вход по аккаунту


Autoantibodies detected to expressed K+ channels are implicated in neuromyotonia.

код для вставкиСкачать
Autoantibodies Detected to Expressed K+
Channels Are Implicated in Neuromyotonia
Ian K. Hart, MRCP,* Caroline Waters, BSc,* Angela Vincent, MRCPath," Claire Newland, PhD,*
David Beeson, PhD,* Olaf Pongs, PhD,t Christine Morris, PhD,$ and John Newsom-Davis, FRS*
Antibody-mediated autoimmunity underlies a diverse range of disorders, particularly in the nervous system where the
extracellular domains of ion channels and receptors are especially vulnerable targets. We present here a novel means of
detecting autoantibodies where the genes of the suspected target proteins are known, and use it to detect specific
autoantibodies in acquired neuromyotonia (Isaacs' syndrome), a disorder characterized by hyperexcitable motor nerves
and sometimes by central abnormalities.We expressed different human brain voltage-gated potassium channels in Xeizopus oocytes by injecting the relevant a-subunit complementary RNA, and detected antibody binding by immunohistochemistry on frozen sections. Antibodies were detected to one or more human brain voltage-gated potassium channel
in 12 of 12 neuromyotonia patients and none of 18 control subjects. The results establish neuromyotonia as a new
antibody-mediatedchannelopathy and indicate the investigative potential of this molecular immunohistochemical assay.
Hart IK, Waters C, Vincent A, Newland C, Beeson D, Pongs 0, Morris C, Newsom-Davis J.
Autoantibodies detected to expressed K+ channels are implicated in neuromyotonia.
Ann Neurol 1997;41:238-246
Antibody-mediated autoimmunity underlies a diverse
range of disorders, particularly in the nervous and endocrine systems where the extracellular domains of
membrane molecules (such as receptors and ion channels) are especially vulnerable to attack. In autoimmune
neurological disease, defining the antibody target at a
molecular level and detecting the presence of circulating autoantibodies have largely depended on the availability of target-specific neurotoxins. For example, the
use of a-bungarotoxin (a component of the venom of
Bungarus multicinctus) established the muscle nicotinic
acetylcholine receptor (AChR) as the primary target of
autoantibodies in myasthenia gravis (MG), and continues to be the basis of a sensitive radioimmunoassay
for the circulating autoantibodies [ 11. Ligands are not
available for many potential targets, however, and for
others low affinity limits their usefulness. Here, we
present a novel means of detecting autoantibodies
where the genes for the suspected target proteins are
known. We use this technique to detect pathogenic
autoantibodies in acquired neuromyotonia (NMT,
Isaacs' syndrome) and confirm the effectiveness of this
diagnostic approach by showing subunit specificity of
anti-AChR antibodies in MG.
Acquired N M T presents with spontaneous and continuous muscle fiber activity [2]. Patients may experi-
ence muscle stiffness, myokymia (muscle twitching),
cramps, weakness, pseudomyotonia (delayed relaxation), pseudotetany, and increased sweating [ 3 ] .Disability can be very severe. Some patients also have
central nervous system (CNS) symptoms, including insomnia, mood changes, and hallucinations [4].The
electrophysiological hallmark is the spontaneous firing
of single motor units as doublet, triplet, or multiplet
discharges that have a high intraburst frequency (40400 sec-') and occur at irregular intervals [2, 41. Their
persistence following proximal nerve block by local anesthetic established their peripheral nerve origin, and
suggested that in such cases the discharges were arising
in the terminal arborization of motor nerves [2].
Several clinical features have suggested that acquired
N M T is immunologically mediated, including its association with thymoma [4]and other autoimmune diseases, especially MG [ 3 ] . Furthermore, as in MG,
plasma exchange (which removes circulating antibodies) can result in short-term clinical and electromyographic improvement [4],suggesting the presence of a
serum antibody that interferes with the neuronal mechanisms normally controlling peripheral nerve excitability.
A likely candidate for the target antigen in acquired NMT is a neuronal Shaker-related voltage-gated
From the *Neurosciences Group, Institute of Molecular Medicine,
Received Mar 5 , 1996, and in revised form Jul 26. Accepted for
publication Aug 5 , 1996.
John Radcliffe Hospital, Oxford; tZentrum fur Molekulare Neurobiologie, Univrrsitat Hamburg, Germany; and
Clinical Neurology, Radcliffe Infirmary, Oxford, United Kingdom.
science, The
to Dr
for Hart,
and N of Neurological
Lane, Liverpool L9 lAE, United Kingdom.
Copyright 0 1997 by the American Neurological Association
potassium channel (VGKC) with either A-type or delayed-rectifier electrophysiological properties [5-7].
Immunocytochemical studies of mouse brain using antibodies raised against amino acid sequences specific to
one or the other of two of these VGKCs suggested that
these channels are concentrated in the paranodal and
terminal regions of myelinated axoils [8]. Furthermore,
electrophysiological studies of isolated human [9] and
rodent [lo, 111 peripheral nerve reported Kt currents in the paranodal region with electrophysiological
and pharmacological properties similar to cloned
Shaker-related VGKCs expressed in Xenopus oocytes
18, 121.
These considerations led us to look for antibodies
to VGKCs in patients with acquired NMT. VGKCs
are the most heterogeneous of the family of voltagegated cation channels [12]. At least six Shaker-related
VGKC a-subunits have been isolated from human
brain, each encoded by a different gene [12]. A functional VGKC consists of four transmembrane asubunits that combine as homomultimeric or heteromultimeric tetramers [8] to interact with intracellular
P-subunits, which are also thought to form a tetramer
[ 131. As specific ligands, for example, a-dendrotoxin
(a-Dtx), have been identified for only some of the
VGKC subtypes, we achieved target specificity by injecting Xenopus oocytes with the complementary RNA
(cRNA) for a particular VGKC a-subunit, and used
both radioimmunoprecipitation and immunohistochemistry to detect antibody binding to the expressed
Materials and Methods
Clinical Material
Sera from 12 patients with acquired N M T were tested. Clinical details have been described for 6 of them (Patients 1, 3,
5, 8, 9, and 12) [4, 141. None of the 12 had a family history
of myokymia. Three patients had MG, thymoma, and raised
anti-AChR antibody titers; all other patients were anti-AChR
antibody negative and had no evidence of thymoma. All except 1 patient (Patient 9) had myokymia and 10 of the 12
reported excessive muscle cramps. Several patients had muscle weakness, including the 3 with MG. All patients had the
characteristic electromyographic discharges seen in NMT,
consisting of doublet, triplet, or multiplet single motor unit
discharges having a high intraburst frequency (40-400/second). Five patients had evidence of a mild neuropathy; this was
a motor-sensory demyelinating neuropathy in 3, a multifocal
neuropathy in 1, and z o n a l neuropathy in 1. Some of these
patients experienced occasional distal paresthesias.
In the immunoprecipitation assays, 18 control sera were
used: 6 from healthy subjects, 6 from patients with antiAChR antibody-positive MG, and 6 from patients with
anti-voltage-gated calcium channel antibody-positive Lambert-Eaton myasthenic syndrome. In the molecular immunohistochemical assay some of the above control sera were used,
and also sera from 4 patients with amyotrophic lateral
sclerosis, 3 with acute inflammatory demyelinating polyneu-
ropathy (high titer of anti-GM, antibodies in I), and 1 with
systemic lupus erythematosus (antinuclear antibody titer
1 : 1,600).
Isolation o f VGKC a-Subunit Complementay
DNAs and Production of Anti- VGKC
Peptide Antibodies
Complementary DNAs (cDNAs) for VGKC a-subunit for
human K"1.2 and 1.6 (KCNA2 and KCNAG) [15] and for
rat K"1.4 and 2.1 (RCK4 and DRKl), and AChR a-subunit
cDNA [16] were isolated as described previously. KCNAla
cDNA was isolated using polymerase chain reaction (PCR)
and oligonucleotide primers against published, highly conserved sequences of rodent brain VGKC genes. We amplified
a product from cDNA derived from a human small-cell lung
carcinoma (SCLC) cell line [17] and used it to probe a
SCLC hgtl 1 cDNA library from which we obtained a fulllength 1,488-bp cDNA clone for KCNAla. The sequence
of this cDNA shows only two nucleotide changes from the
VGKC cDNA KCNAl that has been isolated from human
brain [18]. Compared with the predicted KCNAl amino
acid sequence, KCNAl a has isoleucine instead of methionine
at position 324 and serine instead of tyrosine at position
To produce polyclonal anti-VGKC peptide antibodies,
we injected a New Zealand White rabbit with an oligopeptide with the sequence LGTEIAEQEGNQKGEQATSLAILRV corresponding to the predicted S3-4 extracellular domain of both KCNAl and KCNAla using a method described previously [ 191.
Complementaty RNA Synthesis and Expression
in Oocytes
cRNA was produced from KCNA2, KCNAG, RCK4,
DRKl, and AChR a-subunit cDNAs as described previously
[15, 161. KCNAla cDNA was subcloned into the polylinker
of the pGEM-3Z plasmid (Promega) transcription vector
containing poly A sequences detailed elsewhere [ 161. Capped
transcription reactions were performed with 5 to 10 pg of
EcoR1-linearized template D N A 625 pM adenosine (ATP),
cycosine (CTP), and uridine (UTP) triphosphates; 125 pM
guanosine triphosphate (GTP); 625 p M 'G(S')ppp(S')G cap
analogue; and SO units of SP6 RNA polymerase in standard
transcription buffer (Promega).
Stage 4 and 5 oocytes were dissected from anesthetized
adult Xenopus leavis, injected with 50 nl of cRNA (0.5 gm/
liter) or water, and incubated with or without j5S-methionine
(Promega, 10 pCi/ml) as described previously [20]. Oocytes
from batches injected with each type of cRNA were tested
for channel expression by the two-electrode voltage-clamp
Molecular Immunohistochemical Assay
Oocytes injected with one type of cRNA were placed in embedding medium (Cryo-M-bed, Bright Instruments), frozen
in dry ice, and stored at -70°C. Eight-micrometer-thick sections of these oocytes were cut on a cryostat (Bright Instruments), mounted onto glass slides coated with diaminosilane
(BDH Ltd), fixed in paraformaldehyde (2Yo in phosphatebuffered saline [PBS] solution, pH 7.4) for 5 minutes at
4"C, and stored at -70°C until use.
Hart et' al: Kt Channel Antibodies in N M T
The frozen, fixed section5 were slowly warmed to room
temperature, rehydrated in PBS solution for 10 minutes,
treated with hydrogen peroxide (H,O,; 1.5% in distilled
H,O) for 10 minutes, and washed once in d H 2 0 and three
times in PBS solution. Human serum diluted in PBS solution was applied to the sections, followed by biotin-coupled
goat anti-human or sheep anti-rabbit IgG (Dako, diluted
1 :200 in PBS solution containing 1% fetal calf serum [PBSFCS]), followed by streptavidin-coupled horseradish peroxidase (Dako, 1 :200 in PBS-FCS) for 30 minutes each at
room temperature. The sections were washed three times in
PBS after each step. After incubation in 3-amino, 9-ethylcarbazole (AEC) (Dako) substrate (20 pl of AEC and 15 p1 of
H z 0 2in 20 ml of citrate buffer, p H 5.5; filtered through a
25-pm Acrodisc to remove any precipitate) for 15 to 20
minutes at 37"C, the sections were counterstained with Myer's hematoxylin/eosin stain for 60 to 90 seconds at room
temperature and washed in d H 2 0 and then PBS solution.
Slides were mounted in a glycerol-based medium (Glycergel;
Dako), examined in a light microscope (Olympus) using a
10X or 40X objective and photographed using ASA 400
Tri-X film (Easrman Kodak).
T o prepare membranes from human brain, 2- to 3-cm' sections of frontal cortex were added to 20 ml of ice-cold buffer
(25 mM Tris-hydrochloric acid [HCI], 5 m M HEPES, containing 0.32 M sucrose, 2 p M leupeptin, 1 pM pepstatin,
0.1 m M phenylmethylsulfonylfluoride, and 20 pg/ml of
soyabean trypsin inhibitor). The sections were homogenized
with a Polytron PT10-35 homogenizer (Kinematica GmbH)
and centrifuged for 10 minutes at 1,000 rpm. The supernatants were removed and recentrifuged for 10 minutes at
18,000 rpm and the pellets were resuspended at a final protein concentration of 20 to 25 mglml and stored at -70°C.
For use in immunoprecipitation assays, the membrane preparation was diluted 1 :5 in buffer (20 m M Tris, 100 m M
sodium chloride [NaCl], 5 mM potassium chloride [KCI],
and 1 mg/ml of bovine serum albumin, p H adjusted to 7.4
with HC1) containing digitonin (1% final wtlvol), incubated
at 37°C for 15 minutes, and centrifuged for 30 minutes at
15,000 rpm. The supernatant was removed and used as a
source of VGKCs in the assay.
T o extract recombinant VGKCs, cRNA-injected oocytes
were homogenized as above in buffer (20 m M Tris, 100 m M
NaCI, 5 mM KCI, p H adjusted to 7.4 with HCI, containing
1% wt/vol digitonin, 2 pM leupeptin, 1 p M pepstatin, 0.1
mM phenylmethylsulfonylfluoride, and 20 pg/ml of soyabean trypsin inhibitor; 50 pl/oocyte). After centrifugation
for 10 minutes at 10,000 rpm, the resulting middle layer
was removed and used as VGKC extract in the assay.
Aliquots of brain or oocyte extract were incubated with a
saturating concentration of ' Wabeled-a-Dtx ( I nM, Amersham) with or without a 100-fold excess of cold a-Dtx (Latoxan) at room temperature for 1 hour. Five microliters of
serum or IgG diluted 1 : 10 in 20 mM phosphate buffer, p H
7.4, and 0.1% Triton X-100 (PTX buffer) was added to 50
pl of brain or 150 pl of oocyte labeled extract. After incubation for 2 hours at rooni temperature, 100 pl of goat anti-
Annals of Neurology
The results of the immunoprecipitation assays are expressed
as the mean 2 standard error of the mean (SEM). The unpaired Student's t test was used to analyze possible statistical
Extraction of VGKCs and
Immunoprecipitation Assays
human IgG antibody (diluted 1:1 in PTX) was added to
each sample. One milliliter of PTX was added after a precipitate had formed (20-30 minutes) and the samples were centrifuged at 13,000 rpm for 3 minutes. The pellets were
washed twice in PTX and counted in a gamma counter
(Canberra Packard). Results were expressed as picomoles of
'*'I-labeled a-Dtx binding sites precipitated per liter of serum. Each serum sample was assayed blind three times.
For the non-Dtx-binding VGKCs (KCNAl a, RCK4, and
DRKl), aliquots of 3iS-methionine-labeled VGKCs extracted from oocytes injected with the relevant cRNA were
immunoprecipitated using the method described above, except that 30 pl of St~~ph~dororcw
azmw protein A (GibcoBIU; washed, spun, and resuspended in an equal volume of
20 mM Tris, 100 m M NaCI, and 5 m M KCI, with pH
adjusted to 7.4with HCI) was used in place of goat antihuman IgG antibody. The final precipitate was resuspended
in 1 vol Laemmli sample buffer, incubated at 60°C for 2
minutes, and loaded onto a 10% sodium dodecyl sulfatepolyacrylamjde gel (SDS-PAGE).
Vol 41
No 2
February 1997
Serum Antibodies to '251-a-Dendrotoxin-LabeLed
Brain VGKCs Detected by Immunopreripitation
W e tested NMT sera for antibodies to human brain
VGKCs using immunoprecipitation of VGKCs extracted from human frontal cortex and labeled with
'151-a-Dtx. Six of 12 NMT sera showed precipitation
of the labeled complex (Fig la) that was greater than
the mean plus 3 standard deviations (SDs) for control
sera (>115 pmol of a-Dtx binding sitedliter of serum). Twelve sera from patients with other autoimmune neurological diseases and 6 from healthy individuals were negative. These results suggest that some
NMT patients have anti-VGKC antibodies. However,
the titers of anti-VGKC antibodies in NMT were low
and 50% of patients' sera were negative, suggesting
that a more sensitive technique was needed.
VGKC Genes Expressed in Xenopus Oocytes as a
Source of Recombinant Antigen
To characterize the anti-VGKC antibodies further, and
to try to improve the detection rate of the assay, we
first injected Xenopus oocytes with the cRNA for the
KCNAG a-subunit, a human VGKC gene whose product binds a-Dtx. The homomultimeric KCNAG channels were solubilized, labeled with Iz5I-a-Dtx,and used
as antigen in the iinmunoprecipitation assay. Four of
the 6 NMT sera positive with '251-a-Dtx-labeled native brain VGKCs were also positive (>30 pmol of
a - D t x binding sitedliter) using KCNAG (Fig Ib).
Although titers were lower using the recombinant
I '
u $4,
k 0 >
K J .4
Fig I . "jI-a-Dendrotoxin immunoprecipitation assays for anti-voltage-gated potassium channel (VGKC) antibodies (Ab) in
patients with neuromyotonia (NMT) and in controls. (a) The antigen is native VGKC extractedfiom human brain. (b, c) The
antigen is recombinant VGKC extracted fiom Xenopus oocytes injected with the relevant complementaiy RNA. Antibody titers
were considered positive fi more than the mean plus 3 standard deviations for the control sera (>I15 pmol in [a] or >30 pmol
in [b]). (c) The results are expressed as means 2 SEM of three assays: p < 0.001 for NMT IgG binding to KCNA6 and p <
0.01 for binding to KCNA2 compared with control IgG (Student? t test). Nonspecifir binding was determined by a parallel assay
in which a IOO-jold excess of unhbeled a-dendrotoxin was added initially. The results shown are total counts minus nonspeczfir
counts. MG = myasthenia gravis; LEMS = Lam bert-Eaton myasthenic syndrome; HC = healthy control.
antigen, this experiment nevertheless suggested that
KCNAG VGKC a-subunits may be a target for autoantibodies in some NMT patients.
To extend the range of VGKC antigens to be tested,
we used expressed proteins from four different VGKC
genes in the immunoprecipitarion assay. Two of these
(KCNA2 and KCNAG) code for human brain VGKC
a-subunits that bind a-Dtx, and two (RCK4 and
DRK1) code for rat brain VGKC a-subunits that do
not bind a-Dtx. These antigens were evaluated using
the IgG fraction of the N M T serum that had shown
the highest titer in the previous assays (Serum 1, Table). Titers for the N M T IgG with KCNAG and
KCNA2 gene products were 93 pmol and 41 pmol,
respectively, significantly higher than control titers of
9.7 pmol and 7.5 pmol ( p < 0.001 and p < 0.01, n
= 3, Fig lc). These results suggested that the N M T
patient had antibodies to more than one type of
VGKC and may explain why two N M T sera bound
to a mixed population of brain VGKCs (see Fig la)
but not to recombinant KCNAG product alone (see
Fig lb).
To determine whether N M T sera contain antibodies
to types of VGKCs that do not bind a-Dtx, we first
used 35S-methionine-labeled recombinant VGKC asubunits expressed from RCK4 and DRKl, and also
from a human VGKC a-subunit cRNA (KCNAla)
isolated from a SCLC cell line. Homomultimeric
Hart et al: K+ Channel Antibodies in NMT
Serum Antibodies to Voltage-Gated Potassium Channels (VGKC) in Acquired Neuromyotonia (NMT):
Comparison of Immunoprecipitation and Molecular lmmunohistochemical Assays
Imrnunoprecipitation Assay
Antibody Titer (pmol)
Native VGKC
Immunohistochemical Assay
Imniunoreactivity Pattern
NMT serum no
- 10
- 24
Rabbit anti-KCNAla peptide antibody
O N D controls (mean, n = 12)
Healthy controls (mean, n = 6)
OND = other neurological diseases; ND = not done
VGKCs were extracted from oocytes that had been injected with the relevant cRNA and incubated in 3iSmethionine for 3 days. Autoradiography of precipitates
run on SDS-PAGE showed that none of the 12 NMT
and 12 control sera bound to any of the three VGKC
a-subunits (not shown). Under the same conditions,
serum from a rabbit immunized with a synthetic peptide encoding the amino acid sequence of the S3-4
extracellular domain unique to KCNAl a precipitated
KCNAla product but not RCK4 or DRKl (not
shown). One possible explanation for these results is
that only VGKCs binding a-Dtx are targets for serum
anti-VGKC antibodies in NMT. However, it is also
possible that this assay, which uses antigen denatured
by detergent extraction, while able to demonstrate antibodies raised against peptides, cannot detect antibodies
that recognize conformationally dependent determinants prescnt on full-length VGKCs in vivo.
Serum Antibodies to Human Recombinant VGKC
Subtypes Detected by hnmunohistochmisty
As an alternative approach, we developed an assay in
which immunohistochemical staining was used to de-
tect autoantibody binding to human VGKC a-subunit
proteins expressed in Xenopus oocytes. NMT sera were
incubated with frozen sections of oocytes injected 3 to
5 days previously with KCNA2, KCNA6, or KCNAla
cRNA. Oocytes injected with human AChR a-subunit
cRNA, or with water, served as controls. The functional membrane expression in oocytes of the channel
protein from each cRNA batch was confirmed electrophysiologically (not shown). Antibody binding was detected by treating the sections with biotin-coupled goat
anti-human or sheep anti-rabbit IgG followed by
streptavidin coupled to horseradish peroxidase. The results were confirmed independently in coded samples.
First, we tested a polyclonal antiserum raised in a
rabbit to a synthetic peptide encoding the S3-4 extracellular domain unique to KCNAl a against sections
of KCNAl a-expressing oocytes. Specific staining was
confined to the cytoplasmic vesicles of the oocyte and
was strongest in the vesicles’ surface membrane (Fig
2b). In addition, staining was maximal at the vegetal
pole of the oocyte around the site where cRNA was
injected. This pattern was similar to that reported in
another oocyte assay designed to detect monoclonal an-
Fig 2. Molecular irnmunohistochemical assay f i r antibodies to voltage-gated potassium channels (VGKCs) or acetylcholine receptor
(AChR) a-subunits expressed in Xenopus oocytes previously injected with the relevant complementavy RNA (cRNA). (a, c-f)
cRNA for human brain VGKC (KCNAC;)injected. (6) cRNA for human small-cell VGKC (KCNAla) injected. (g. h) cRNA f i r
ACbR a-subunit injected. (a, 6) Rabbit anti-KCNAla S3-4 peptide antibody. (d j g) Neuromyotonia serum. (c, e) LambertEaton myasthenic syndrome serum. (h) Anti-AChR antibody-positive myasthenia gravis serum. (c, d) Low power (X 10).
(a, b, e-h) High power (X 40). Note the redlbrown horseradish peroxidase staining of the cytoplasmic vesicles in (b), (d), cf,,
and (b) and its absence in sections (a), (c), (e), and (g). Scale bar = 100 p m in (c) and (4,
and 25 p m in (a), (b), and (e)
through (h).
Annals of Neurology
Vol 41
No 2
February 1937
Hart et al: K' Channel Antibodies in NMT
tibodies to rat VGKCs [20].Rabbit anti-KCNAla peptide serum did not bind to the cytoplasm of sections
of oocytes that had been injected with KCNA2,
KCNAG, or AChR a-subunits or with water (e.g., see
Fig 2a). By contrast, rabbit serum (and all human sera,
see below) reacted strongly with the vitellin envelope
of all the oocytes studied, even those injected with water. This eiivelope surrounds the plasma membrane and
contains hydrophobic proteins that may bind antibodies nonspecifically [20].
T o minimize any effect of differing expression efficiency of VGKCs between oocytes on serum binding,
we tested all sera against sections from 2 oocytes per
a-subunit. Binding of the two NMT sera with the
strongest immunoreactivity (Sera 1 and 2, see Table)
was detected down to the same serum dilutions in sections from each of the 2 oocytes for all three VGKC
a-subunits, suggesting that there was equal expression
within oocyte pairs. The results were reproducible as
we obtained similar serum binding patterns in 50 to
I00 oocytes per a-subunit (not shown).
Each of the 12 N M T sera tested bound to at least
one type of expressed VGKC and shared the same
staining characteristics as the rabbit anti-KCNAla peptide antisera (see Fig 2). Immunoreactivity to KCNAG
was found in 11 of 12 NMT sera and persisted to
serum dilutions ranging from 1 : 128 to 1:2,500 (see
Figs 2d, 2f). Ten of 12 N M T sera bound to KCNA2
product at dilutions of 1 :32 to 1 : 1,024 and 5 of
12 sera bound to KCNAl a at dilutions of 1 : 32 to
1:512. None of the 12 NMT sera tested bound to
water-injected oocytes. None of 14 sera from patients
with various neurological diseases and none of 6 sera
from healthy controls bound to sections of oocytes that
had been injected with cRNAs for KCNA6, KCNA2,
or KCNAla, or AChR a-subunit or with water (e.g.,
see Figs 2c, 2e).
To establish whether the assay could be applied to
other autoantibodies, we tested 6 sera from M G patients known to be positive for anti-AChR antibodies
in the standard '"I-a-bungarotoxin radioinimuiioassay
[I]. Three of these patients also had NMT (Sera 4,
5 , and 7,see Table). All 6 sera bound to sections expressing AChR a-subunits (see Fig 2h), but only the
sera from the 3 patients who also had N M T bound to
sections expressing KCNA6, KCNA2, or KCNAla.
Sera from the patients with MG alone did not bind
to sections of oocytes expressing any of the VGKCs,
and sera from the 9 patients with NMT alone did not
bind to sections expressing AChR a-subunits (e.g., see
Fig 2g).
We detected serum autoantibodies to recombinant human brain VGKCs in 12 of 12 patients with clinically
244 Annals of Neurology Vol 41
No 2
February 1997
definite, acquired NMT using a novel adaptation of a
molecular immunohistochemical assay. This is an important step in establishing anti-VGKC antibodies as
the causative factor in this disorder. This approach, in
which the injected cRNA provides the target specificity
for the assay, is also a means of detecting and characterizing antibodies in other autoimmune diseases where
the gene for the putative antigen is known, as we have
shown here for anti-AChR antibodies in MG. Our previous in vitro studies of neuromuscular transmission in
mice injected with NMT IgG showed an increase in
the number of quanta of acetylcholine released by a
nerve impulse compared with controls, implying prolongation of the action potential [14, 211. This functional effect of N M T IgG mimics agents that block
VGKCs such as 4-aminopyridine [22] and a-Dtx [ 11,
231, which thereby interfere with nerve repolarization.
Taken together, the findings provide persuasive evidence chat a reduction in the number of functional
VGKCs caused by anti-VGKC antibodies underlies the
peripheral motor nerve hyperexcitability that characterizes acquired NMT.
Some NMT patients have oligoclonal bands in their
cerebrospinal fluid (CSF) [4].In preliminary experiments, we found that CSF IgG from several of these
patients binds to neurons in the dentate nucleus of
the human cerebellum (not shown). Since VGKCs are
present in both neurons [24] and glial cells [25]
throughout the CNS, we speculate that anti-VGKC
antibodies also contribute to the central disorders seen
in some NMT patients.
The ability of the molecular immunohistochemical
assay to test different gene products individually allows
precise identification of the antigenic target. We found
that 10 of 12 NMT sera contained antibodies to more
than one VGKC subtype. Four sera bound to all 3
VGKC gene products tested (KCNAla, KCNA2, and
KCNA6), 6 bound to various combinations of two of
these, and 2 only to one (KCNAG). In addition, these
findings demonstrate that NMT sera vary widely in
their immunoreactivity to the three VGKCs studied
and suggest that anti-VGKC antibodies in N M T may
be heterogeneous in their fine specificities (as antiAChR antibodies are in MG), binding not only to determinants found on one particular VGKC subtype
but perhaps also to several determinants on the same
VGKC subtypes.
The presence of antibodies with different target
specificities in N M T patients may also provide a molecular explanation for the clinical diversity of this syndrome. NMT patients by definition have motor nerve
hyperexcitability. However, 30% also have symptoms
of sensory nerve overactivity such as paresthesia [26]
and some have CNS disorders. There is evidence from
both the central and the peripheral nervous system that
neurons express K+ channels in a cell-type specific
manner. In the rodent brain, in situ hybridization and
immunocytochemical studies [24, 27, 281 suggest that
neurons vary widely in the types of VGKC a-subunits
that they express. Electrophysiological recordings also
show that peripheral motor nerves display different K' currents from sensory nerves [29, 301. Thus,
the VGKC population may be subject to antibodymediated attack that is subtype specific or subtype
dominant. Some NMT patients, for example, may
have antibodies that bind only to VGKCs expressed
on motor nerves, while others may also have antibodies
to VGKCs expressed on sensory nerves or dorsal root
ganglion cell bodies, or on CNS neurons or glial cells.
It seems possible that in vivo there may be even
greater anti-VGKC antibody heterogeneity than we
demonstrated here in the molecular immunohistochemical assay. Functional VGKCs contain four asubunits that combine as homomultimers and heteromultimers [8], whereas we studied antibody binding to
cytoplasmic-expressed a-subunits that exist mainly as
monomers. Therefore, we may have failed to detect
serum antibodies that, for example, bind to conformationally dependent epitopes arising at the interface
of a-subunits. Furthermore, the important question
whether antibodies to epitopes on a single a-subunit
can bind to one or more classes of heteromultimeric
VGKC containing that subunit remains to be answered.
Although NMT is usually an acquired disorder, it
may also occur in hereditary diseases affecting peripheral nerve function [31]. Of particular relevance to the
present study is the autosomal dominant syndrome of
familial episodic ataxia that is caused by mutations of
the neuronal VGKC a-subunit gene KCNAl located
on chromosome 12 [32]. The predicted amino acid
sequence of the product of this gene [18] differs in
only two amino acids from KCNAla. We showed that
5 of 12 NMT sera bind to this protein expressed in
oocyte sections. Therefore, it appears that antibodies
binding to VGKCs in acquired NMT have the same
effect on peripheral nerve function that KCNAl missense mutations have in this inherited form of NMT.
The molecular immunohistochemical assay detected
anti-VGKC antibodies in all of the NMT sera tested,
whereas the immunoprecipitation assays were only able
to demonstrate antibodies to a-Drx binding VGKCs
in about half these sera (see Table). The new assay
differs from existing methods of autoantibody detection in several ways. It allows study of antibody binding to a single isoform of a target antigen that is expressed at very high level; the antigen is presented
within the oocyte cytoplasm in a more natural state as
it is not denatured by detergent extraction; and a ligand is not required to label the expressed protein as
binding specificity is provided at the molecular level.
These advantages help explain both the increased sensitivity of the new assay and its ability to discriminate
between antibody binding to closely related antigens.
Cation channels are fundamental to the normal
working of the nervous system and functional abnormalities of all the major types of these channels are
associated with neurological disease [33].In the peripheral nervous system these channels are particularly
vulnerable to autoantibody attack: AChRs are the target in MG and voltage-gated calcium channels in the
Lambert-Eaton myasthenic syndrome. More recently,
anti-GM, antibodies detected in some patients with the
Guillain-Bar& syndrome were shown in vitro to both
inhibit voltage-gated Na' currents [34] and block
nerve conduction [35]. Our results identifying antiVGKC antibodies in NMT, together with our earlier
studies [5-7, 14, 211, establish this disorder as another
antibody-mediated ion channel syndrome, or channelopathy. The molecular immunohistochemical assay
that we developed to show this may similarly elucidate
other autoimmune diseases where the antibody specificities are unknown.
This work was supported by the Medical Research Council (MRC).
Dr Hart holds an MRC Clinician Scientist Fellowship.
We thank Leslie Jacobsen for synthesizing the S3-4 peptide and
Katherine Leys and Beth Lang for advice. We are grateful to the
many colleagues who allowed us to study their cases, particularly
Drs A. Bajhlan, G. Chauylannaz, K. Cumming, F. Heidenreich, A.
Papadernetriou, W. Schady, and A. Winrzen.
1. Vincent A, Newsom-Davis J. Acetylcholine receptor antibody
as a diagnostic test for rnyasthenia gravis: results in 153 validated cases and 2967 diagnostic assays. J Ncurol Neurosurg
Psychiatry 1985;48:1246-1252
2. Isaacs H. A syndrome of continuous muscle-fibre activity. J
Neurol Neurosurg Psychiatry 1961;24:319-325
3. Hart IK, Newsom-Davis J. Neuromyotonia (Isaacs' syndrome).
In: Lane RJM, ed. Handbook of muscle diseases. New York:
Marcel Dekker, 1996:355-363
4. Newsom-Davis J, Mills KR. Immunological associations of acquired neuromyotonia (Isaacs' syndrome). Report of five cases
and literature review. Brain 1993;116:453-469
5. Hart IK, Leys K, Vincent A, et al. Autoantibodies to voltagegated potassium channels in acquired neurornyotonia. Neuromuscul Disord 1394;4:535
6. Hart IK, Vincent A, Leys K, et al. Serum autoantibodies bind
to voltage-gated potassium channels in acquired neuromyotonia. Ann Neurol 1995;36:325
7. Hart IK, Waters C, Vincent A, et al. Serum anti-voltage-gated
potassium channels detected in acquired neuromyotonia. J
Neurol Neurosurg Psychiatry 1996;60:113
8. Wang H, Kunkel DD, Martin TM, et al. Heterornultimeric
potassium channels in terminal and juxtaparanodal regions of
neurons. Nature 1993;365:75-79
9. Bostock H, Baker M. Evidence for two types of potassium
Hart et al:
K+ Channel Antibodies
in NMT
channel in human motor axons in vivo. Brain Res 1988;462:
10. Baker M , Bostock H , Grafe H , Martius P. Function and distribution of three types of rectifying channels in rat spinal roor
myelinated axons. J Physiol (Lond) 1987;383:45-67
11. Roeper J, Schwarz JR. Heterogeneous distribution of fast and
slow potassium channels in myelinated rat nerve fibres. J Physiol (Lond) 1989;416:93-110
12. Pongs 0 . Molecular biology of voltage-dependent potassium
channels. Physiol Rev 1992;72(suppl 4):S69-S88
13. Rettig J, Heinemann SH, Wunder F, et al. Inactivation properties of voltage-gated potassium channels altered by presence of
P-subunit. Nature 1994;369:289-294
14. Shillito P, Molenaar PC, Vincent A, et al. Acquired neuromyotonia: evidence for autoantibodies directed against Kt channels
of peripheral nerves. Ann Neurol 1995;38:714-722
15. Grupe A, Schrtiter KH, Ruppersberg M, et al. Cloning and
expression of a human voltage-gated potassium channel. A
novel member of the RCK potassium channel family. EMBO
J 1990;9:1749-1756
16. Beeson D, Morris A, Vincent A, Newsom-Davis J. T h e human
muscle nicotinic acetylcholine receptor alpha-subunit exists as
two isoforms: a novel exon. EMBO J 1990;9:2101-2106
17. Johnston I, Lang 8 , Leys K, Vincent A. Heterogeneity of calcium channel autoantibodies detected using a small-cell lung
carcinoma line derived from a Lamberr-Eaton myasthenic syndrome patient. Neurology 1994;44:334-338
18. Ramaswami R, Gautam M , Kamb A, et al. Human potassium
channel genes: molecular cloning and functional expression.
Mol Cell Neurosci 1990;1:214-223
19. Vincent A, Jacobsen L, Shillito P. Response to acetylcholine
receptor a 1 3 8 to 199 determinant spreading initiates autoimmunicy to self antigen in rabbits. Inimunol Lett 1994;39:269275
20. Reinhardt-Maelicke S, Kurz S, Sewing S, et al. Application
of an ectopic expression system for the selection of proteinisoform-specific antibodies. Eur J Biochem l993;216:871a77
21. Sinha S, Newsom-Davis J, Mills KR, et al. Autoimmune aetiology for acquired neuromyotonia (Isaacs' syndrome). Lancet
22. Fu W M , Lin RH, Lin-Shiau SY. Study on the ncuromuscular
246 Annals of Neurology
Vol 41
No 2
February 1997
action of 4-aminopyridine in the mouse diaphragm. Arch Int
Pharmacodyn Ther 1987;289:198-211
23. Harvey AL, Anderson AJ. Dendroroxins: snake toxins that
block potassium channels and facilitate neurotransmitter release. In: Harvey AL, ed. Snake toxing. New York: Pergamon,
1991: 1 3 1-164
24. Drewe JA, Verma S, Frech G, Joho RH. Distinct spatial and
temporal patterns of potassium channel mRNAs from different
subfamilies. J Neurosci 1992;12:538-548
25. Barres BA, Chun LLY, Corey DP. Ion channel expression by
white matter glia: type-2 astrocytes and oligodendrocyres. Glia
1988; 1: 10-30
26. Jamieson PW, Katirji MB. ldiopathic generalized myokymia.
Muscle Nerve 1994;17:42-51
27. Sheng M, Tsaur M-L, Jan YN, Jan LY. Subcellular segregation
of two A-type Kt channel proteins in rat central neurons.
Neuron 1992;9:271-284
28. Veh RW, Lichtinghagen K, Sewing S, et al. Immunohistochemical localization of five members of the K, 1 channel subunits: contrasting subcellular locarions and neuron-specific colocalizations in rat brain. Eur J Neurosci 1995;7:2189-2205
23. Applegate C, Burke D. Changes in excitability of human cutaneous afferents following prolonged high frequency stimulation. Brain 1989;12:147-164
30. Kocsis J D , Bowe C M , Waxman SC. Different effects of 4aminopyridine on sensory and motor fibers: pathogenesis of
paresrhesias. Neurology 1986;36:117- 120
31. Hahn AF, Parkes AW, Bolton CF, Stewart SA. Neuromyotonia
in hereditary motor neuropathy. J Neurol Neurosurg Psychiatry 199 1;54:230-235
32. Browne DL, Gancher ST, Nutt JG, et al. Episodic ataxialmyokymia syndrome is associated with point mutations in the human potassium channel gene KCNAI. Nat Genet 1994;s:
33. Ptacek L. Ion channel Shakedown. Nat Genet 1994;8:111112
34. Takigawa T, Yasuda H, Kikkawa R, et al. Antibodies against
GM, ganglioside affect K' and Na' currents in isolated rat
myelinated nerve fibers. Ann Neurol 1995;37:436-442
35. Roberts M , Willison HJ, Vincent A, Newsom-Davis J.
Multifocal motor neuropathy sera block distal motor nerve
conduction in mice. Ann Neurol 1995;38:111-118
Без категории
Размер файла
1 054 Кб
channel, detected, autoantibodies, expressed, implicated, neuromyotonia
Пожаловаться на содержимое документа