OUP user menu

Astrocytes within multiple sclerosis lesions upregulate sodium channel Nav1.5

Joel A. Black, Jia Newcombe, Stephen G. Waxman
DOI: http://dx.doi.org/10.1093/brain/awq003 835-846 First published online: 10 February 2010

Summary

Astrocytes are prominent participants in the response of the central nervous system to injury, including neuroinflammatory insults. Rodent astrocytes in vitro have been shown to express voltage-gated sodium channels in a dynamic manner, with a switch in expression of tetrodotoxin-sensitive to tetrodotoxin-resistant channels in reactive astrocytes. However, the expression of sodium channels in human astrocytes has not been studied, and it is not known whether there are changes in the expression of sodium channels in reactive astrocytes of the human central nervous system. Here, we demonstrate a focal and robust upregulation of sodium channel Nav1.5 in reactive astrocytes at the borders of, and within, active and chronic multiple sclerosis lesions. Nav1.5 was only detectable at very low levels in astrocytes within multiple sclerosis macroscopically normal-appearing white matter or in normal control brain. Nav1.1, Nav1.2, Nav1.3 and Nav1.6 showed little or no expression in astrocytes within normal control tissue and limited upregulation in active multiple sclerosis lesions. Nav1.5 was also expressed at high levels in astrocytes in tissue surrounding new and old cerebrovascular accidents and brain tumours. These results demonstrate the expression of Nav1.5 in human astrocytes and show that Nav1.5 expression is dynamic in these cells. Our observations suggest that the upregulated expression of Nav1.5 in astrocytes may provide a compensatory mechanism, which supports sodium/potassium pump-dependent ionic homoeostasis in areas of central nervous system injury.

  • sodium channels
  • multiple sclerosis
  • reactive astrocytes
  • brain tumour
  • stroke

Introduction

Although astrocytes have traditionally been considered to be electrically inexcitable, multiple studies have demonstrated that these cells, which outnumber neurons by a large margin within the brain and spinal cord, can express voltage-gated sodium channels (Ritchie, 1987; Black and Waxman, 1996; Sontheimer et al., 1996). The sodium channel family includes nine members (Nav1.1–Nav1.9) and, while substantial homology exists between the isoforms, amino acid sequence differences confer distinct voltage-dependence, kinetic and pharmacological properties on each of the isotypes (Catterall et al., 2005). Thus far, studies on the expression of sodium channels in astrocytes have been confined to non-human species. These studies have demonstrated that, depending on a number of factors, rodent astrocytes in vivo or in vitro can express the Nav1.1, Nav1.2, Nav1.3 (Black et al., 1994a, b), Nav1.5 (Black et al., 1998) and/or Nav1.6 (Schaller et al., 1996) isoforms of sodium channels. Importantly, these studies indicate that expression of sodium channels by astrocytes in rodents is a dynamic process, changing in response to age of the astrocytes (Sontheimer et al., 1991), exposure to extracellular factors (Thio and Sontheimer, 1993; Thio et al., 1993) and injury (MacFarlane and Sontheimer, 1998).

Astrocytes subserve a number of functions in the normal CNS, including ionic homoeostasis; metabolic support of neurons; participation in formation and maintenance of the blood–brain barrier; and becoming reactive to form a glial scar in response to injury. Critical immunomodulatory roles for astrocytes in the injured CNS have recently been recognized (Dong and Benveniste, 2001; Prat and Antel, 2005; Nair et al., 2008; Okun et al., 2009). Whether the activity of astrocyte sodium channels contributes to their response to CNS injury is not known. Blockade of sodium channels in microglia, which are also non-excitable cells, has been shown to attenuate phagocytosis, cytokine/chemokine release and migration of these CNS resident immune cells (Craner et al., 2005; Black et al., 2009).

Sodium channel expression within human astrocytes has not been studied, and it is not known whether sodium channel expression changes within these glial cells in multiple sclerosis. Here, we demonstrate that Nav1.5 sodium channels, while not detectable in astrocytes within normal control white matter, are focally upregulated in astrocytes within multiple sclerosis lesions. We also demonstrate robust expression of Nav1.5 in reactive astrocytes surrounding cerebral infarcts and brain tumours, consistent with a commonality of upregulated Nav1.5 expression by astrocytes following CNS tissue insult.

Methods

Multiple sclerosis tissue

Post-mortem CNS tissue, acquired via a rapid autopsy protocol from subjects with disabling secondary progressive multiple sclerosis (Table 1: n = 6 active lesion samples from four subjects, age 45.3 ± 12.3 years, disease duration 17.5 ± 6.4 years; and n = 7 chronic lesion samples from five subjects, age 56.2 ± 7.5 years, disease duration 20.8 ± 5.1 years) and from controls with no clinical or neuropathological evidence of CNS disease (Table 1: n = 7 samples from seven subjects, age 59.1 ± 6.5 years), was obtained from the NeuroResource tissue bank at UCL Institute of Neurology (Newcombe and Cuzner, 1993). Tissue was rapidly frozen and stored at −80°C as previously described (Craner et al., 2004); active plaques were characterized by the loss of myelin proteolipid protein (PLP) immunostaining and substantial infiltrates of Ricinus communis agglutinin I positive macrophages/microglia (Bagasra et al., 1995).

View this table:
Table 1

Characteristics of controls and patients with multiple sclerosis

Case numbersAge (year)/sexClinical statusTissue typeDisease duration (year)Post-mortem interval (h)Cause of death
129/FSPAQ,O,Sv811Bronchopneumonia
29/FSPAQ,Pons811Bronchopneumonia
247/FSPAQ,F,Sv209Bronchopneumonia
47FSPAQ,F,Sv209Bronchopneumonia
359/FSPAQ,F,Scort2013Bronchopneumonia
446/FSPAQ,O,Scort228Bronchopneumonia
567/MSPCQ,SC,Cv2915Bronchopneumonia
671/FSPCQ,SC,Cv3219Bronchopneumonia
749/FSPCQ,SC,Cv1116Bronchopneumonia
865/MSPCQ,SC,Cv2520Bronchopneumonia
929/FSPCQ,Pons714Bronchopneumonia
29/FSPCQ,SC,Cv714Bronchopneumonia
29/FSPCQ,Pons714Bronchopneumonia
1053/MNC,O,Sv19Cardiac arrest
1134/MNC,P,Sv21Myocardial infarction
1268/FNC,F,V23Colo-rectal cancer
1375/FNC,Pons7Cardiac arrest
1469/MNC,F,Sv10Coronary disease; pulmonary embolism
1575/FNC,P,V7Cardiac arrest
1640/MNC,P,V6Asphyxia
  • AQ = active plaque; CQ = chronic plaque; Cv = cervical; F = frontal; NC = normal control; O = occipital; P = parietal; SC = spinal cord; Scort = subcortical; SP = secondary progressive; Sv = subventricular; V = ventricular.

Stroke and brain tumour tissue

For comparison with multiple sclerosis tissue, samples were obtained from subjects with cerebrovascular accident (Table 2: n = 3 samples from three subjects, age 78.1 ± 5.9 years) and brain tumours (n = 3 samples from three subjects, age 54.1 ± ± 6.9 years) via the same procedures as for multiple sclerosis samples.

View this table:
Table 2

Characteristics of patients with cerebrovascular accident or tumour

Case numbersAge (year)/sexTissueCause of deathPost-mortem interval (h)
1789/FW near recent haem CVA; P,VAcute cardiac failure16
1868/MSclerotic W&G near old CVA; P,CoBronchopneumonia10
1977/FSclerotic W&G near old CVA; F,CoBronchopneumonia8
2044/MW&G near metastatic tumour; F,CoBronchial carcinoma; brain metastasis13
2167/MSclerotic W near tumour; O,SvGlioblastoma multiforme14
2251/MW&G near tumour; F,CoGlioblastoma multiforme15
  • Co = cortical; CVA = cerebrovascular accident; F = frontal; G = grey matter; haem = haemorrhagic; O = occipital; P = parietal; Sv = subventricular; V = ventricular; W = white matter.

Immunocytochemistry

Ten micrometres of cryosections were processed as described previously (Craner et al., 2004; Black et al., 2007). Briefly, sections were initially fixed for 5 min in 4% paraformaldehyde in 0.14 M Sorensen’s phosphate buffer, pH 7.4, rinsed several times in phosphate-buffered saline and incubated in blocking solution (phosphate-buffered saline with 5% fish skin gelatin, 0.3% Triton X-100, 0.02% sodium azide and 0.1 mg/ml human immunoglobulin G) for 30 min at room temperature. Sections were then incubated with primary antibodies [mouse Nav1.1, 1:100, Antibodies, Inc., Davis, CA; mouse Nav1.2, 1:100, Antibodies, Inc.; rabbit Nav1.3, 1:1000, no. 16153 (Hains et al., 2002); rabbit Nav1.5, 1:100, Alomone, Jerusalem, Israel; rabbit Nav1.6 (PN4), 1:200, Sigma, St. Louis, MO; chicken glial fibrillary acidic protein (GFAP), 1:1000, Encor, Gainesville, FL; mouse GFAP (SMI22) 1:1000, Covance, Emeryville, CA; R. communis agglutinin I-biotin, 1:500, Vector Lab, Burlingame, CA; and mouse PLP, 1:1000, Abcam, Cambridge, MA] for 24–48 h at 4°C, rinsed several times with phosphate-buffered saline and incubated with appropriate secondary antibodies (donkey anti-mouse immunoglobulin G-Alexa Fluor 488 or 546, 1:1000, Invitrogen, Carlsbad, CA; donkey anti-mouse immunoglobulin G DyLight-649, 1:200, Jackson ImmunoResearch, West Grove, PA; donkey anti-rabbit immunoglobulin G-Alexa Fluor 546, 1:1000, Invitrogen; donkey anti-chicken-Cy5, 1:200, Chemicon, Temecula, CA; and, StreptAvidin-Alexa Fluor 633, 1:200, Invitrogen; donkey anti-rabbit immunoglobulin G-biotin, 1:1000, Jackson ImmunoResearch; ExtrAvidin-horseradish peroxidase, 1:1000, Sigma) for 12–24 h at 4°C. Sections were rinsed with phosphate-buffered saline and mounted with Aqua Poly mount (Polysciences, Warrington, PA).

Control experiments were performed with the omission of the primary antibodies and, for the Nav1.5 antibody, pre-adsorption with the cognate peptide (∼48 M peptide to antibody molar concentration ratio) for 3 h at room temperature. In both control experiments, only background labelling was observed.

Tissue analysis

For analysis of control tissues, active and chronic multiple sclerosis plaques and macroscopically normal-appearing white matter surrounding multiple sclerosis lesions, cerebrovascular accident and brain tumours, multiple images were accrued with a Nikon C1si confocal microscope (Nikon USA, Melville, NY) operating under identical gain settings with frame lambda (sequential) mode and saturation indicator activated to prevent possible bleed-through between channels. Low-magnification montage images were acquired with a Nikon E800 light microscope equipped with epi-fluorescent optics and CoolSNAP HQ camera (Photometrics, Tucson, AZ) or a Nikon C1si confocal microscope. Images were composed and processed to enhance contrast in the figures in Adobe Photoshop, with identical settings for the different conditions.

For quantification of Nav1.5 signal in astrocytes within control, normal-appearing white matter and active plaque tissues, images of six non-overlapping fields were acquired for each condition (control = four samples from four subjects, normal-appearing white matter = three samples from three subjects and active plaque = six samples from four subjects). A threshold for Nav1.5 signal intensity was established and the mean ± SEM Nav1.5+ pixel area of GFAP positive astrocytes was determined with MetaMorph software (Molecular Devices, Downingtown, PA).

Results

Acute multiple sclerosis lesions

Multiple isoforms of sodium channels (Nav1.1, Nav1.2, Nav1.3, Nav1.5 and Nav1.6) have been identified in astrocytes in vitro and in situ in infra-human species. We examined the expression and distribution of all these sodium channels in cryosections of active multiple sclerosis plaques and from normal control CNS tissue. We detected limited immunofluorescence for Nav1.1 and Nav1.2 in astrocytes within control tissue, and only background levels of labelling for Nav1.3 and Nav1.6 (Fig. 1). Reactive astrocytes within active multiple sclerosis lesions are hypertrophic and exhibit increased GFAP labelling, consistent with previous descriptions (Lee et al., 1990; Van Der Voorn et al., 1999). We observed a slight upregulation of Nav1.2 signal in reactive astrocytes within active plaques from multiple sclerosis subjects, but no change in labelling for Nav1.1 (Fig. 1). We did not detect enhancement of the signals for Nav1.3 and Nav1.6 in reactive astrocytes within active lesions compared to control tissue.

Figure 1

Sodium channel expression in astrocytes within control tissue and active multiple sclerosis lesions. Sections of control and multiple sclerosis tissue were reacted with antibodies specific for sodium channels Nav1.1, Nav1.2, Nav1.3 and Nav1.6. Astrocytes in control tissue have small cell bodies and slender processes that display GFAP immunolabelling (green). Control astrocytes exhibit limited Nav1.1 and Nav1.2 immunolabelling (red), and Nav1.3 and Nav1.6 are not detectable. Astrocytes within active multiple sclerosis lesions exhibit hypertrophic cell bodies with pronounced GFAP labelling; processes of reactive astrocytes are prominent. Reactive astrocytes in active lesions display slight upregulation of Nav1.2, in contrast to Nav1.1, Nav1.3 and Nav1.6, which do not change expression between control tissue and multiple sclerosis lesions.

In contrast to the expression of Nav1.1, Nav1.2, Nav1.3 and Nav1.6, there was a substantial increase in the expression of Nav1.5 in astrocytes within active multiple sclerosis plaques (Fig. 2). Astrocytes within control tissue exhibited small cell bodies with slender processes and did not display Nav1.5 immunolabelling. Astrocytes within active multiple sclerosis lesions were hypertrophic with prominent GFAP expression and typically possessed multiple thick processes. Reactive astrocytes within the core and borders of the active lesions expressed robust Nav1.5 immunolabelling. Nav1.5 labelling was not confined to the hypertrophic cell bodies, but was also expressed within the extensive network of astrocyte processes formed by these cells in the active lesions (Fig. 3). Enhanced levels of Nav1.5 immunolabelling in reactive astrocytes were observed in all active lesions examined (n = 6). As a control to ensure that we were assessing a specific upregulation of Nav1.5 in reactive astrocytes, we examined tissue after preadsorption of the Nav1.5 antibody with the cognate peptide and observed substantially attenuated signal in reactive astrocytes (Fig. 4).

Figure 2

Nav1.5 expression in astrocytes within control and multiple sclerosis lesion (MS). GFAP positive astrocyte (green) within control tissue does not exhibit Nav1.5 labelling. In contrast, reactive astrocyte with prominent GFAP labelling within an active multiple sclerosis lesion displays robust Nav1.5 immunolabelling (red). Merged image of GFAP and Nav1.5 is yellow. R. communis agglutinin I (RCA) positive macrophages (blue) are present adjacent to the reactive astrocyte.

Figure 3

Nav1.5 expression in astrocytic processes within active multiple sclerosis lesion. GFAP positive astrocyte cell bodies and processes (green) within active lesions exhibit Nav1.5 immunolabelling (red). At increased magnification (insets), co-localization of GFAP and Nav1.5 (yellow) is apparent within astrocytic processes (arrowheads).

Figure 4

Preadsorption of Nav1.5 antibody attenuates signal within reactive astrocytes. Sections of active multiple sclerosis lesions were incubated with Nav1.5 antibody (Nav1.5) or Nav1.5 antibody preadsorbed with the cognate peptide (Nav1.5 + p). Preadsorption of the Nav1.5 antibody substantially attenuates the fluorescent signal for Nav1.5 in reactive astrocytes.

In light of the striking upregulation of Nav1.5 in reactive astrocytes within active multiple sclerosis plaques in comparison with astrocytes within control tissue, we next examined whether astrocytes within normal-appearing white matter, several millimetres from the border of the active lesion, also exhibited Nav1.5 upregulation. A low-magnification montage of images extending from an active lesion through several millimetres of normal-appearing white matter is shown in Fig. 5. There is a sharp demarcation at the lesion border at which myelin labelling (blue) with anti-PLP antibody is lost in the plaque. There is also a sharp boundary between reactive astrocytes with enhanced GFAP labelling (green) within the lesion and non-reactive astrocytes with attenuated GFAP signal in the surrounding tissue. Likewise, there is an abrupt fall-off of Nav1.5 immunostaining (red) at the edge of the lesion. From the same sections in which astrocytes within active lesions displayed robust Nav1.5 immunoreactivity (Fig. 6), we also imaged astrocytes in normal-appearing white matter. Within normal-appearing white matter several millimetres from the lesion border, astrocytes exhibited small cell bodies with a limited number of short, slender processes, similar to that observed in control white matter. Astrocytes within this region of normal-appearing white matter displayed minimal levels of Nav1.5 labelling (Fig. 6).

Figure 5

Montage of images of Nav1.5, GFAP and PLP expression in normal-appearing white matter and active lesion. Low-magnification montage extending from active lesion through several millimetres of normal-appearing white matter demonstrates sharp boundary of myelin PLP labelling (blue) between active plaque and surrounding white matter. Labelling patterns for both Nav1.5 (red) and GFAP (green) exhibit an abrupt border between active lesion and surrounding white matter, with pronounced Nav1.5 and GFAP labelling within the lesion and limited signal in the white matter.

Figure 6

Nav1.5 expression in normal-appearing white matter and active lesions. Within normal-appearing white matter (NAWM) several millimetres from an active lesion, which displays pronounced myelin PLP labelling (blue), astrocytes (GFAP, green) have small cell bodies and slender process that exhibit only minimal Nav1.5 immunolabelling. In contrast, hypertrophic astrocytes within active multiple sclerosis lesion (MS) with abundant GFAP display robust Nav1.5 labelling (red).

Within white matter that retained strong myelin PLP staining but showed a low level of macrophage infiltration, a small number of astrocytes showed Nav1.5 immunoreactivity. For example, Fig. 7 (panel 2) shows a moderately hypertrophic astrocyte, located within the white matter between two closely apposed lesions exhibiting macrophage infiltration. This cell displays a clear upregulation of Nav1.5, although less marked than in astrocytes at the lesion border (panel 3) or within the lesion (panel 4).

Figure 7

Nav1.5 expression in astrocytes within multiple sclerosis tissue. A montage of greyscale images of R. communis agglutinin I (RCA1) positive macrophages (white) throughout a portion of a section of multiple sclerosis tissue was constructed to provide orientation for localization of astrocytes within the tissue. Two regions of substantial macrophage infiltration, at the top and bottom of the figure, are separated by white matter that exhibits limited macrophage infiltration. In a serial section to that stained for R. communis agglutinin I, labelling for Nav1.5 (red), GFAP (green) and PLP (blue) was performed and high-magnification images acquired (right-hand panels). An astrocyte within a normal-appearing white matter region of the section not included in the montage and which exhibits strong PLP labelling displayed minimal Nav1.5 labelling (top right panel). In contrast, an astrocyte in the white matter between two areas of macrophage infiltration (panel 2) is hypertrophic and also shows an upregulation of Nav1.5. Note that the PLP labelling (blue) is more disrupted compared to that in normal-appearing white matter (top right panel). Astrocytes at the border of (panel 3) and within (panel 4) the active lesion display intense Nav1.5 immunolabelling. Note the substantial disruption of PLP labelling at the border of the lesion (panel 3) and the absence of PLP labelling within the lesion (panel 4). PLP = blue; GFAP = green; Nav1.5 = red; co-localization of GFAP and Nav1.5 = yellow.

Quantification of Nav1.5 signals in astrocytes within control, normal-appearing white matter and active plaques is provided in Fig. 8. Astrocytes in control tissue and normal-appearing white matter both exhibit extremely low levels of Nav1.5 immunoreactivity that are not significantly different from each other. In contrast, reactive astrocytes from active lesions display an >10-fold increase in Nav1.5 signal.

Figure 8

Quantification of Nav1.5 immunolabelling in astrocytes within control tissue, normal-appearing white matter and active lesions. Astrocytes within both control tissue and normal-appearing white matter (NAWM) exhibit minimal percentages of pixels with Nav1.5 fluorescence. In contrast, reactive astrocytes within active multiple sclerosis lesions have an ∼10-fold increase in the Nav1.5 fluorescence signal. *P < 0.05 compared to control and normal-appearing white matter.

Chronic multiple sclerosis lesions

To determine whether the upregulation of Nav1.5 in reactive astrocytes within acute multiple sclerosis lesions was confined to this stage of the disease, we also examined chronic multiple sclerosis plaques, which lacked PLP labelling and displayed a paucity of macrophages/microglia. Within chronic lesions, hypertrophic reactive astrocytes are rarely encountered, but the lesion is sclerotic with a dense network of thin processes emanating from astrocytes with very small cell bodies (Holley et al., 2003). Seven chronic multiple sclerosis lesions from five subjects were examined and, in each chronic plaque, the GFAP-positive astrocyte processes displayed robust Nav1.5 immunolabelling (Fig. 9), similar to that exhibited by astrocyte processes within active lesions.

Figure 9

Nav1.5 immunolabelling in astrocytes within a chronic multiple sclerosis plaque. Nav1.5 labelling is not exhibited by GFAP-positive astrocytes within normal-appearing white matter (NAWM), which displays substantial PLP labelling. In contrast, the dense network of slender GFAP-positive astrocytic processes within the chronic multiple sclerosis plaque (chronic) exhibits robust Nav1.5 immunolabelling. Note the absence of PLP labelling within the chronic lesion.

Cerebrovascular accident lesions

We also examined the expression of Nav1.5 in astrocytes adjacent to recent and old cerebrovascular accident lesions as well as brain tumours (glioblastoma multiforme and metastatic carcinoma) (Fig. 10). Astrocytes adjacent to a recent haemorrhagic cerebrovascular accident form a dense network of slender processes, similar to that exhibited within chronic multiple sclerosis lesions, and these processes are strongly Nav1.5-positive (Fig. 10B). Cell bodies of the astrocytes adjacent to the recent cerebrovascular accident also are Nav1.5 immunolabelled. Tissue near to old cerebrovascular accident was examined. In each sample, Nav1.5 expression was maintained in the reactive astrocytes (Fig. 10C). Both astrocyte cell bodies and processes exhibited Nav1.5 labelling, although the network of processes adjacent to old cerebrovascular accident were generally not as dense as that near the recent cerebrovascular accident.

Figure 10

Nav1.5 immunolabelling in astrocytes adjacent to cerebrovascular accident and brain tumours. (A) For comparison purposes, the Nav1.5 labelling of two hypertropic reactive astrocytes within an acute multiple sclerosis lesion is shown. Images in panels (B–F) were acquired and processed with conditions identical to panel A. (B) Reactive astrocytes adjacent to (∼1 cm) a recent cerebrovascular accident exhibit numerous, slender GFAP-positive processes (green) that exhibit Nav1.5 immunoreactivity (red). Note the infiltration of R. communis agglutinin-1 (RCA) positive macrophages (blue) in this region. (C) A hypertrophic reactive astrocyte adjacent to an old cerebrovascular accident lesion displays Nav1.5 labelling within its cell body and processes. (D) Pleomorphic GFAP-positive cells in tissue near a glioblastoma multiforme exhibit intense Nav1.5 immunoreactivity. (E) Hypertrophic reactive astrocyte adjacent to a glioblastoma multiforme displays Nav1.5 immunolabelling within its cell body and processes. (F) Reactive astrocytes adjacent to a metastatic carcinoma exhibit Nav1.5 immunoreactivity. Panels in first and third columns from left show only red (Nav1.5) channel; panels in second and fourth columns from left show merged channels. Red = Nav1.5; green = GFAP; blue = R. communis agglutinin I; yellow = merged image of Nav1.5 (red) and GFAP (green).

Brain tumour lesions

We next examined the expression of Nav1.5 within astrocytes in tissue close to brain tumours to determine its expression in the absence of inflammatory macrophages. Pleomorphic GFAP-positive cells adjacent to glioblastoma multiforme exhibited robust Nav1.5 immunolabelling (Fig. 10D). Likewise, hypertrophic reactive astrocytes near glioblastoma multiforme displayed prominent Nav1.5 labelling within their cell bodies and processes (Fig. 10E). Similar to that observed for gliomas, reactive astrocytes adjacent to a metastatic brain tumour also exhibited robust Nav1.5 immunolabelling (Fig. 10F).

Discussion

Previous studies have demonstrated voltage-gated sodium channels in astrocytes in several non-human species (for reviews, see Black and Waxman, 1996; Sontheimer et al., 1996; Verkhratsky and Steinhäuser, 2000). These studies have demonstrated, in rodent astrocytes under varying conditions, the expression of Nav1.1, Nav1.2, Nav1.3 (Black et al., 1994a, b), Nav1.6 (Schaller et al., 1996) and, notably, Nav1.5 sodium channels (Black et al., 1998). The peripheral nervous system-specific sodium channel isoforms, Nav1.7, Nav1.8 and Nav1.9, have not been identified in astrocytes (Akopian et al., 1996; Toledo-Aral et al., 1997; Dib-Hajj et al., 1998). In the present study, we demonstrate a strong upregulation of expression of Nav1.5, but not of other sodium channel isoforms, in astrocytes within multiple sclerosis lesions. In addition, we show robust expression of Nav1.5 in astrocytes adjacent to new and old cerebral infarcts and surrounding gliomas and a metastatic brain tumour. The low level of Nav1.5 expression in astrocytes within normal-appearing white matter of the same tissue sections in which reactive astrocytes exhibit strong Nav1.5 labelling, and the attenuation of the Nav1.5 signal following preadsorption with the cognate peptide, support the interpretation that the Nav1.5 upregulation is a biological response of astrocytes within lesions in the human CNS.

Although the expression of Nav1.5 sodium channels in human astrocytes has not been previously studied, Nav1.5 mRNA and protein have been described in rodent astrocytes in vitro and in situ (Black et al., 1998). Our demonstration of upregulation of Nav1.5, but not of other sodium channel isoforms, in reactive astrocytes within the human CNS indicates that upregulation of Nav1.5 is not part of a global upregulation of sodium channels in these glial cells. Consistent with the idea that Nav1.5 expression within astrocytes is dynamic and not coupled to expression of other sodium channel isoforms, Nav1.5 is known to be tetrodotoxin (TTX)-resistant (Rogart et al., 1989; Satin et al., 1992), and TTX-resistant sodium currents, with biophysical properties similar to those of Nav1.5, display a labile pattern of expression in astrocytes in vitro, changing markedly depending on time in culture, culture condition and age of the animal from which they are derived (Sontheimer and Waxman, 1992; Sontheimer et al., 1992a, b; Thio et al., 1993). Bevan et al. (1987) described voltage-gated sodium currents in reactive astrocytes cultured from adult rat brain, and, more recently, MacFarlane and Sontheimer (1998) reported a switch from TTX-sensitive sodium currents to TTX-resistant sodium currents with properties attributable to Nav1.5 in rodent astrocytes in an in vitro model of gliosis.

The function of sodium channels in astrocytes is not yet clear. Similar to microglia/macrophages where the level of sodium channel expression increases with activation (Craner et al., 2005), we found that Nav1.5 expression was upregulated in reactive astrocytes in acute multiple sclerosis lesions and surrounding a recent cerebrovascular accident. We also found, however, that Nav1.5 expression remains elevated in astrocytes within chronic multiple sclerosis lesions and surrounding old cerebrovascular accident. Several studies have shown that sodium channels in astrocytes can be deployed to the cell membrane where they are functional in terms of generating sodium currents (Bevan et al., 1985; Barres et al., 1989; Sontheimer and Waxman, 1992; Sontheimer et al., 1992a). The density of TTX-resistant (putative Nav1.5) sodium channels within the membranes of cultured spinal cord astrocytes is sufficient to support the generation of action potential-like responses when inactivation of these channels is removed by hyperpolarization (Sontheimer et al., 1992a). Na+/K+ ATPase activity within astrocytes appears to depend on a standing Na+ influx through sodium channels, and this role of sodium channels may be a prerequisite for the viability of these cells (Sontheimer et al., 1994). Total tissue sodium levels are elevated in acute and chronic multiple sclerosis lesions (Inglese et al., in press), in stroke (Thulborn et al., 2005) and in brain tumour tissues (Ouwerkerk et al., 2003). Consistent with this finding and the more general observation of perturbed extracellular ion concentrations in regions of CNS injury, it might be speculated that the upregulated expression of sodium channels within astrocytes may provide a compensatory mechanism, which supports Na+/K+ ATPase-dependent ionic homoeostasis in areas of CNS injury.

Sodium channels are expressed and functional within intracellular membranes in a number of non-excitable cell types. For example, Nav1.6 sodium channels are expressed in association with intracellular membranes within macrophages and appear to regulate motility of these cells (Carrithers et al., 2009), as they do within microglia (Black et al., 2009). Recent studies have reported the expression of Nav1.5 in gastric epithelial cells (Wu et al., 2006) and human breast cancer cells (Brackenbury et al., 2007), and Nav1.5-specific knock-down experiments indicate that Nav1.5 participates in regulation of proliferation (Wu et al., 2006) and invasive behaviour (Brackenbury et al., 2007) of these inexcitable cell-types. Importantly, Nav1.5 sodium channels have been shown to be expressed by human macrophages, where they are localized in the membranes of endosomes, in which they regulate phagocytosis by providing a route for sodium efflux that offsets proton influx during acidification, a late stage of phagocytosis (Carrithers et al., 2007). While there have been reports of phagocytosis by astrocytes (Vinores and Herman, 1993; Al-Ali and Al-Hussain, 1996; Kalmar et al., 2001), the degree to which Nav1.5 regulates this activity is not understood at this time.

Our observations of upregulated Nav1.5 expression in hypertrophic astrocytes in a narrow bridge of white matter between adjacent active multiple sclerosis lesions suggest that Nav1.5 channels are produced during an early stage of activation. The robust expression of Nav1.5 within astrocytes surrounding a recent cerebral infarct is consistent with this interpretation. It is noteworthy that Nav1.5 expression persists in astrocytes within chronic multiple sclerosis lesions and surrounding old cerebrovascular accident. Both lesions exhibit a paucity of infiltrating macrophages, indicating that the continued expression of Nav1.5 within reactive astrocytes is not dependent upon these immune cells.

Accumulating evidence indicates that sodium channel blockade reduces the number of activated microglia/macrophages in experimental autoimmune encephalomyelitis (EAE) (Craner et al., 2005; Black et al., 2007). In addition, several studies have demonstrated a reduction of the degree of axonal loss and improved clinical status, in EAE, following treatment with sodium channel blockers (Bechtold et al., 2002, 2004, 2006; Lo et al., 2002, 2003). It is interesting, in this regard, that one of the drugs with a protective effect in EAE, flecainide (Bechtold et al., 2002, 2004), has strong effects on the cardiac Nav1.5 channel (Ramos and O’Leary, 2004). Two other drugs that show protective effects in EAE, phenytoin and carbamazepine (Lo et al., 2003; Black et al., 2007), can affect cardiac myocytes at clinical therapeutic levels (Durelli et al., 1985; Kennebäck et al., 1995), with phenytoin displaying use-dependent block of cardiac sodium channels at therapeutic levels (Barber et al., 1991). It has been shown that sodium channel blockade can attenuate multiple microglial activities, including phagocytosis, motility and the release of the proinflammatory cytokines interleukin-1 and tumour necrosis factor-α by activated microglia (Black et al., 2009), and it is tempting to speculate that blockade of astrocytic sodium channels may contribute to the protective effects of sodium channel blockers in EAE. Clinical studies of sodium channel blockers in multiple sclerosis are currently underway (Kapoor et al., 2006; Waxman, 2008), and it would be of interest to determine, in humans, whether sodium channel blockade can influence astrocytic functions, such as cytokine/chemokine release and/or the formation of glial scars.

Funding

National Multiple Sclerosis Society (SGW: RG1912); Medical Research Service and Rehabilitation Service, Department of Veterans Affairs (SGW); Nancy Davis Foundation. The Centre for Neuroscience and Regeneration Research is a Collaboration of the Paralyzed Veterans of America and the United Spinal Association with Yale University.

Footnotes

  • Abbreviations:
    Abbreviations
    EAE
    experimental autoimmune encephalomyelitis
    GFAP
    glial fibrillary acidic protein
    PLP
    proteolipid protein
    TTX
    tetrodotoxin

References

View Abstract