There was a significant difference in the number of phospho-ERK1/2-positive cells at the L3–L6 sections of the lumbar spinal cord after a formalin injection on the side ipsilateral to paw injection. Although formalin stimulation significantly increased phospho-ERK1/2-positive cells in both control (n = 6) and Erk2 CKO littermates (n = 6) on the side ipsilateral to paw injection compared with the contralateral sides, there was a significant difference between control and Erk2 CKO mice littermates. No significant difference was detected between control and Erk2 CKO mice on the side contralateral to paw injection (2-way ANOVA with Bonferroni post hoc test; ⁄P < 0.05; ⁄⁄P < 0.01; ⁄⁄⁄ P < 0.001).

may have a predominant and/or specific role in the induction of robust spontaneous behaviors following formalin injection. 3.4. ERK2 but not ERK1 was required for mechanical allodynia in a PSNL model To examine the contribution of ERKs to persistent pain, we tested whether abrogation of ERK1 or 2 modified mechanical allodynia in the neuropathic pain model. In an experiment with

Erk1 KO mice, PSNL caused a significant increase in the response to a mechanical stimulus only on the side ipsilateral to ligation compared to the contralateral side (Fig. 7A; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 27.06, P < 0.0001), similar to control mice (Fig. 7A; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 27.58, P < 0.0001). We found no significant difference on the side ipsilateral to ligation between Erk1 KO and littermate controls (Fig. 7A; 2-way RM ANOVA, Erk1 KO vs. control, F = 0.11, P > 0.05).

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

2247

Fig. 5. Extracellular signal-regulated kinase (ERK) mice were phosphorylated after a formalin injection. (A) Western blot analysis of ERK protein expression and phosphorylation levels of spinal cord from control (n = 6) and Erk2 conditional knockout (CKO) (n = 6) mice 0 or 5 min after a formalin injection. b-actin was used as an internal control. (B, C) The expression levels of both ERK1 (B) and ERK2 (C) were not altered at 5 min after formalin injection compared to 0 min both in control and Erk2 CKO mice. (D) ERK1 was significantly phosphorylated after formalin injection both in control and Erk2 CKO mice. (E) ERK2 was also significantly phosphorylated after formalin injection in control mice. To evaluate the expression and phosphorylation, band levels were divided by their corresponding loading control (b-actin). We used a 2-way analysis of variance with a Bonferroni post hoc test; #P < 0.05; ###P < 0.001 compared with corresponding genotype at 0 min. ⁄P < 0.05; ⁄⁄P < 0.01; ⁄⁄⁄P < 0.001.

In an experiment with Erk2 CKO mice, PSNL caused a small but significant increase in the response to a mechanical stimulus on the side ipsilateral to ligation compared to the contralateral side (Fig. 7B; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 9.92, P < 0.01), similar to control mice (Fig. 7B; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 37.99, P < 0.0001). However, we found that mechanical allodynia was significantly restored in Erk2 CKO mice compared to control mice on the side ipsilateral to ligation (Fig. 7B). Two-way RM ANOVA confirmed these differences, indicating significant main effect of genotypes on the side

ipsilateral to ligation (Erk2 CKO vs. control, F = 101.1, P < 0.0001). On the side contralateral to ligation, neither control nor Erk2 CKO mice developed mechanical allodynia (Fig. 7B). Importantly, neither Erk1 KO nor Erk2 CKO mice exhibited altered baseline withdrawal responses to mechanical stimulation compared with control mice (Fig. 7A, B), indicating that baseline function remains normal in Erk1 KO and Erk2 CKO mice. Overall, these results strongly suggest that ERK2, but not ERK1, in the spinal cord may have a specific role in the induction of mechanical allodynia by PSNL.

2248

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

Fig. 6. Pharmacological blockade of Erk1 in Erk2 conditional knockout (CKO) mice did not produce additional effects in the formalin test. (A) A Western blot analysis reveals that (50 mg/kg) blocked the phosphorylation of extracellular signal-regulated kinase (ERK)1 and ERK2 in the spinal cord of both control and Erk2 CKO mice, while expression levels of these proteins did not change after formalin injection. On the other hand, phosphorylation as well as expression levels of ERK5 were not changed after formalin injection in both mice. Samples were sacrificed 5 min after the formalin injection. (B, C) To evaluate phosphorylation, phospho-ERK1 (B) and phspho-ERK2 (C) bands were divided by their corresponding loading control (b-actin). Then, intensities were normalized to the mean control with vehicle injection. SL327 reduced phosphorylation of ERK1 both in control (n = 6) and Erk2 CKO mice (n = 6). We used a 2-way analysis of variance (ANOVA) with a Bonferroni post hoc test; ###P < 0.001 compared with vehicle of same genotype. ⁄⁄⁄P < 0.001. (D) Treatment with SL327 (50 mg/kg) reduced spontaneous nociceptive behaviors in the second phase following formalin injection in control mice (n = 10) compared to vehicle-treated mice (n = 10). Note that nociceptive behaviors of the first phase were not changed by SL327 treatment. (E) No significant differences were observed between vehicle-treated Erk2 CKO (n = 9) and SL327-treated Erk2 CKO mice (n = 9). We used a 2-way repeated-measures ANOVA with a Bonferroni post hoc test; ⁄⁄⁄P < 0.001.

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

2249

Fig. 7. Extracellular signal-regulated kinase (ERK)2 was important for inducing mechanical allodynia in a partial sciatic nerve ligation (PSNL) model. To measure neuropathic mechanical allodynia, mechanical stimuli were investigated using von Frey filaments. (A) ERK1 was not required for allodynia in a PSNL model. In all cases, there was no statistically significant difference between Erk1 knockout (KO) mice and control mice. Similar to controls (n = 9), Erk1 KO mice (n = 9) showed significant increases in paw withdrawal in response to a mechanical stimulus only on the side ipsilateral to ligation. On the contralateral side, neither control nor Erk1 KO mice developed mechanical allodynia. (B) ERK2 was required for allodynia in a PSNL model. Allodynia was attenuated in Erk2 conditional knockout (CKO) mice (n = 9) on the side ipsilateral to ligation compared to controls (n = 9). On the contralateral side, neither control nor Erk2 CKO mice developed mechanical allodynia. We used a 2-way repeated-measures analysis of variance with a Bonferroni post hoc test; ⁄P < 0.05; ⁄⁄⁄P < 0.001 compared with the corresponding side in control.

3.5. ERK2 in dorsal horn neurons was not necessary for thermal hyperalgesia in a PSNL model Next, we tested whether abrogation of ERK1 or ERK2 modified thermal hyperalgesia in a neuropathic pain model. We found no differences in baseline paw withdrawal latencies in the plantar test between control, Erk1 KO, and Erk2 CKO mice before surgery (Fig. 8A, B), indicating that baseline function to heat stimulation remains normal in Erk1 KO and Erk2 CKO mice. In an experiment with Erk1 KO mice, PSNL caused a significant decrease in paw withdrawal time from a heat stimulus only on the side ipsilateral to ligation compared to the contralateral side (Fig. 8A; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 28.85, P < 0.0001), similar to

control mice (Fig. 8A; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 17.10, P < 0.001). We found no significant difference on the side ipsilateral to ligation between Erk1 KO and littermate controls (Fig. 8A; 2-way RM ANOVA, Erk1 KO vs. control, F = 0.48, P > 0.05). In an experiment with Erk2 CKO mice, PSNL caused a significant decrease in paw withdrawal time from a heat stimulus only on the side ipsilateral to ligation compared to the contralateral side (Fig. 8B; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 23.15, P < 0.0001), similar to control mice (Fig. 8B; 2-way RM ANOVA, ipsilateral vs. contralateral, F = 16.03, P < 0.001). We found no significant difference on the side ipsilateral to ligation between Erk2 CKO and littermate controls (Fig. 8B; 2-way RM ANOVA, Erk2

Fig. 8. Thermal hyperalgesia in Erk2 conditional knockout (CKO) mice was indistinguishable from that in control mice in a partial sciatic nerve ligation (PSNL) model. Thermal hyperalgesia was evaluated by a plantar test. (A) Extracellular signal-regulated kinase (ERK)1 was not required for thermal hyperalgesia in a PSNL model. In all cases, there was no statistically significant difference between Erk1 knockout (KO) (n = 13) and control mice (n = 13). Both Erk1 KO and control mice showed significant decreases in paw withdrawal time to a heat stimulation only on the side ipsilateral to ligation. On the contralateral side, neither control nor Erk1 KO mice developed hyperalgesia. (B) Erk2 CKO mice also developed thermal hyperalgesia in a PSNL model. No significant difference in paw withdrawal time was observed on the side ipsilateral to ligation between Erk2 CKO (n = 12) and littermate controls (n = 12). On the contralateral side, neither control nor Erk2 CKO mice developed hyperalgesia. We used a 2-way repeated-measures analysis of variance with a Bonferroni post hoc test.

2250

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

CKO vs. control, F = 0.37, P > 0.05). These results reveal that, unlike mechanical allodynia, thermal hyperalgesia in Erk2 CKO mice was indistinguishable from that in control mice in a PSNL model. We next explored the phosphorylation levels of ERK1 and ERK2 following PSNL using Western blot analysis. In control mice, the phosphorylation levels of both ERK1 and ERK2 increased the first day after ligation and then declined to basal levels by the 21st day after ligation, although the total expression level of proteins did not change (Fig. 9). Similar to control mice, Erk2 CKO mice exhibited elevated phosphorylation of ERK1 on the first day after ligation, which then declined to basal level by the 21st day after ligation (Fig. 9). Phosphorylation level of ERK1 on the first day after ligation was higher in Erk2 CKO mice than in control mice, suggesting a compensatory action of ERK1 in the absence of ERK2. On the other hand, total ERK1 protein level in Erk2 CKO mice was not significantly changed after ligation, indicating no compensatory action of ERK1 for ERK2 in expression levels. In Erk2 CKO mice, although the total ERK2 protein was significantly abrogated, slight expression of ERK2 remained (Fig. 9). This remaining ERK2 was phosphorylated, with peak on the first day after ligation, although the phosphorylation level of ERK2 was much smaller than in control mice (Fig. 9). 4. Discussion In the present study, we show that a selective Erk2 deletion that leaves Erk1 intact in the central nervous system (including the spinal cord) resulted in a reduced response to formalin injections and attenuation of mechanical allodynia in a mouse PSNL model. Our results complement and expand the findings of a report by Alter et al. [2] describing the characteristics of Erk1 KO mice. They found that Erk1 KO mice exhibited normal responses in a formalin test, but the additional inhibition of ERK2 by an MEK inhibitor resulted in significant decrease of response in the second phase. These results suggested a predominant role of ERK2 in the induction of robust spontaneous behaviors following formalin injection. Our result shows that abrogation of ERK2 induced significant decrease of response in the formalin test. Furthermore, additional inhibition of ERK1 phosphorylation by an MEK inhibitor SL327 in Erk2 CKO mice did not produce further change in the response in the formalin test. Collectively, these results indicate that ERK2, but not ERK1, has critical and/or specific roles in pain hypersensitivity.

Fig. 9. Extracellular signal-regulated kinase (ERK) mice were phosphorylated in a partial sciatic nerve ligation (PSNL) model. In control mice, the phosphorylation levels of both ERK1 and ERK2 increased within the first day after ligation and had declined to basal level by the 21st day after ligation, while the total expression level was unchanged. Similar to control mice, Erk2 conditional knockout (CKO) mice exhibited elevated phosphorylation of ERK1 on the first day after ligation that declined to basal level by the 21st day after ligation. Phosphorylation level of ERK1 in Erk2 CKO mice on the first day after ligation was higher than that in control mice. Note the slight increase in ERK2 phosphorylation on the first day after ligation. bactin was used as an internal control.

Alternatively, one might speculate that the roles of ERK1/2 are redundant such that total ERK1/2 phosphorylation actually relates directly to pain hypersensitivity. However, we do not favor this explanation because additional inhibition of ERK1 phosphorylation in Erk2 CKO mice did not result in further attenuation of nociceptive responses in the formalin test (Fig. 6E). Interestingly, the band intensity of phosphorylated ERK1 in the spinal cords of control mice is weaker than that of ERK2, although the total ERK1 protein expression level is stronger than that of ERK2 (Figs. 1B, 5A). This suggests the predominant phosphorylation of ERK2, rather than ERK1, by MEK in the spinal cord. The notion of a predominant and/or specific role for ERK2 and a limited role for ERK1 in pain hypersensitivity is related to findings regarding long-term memory in learning abilities. Accumulating evidence suggests that the role of each isoform in long-term memory may not be functionally redundant. Erk1 KO mice did not show a significant impairment in learning ability [28], although the treatment of rodents with an MEK inhibitor impaired memory formation [30]. Furthermore, Erk2 knockdown mice showed marked deficits in long-term memory [25]. These results suggest a central contribution of the ERK2 isoform to learning and memory. In this study, we used an MEK inhibitor, SL327, to investigate the role of ERK1/2 in the formalin test, similar to the previous study [2]. Some reports indicated that U0126 and PD98059 blocked activation of another isoform, ERK5, in addition to ERK1 and 2 [17,18]. Since SL327 is a structural analogue of U0126, it would be possible that this drug also blocked the phosphorylation of ERK5. If so, one could not exclude the possibility that the effects of SL327 in the formalin test were due to reduction of ERK5 phosphorylation. However, we found that SL327 (50 mg/kg) did not block the phosphorylation of spinal ERK5, contrary to other MEK inhibitors. In this relation, it is worth noting that additional inhibition of ERK2 by the MEK inhibitor SL327 in Erk1 KO mice resulted in significant decrease of response in the second phase [2], since this effect might be only due to reduction of ERK2 phosphorylation and not participation of ERK5 phosphorylation. We found that thermal hyperalgesia was indistinguishable between Erk2 CKO and control mice in a PSNL model, although Erk2 CKO mice showed restored mechanical allodynia compared to control mice. Mechanical allodynia and hyperalgesia are well-used indices for neuropathic pain, but the mechanisms underlying them are not necessarily the same [8,21,29]. For example, mice lacking the sigma-1 receptor, which is involved in regulating neuropathic pain after PNSL, did not exhibit significant differences in hyperalgesia compared to wild-type mice. However, mechanical allodynia did not develop in these mutant mice, contrary to wild-type mice [8]. It is worth noting that these mutant mice did not show increased phosphorylation of ERK in the spinal cord after PNSL, although wild-type mice did show this phenomenon. Further investigation is required to understand the underlying roles of ERK in mechanical allodynia and hyperalgesia. A previous paper reported the characteristics of an Erk2 knockdown in the spinal cord using siRNA [33]. However, approximately half of the ERK2 protein remained, and a significant level of phosphorylated ERK2 protein was detected (approximately 70% of the level found in the control). Furthermore, since the knockdown mediated by siRNA is relatively transient, it is difficult to analyze mechanical allodynia in a PSNL model, which develops over a period of 3 weeks. Thus, another study using a gene-targeting method is required. Nonetheless, mice with siRNA-mediated knockdown showed significantly attenuated mechanical allodynia, suggesting an important role of ERK2 in pain plasticity. However, these mice also showed significantly attenuated hyperalgesia, which is in contrast to our results. Although we do not know the reason for this discrepancy, different levels of expression of ERK2 in microglia might result in dissimilar responses. Although the expression of

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

ERK2 within the spinal microglia would be downregulated using siRNA method, ERK2 protein is intact within the spinal microglia of Erk2 CKO mice, which might contribute to the induction of hyperalgesia. In this study, we used nestin promoter-driven cre transgenic mice to induce recombination in Erk2 floxed mice. The nestin promoter drives cre expression in neuronal and glial progenitors, but not in microglia, because microglia are of hematopoietic origin [5,12,14]. Thus, the Erk2 CKO mice in this study are expected to have normal microglia. Accumulating evidence indicates that microglia play an important role in neuropathic pain [20]. It has been reported that ERK activation mediates pain plasticity, through different mechanisms in different cell types at different times following SNL in the rat [34]. ERK was sequentially activated in dorsal horn neurons, in microglia, and then in astrocytes, reflecting distinct roles for these cell types in the temporal evolution of neuropathic pain [34]. They reported that SNL induced an immediate, but transient (<6 h), induction of phospho-ERK, which was restricted to neurons in the SCDH. This was followed by an induction of phospho-ERK in microglia, which peaked between 1 and 3 days after ligation. By day 21, phospho-ERK was expressed predominantly in astrocytes [34]. In this context, it is worth noting that the remaining ERK2 in the Erk2 CKO mice was phosphorylated, with a peak 1 day after ligation (Fig. 9). Thus, we speculate that ERK2 activation in microglia might have important roles in thermal hyperalgesia in the PSNL model. It has been reported that ERK1/2 was phosphorylated in the DRG in addition to the spinal cord after sciatic nerve ligation [22,34]. Our results cannot exclude the possibility that ERKs in this region have critical roles in pain plasticity. Furthermore, in Erk2 CKO mice, ERK2 is abrogated in other brain regions as well as in the spinal cord, and so it is difficult to distinguish the contribution of ERK2 in these brain regions to pain responses. However, our results clearly indicate that ERK2 in these regions is not the primary determinant for the induction of hyperalgesia in a PSNL model. Further analysis using mice deficient in Erk2 in other regions would be useful for understanding the mechanisms of pain responses. In conclusion, our results indicate that ERK2 plays a predominant and/or specific role in pain plasticity, but the contribution of ERK1 to pain plasticity is limited. Together with the previous result that ERK activation mediates pain plasticity through different mechanisms in different cell types, the different contributions of ERK2 to mechanical allodynia and thermal hyperalgesia suggest the complicated role of ERK signaling in pain plasticity. Further study is necessary to more fully understand the mechanisms that regulate ERK1/2 phosphorylation in pain responses. Conflict of interest There are no conflicts of interest. Acknowledgments This study was, in part, performed with support from the MEXT (Ministry of Education, Culture, Sports, Science and Technology) of Japan, the Naito Foundation, and Japan Foundation for Aging and Health. We thank Kiyoko Takamiya, Yuko Ogura (Department of Anesthesiology, National Defense Medical College, Tokorozawa, Saitama, Japan), Masako Suzuki (Aging Regulation Research Team, Tokyo Metropolitan Geriatric Hospital and Institute of Gerontology) and Tatsuyo Harasawa (Central Research Laboratory, National Defense Medical College, Tokorozawa, Saitama, Japan) for excellent technical help in this study; Dr. Kouichi Fukuda (Center for Laboratory Animal Science, National Defense Medical College, Tokorozawa, Saitama, Japan) for the assistance in animal administration.

2251

The nestin-cre transgenic mice were a kind gift from Dr. Kageyama (University of Kyoto). References [1] Abbadie C, Taylor BK, Peterson MA, Basbaum AI. Differential contribution of the two phases of the formalin test to the pattern of c-fos expression in the rat spinal cord: studies with remifentanil and lidocaine. PAINÒ 1997;69:101–10. [2] Alter BJ, Zhao C, Karim F, Landreth GE, Gereau RW. Genetic targeting of ERK1 suggests a predominant role for ERK2 in murine pain models. J Neurosci 2010;30:11537–47. [3] Atkins CM, Selcher JC, Petraitis JJ, Trzaskos JM, Sweatt JD. The MAPK cascade is required for mammalian associative learning. Nat Neurosci 1998;1:602–9. [4] Boulton TG, Nye SH, Robbins DJ, Ip NY, Radziejewska E, Morgenbesser SD, DePinho RA, Panayotatos N, Cobb MH, Yancopoulos GD. ERKs: a family of protein-serine/threonine kinases that are activated and tyrosine phosphorylated in response to insulin and NGF. Cell 1991;65:663–75. [5] Chan WY, Kohsaka S, Rezaie P. The origin and cell lineage of microglia: new concepts. Brain Res Rev 2007;53:344–54. [6] Coderre TJ, Melzack R. The role of NMDA receptor-operated calcium channels in persistent nociception after formalin-induced tissue injury. J Neurosci 1992;12:3671–5. [7] Cruz CD, Neto FL, Castro-Lopes J, McMahon SB, Cruz F. Inhibition of ERK phosphorylation decreases nociceptive behaviour in monoarthritic rats. PAINÒ 2005;116:411–9. [8] de la Puente B, Nadal X, Portillo-Salido E, Sanchez-Arroyos R, Ovalle S, Palacios G, Muro A, Romero L, Entrena JM, Baeyens JM, Lopez-Garcia JA, Maldonado R, Zamanillo D, Vela JM. Sigma-1 receptors regulate activity-induced spinal sensitization and neuropathic pain after peripheral nerve injury. PAINÒ 2009;145:294–303. [9] De Leo JA, Tawfik VL, LaCroix-Fralish ML. The tetrapartite synapse: path to CNS sensitization and chronic pain. PAINÒ 2006;122:17–21. [10] Dickenson AH, Sullivan AF. Subcutaneous formalin-induced activity of dorsal horn neurones in the rat: differential response to an intrathecal opiate administered pre or post formalin. PAINÒ 1987;30:349–60. [11] Dubuisson D, Dennis SG. The formalin test: a quantitative study of the analgesic effects of morphine, meperidine, and brain stem stimulation in rats and cats. PAINÒ 1977;4:161–74. [12] Enquist IB, Lo Bianco C, Ooka A, Nilsson E, Mansson JE, Ehinger M, Richter J, Brady RO, Kirik D, Karlsson S. Murine models of acute neuronopathic Gaucher disease. Proc Natl Acad Sci USA 2007;104:17483–8. [13] Hargreaves K, Dubner R, Brown F, Flores C, Joris J. A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. PAINÒ 1988;32:77–88. [14] Hess DC, Abe T, Hill WD, Studdard AM, Carothers J, Masuya M, Fleming PA, Drake CJ, Ogawa M. Hematopoietic origin of microglial and perivascular cells in brain. Exp Neurol 2004;186:134–44. [15] Impey S, Obrietan K, Storm DR. Making new connections: role of ERK/MAP kinase signaling in neuronal plasticity. Neuron 1999;23:11–4. [16] Ji RR, Baba H, Brenner GJ, Woolf CJ. Nociceptive-specific activation of ERK in spinal neurons contributes to pain hypersensitivity. Nat Neurosci 1999;2:1114–9. [17] Kamakura S, Moriguchi T, Nishida E. Activation of the protein kinase ERK5/ BMK1 by receptor tyrosine kinases. Identification and characterization of a signaling pathway to the nucleus. J Biol Chem 1999;274:26563–71. [18] Mody N, Leitch J, Armstrong C, Dixon J, Cohen P. Effects of MAP kinase cascade inhibitors on the MKK5/ERK5 pathway. FEBS Lett 2001;502:21–4. [19] Narita M, Usui A, Niikura K, Nozaki H, Khotib J, Nagumo Y, Yajima Y, Suzuki T. Protease-activated receptor-1 and platelet-derived growth factor in spinal cord neurons are implicated in neuropathic pain after nerve injury. J Neurosci 2005;25:10000–9. [20] Narita M, Yoshida T, Nakajima M, Miyatake M, Takagi T, Yajima Y, Suzuki T. Direct evidence for spinal cord microglia in the development of a neuropathic pain-like state in mice. J Neurochem 2006;97:1337–48. [21] Nieto FR, Entrena JM, Cendan CM, Pozo ED, Vela JM, Baeyens JM. Tetrodotoxin inhibits the development and expression of neuropathic pain induced by paclitaxel in mice. PAINÒ 2008;137:520–31. [22] Obata K, Yamanaka H, Kobayashi K, Dai Y, Mizushima T, Katsura H, Fukuoka T, Tokunaga A, Noguchi K. Role of mitogen-activated protein kinase activation in injured and intact primary afferent neurons for mechanical and heat hypersensitivity after spinal nerve ligation. J Neurosci 2004;24:10211–22. [23] Pages G, Guerin S, Grall D, Bonino F, Smith A, Anjuere F, Auberger P, Pouyssegur J. Defective thymocyte maturation in p44 MAP kinase (Erk 1) knockout mice. Science 1999;286:1374–7. [24] Puig S, Sorkin LS. Formalin-evoked activity in identified primary afferent fibers: systemic lidocaine suppresses phase-2 activity. PAINÒ 1996;64:345–55. [25] Satoh Y, Endo S, Ikeda T, Yamada K, Ito M, Kuroki M, Hiramoto T, Imamura O, Kobayashi Y, Watanabe Y, Itohara S, Takishima K. Extracellular signalregulated kinase 2 (ERK2) knockdown mice show deficits in long-term memory; ERK2 has a specific function in learning and memory. J Neurosci 2007;27:10765–76. [26] Satoh Y, Endo S, Nakata T, Kobayashi Y, Yamada K, Ikeda T, Takeuchi A, Hiramoto T, Watanabe Y, Kazama T. ERK2 contributes to the control of social behaviors in mice. J Neurosci 2011;31:11953–67.

2252

Ò

Y. Otsubo et al. / PAIN 153 (2012) 2241–2252

[27] Satoh Y, Kobayashi Y, Takeuchi A, Pages G, Pouyssegur J, Kazama T. Deletion of ERK1 and ERK2 in the CNS causes cortical abnormalities and neonatal lethality: Erk1 deficiency enhances the impairment of neurogenesis in Erk2deficient mice. J Neurosci 2011;31:1149–55. [28] Selcher JC, Nekrasova T, Paylor R, Landreth GE, Sweatt JD. Mice lacking the ERK1 isoform of MAP kinase are unimpaired in emotional learning. Learn Mem 2001;8:11–9. [29] Shibata K, Sugawara T, Fujishita K, Shinozaki Y, Matsukawa T, Suzuki T, Koizumi S. The astrocyte-targeted therapy by Bushi for the neuropathic pain in mice. PLoS One 2011;6:e23510. [30] Thomas GM, Huganir RL. MAPK cascade signalling and synaptic plasticity. Nat Rev Neurosci 2004;5:173–83. [31] Vernay B, Koch M, Vaccarino F, Briscoe J, Simeone A, Kageyama R, Ang SL. Otx2 regulates subtype specification and neurogenesis in the midbrain. J Neurosci 2005;25:4856–67.

[32] Woolf CJ, Thompson SW. The induction and maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states. PAINÒ 1991;44:293–9. [33] Xu Q, Garraway SM, Weyerbacher AR, Shin SJ, Inturrisi CE. Activation of the neuronal extracellular signal-regulated kinase 2 in the spinal cord dorsal horn is required for complete Freund’s adjuvant-induced pain hypersensitivity. J Neurosci 2008;28:14087–96. [34] Zhuang ZY, Gerner P, Woolf CJ, Ji RR. ERK is sequentially activated in neurons, microglia, and astrocytes by spinal nerve ligation and contributes to mechanical allodynia in this neuropathic pain model. PAINÒ 2005;114:149–59.