Reactive astrocytes express the potassium channel Kir4.1 in active multiple sclerosis lesions



Jun-ichi Satoh, MD, Department of Bioinformatics and Molecular Neuropathology, Meiji Pharmaceutical University, 2-522-1 Noshio, Kiyose, Tokyo 204-8588, Japan.

Tel: +81-42-495-8678

Fax: +81-42-495-8678




Kir4.1, an inwardly rectifying potassium channel expressed on perivascular and perisynaptic end-feet of astrocytes, plays a pivotal role in the spatial buffering of the potassium in the brain. A recent study showed that autoantibodies directed to Kir4.1 are detectable in the serum derived from approximately half of the patients with multiple sclerosis (MS) and clinically isolated syndrome, although their pathogenic roles should be elucidated.


We studied Kir4.1 expression in MS and control brains by immunohistochemistry.


We found that reactive astrocytes expressed an intense immunoreactivity for Kir4.1 in active demyelinating lesions of MS, active lesion edges of neuromyelitis optica, ischemic lesion edges of cerebral infarction and neurodegenerative lesions of Alzheimer's disease. Reactive astrocytes accumulated in active MS lesions coexpressed Kir4.1 and AQP4. A subset of amyloid plaques in Alzheimer's disease brains also expressed Kir4.1. In contrast, infiltrating macrophages, activated microglia and surviving oligodendrocytes in active MS lesions did not express Kir4.1. Furthermore, cultured human astrocytes expressed Kir4.1, and the expression levels were not altered by exposure to tumor necrosis factor-α or interleukin-1β, but were elevated by transforming growth factor-β1.


These results show that reactive astrocytes abundantly express Kir4.1, and Kir4.1 immunoreactivity is not lost in active demyelinating lesions of MS.


Kir4.1 is an adenosine triphosphate-dependent, inwardly rectifying potassium channel essential for the homeostasis of extracellular potassium, which forms a tetramer expressed on renal epithelial cells, inner ear cells, and perivascular and perisynaptic end-feet of astrocytes.[1, 2] Kir4.1 is colocalized with the water channel, aquaporin-4 (AQP4), in the dystrophin-associated glycoprotein complex (DGC) at the interface of astrocytes and small blood vessels, where Kir4.1 cooperates with AQP4 for the spatial buffering of potassium and water transport.[3, 4] Furthermore, the electrochemical gradient generated by K+ buffering across the cell membrane of perisynaptic astrocytes regulates the uptake of extracellular glutamate.[5] Kir4.1 is expressed on cell bodies of oligodendrocytes in the rat optic nerve during development.[6] Kir4.1 knockout mice show motor impairment as a result of hypomyelination and axonal degeneration in the spinal cord, suggesting that Kir4.1 is indispensable for oligodendrocyte development.[7] The conditional knockout of astrocytic Kir4.1 leads to glial membrane depolarization, along with reduced clearance of the excitatory neurotransmitter glutamate, suggesting that Kir4.1 plays a key role in the regulation of neuronal excitability.[8] Supporting this view, loss of function mutations in the human KCNJ10 gene encoding Kir4.1 cause the epilepsy, ataxia, sensorineural deafness, and tubulopathy (EAST)/seizure, sensorineural deafness, ataxia, mental retardation, and electrolyte imbalance (SeSAME) syndrome, characterized by epilepsy, ataxia, sensorineural deafness and renal tubulopathy.[9] Furthermore, the expression of Kir4.1, α-syntrophin and dystrophin is concurrently lost in perivascular end-feet of astrocytes in hippocampal sclerosis lesions of the patients with temporal lobe epilepsy.[10]

Accumulating evidence has indicated a pivotal role of B cells in the immunopathogenesis of multiple sclerosis (MS), an inflammatory demyelinating disease affecting mainly the human central nervous system (CNS) white matter.[11] More than 90% of MS patients show oligoclonal bands (OCB) in the cerebrospinal fluid (CSF), suggesting an intrathecal clonal expansion of B cells.[12] The meninges of MS patients often contain ectopic B cell follicles having germinal centers.[13] A subtype of acute demyelinating lesions of MS shows prominent deposition of immunoglobulins and complement activation.[14] Treatment with B cell-targeting monoclonal antibodies dramatically suppresses formation of new lesions in relapsing-remitting MS patients.[15]

A recent study showed that immunoglobulin G (IgG) autoantibodies directed to Kir4.1 are detectable in the serum derived from approximately half of the patients with MS, regardless of clinical subtypes, and clinically isolated syndrome (CIS), whereas they are barely detectable in the serum of the patients with other neurological diseases.[16] Intrathecal injection of serum anti-Kir4.1 IgG, in combination with human complement, reduced Kir4.1 expression on subpial and cortical astrocytes, accompanied by complement activation at antibody-binding sites in mouse brains in vivo.[16] However, the question remains to be clarified whether anti-Kir4.1 antibodies are pathogenic in the human brain. It is well known that anti-AQP4 antibodies are found in the serum derived from the great majority of patients with neuromyelitis optica (NMO), whereas they are rarely detectable in the serum of MS patients.[17] The active lesions of NMO show a profound loss of AQP4-expressing astrocytes.[18] Previously, we found that AQP4 expression is greatly enhanced in astrocytes and glial scars in active MS lesions.[19] Therefore, it is highly important to address the question whether or not Kir4.1 expression is lost in active demyelinating lesions of MS. The present study was designed to characterize Kir4.1 expression in active demyelinating lesions of MS by immunohistochemistry.


Human brain tissues

For immunohistochemistry, 10-μm thick serial sections of the cerebral cortex were prepared from autopsied brains of four MS patients and 14 non-MS subjects (Table 1). All four MS cases were clinically diagnosed as chronic progressive MS.[20]

Table 1. Human brain tissues examined in the present study
Case no.Age and sexClinical diagnosisCause of death
  1. The present study includes four cases of multiple sclerosis (MS), one case of neuromyelitis optica (NMO), four cases of neurologically normal controls (NC), four cases of cerebral infarction (CI), and five cases of Alzheimer's disease (AD).

MS-129 FemaleSecondary progressive multiple sclerosisAsphyxia
MS-240 FemaleSecondary progressive multiple sclerosisRespiratory failure
MS-343 FemalePrimary progressive multiple sclerosisHyperglycemia
MS-433 MaleSecondary progressive multiple sclerosisSepsis and multiple organ failure
NMO-170 FemaleDevic's diseasePneumonia
NC-188 FemaleNeurologically normal subjectMyocardial infarction
NC-284 MaleNeurologically normal subjectMmyocardial infarction
NC-377 MaleNeurologically normal subjectLung cancer
NC-467 MaleNeurologically normal subjectDissecting aortic aneurysm
CI-147 MaleAcute cerebral infarctionSepsis
CI-284 MaleAcute cerebral infarctionDisseminated intravascular coagulation
CI-362 MaleOld cerebral infarctionPancreatic cancer
CI-456 MaleOld cerebral infarctionMyocardial infarction
AD-159 MaleAlzheimer's diseasePneumonia
AD-268 FemaleAlzheimer's diseaseMultiple organ failure
AD-380 MaleAlzheimer's diseasePneumonia
AD-472 MaleAlzheimer's diseasePneumonia
AD-577 FemaleAlzheimer's diseasePulmonary infarction

Non-MS cases included a previously reported case of neuromyelitis optica (NMO),[21] four neurologically normal control (NC) subjects, four patients with cerebral infarction (CI) and five patients with Alzheimer's disease (AD). All AD cases were satisfied with the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) criteria for diagnosis of definite AD, and were categorized into the stage C of amyloid deposition and the stage VI of neurofibrillary degeneration, following the Braak staging system, as described previously.[22]

Autopsies were carried out at the National Center Hospital, National Center of Neurology and Psychiatry (NCNP), Kohnodai Hospital, National Center for Global Health and Medicine (NCGM), or the Nishitaga National Hospital. Comprehensive examination by four established neuropathologists (KA, YS, TI, HK) validated the pathological diagnosis. Written informed consent was obtained from all the cases. The Ethics Committee of the corresponding institutions approved the present study. The serum samples were not available from any of these cases.

Human astrocytes in culture

Cultured human astrocytes (AS) were established from primary cultures of neuronal progenitor (NP) cells isolated from human fetal brains (Cambrex, Walkersville, MD, USA), as described previously.[23] NP cell cultures were maintained in Dulbecco's modified Eagle's medium (DMEM)/F-12 medium (Invitrogen, Carlsbad, CA, USA) supplemented with an insulin-transferrin-selenium (ITS) supplement (Invitrogen), 20 ng/mL recombinant human epidermal growth factor (EGF) (Higeta, Tokyo, Japan), 20 ng/mL recombinant human basic fibroblast growth factor (bFGF) (PeproTech, London, UK) and 10 ng/mL recombinant human leukemia inhibitory factor (LIF) (Millipore, Temecula, CA, USA). For the induction of astrocyte differentiation, NP cells were incubated for several weeks in DMEM (Invitrogen) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin and 100 μg/mL streptomycin (feeding medium). The purity of astrocytes exceeded 98% by glial fibrillary acidic protein (GFAP) immunolabeling.[23] In some experiments, human astrocytes were incubated in the feeding medium supplemented with 50 ng/mL recombinant human transforming growth factor-β1 (TGF-β1), tumor necrosis factor-α (TNF-α) or interleukin-1β (IL-1β) (all from PeproTech).

Immunohistochemistry and immunocytochemistry

After deparaffination, tissue sections were heated in 10 mmol/L citrate sodium buffer, pH 6.0 by autoclave at 110°C for 15 min in a temperature-controlled pressure chamber (Biocare Medical, Concord, CA, USA). They were treated at room temperature (RT) for 15 min with 3% hydrogen peroxide-containing methanol to block the endogenous peroxidase activity. The tissue sections were then incubated with phosphate-buffered saline (PBS) containing 10% normal goat serum at RT for 15 min to block non-specific staining. They were incubated in a moist chamber at 4°C overnight with a rabbit antibody against the C-terminal peptide of Kir4.1, whose sequence is conserved among the human, mouse and rat (1:2000; APC-035; Alomone Labs, Jerusalem, Israel), a rabbit antibody against the C-terminal peptide of human AQP4 (1:1000; sc-20812; Santa Cruz Biotechnology, Santa Cruz, CA, USA) or a rabbit anti-GFAP antibody (prediluted; N1506; Dako, Tokyo, Japan). After washing with PBS, the tissue sections were labeled at RT for 30 min with horseradish peroxidase (HRP)-conjugated secondary antibodies (Nichirei, Tokyo, Japan), followed by incubation with diaminobenzidine tetrahydrochloride (DAB) substrate (Vector, Burlingame, CA, USA). They were processed for a counterstain with hematoxylin. For negative controls, tissue sections were incubated with the Kir4.1 antibody (APC-035) preabsorbed with the specific peptide (Almone Labs). Non-astroglial cell types were identified by immunolabelling with cell type-specific antibodies.[24]

For double labeling, tissue sections and human astrocytes on glass slides were initially stained with a mixture of Kir4.1 antibody (APC-035) and a mouse monoclonal anti-GFAP antibody (GA5; Nichirei). They were then incubated with a mixture of Alexa Fluor 488-conjugated anti-rabbit IgG and Alexa Fluor 568-conjugated anti-mouse IgG (Invitrogen), followed by nuclear staining with 6′-diamidino-2-phenylindole (DAPI; Invitrogen). They were examined under the Olympus BX51 fluorescent microscope. Negative controls were processed following all the steps except for exposure to primary antibody.

Reverse transcription polymerase chain reaction analysis

DNase-treated total cellular RNA was processed for cDNA synthesis using oligo(dT)12-18 primers and SuperScript II reverse transcriptase (Invitrogen). Then, cDNA was amplified by polymerase chain reaction (PCR) using HotStar Taq DNA polymerase (Qiagen, Valencia, CA, USA), and a panel of sense and antisense primer sets following: 5′gttgtgaaagtggcctctcctagt3′ and 5′tcagacattgctgatgcgcacact3′ for a 144 bp product of Kir4.1; 5′cctcgctggtggcctttatgagta3′ and 5′gtctttccccttcttctcctctcc3′ for a 218 bp product of AQP4; and 5′ccatgttcgtcatgggtgtgaacca3′ and 5′gccagtagaggcagggatgatgttc3′ for a 251 bp product of the glyceraldehyde-3-phosphate dehydrogenase (G3PDH) gene, serving as an internal control. The amplification program consisted of an initial denaturing step at 95°C for 15 min, followed by a denaturing step at 94°C for 1 min, an annealing step at 60°C for 40 s and an extension step at 72.9°C for 50 s for 35 cycles, except for G3PDH amplified for 28 cycles. For the positive control, total RNA of the human frontal cerebral cortex (Clontech, Mountain View, CA, USA) was processed in parallel for reverse transcription polymerase chain reaction (RT–PCR).

For quantitative RT–PCR (qPCR), cDNA was amplified by PCR in LightCycler ST300 (Roche Diagnostics, Tokyo, Japan) using SYBR Green I, and the set of sense and antisense primers described earlier. The expression levels of target genes were standardized against the levels of G3PDH detected in the corresponding cDNA samples. All the assays were carried out in triplicate.

Transient expression of Kir4.1 and AQP4 in HEK293 cells

Open reading frames (ORF) of the human KCNJ10 gene encoding Kir4.1 (GenBank NM_002241) and the human AQP4 gene (GenBank NM_001650) were amplified by PCR using PfuTurbo DNA polymerase (Stratagene, La Jolla, CA, USA), and sense and antisense primer sets following: 5′acgtcagttgccaaggtgtattac3′ and 5′tcagacattgctgatgcgcacact3′ for KCNJ10 and 5′agtgacagacccacagcaaggcgg3′ and 5′tcatactgaagacaatacctctcc3′ for AQP4. Then, they were cloned in an expression vector named pcDNA4/HisMax-TOPO (Invitrogen). Then, the vectors were transfected into HEK293 cells by using Lipofectamine 2000 reagent (Invitrogen). At 24 h after transfection, the cells were processed for western blot analysis.

Western blot analysis

To prepare total protein extract, the cells were homogenized in a non-denaturing lysis buffer composed of 20 mmol/L Tris-HCl, pH 8.0, 137 mmol/L NaCl, 1% Nonidet P40, 2 mmol/L ethylenediaminetetraacetic acid, and a cocktail of protease inhibitors, followed by centrifugation at 13200 g for 5 min at RT. The supernatant was separated on a 12% sodium dodecylsulfate-polyacrylamide gel electrophoresis gel. After gel electrophoresis, the protein was transferred onto nitrocellulose membranes, and immunolabeled at RT overnight with anti-Kir4.1 antibody (APC-035) or anti-AQP4 antibody (sc-20812). Then, the membranes were incubated at RT for 30 min with HRP-conjugated secondary antibodies (Santa Cruz Biotechnology). The specific reaction was visualized by using a chemiluminescent substrate (Pierce, Rockford, IL, USA). After the antibodies were stripped by incubating the membranes at 50°C for 30 min in a stripping buffer composed of 62.5 mmol/L Tris-HCl, pH 6.7, 2% sodium dodecylsulfate and 100 mmol/L 2-mercaptoethanol, the membranes were relabeled with a goat anti-heat shock protein HSP60 antibody (sc-1052; Santa Cruz Biotechnology), serving as an internal control for protein loading, or a mouse anti-Xpress antibody (R91025,; Invitrogen).


Reactive astrocytes expressed Kir4.1 in chronic active demyelinating lesions of multiple sclerosis brains

First, to verify the specificity of anti-Kir4.1 antibody (APC-035), the ORF of either the human KCNJ10 gene or the human AQP4 gene cloned in the expression vector was expressed transiently as an Xpress-tagged protein in HEK293 cells. APC-035 reacted specifically with the recombinant Kir4.1 protein, but not with the AQP4 protein (Fig. 1a, lanes 1–3), whereas anti-AQP4 antibody (sc-20812) reacted with the recombinant AQP4 protein, but not with the Kir4.1 protein (Fig. 1b, lanes 1–3), validating the specificity of these antibodies.

Figure 1.

Validation of the specificity of anti-Kir4.1 antibody. The full-length protein of Kir4.1 or AQP4 fused with an Xpress tag was expressed in HEK293 cells, and the cellular protein extract was processed for western blot. (a) Kir4.1, (b) AQP4, (c) Xpress and (d) Hsp60, an internal control for protein loading. Lanes 1–3 indicate 15 μg of protein isolated from (1) non-transfected cells, (2) Kir4.1 expression vector-transfected cells and (3) AQP4 expression vector-transfected cells. The position of molecular weight size markers is shown on the left.

Next, the expression of Kir4.1 was studied in brain tissue sections of four MS and 11 non-MS cases by immunohistochemistry using APC-035. In the brains of neurologically normal control (NC) subjects, an intense Kir4.1 immunoreactivity was identified chiefly in the vascular walls, in addition to the choroid plexus, the ependymal lining and the pial basement membrane (Fig. 2a). In NC brains, Kir4.1 immunoreactivity was almost absent in the capillaries, whereas vascular smooth muscle cells of larger vessels showed an intense immunoreactivity for Kir4.1.

Figure 2.

Kir4.1 expression in non-multiple sclerosis brains. The brain tissue sections of neurologically normal controls (NC), cerebral infarction (CI) and Alzheimer's disease (AD) were processed for immunohistochemistry with anti-Kir4.1 antibody. (a) NC, the temporal cortex, the vascular walls express Kir4.1. (b) Acute CI, the parietal cortex, Kir4.1-negative foamy macrophages on the left and Kir4.1-positive reactive astrocytes on the right. (c) Acute CI, the parietal white matter penumbra. (d) AD, the frontal cortex, Kir4.1-positive amyloid plaques (arrows). (e) AD, the frontal cortex, some neurons express Kir4.1. (f) AD, the frontal cortex, perivascular deposits of amyloid express Kir4.1.

In chronic active demyelinating lesions of MS, GFAP-positive reactive astrocytes expressed a very intense Kir4.1 immunoreactivity, whereas infiltrating myelin-phagocytosing macrophages, reactive microglia and surviving oligodendrocytes did not express Kir4.1 (Fig. 3a–e). Double labeling validated coexpression of Kir4.1 and GFAP in reactive astrocytes accumulated in active MS lesions (Fig. 4a–c). Reactive astrocytes in chronic active lesions of MS also coexpressed AQP4 (Fig. 3d–f). In contrast, in inactive MS lesions without accumulation of macrophages, Kir4.1 immunoreactivity on astrocytes was fairly weak or often absent. However, glial scar-forming astrocytes surrounding active demyelinating lesions and the glial scar were often intensely labeled with APC-035 (Fig. 3g). In contrast, Kir4.1 immunolabeling was totally absorbed by preincubation of APC-035 with the specific peptide (Fig. 3h). Notably, numerous hypertrophic reactive astrocytes accumulated in active lesion edges of NMO brains expressed an intense Kir4.1 immunoreactivity (Fig. 3i).

Figure 3.

Kir4.1 expression in chronic active demyelinating lesions of multiple sclerosis (MS) brains. The brain tissue sections of MS and neuromyelitis optica (NMO) were processed for immunohistochemistry with anti-Kir4.1 antibody, anti-AQP4 antibody or anti-glial fibrillary acidic protein (GFAP) antibody. (a) Kir4.1, MS, the active demyelinating lesion center in the frontal white matter. (b) Kir4.1, the higher magnification of the same area as (a). (c) Kir4.1, MS, the active demyelinating lesion in the frontal white matter, perivascular foamy macrophages (arrows) do not express Kir4.1. (d) Kir4.1, MS, the frontal subcortical white matter. (e) GFAP, the same area as (d). (f) AQP4, the same area as (d), perivascular end-feet of reactive astrocytes express AQP4. (g) Kir4.1, MS, the periventricular white matter and glial scar-forming astrocytes express Kir4.1. (h) Kir4.1 absorbed by the specific peptide, the same area as (g). (i) Kir4.1, NMO, the active lesion edge in the frontal subcortical white matter and hypertrophic reactive astrocytes on the left express Kir4.1.

Figure 4.

Coexpression of Kir4.1 and glial fibrillary acidic protein (GFAP). The (a–c) brain tissue section of active multiple sclerosis lesions and (d–f) cultured human astrocytes (AS) were processed for double labeling with anti-Kir4.1 antibody and anti-GFAP antibody. (a,d) Kir4.1, (b,e) GFAP and (c,f) merge with 6′-diamidino-2-phenylindole (DAPI) nuclear staining.

Reactive astrocytes expressed Kir4.1 in ischemic lesions and Alzheimer's disease brains

Reactive astrocytes accumulated in ischemic lesion edges, and the penumbra of acute and chronic cerebral infarction also expressed an intense Kir4.1 immunoreactivity, whereas foamy macrophages, reactive microglia and surviving oligodendrocytes did not express Kir4.1 (Fig. 2b,c). In neurodegenerative lesions of AD brains, reactive astrocytes, some neurons, the neuropil, the vascular walls and capillaries expressed Kir4.1 immunoreactivity at variable intensities (Fig. 2d–f). Notably, a subset (less than 5%) of amyloid plaques and perivascular deposits of amyloid were labeled with APC-035 (Fig. 2d,f).

Cultured human astrocytes constitutively expressed Kir4.1

Finally, we studied the expression of Kir4.1 in cultured human AS. By RT–PCR, we identified mRNA expression of Kir4.1 and AQP4 in both AS and NP cells, in addition to human brain tissues (Fig. 5a,b, lanes 1, 3 and 4). The levels of G3PDH, an internal control, were almost constant in the cells and tissues examined, whereas no products were amplified when the reverse transcription step was omitted (Fig. 5c, lanes 1–4). GFAP-immunoreactive AS intensely expressed Kir4.1 chiefly located on the cell surface by double-labeling immunocytochemistry (Fig. 4d–f). Next, we quantitatively studied Kir4.1 and AQP4 mRNA levels in AS after exposure to TNF-α, IL-1β or TGF-β1. The levels of expression of Kir4.1 and AQP4 were not altered in AS exposed to TNF-α or IL-1β (Fig. 5d,e) Although a trend for upregulation of Kir4.1 and downregulation of AQP4 was found in AS after exposure to TGF-β1, these alterations did not reach statistical significance. However, AS exposed to TGF-β1 showed a 1.49-fold increase in Kir4.1 protein levels by western blot, suggesting that TGF-β1 acts as a potential inducer of Kir4.1 expression in AS (Fig. 5f,g, lane 8).

Figure 5.

Expression of Kir4.1 in cultured human astrocytes. cDNA prepared from cultured human astrocytes (AS) and neuronal progenitor (NP) cells was processed for reverse transcription polymerase chain reaction (RT–PCR). (a) Kir4.1, (b) AQP4 and (c) G3PDH, an internal control. Lanes 1–4 indicate (1) the human frontal cerebral cortex (CBR) with inclusion of the reverse transcription (RT) step, (2) CBR omitting the RT step, (3) AS and (4) NP cells. AS exposed to 50 ng/mL tumor necrosis factor-α TNF-α, interleukin-1β (IL-1β) or transforming growth factor-β1 (TGF-β1), or untreated (CNT) were processed for quantitative RT–PCR (qPCR) and western blot. (d) Kir4.1 and (e) AQP4 expression on qPCR in AS after a 24-h exposure. No statistically significant differences (ns) are found among the treatments by one-way anova with post-hoc Turkey's test. (f) Kir4.1 and (g) Hsp60 expression on western blot of 15 μg of protein isolated from AS after a 48-h exposure. Kir4.1 expression levels normalized by Hsp60 levels, an internal control for protein loading, are indicated in the panel (f). Lanes 5–8 indicate AS (5) untreated or exposed to (6) TNF-α, (7) IL-1β or (8) TGF-β1.


Kir4.1, an inward rectifying potassium channel expressed in renal epithelial cells, inner ear cells and glial cells in the CNS regulates the spatial buffering of extracellular potassium pivotal for glial function and neuronal excitability.[1, 2, 25] Recently, IgG autoantibodies directed to Kir4.1, binding to the first extracellular loop of Kir4.1, were detected in the serum derived from approximately half of MS and CIS patients, although their pathogenic roles remain unknown.[16] The present study, by using immunohistochemistry, showed that reactive astrocytes express Kir4.1 in active demyelinating lesions of MS, active lesion edges of NMO, ischemic lesion edges of cerebral infarction and neurodegenerative lesions of AD. In contrast, infiltrating macrophages, activated microglia and surviving oligodendrocytes do not express Kir4.1. Thus, these observations show that Kir4.1 expression is not lost in active lesions of MS.

We found that cultured human AS express Kir4.1, and the levels of Kir4.1 on AS are not altered by exposure to pro-inflammatory cytokines, TNF-α and IL-1β, but upregulated modestly by TGF-β1, one of the gliosis-inducing cytokines.[26] In the mainstream of post-receptor signaling pathways, TGF-β1 utilizes SMAD proteins as a central signal transducer, whereas both TNF-α and IL-1β activate nuclear factor-κB (NF-κB), being responsible for the differential effects of these cytokines. In contrast to the present results, a recent study showed that IL-1β reduces Kir4.1 expression in cultured human AS isolated from a mixture of aborted fetal brains.[27] The apparent inconsistency is attributable to differences in cultured cells utilized. We found that reactive astrocytes accumulated in active MS lesions coexpress Kir4.1 and AQP4. A recent study showed that reactive astrocytes surviving in NMO brain lesions express AQP4 located on the cells surface and in cytoplasmic vesicles.[28]

Previous studies showed that Kir4.1 is expressed in both astrocytes and oligodendrocytes in the rat optic nerve,[6] and Kir4.1 knockout mice show an impaired ability to develop oligodendrocytes.[7] In contrast to the observations on rodent models, we did not identify Kir4.1 expression on oligodendrocytes existing in the white matter of MS and non-MS brains. This discrepancy is attributable to differences in the species, regions, and the stages of development and maturation. Supporting our observations, a previous study showed that Kir4.1 is expressed mainly in astrocytes surrounding synapses and blood vessels, but undetectable in the white matter in the adult rat brain.[1] In MS and non-MS brains, we identified Kir4.1 expression in vascular smooth muscle cells of larger vessels, in addition to vascular end-feet of astrocytes, being consistent with a recent study carried out on Kir4.1 expression in mouse mesenteric vascular smooth muscle cells.[29]

Increasing evidence indicates that Kir4.1 expression is regulated in a cell type-specific manner dependent on cellular growth and differentiation, and in response to stress-inducing insults. A previous study showed that Kir4.1 channels are absent in immature, proliferating glial cells.[30] In these cells, upregulated expression of Kir4.1 correlates with astrocyte differentiation characterized by the establishment of a negative membrane potential and cell growth arrest. The levels of Kir4.1 expression are elevated in human astrocytic tumors with an increasing pathological grade.[31] Reactive gliosis induces redistribution of Kir4.1 in Müller cells from end-feet to cellular fibers, accompanied by inactivation of Kir4.1 channel function in a rabbit model of proliferative vitreoretinopathy.[32] Spinal cord injury causes a widespread loss of Kir4.1 and glutamate transporter 1 expression on astrocytes in situ.[33] Downregulation of Kir4.1 by RNAi in cultured rat astrocytes inhibits not only the potassium transfer, but also the glutamate uptake.[5] Progressive loss of Kir4.1 is identified in the spinal cord of the SOD1-G93A transgenic mouse model of amyotrophic lateral sclerosis (ALS), suggesting that higher extracellular concentrations of potassium and glutamate are responsible for induction of motor neuron cell death.[34]

We identified coexpression of Kir4.1 and AQP4 in reactive astrocytes in active MS lesions, supporting a functional interaction between Kir4.1 and AQP4 in the regulation of potassium and water transport.[3, 4] A previous study showed that cerebral amyloid angiopathy (CAA) causes a loss of expression of both Kir4.1 and AQP4 from the astrocyte end-feet.[35] However, we found that some amyloid plaques and perivascular amyloid deposits moderately express Kir4.1. Importantly, Kir4.1 prevents swelling of astroglial processes in experimental spinal cord edema caused by hypotonic osmotic stress, suggesting a protective role of Kir4.1 expressed on astrocyte end-feet against aberrant water transport.[36]

At present, the precise mechanisms remain unknown as to how MS patients generate autoantibodies against KiR4.1, and how these antibodies induce pathogenic effects on demyelinated axons through defective transport of potassium, glutamate and water. Because astrocyte damage, called astrocytopathy, occurs not only in NMO lesions, but also in a subtype of active MS lesions,[37] we could put forward a possible scenario that Kir4.1-expressing reactive astrocytes damaged in active and aggressive MS lesions serve as a continuous source of immunogens, leading to sensitization of autoreactive lymphocytes in MS.


All autopsied brain samples were obtained from the Research Resource Network (RRN), Japan. This work was supported by grants to J-IS from Research on Intractable Diseases (H21-Nanchi-Ippan-201 and H22-Nanchi-Ippan-136), the Ministry of Health, Labour and Welfare (MHLW), Japan, and the High-Tech Research Center (HRC) Project (S0801043) and the Genome Research Center (GRC) Project and a Grant-in-Aid (C22500322), the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.