Acta Oto-Laryngologica. 2015; Early Online, 1–5
ORIGINAL ARTICLE
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
Expression of aquaporin 2 following facial nerve crush in rats
FELIPE COSTA NEIVA1*, ANDREI BORIN1*, KIL SUN LEE2*, MARCOS VINICIUS SALLES DIAS3*, BRUNA ROZ RODRIGUES3, JOSÉ RICARDO GURGEL TESTA1, OSWALDO LAÉRCIO MENDONÇA CRUZ1 & LUCIENE COVOLAN4 1
Department of Otorhinolaryngology, 2Department of Biochemistry, Universidade Federal de São Paulo (UNIFESP)/ Escola Paulista de Medicina,, 3AC Camargo Cancer Center (ACCCC) and 4Discipline of Neurophysiology, UNIFESP, São Paulo, SP, Brazil
Abstract Conclusion: We demonstrated an early increase in aquaporin 2 (AQP2) expression in a motor nerve (extratemporal facial nerve, FN) following acute peripheral compression (crush), concomitant to effective development of motor dysfunction (facial palsy). The early increase in AQP2 expression that occurred concomitantly with the appearance of a deficit in a peripheral motor nerve suggests that this protein is involved in the physiological events associated with post-injury edema, similar to the already demonstrated behavior of AQP4 in the central nervous system (CNS). Objective: The aim of this study was to assess the expression of AQP2 in the FN of rats up to 7 days after crush. Methods: The extratemporal trunk of the right FN of rats was subjected to mechanical crush, and the expression of AQP2 in the affected (right) and non-affected (left) FN was measured by means of western blotting at days 1, 3, and 7 after injury. Behavioral analysis of the development of facial palsy was also performed over the same time period. Results: Increased expression of AQP2 was shown in the affected FN compared with its corresponding control at day 1 after compression, simultaneously with the appearance of facial palsy.
Keywords: Aquaporin, facial palsy, western blotting, animals, nerve crush, molecular biology, peripheral nervous system
Introduction The description of aquaporins (AQPs) in the 1990s added a new perspective to the study of physiological events involving water flow and granted Professor Peter Agre a Nobel Prize in 2003. At least 13 similar proteins (AQPs 0–12) are currently described in mammals. Most AQPs are located in the cell membrane, have a mass of approximately 30 kDa (monomers), and form 3 nm pores when oligomerized as homotetramers [1,2]. As a function of their spatial conformation and of the distribution of the electric charges, the pores are permeable to water, as well as to small solutes such as glycerol and urea (AQPs 3, 7, 9, and 10), but not to ions [1,3]. Using the specific tissue and cell distributions
(both phenotypic and topographic) of various AQPs, some studies have demonstrated that AQPs participate in various physiological events such as transbarrier osmosis, urine concentration, fluid secretion, cell migration, neural signaling, skin hydration, cell proliferation, and control of fat deposition [1]. AQP2 is located in the intracellular organelles, and it is rapidly transported to and fused with the cell membrane in response to hormonal stimuli (vasopressin/ antidiuretic hormone, ADH) [1]. The role of AQP2 in kidney physiology is very well established because changes in its expression are associated with the occurrence of nephrogenic diabetes insipidus [1]. Other reports have demonstrated the expression of AQP2 in other tissues, including the male reproductive
Correspondence: Andrei Borin, Rua Rio Grande 551 ap 233B, São Paulo, SP, CEP 04018-001, Brasil. Tel: +55 11 981263544. E-mail: [email protected] *These authors equally contributed to this work.
(Received 30 November 2014; accepted 6 January 2015) ISSN 0001-6489 print/ISSN 1651-2251 online 2015 Informa Healthcare DOI: 10.3109/00016489.2015.1010104
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
2
A. Borin et al.
system, inner ear, central nervous system (CNS), pancreatic islets, and gastrointestinal epithelium [1,3,4]. AQPs 1–5, 8, and 9 have been detected in the CNS, mostly in glial cells such as astrocytes and ependymal cells [1,4]. Several physiological functions of the AQPs in the CNS have been established, and remarkable advances have been made in the identification of correlations between those proteins and clinical conditions such as traumatic and vascular injury, tumor proliferation, autism, dementia associated with infection by the human immunodeficiency virus (HIV), epilepsy, neuromyelitis optica, and transverse myelitis [1,4]. In contrast to such remarkable advances in the knowledge of the mechanism of action of AQPs in the CNS, the results of studies addressing the peripheral nervous system (PNS) are still conflicting and do not form a coherent model [2,4,5]. Using a peripheral nerve chronic constriction injury model, Buffoli et al. [6] reported the absence of AQPs 1 and 2 in the rat sciatic nerve following chronic compression. However, the presence of those proteins was reported in the dorsal ganglia, neurons, and Schwann cells (SCs). Borsani et al. [3] reported the presence of AQP1 in the neurons but not the SCs in the trigeminal ganglion, whereas AQP2 was found in both neurons and SCs. In an in vitro study, Zhang et al. [7] showed the presence of AQP1 (and the absence of AQP4) in SCs from the facial nerve (FN). Although this requires further confirmation, it is believed that the two main AQPs expressed in the PNS are AQPs 1 and 2, which are mainly concentrated in SCs, and perhaps also in some neuron subpopulations [4], whereas other AQPs (3–9) were absent from the PNS according to previous reports [2,4,5,8,9]. Like the results reported for the CNS, AQP expression following peripheral nerve injury is relevant for the understanding of neural physiology and for therapeutic efforts targeting clinical conditions involving edema, Wallerian degeneration, and later regeneration of the FN. These clinical conditions include the following: Bell’s palsy (idiopathic), Ramsay Hunt syndrome (caused by the herpes zoster virus), traumatic facial palsy, and Melkersson-Rosenthal syndrome. Based on previous findings from the literature that suggested for this isoform of AQP, in this initial report our aim was to investigate the possible presence of AQP2 in the FN of rats and the temporal profile of its expression up to 7 days after acute nerve compression (crush). Material and methods The present study was approved by the research ethics committee of the Federal University of São Paulo
(0314/10). The study sample comprised 15 adult male Wistar rats (250–280 g), which were kept in shared cages (30 40 20 cm cages; up to five animals per cage) in a vivarium, under a 12 h lightdark cycle, with free access to food and water, and a climate-controlled environment (23 ± 1 C). The animals were allocated to three groups as a function of the time elapsed between FN crush and its removal (five animals per group). Following anesthesia with intraperitoneal ketamine (100 mg/kg) and xylazine (10 mg/kg), the extratemporal trunk of the right FN was subjected to transient compression (crush) for 5 min using a metal clamp. Briefly, through a retroauricular incision, the FN trunk was easily exposed between its departure from the temporal foramen and its entry in the parotid gland, based on an experimental protocol previously published by our group [10]. The animals’ facial movements were recorded with a digital video camera for behavioral analysis of the development of palsy on the right side of the face on days 1, 3, and 7 after surgery, based on a previously validated method [10,11]. Briefly, we performed dynamic analysis of the facial movements based on the videos, and we classified the vibrissal movement and eye closing during the blinking reflex into five grades (from 1, no movement, to 5, similar to the contralateral/control movement). Next, based on static images (Figure 1), we performed semiautomatic metric analysis of the right eyelid slit closure during the blinking reflex (observed on the contralateral eye) using ImageJ software (National Institutes of Health, USA), and the percentage of closure was calculated [11]. For each group, on days 1, 3, and 7 after induction of deep anesthesia, 3–6 mm samples of the right (injury/crush) and left (control/not crushed) FNs were rapidly removed. These samples were then homogenized in lysis buffer (100 mM Tris, pH 8.0, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100,
A
B
Figure 1. Example of static behavioral analysis to calculate the percentage of eyelid slit closure. (A) Left eye (contralateral/control/ not crushed) is open; measurement of the eyelid slit height (yellow line) on the right eye (injury/crush). (B) Left eye is closed; measurement of the right slit (yellow line)
3
Expression of AQP2 after facial nerve crush in mice Behavioral assessment Dynamic analysis
Static analysis
movement grades
percentual of eyelid slit closure 1.0 0.9 0.8 0.7
5
4
0.6 0.5 0.4 0.3 0.2 0.1 0.0
3
1 D1
D3
D7
Eyelid movement
Vibrissa movement
Movement average
Normal condition
D1
D3
D7
Normal condition
Figure 2. Results of behavioral assessment via dynamic and static methods, showing significant facial palsy on the first day after induction of injury and partial recovery until the seventh day after injury induction
0.5% sodium deoxycholate, and protease inhibitor cocktail; Roche, ????, ????). Protein samples (10 mg) were subjected to SDSPAGE and transferred to polyvinylidene difluoride (PVDF) membranes. AQP2 was detected using antiAQP2 (sc-28629; Santa Cruz Biotechnology, ????, ????) and a peroxidase-conjugated secondary antibody. Membranes were imaged, and images were quantified using a photo-documentation system (Alliance Mini 4m; UVITEC, Cambridge, ????). The intensity of AQP2 was normalized to that of a-tubulin. All data are presented as the mean ± SE (standard error of mean). The one- or two-way analysis of variance (ANOVA) was followed by the TukeyKramer post hoc test with the statistical significance set to p £ 0.001 for each analysis. Results All of the animals developed the expected right-sided facial palsy, which was classified as nearly total on day 1, and partial recovery was observed on days 3 and 7 on dynamic and static analysis (Figure 2). The expression levels of AQP2 were analyzed in five samples per group by WB (Figure 3). AQP2 was detected within the 29–32 kDa range in all the control FN samples (left/not crushed). The two-way ANOVA indicated that AQP2 intensity was under the effect of time (F[2,16] = 59.3; p = 0.000) and lesion F[1,8] = 25.3; p = 0.001) with interaction between both factors (F[2,16] = 14.1; p = 0.001). The profile of AQP2 intensity was transiently increased 1 day after FN injury (3.4 ± 0.3) and
slowly decreased on day 3 (0.7 ± 0.2) up to day 7 (0.1 ± 0.04). Differences between lesion and control groups were detected only in the day 1 group (p = 0.001). In the FNs subjected to injury (right/crushed), the same bands showed increased intensity of AQP2 on Lesion Days
1
3
CTR 7
1
3
7
AQP2
α-Tubulin
Relative expression
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
2
4 3.5
*
#
3 2.5 2 1.5 1 0.5 0
Lesion
1
3
CTR
7
Days post lesion Figure 3. Aquaporin 2 (AQP2) expression. Representative western blot showing increased AQP2 expression in the affected nerve (Lesion, crushed) compared with the contralateral control (CTR, not crushed) and the combined results of three blots shown in a bar graph. Data are expressed as mean ± SE. *Indicates difference from control; # indicates difference from D3 and D7 groups (p < 0.001).
4
A. Borin et al.
day 1 compared to the contralateral side, with progressive decreases on days 3 and 7.
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
Discussion In the present study, we observed increased AQP2 expression following acute compression of the FN in rats, and this coincided with the effective establishment of facial palsy. In a recent study, Zhang et al. [7] reported a similar increase in the expression of AQP1 following injury to CN VII in mice, beginning 6 h after induction of injury, peaking 72 h later, followed by a decrease at day 7. Those findings may suggest that increased expression of AQPs 1 and 2 is part of the physiological response underlying edema formation and resolution in the PNS. In mixed (sensory and motor) nerve models, Borsani et al. [3] found increased expression of AQP2 in the neurons and SCs of the trigeminal ganglion in response to a facial injection of formalin, after a period of 4 h. In addition, increased expression of AQP2 was observed in the dorsal ganglion 14 days after sciatic nerve compression [6]. We believe that our results support the occurrence of such changes in AQP2 expression in a motor nerve model, which had not previously been reported. Our results appear to contradict those of Buffoli et al. [6], who did not find AQP2 expression in the sciatic nerve following compression (but only in its sensory ganglion). However, we believe that there are highly significant differences between the two experimental models. Buffoli et al. [6] applied chronic and discrete compression (‘... ligatures were then loosely tied around the nerve. ’) to a mixed sensory and motor nerve (the sciatic nerve) and analyzed the results only at 14 days later (‘chronic constriction injury’ model). In the present study, we subjected a motor nerve (the extratemporal FN) to transient and intense crush (by means of a metal clamp) to induce its total dysfunction, and we analyzed the expression of AQP2 as early as 24 h after the induction of injury (‘acute crush injury’ model). Therefore, we believe that whereas our model caused local functional and structural injury, that used by Buffoli et al. [6] induced steady nerve stimulation. In addition, the timeframes used in the two studies also caused substantial differences in the events observed. Interestingly, in the study by Buffoli et al. [6], AQP2 expression was not observed in the animals in the control (not subjected to surgery) and naïve (not subjected to compression) groups but was only shown in the dorsal ganglia of the animals subjected to compression. Those findings suggest that under normal conditions, AQP2 may not be expressed in the
SNP, instead manifesting at times of ‘stress, injury and/or exaggerated nerve activity,’ as was the local situation in the FN extratemporal trunk in our model and in the sensory ganglion in that of Buffoli et al. [6]. The functional meaning of the increase of AQP2 expression following FN injury shown here may also be established in the present study; however, the participation of AQP2 in the events that follow transient nerve compression may be related to control of swelling, potassium reuptake, and/or regenerative axon elongation. Several AQPs are involved in the control of tissue swelling in the brain (AQP4), spinal cord (AQP1), retina (AQP4), and eye lens (AQPs 1 and 5) [1,6]. AQPs also participate in the reuptake of potassium released in the course of neural activity, in association with the Kir channels, thus achieving osmotic rebalancing of the tissue microenvironment in the brain (AQP4), spinal cord (AQP1), dorsal ganglion (AQP1), retina (AQP4), and inner ear (AQP4) [1,3,4]. The PNS SCs express AQPs and Kir channels, which concentrate close to the nodes of Ranvier; this may suggest that the mechanism in the PNS is similar to that in the CNS [7,12]. AQP1 is also co-expressed with GAP-43 in the posterior horn of the spinal cord during the initial elongation of axons in the course of embryogenesis, as well as during synaptic remodeling and in the process of regeneration following injury [4]. Based on the specific and heterogeneous tissue distribution of AQPs, we might speculate that similar functions in different tissues may involve the participation of different AQPs in each tissue [1,4,6,7]. Because there is no consistent description of the expression of AQP4 in the PNS, one of its functions in the CNS (control of tissue swelling and/or osmotic rebalancing of the tissue microenvironment) may be performed by another tissue-specific AQP (perhaps AQPs 1 and 2). Based on this line of reasoning, Zhang et al. [7] used in vitro experiments to investigate the suppression (knockdown Schwann cells – SC-KDAqp1) and overexpression of AQP1 in SCs from the FN. They showed that AQP1 suppression resulted in cell ‘shrinkage,’ whereas stimulation caused cell ‘swelling,’ thus pointing to the participation of AQP1 in the phenotype manifested by SCs [7]. Those authors further showed that facing in vitro induction of hypoxia (to mimic nerve injury), the SC-KDAqp1 were not associated with cell edema, which the normal SCs are, suggesting a response similar to that exhibited by astrocytes in association with APQ4 in the CNS [7]. That the cell group that exhibited increased AQP2 expression is uncharacterized is a limitation of the present study. The specimen we used, a segment of the FN extratemporal trunk, contains
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
Expression of AQP2 after facial nerve crush in mice motor nerve fibers from the facial nerve nuclei, SCs, connective tissue, epidural blood vessels, and eventually also injury-recruited inflammatory cells. Classically, only AQP1 is described as present in the vascular endothelium, fibroblasts, and leukocytes [1], to the exclusion of AQP2, making those cell populations the least likely candidates to account for the expression of AQP2 observed in the present study. Nevertheless, there are conflicting reports in the literature. For instance, according to the study by Buffoli et al. [6], AQP2 and CD-31 (an endothelial cell marker) co-localize in the small veins of the dorsal ganglion. Neuronal expression of AQP2 has been described in sensory ganglia following nociceptive stimulation [3,6]. However, those studies assessed neuronal cell bodies, which differ considerably from our model. Based on the typical slowing of the somatic-axonal transport following peripheral nerve injury and the lack of a large protein synthesis apparatus in the axon, we believe that those cells are unlikely to express the consistent increase in AQP2. Thus, we believe that the SCs are the main candidates for the AQP2 increase, given that glial cells have been shown to express this protein [3,4,6]. Further studies are needed to achieve a more thorough understanding of the role of AQPs in the PNS, and more particularly in FN peripheral compression models, which are typical models of motor nerve injury. Expression of other AQPs, both earlier (during the first hours following injury induction) and later (after the seventh day following injury induction) observation windows, other methods of analysis, identification of the cell type that expresses AQP, analysis of the central nucleus, use of gene knockdown models, and investigation of possible pharmacological influences (inhibitors and stimulators of identified AQPs) should be considered in future studies. Even by limitation in our study, at this time we cannot explain the increase in AQP2 observed on the control side (left FN) that was observed on the first day. In conclusion, we demonstrated an early increase in AQP2 expression in a motor nerve (extratemporal
5
FN) following acute peripheral crush, concomitant with effective development of motor dysfunction (facial palsy). Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References [1] Verkman AS. Mammalian aquaporins: diverse physiological roles and potential clinical significance. Expert Rev Mol Med 2008;10:e13. [2] Gao H, He C, Fang X, Hou X, Feng X, Yang H, et al. Localization of aquaporin-1 water channel in glial cells of the human peripheral nervous system. Glia 2006;53:783–7. [3] Borsani E, Bernardi S, Albertini R, Rezanni R, Rodella RF. Alterations of AQP2 expression in trigeminal ganglia in a murine inflammation model. Neurosci Lett 2009;449:183–8. [4] Albertini R, Bianchi R. Aquaporins and glia. Curr Neuropharmacol 2010;8:84–91. [5] Matsuzaki T, Ablimit A, Tajika Y, Suzuki T, Aoki T, Hagiwara H, et al. Water channel aquaporin 1 (AQP1) is present in the perineurium and perichondrium. Acta Histochem Cytochem 2005;38:37–42. [6] Buffoli B, Borsani E, Rezzani R, Rodella LF. Chronic constriction injury induces aquaporin-2 expression in the dorsal root ganglia of rats. J Anat 2009;215:498–505. [7] Zhang J, Xiong Y, Lu LX, Wang H, Zhang YF, Fang F, et al. Aquaporin 1 expression alterations affect morphology and water transport in Schwann cells and hypoxia-induced upregulation of AQP1 occurs in a HIF-1alpha-dependent manner. Neuroscience 2013;252:68–79. [8] Thi MM, Spary DC, Hanani M. Aquaporin-4 water channels in enteric neurons. J Neurosci Res 2008;86:448–56. [9] Mobasheri A, Wray S, Marples D. Distribution of AQP 2 and AQP3 water channels in human tissue microarrays. J Mol Histol 2005;36:1–14. [10] Faria SD, Testa JR, Borin A, Toledo RN. Standardization of techniques used in facial nerve section and facial movement evaluation in rats. Braz J Otorhinolaryngol 2006;72:341–7. [11] Borin A, Mello LEA, Neiva FC, Testa JRG, Cruz OLM. Experimental video analysis of eye blink reflex in a primate model. Otol Neurotol 2012;33:1625–9. [12] Huaiyu M, Deerinck TJ, Jones M, Ellisman MH, Schwarz TL. Inwardly rectifying K+ channels that may participate in K+ buffering are localized in microvilli of Schwann cells. J Neurosci 1996;16:2421–9.
ORIGINAL ARTICLE
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
Expression of aquaporin 2 following facial nerve crush in rats
FELIPE COSTA NEIVA1*, ANDREI BORIN1*, KIL SUN LEE2*, MARCOS VINICIUS SALLES DIAS3*, BRUNA ROZ RODRIGUES3, JOSÉ RICARDO GURGEL TESTA1, OSWALDO LAÉRCIO MENDONÇA CRUZ1 & LUCIENE COVOLAN4 1
Department of Otorhinolaryngology, 2Department of Biochemistry, Universidade Federal de São Paulo (UNIFESP)/ Escola Paulista de Medicina,, 3AC Camargo Cancer Center (ACCCC) and 4Discipline of Neurophysiology, UNIFESP, São Paulo, SP, Brazil
Abstract Conclusion: We demonstrated an early increase in aquaporin 2 (AQP2) expression in a motor nerve (extratemporal facial nerve, FN) following acute peripheral compression (crush), concomitant to effective development of motor dysfunction (facial palsy). The early increase in AQP2 expression that occurred concomitantly with the appearance of a deficit in a peripheral motor nerve suggests that this protein is involved in the physiological events associated with post-injury edema, similar to the already demonstrated behavior of AQP4 in the central nervous system (CNS). Objective: The aim of this study was to assess the expression of AQP2 in the FN of rats up to 7 days after crush. Methods: The extratemporal trunk of the right FN of rats was subjected to mechanical crush, and the expression of AQP2 in the affected (right) and non-affected (left) FN was measured by means of western blotting at days 1, 3, and 7 after injury. Behavioral analysis of the development of facial palsy was also performed over the same time period. Results: Increased expression of AQP2 was shown in the affected FN compared with its corresponding control at day 1 after compression, simultaneously with the appearance of facial palsy.
Keywords: Aquaporin, facial palsy, western blotting, animals, nerve crush, molecular biology, peripheral nervous system
Introduction The description of aquaporins (AQPs) in the 1990s added a new perspective to the study of physiological events involving water flow and granted Professor Peter Agre a Nobel Prize in 2003. At least 13 similar proteins (AQPs 0–12) are currently described in mammals. Most AQPs are located in the cell membrane, have a mass of approximately 30 kDa (monomers), and form 3 nm pores when oligomerized as homotetramers [1,2]. As a function of their spatial conformation and of the distribution of the electric charges, the pores are permeable to water, as well as to small solutes such as glycerol and urea (AQPs 3, 7, 9, and 10), but not to ions [1,3]. Using the specific tissue and cell distributions
(both phenotypic and topographic) of various AQPs, some studies have demonstrated that AQPs participate in various physiological events such as transbarrier osmosis, urine concentration, fluid secretion, cell migration, neural signaling, skin hydration, cell proliferation, and control of fat deposition [1]. AQP2 is located in the intracellular organelles, and it is rapidly transported to and fused with the cell membrane in response to hormonal stimuli (vasopressin/ antidiuretic hormone, ADH) [1]. The role of AQP2 in kidney physiology is very well established because changes in its expression are associated with the occurrence of nephrogenic diabetes insipidus [1]. Other reports have demonstrated the expression of AQP2 in other tissues, including the male reproductive
Correspondence: Andrei Borin, Rua Rio Grande 551 ap 233B, São Paulo, SP, CEP 04018-001, Brasil. Tel: +55 11 981263544. E-mail: [email protected] *These authors equally contributed to this work.
(Received 30 November 2014; accepted 6 January 2015) ISSN 0001-6489 print/ISSN 1651-2251 online 2015 Informa Healthcare DOI: 10.3109/00016489.2015.1010104
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
2
A. Borin et al.
system, inner ear, central nervous system (CNS), pancreatic islets, and gastrointestinal epithelium [1,3,4]. AQPs 1–5, 8, and 9 have been detected in the CNS, mostly in glial cells such as astrocytes and ependymal cells [1,4]. Several physiological functions of the AQPs in the CNS have been established, and remarkable advances have been made in the identification of correlations between those proteins and clinical conditions such as traumatic and vascular injury, tumor proliferation, autism, dementia associated with infection by the human immunodeficiency virus (HIV), epilepsy, neuromyelitis optica, and transverse myelitis [1,4]. In contrast to such remarkable advances in the knowledge of the mechanism of action of AQPs in the CNS, the results of studies addressing the peripheral nervous system (PNS) are still conflicting and do not form a coherent model [2,4,5]. Using a peripheral nerve chronic constriction injury model, Buffoli et al. [6] reported the absence of AQPs 1 and 2 in the rat sciatic nerve following chronic compression. However, the presence of those proteins was reported in the dorsal ganglia, neurons, and Schwann cells (SCs). Borsani et al. [3] reported the presence of AQP1 in the neurons but not the SCs in the trigeminal ganglion, whereas AQP2 was found in both neurons and SCs. In an in vitro study, Zhang et al. [7] showed the presence of AQP1 (and the absence of AQP4) in SCs from the facial nerve (FN). Although this requires further confirmation, it is believed that the two main AQPs expressed in the PNS are AQPs 1 and 2, which are mainly concentrated in SCs, and perhaps also in some neuron subpopulations [4], whereas other AQPs (3–9) were absent from the PNS according to previous reports [2,4,5,8,9]. Like the results reported for the CNS, AQP expression following peripheral nerve injury is relevant for the understanding of neural physiology and for therapeutic efforts targeting clinical conditions involving edema, Wallerian degeneration, and later regeneration of the FN. These clinical conditions include the following: Bell’s palsy (idiopathic), Ramsay Hunt syndrome (caused by the herpes zoster virus), traumatic facial palsy, and Melkersson-Rosenthal syndrome. Based on previous findings from the literature that suggested for this isoform of AQP, in this initial report our aim was to investigate the possible presence of AQP2 in the FN of rats and the temporal profile of its expression up to 7 days after acute nerve compression (crush). Material and methods The present study was approved by the research ethics committee of the Federal University of São Paulo
(0314/10). The study sample comprised 15 adult male Wistar rats (250–280 g), which were kept in shared cages (30 40 20 cm cages; up to five animals per cage) in a vivarium, under a 12 h lightdark cycle, with free access to food and water, and a climate-controlled environment (23 ± 1 C). The animals were allocated to three groups as a function of the time elapsed between FN crush and its removal (five animals per group). Following anesthesia with intraperitoneal ketamine (100 mg/kg) and xylazine (10 mg/kg), the extratemporal trunk of the right FN was subjected to transient compression (crush) for 5 min using a metal clamp. Briefly, through a retroauricular incision, the FN trunk was easily exposed between its departure from the temporal foramen and its entry in the parotid gland, based on an experimental protocol previously published by our group [10]. The animals’ facial movements were recorded with a digital video camera for behavioral analysis of the development of palsy on the right side of the face on days 1, 3, and 7 after surgery, based on a previously validated method [10,11]. Briefly, we performed dynamic analysis of the facial movements based on the videos, and we classified the vibrissal movement and eye closing during the blinking reflex into five grades (from 1, no movement, to 5, similar to the contralateral/control movement). Next, based on static images (Figure 1), we performed semiautomatic metric analysis of the right eyelid slit closure during the blinking reflex (observed on the contralateral eye) using ImageJ software (National Institutes of Health, USA), and the percentage of closure was calculated [11]. For each group, on days 1, 3, and 7 after induction of deep anesthesia, 3–6 mm samples of the right (injury/crush) and left (control/not crushed) FNs were rapidly removed. These samples were then homogenized in lysis buffer (100 mM Tris, pH 8.0, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100,
A
B
Figure 1. Example of static behavioral analysis to calculate the percentage of eyelid slit closure. (A) Left eye (contralateral/control/ not crushed) is open; measurement of the eyelid slit height (yellow line) on the right eye (injury/crush). (B) Left eye is closed; measurement of the right slit (yellow line)
3
Expression of AQP2 after facial nerve crush in mice Behavioral assessment Dynamic analysis
Static analysis
movement grades
percentual of eyelid slit closure 1.0 0.9 0.8 0.7
5
4
0.6 0.5 0.4 0.3 0.2 0.1 0.0
3
1 D1
D3
D7
Eyelid movement
Vibrissa movement
Movement average
Normal condition
D1
D3
D7
Normal condition
Figure 2. Results of behavioral assessment via dynamic and static methods, showing significant facial palsy on the first day after induction of injury and partial recovery until the seventh day after injury induction
0.5% sodium deoxycholate, and protease inhibitor cocktail; Roche, ????, ????). Protein samples (10 mg) were subjected to SDSPAGE and transferred to polyvinylidene difluoride (PVDF) membranes. AQP2 was detected using antiAQP2 (sc-28629; Santa Cruz Biotechnology, ????, ????) and a peroxidase-conjugated secondary antibody. Membranes were imaged, and images were quantified using a photo-documentation system (Alliance Mini 4m; UVITEC, Cambridge, ????). The intensity of AQP2 was normalized to that of a-tubulin. All data are presented as the mean ± SE (standard error of mean). The one- or two-way analysis of variance (ANOVA) was followed by the TukeyKramer post hoc test with the statistical significance set to p £ 0.001 for each analysis. Results All of the animals developed the expected right-sided facial palsy, which was classified as nearly total on day 1, and partial recovery was observed on days 3 and 7 on dynamic and static analysis (Figure 2). The expression levels of AQP2 were analyzed in five samples per group by WB (Figure 3). AQP2 was detected within the 29–32 kDa range in all the control FN samples (left/not crushed). The two-way ANOVA indicated that AQP2 intensity was under the effect of time (F[2,16] = 59.3; p = 0.000) and lesion F[1,8] = 25.3; p = 0.001) with interaction between both factors (F[2,16] = 14.1; p = 0.001). The profile of AQP2 intensity was transiently increased 1 day after FN injury (3.4 ± 0.3) and
slowly decreased on day 3 (0.7 ± 0.2) up to day 7 (0.1 ± 0.04). Differences between lesion and control groups were detected only in the day 1 group (p = 0.001). In the FNs subjected to injury (right/crushed), the same bands showed increased intensity of AQP2 on Lesion Days
1
3
CTR 7
1
3
7
AQP2
α-Tubulin
Relative expression
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
2
4 3.5
*
#
3 2.5 2 1.5 1 0.5 0
Lesion
1
3
CTR
7
Days post lesion Figure 3. Aquaporin 2 (AQP2) expression. Representative western blot showing increased AQP2 expression in the affected nerve (Lesion, crushed) compared with the contralateral control (CTR, not crushed) and the combined results of three blots shown in a bar graph. Data are expressed as mean ± SE. *Indicates difference from control; # indicates difference from D3 and D7 groups (p < 0.001).
4
A. Borin et al.
day 1 compared to the contralateral side, with progressive decreases on days 3 and 7.
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
Discussion In the present study, we observed increased AQP2 expression following acute compression of the FN in rats, and this coincided with the effective establishment of facial palsy. In a recent study, Zhang et al. [7] reported a similar increase in the expression of AQP1 following injury to CN VII in mice, beginning 6 h after induction of injury, peaking 72 h later, followed by a decrease at day 7. Those findings may suggest that increased expression of AQPs 1 and 2 is part of the physiological response underlying edema formation and resolution in the PNS. In mixed (sensory and motor) nerve models, Borsani et al. [3] found increased expression of AQP2 in the neurons and SCs of the trigeminal ganglion in response to a facial injection of formalin, after a period of 4 h. In addition, increased expression of AQP2 was observed in the dorsal ganglion 14 days after sciatic nerve compression [6]. We believe that our results support the occurrence of such changes in AQP2 expression in a motor nerve model, which had not previously been reported. Our results appear to contradict those of Buffoli et al. [6], who did not find AQP2 expression in the sciatic nerve following compression (but only in its sensory ganglion). However, we believe that there are highly significant differences between the two experimental models. Buffoli et al. [6] applied chronic and discrete compression (‘... ligatures were then loosely tied around the nerve. ’) to a mixed sensory and motor nerve (the sciatic nerve) and analyzed the results only at 14 days later (‘chronic constriction injury’ model). In the present study, we subjected a motor nerve (the extratemporal FN) to transient and intense crush (by means of a metal clamp) to induce its total dysfunction, and we analyzed the expression of AQP2 as early as 24 h after the induction of injury (‘acute crush injury’ model). Therefore, we believe that whereas our model caused local functional and structural injury, that used by Buffoli et al. [6] induced steady nerve stimulation. In addition, the timeframes used in the two studies also caused substantial differences in the events observed. Interestingly, in the study by Buffoli et al. [6], AQP2 expression was not observed in the animals in the control (not subjected to surgery) and naïve (not subjected to compression) groups but was only shown in the dorsal ganglia of the animals subjected to compression. Those findings suggest that under normal conditions, AQP2 may not be expressed in the
SNP, instead manifesting at times of ‘stress, injury and/or exaggerated nerve activity,’ as was the local situation in the FN extratemporal trunk in our model and in the sensory ganglion in that of Buffoli et al. [6]. The functional meaning of the increase of AQP2 expression following FN injury shown here may also be established in the present study; however, the participation of AQP2 in the events that follow transient nerve compression may be related to control of swelling, potassium reuptake, and/or regenerative axon elongation. Several AQPs are involved in the control of tissue swelling in the brain (AQP4), spinal cord (AQP1), retina (AQP4), and eye lens (AQPs 1 and 5) [1,6]. AQPs also participate in the reuptake of potassium released in the course of neural activity, in association with the Kir channels, thus achieving osmotic rebalancing of the tissue microenvironment in the brain (AQP4), spinal cord (AQP1), dorsal ganglion (AQP1), retina (AQP4), and inner ear (AQP4) [1,3,4]. The PNS SCs express AQPs and Kir channels, which concentrate close to the nodes of Ranvier; this may suggest that the mechanism in the PNS is similar to that in the CNS [7,12]. AQP1 is also co-expressed with GAP-43 in the posterior horn of the spinal cord during the initial elongation of axons in the course of embryogenesis, as well as during synaptic remodeling and in the process of regeneration following injury [4]. Based on the specific and heterogeneous tissue distribution of AQPs, we might speculate that similar functions in different tissues may involve the participation of different AQPs in each tissue [1,4,6,7]. Because there is no consistent description of the expression of AQP4 in the PNS, one of its functions in the CNS (control of tissue swelling and/or osmotic rebalancing of the tissue microenvironment) may be performed by another tissue-specific AQP (perhaps AQPs 1 and 2). Based on this line of reasoning, Zhang et al. [7] used in vitro experiments to investigate the suppression (knockdown Schwann cells – SC-KDAqp1) and overexpression of AQP1 in SCs from the FN. They showed that AQP1 suppression resulted in cell ‘shrinkage,’ whereas stimulation caused cell ‘swelling,’ thus pointing to the participation of AQP1 in the phenotype manifested by SCs [7]. Those authors further showed that facing in vitro induction of hypoxia (to mimic nerve injury), the SC-KDAqp1 were not associated with cell edema, which the normal SCs are, suggesting a response similar to that exhibited by astrocytes in association with APQ4 in the CNS [7]. That the cell group that exhibited increased AQP2 expression is uncharacterized is a limitation of the present study. The specimen we used, a segment of the FN extratemporal trunk, contains
Acta Otolaryngol Downloaded from informahealthcare.com by University of Massachusetts on 04/06/15 For personal use only.
Expression of AQP2 after facial nerve crush in mice motor nerve fibers from the facial nerve nuclei, SCs, connective tissue, epidural blood vessels, and eventually also injury-recruited inflammatory cells. Classically, only AQP1 is described as present in the vascular endothelium, fibroblasts, and leukocytes [1], to the exclusion of AQP2, making those cell populations the least likely candidates to account for the expression of AQP2 observed in the present study. Nevertheless, there are conflicting reports in the literature. For instance, according to the study by Buffoli et al. [6], AQP2 and CD-31 (an endothelial cell marker) co-localize in the small veins of the dorsal ganglion. Neuronal expression of AQP2 has been described in sensory ganglia following nociceptive stimulation [3,6]. However, those studies assessed neuronal cell bodies, which differ considerably from our model. Based on the typical slowing of the somatic-axonal transport following peripheral nerve injury and the lack of a large protein synthesis apparatus in the axon, we believe that those cells are unlikely to express the consistent increase in AQP2. Thus, we believe that the SCs are the main candidates for the AQP2 increase, given that glial cells have been shown to express this protein [3,4,6]. Further studies are needed to achieve a more thorough understanding of the role of AQPs in the PNS, and more particularly in FN peripheral compression models, which are typical models of motor nerve injury. Expression of other AQPs, both earlier (during the first hours following injury induction) and later (after the seventh day following injury induction) observation windows, other methods of analysis, identification of the cell type that expresses AQP, analysis of the central nucleus, use of gene knockdown models, and investigation of possible pharmacological influences (inhibitors and stimulators of identified AQPs) should be considered in future studies. Even by limitation in our study, at this time we cannot explain the increase in AQP2 observed on the control side (left FN) that was observed on the first day. In conclusion, we demonstrated an early increase in AQP2 expression in a motor nerve (extratemporal
5
FN) following acute peripheral crush, concomitant with effective development of motor dysfunction (facial palsy). Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References [1] Verkman AS. Mammalian aquaporins: diverse physiological roles and potential clinical significance. Expert Rev Mol Med 2008;10:e13. [2] Gao H, He C, Fang X, Hou X, Feng X, Yang H, et al. Localization of aquaporin-1 water channel in glial cells of the human peripheral nervous system. Glia 2006;53:783–7. [3] Borsani E, Bernardi S, Albertini R, Rezanni R, Rodella RF. Alterations of AQP2 expression in trigeminal ganglia in a murine inflammation model. Neurosci Lett 2009;449:183–8. [4] Albertini R, Bianchi R. Aquaporins and glia. Curr Neuropharmacol 2010;8:84–91. [5] Matsuzaki T, Ablimit A, Tajika Y, Suzuki T, Aoki T, Hagiwara H, et al. Water channel aquaporin 1 (AQP1) is present in the perineurium and perichondrium. Acta Histochem Cytochem 2005;38:37–42. [6] Buffoli B, Borsani E, Rezzani R, Rodella LF. Chronic constriction injury induces aquaporin-2 expression in the dorsal root ganglia of rats. J Anat 2009;215:498–505. [7] Zhang J, Xiong Y, Lu LX, Wang H, Zhang YF, Fang F, et al. Aquaporin 1 expression alterations affect morphology and water transport in Schwann cells and hypoxia-induced upregulation of AQP1 occurs in a HIF-1alpha-dependent manner. Neuroscience 2013;252:68–79. [8] Thi MM, Spary DC, Hanani M. Aquaporin-4 water channels in enteric neurons. J Neurosci Res 2008;86:448–56. [9] Mobasheri A, Wray S, Marples D. Distribution of AQP 2 and AQP3 water channels in human tissue microarrays. J Mol Histol 2005;36:1–14. [10] Faria SD, Testa JR, Borin A, Toledo RN. Standardization of techniques used in facial nerve section and facial movement evaluation in rats. Braz J Otorhinolaryngol 2006;72:341–7. [11] Borin A, Mello LEA, Neiva FC, Testa JRG, Cruz OLM. Experimental video analysis of eye blink reflex in a primate model. Otol Neurotol 2012;33:1625–9. [12] Huaiyu M, Deerinck TJ, Jones M, Ellisman MH, Schwarz TL. Inwardly rectifying K+ channels that may participate in K+ buffering are localized in microvilli of Schwann cells. J Neurosci 1996;16:2421–9.
