вход по аккаунту


Anti-Ri An antibody associated with paraneoplastic opsoclonus and breast cancer.

код для вставкиСкачать
Anti-&: An Antibody Associated with
Paraneoplastic Opsoclonus and Breast Cancer
F. Antonio Luque, MD, PhD,“I Henry M. Furneaux, PhD,**$Reuvcn Ferziger, AB,”i Marc K. Rosenblum, MD,?
Shirley H. Wray, MD, PhD,§ S. Clifford Schold, Jr, MD,’ Michael J. Glantz, MD,” Kurt A. Jaeckle, MI>,”*
Haim Biran, MD,ff Martin Lesser, MD,I$ William A. Paulsen, MD,§$ Mary E. River, MD,“’
and Jerome B. Posner, MD“$
The serum and cerebrospinal fluid (CSF) of 8 women with ataxia, 6 of whom also had eye movement abnormalities
believed to be opsoclonus, were found to contain a highly specific antineurondl antibody we call anti-Ri. Seven of the
8 women also had or developed cancer: carcinoma of the breast in 5, adenocarcinoma in an axillary lymph node in 1,
and carcinoma of the fallopian tube in 1. Four patients presented with the neurological disorder; the cancer was
diagnosed first in the other 4. Immunohistochemical studies using serum or CSF from all 8 patients revealed a highly
specific antibody interaction with central nervous system neuronal nuclei but not with glial or other cells; the titer
ranged from 1:5,000 to 1:320,000 in serum and from 1:2,000 to 1:16,000 in CSF. Biotinylated IgG from the patients’
serum reacted with the tumors of 3 of 4 patients with anti-Ri antibody but not with breast cancers from patients
without anti-Ri antibody. Immunoblots against cerebral cortex neuronal extracts identified protein antigens of 55-kd
and 80-kd relative molecular mass. Serum titers by immunoblot ranged from 1:500 to more than 1:40,000 and CSF
titers, from 1 : l O to 1:2,000. The relative amount of anti-Ri was always higher in CSF than in serum. The antibody
was not present in sera from normal individuals; patients with breast cancer without opsoclonus; other patients with
opsoclonus; or patients with other paraneoplastic syndromes related to breast, ovarian, or small-cell lung cancer. We
conclude that the presence of anti-Ri antibody identifies a subset of patients with paraneoplastic ataxia and eye
movement disorders (opsoclonus) who usually suffer from breast or other gynecological cancer; the antibody when
present is a useful marker for an underlying malignancy.
Luque FA, Furneaux HM, Ferziger R, Rosenblum MK, Wrav SH, Schold SC Jr, Glantz MJ, Jaeckle KA,
Biran H, Lesser M, Paulsen WA, Rwer ME, Posner JB. Anti-fi: an antibody associated with paraneoplastic
opsoclonus and breast cancer. Ann Neurol 1991;29:241-25 1
The serum and cerebrospinal fluid (CSF) of some patients with neurological paraneoplastic syndromes contain antibodies that react uniquely with portions of the
nervous system and with the underlying cancer. Such
antibodies have been reported in paraneoplastic cerebellar degeneration associated with gynecological cancers (anti-Yo) [l, 21, in encephalomyelitis-sensory neuronopathy associated with small-cell lung cancer (antiHu) [33, in the Lambert-Eaton myasthenic syndrome
associated with small-cell lung cancer 141, and in retinopathy associated with small-cell lung cancer 151. The
presence of these distinctive antibodies in patients with
a neurological disorder often leads to the early diagnosis of an occult malignancy [GI.
Three years ago we identified a patient with “op-
soclonus” and breast cancer who harbored in her serum
an antibody that reacted with nuclei of neurons in the
central nervous system (CNS) (71. Opsoclonus can be
diagnosed at the bedside by the presence of spontaneous, large-amplitude conjugate saccades occurring in all
directions of gaze without a saccadic interval. Torsional
oscillations may be combined with horizontal and vertical oscillations occurring at the same time. When the
back-to-back saccades without a saccadic interval are
limited to the horizontal direction, the disorder is
called ocular flutter. Ocular flutter often occurs in the
recovery phase from opsoclonus. These criteria allow
t h e observer at the bedside to distinguish opsoclonus
and ocular flutter from other eye movement abnormalities such as square wave jerks. Since our first report,
From the *Department of Neurology and the ?Department of Pathology, and George C. Cotzias Laboratory of Neuro-Oncology Memorial Sloan-Kettering Cancer Center, New York, NY; $Department of bieurology and Seurusciences, Cornell University Medical
College, New York, NY; $Department of Neurology, Massachusetts General Hospital, Boston, MA; TDepartment of Neurology,
Duke Universitv Medical Center. Durham. NC: **Debartment of
Neurology, University of Utah, Salt Lake City, UT; ttDepartmenr
of Oncology, Soroka Medical Center, Beer-Sheba, Israel; %Depart-
ment of Neurology, Florida Medical Center, Ft. Lauderdale, FL;
§#EastTennessee Neurological Associates, Knoxville, TN; and “Visalia Medical Clinic, Inc., Visalia, CA.
Received Jun 25, 1990, and in revised form Sep 5. Accepted for
publication Sep 9, 1990.
Address correspondence to Dr Posner, 1275 York Avenue, New
York, N Y 10021.
Copyright 0 1991 by the American Neurological Association
an identical antibody, which w e have designated antiRi (using the first two letters of the last name of our
first patient), was found in the serum of 7 additional
patients who had a neurological syndrome characterized by ataxia and eye movement disorders (usually
opsoclonus). A n underlying breast cancer was identified in 5 of the 8 patients. The antigen (Ri) recognized
by the anti-Ri antibody was expressed in the tumors of
3 of o u r patients. This report describes the clinical and
immunological findings in these patients. A n abstract
of these findings has been published {8].
Material and Methods
Biological Materials
Human CNS (cerebellum, cerebral cortex, etc.) and other
tissues were obtained at autopsy from individuals without
known neurological disease. Frozen sections were used for
immunohistochemistry and partially purified cerebral cortex
neurons for Western blotting. For controls, serum was obtained from normal volunteers and from blood donors. CSF
was obtained at lumbar puncture performed for diagnostic
purposes in parients subsequently found not to have neurological disease. Samples of serum and CSF and tissue specimens from patients with paraneoplastic opsoclonus were collected by their physicians and sent to the Cotzias Laboratory
for assay.
Serial dilutions of serum and CSF were incubated with 7-krnthick frozen sections of histologically normal human cerebral
cortex and cerebellum, using the avidin-biotin immunohistochemical method as described by Hsu and associates “)I.
Negative control samples included those incubated with
phosphate-buffered saline only (PBS) solution 140 mM (sodium chloride [NaCI], 7.5 mM sodium biphosphate
{NaH2P0J, p H 7.2; Difco, Detroit, MI), normal serum, and
normal CSF. Frozen sections were first fixed with cold acetone (stored at 4°C) for 10 minutes, washed twice for 5 minutes each with PBS reacted with 0.3% hydrogen peroxide
for 5 minutes, washed twice for 5 minutes with PBS, and
incubated at room temperature for 15 minutes with 10%
normal goat serum (Cappel; Cooper Biomedical, Malvern,
PA) to suppress nonspecific antibody binding. The excess
goat serum was removed. Individual sections were incubated
with PBS alone, normal serum, or CSF and with patient’s
serum or CSF at different concentrations obtained by serial
dilutions in PBS containing 2% bovine serum albumin
(BSA). Sections were incubated overnight at 4°C. The next
day, sections were individually washed twice for 5 minutes
with PBS and treated with biotinylated goat antihuman IgG
(Vector Laboratories, Burlingame, CA) diluted 1:2,000 in
PBS for 60 minutes at room temperature, and then stained
by the Vectastain avidin-biotin-peroxidase complex (ABC)
(Vector Laboratories) for 30 minutes at room temperature.
The sections were washed with 0.5% Triton X-100 in PBS
for 30 seconds and the substrate reaction developed with
0.05% diaminobenzidine tetrahydrochloride, 0.5% Triton
X-100, and 0.01% hydrogen peroxide in PBS for 2 minutes.
Sections were dehydrated, fixed, and mounted with a glass
cover for examination under the microscope.
242 Annals of Neurology
Vol 29 No 3 March 1991
Western Blotting
Cortical neurons were prepared using a slight modification
of the method of Blomsuand and Hamberger [lo] and Yaragihara and Hamberger El 11employing Ficoll (Pharmacia, Piscataway, NJ) discontinuous density-gradient centrifugation.
The protein content was determined by the Bic-Rad protein
assay (Bio-Rad, Richmond, CA). Partially purified human
cortical neurons were extracted with 0.5 M NaCl and 0.1%
Nomidet P. 40 (NP.40, Sigma), 0.05 M Tris-hydrochloric
acid (HCI), p H 8.8, and 2 mM phenylmethylsulfonylfluoride
(PMSF) for 20 minutes at 4”C, and then centrifuged at 5,800
g for 30 minutes. The pellet was resuspended in buffer (10
mM Tris-HCI, p H 7.4, and 5 mM magnesium chloride
{MgCI)) and Triton X-100 was added to obtain a final concentration of 0.5%. This suspension was vortexed 3 to 5
minutes and centrifuged at 100 g for 5 minutes. The nuclear
pellet was extracted again with 0.5 M NaCl and 0.1% NP40.
All supernatants were pooled together and used for electrophoresis and affinity chromatography.
Cortical neuronal protein extract (approximately 700 kg)
was electrophoresed on a preparative 8% “Laemmli” sodium
dodecyl sulfate (SDS) polyacrylamide gel with 4% stacking
gels C127, then transferred to nitrocellulose paper in a Hoefer
Transphor cell (Hoefer, San Francisco, CA) by the method
of Towbin and colleagues [131. The nitrocellulose sheets
were then blocked with 5% nonfat dry milk in PBS to saturate any remaining protein-binding sites. The nitrocellulose
paper was cut into 6- to 7-mm strips that were incubated
with different dilutions of serum andor CSF for 2 hours at
room temperature. After a thorough wash with 0.1 M Tris,
p H 8.0, 0.2 M NaCI, 0.5cZ Triton X-100, and 0.1% BSA
(TBST), the nitrocellulose strips were incubated with 1251Protein A (0.1 pCdml) in TBST buffer for 1 hour at room
temperature. After stringent washes with TBST, the nitrocellulose strips were allowed to air dry at room temperature
before being apposed to x-ray film for 10 to 12 hours.
The antigens identified in this report are identical in size
(i.e., 55 and 80 kd) to those reported previously by BuddeSteffen and colleagues 173. In this study, we routinely used
chromophore-conjugated protein molecular mass markers
(Amersham, “Rainbow Markers,” Arlington Heights, IL).
Since chromophore-conjugated markers migrate differently
than unconjugated protein molecular mass markers (Amersham Tech Bulletin), the relative molecular masses of the Ri
antigens identified here appear to differ from those in the
original report. In order to maintain consistency, we refer to
these antigens by the original molecular mass estimation.
Qlluntitative Western Blotting
Serial dilutions of sera or CSF were incubated with the nitrocellulose strips (above) containing identical amounts of Ri
antigen. The conditions chosen were those in which there
was a directly proportional linear binding of IgG to the Ri
anngen. The appropriate bands were cut and counted in a
scintillation counter; corresponding pieces of nitrocellulose
incubated with normal sera or buffer alone were counted to
determine the nonspecific binding and the background. A
unit of activity was defined as 1,000 cpm of ‘251-ProteinA
bound to the Ri antigen, after subtraction of background
Table I : Clinical Characteristics oJ Anti-Ri-positive Patients
Age (yr)
Time From
Diagnosis to Tumor
Diagnosis (mo.)
Dizziness, nausea
Breast, 1986
+ 108
Breast, 1978
Dizziness, nausea
Breast, Stage I, 1983
+ (moderate)
Breast (Stage 11-III),
+ (severe)
+ (severe)
+ (moderate)
Dysesthesias, proximal muscle
Spastic quadriparrsis;
hyperreflexic, decreased hearing,
right ear; swallowing difficulty
Dizziness, occasional
diplopia, dysarthria, dementia, cerebral and cerebellar atrophy on MRI
Dizziness, nausea
Dizziness, blepharospasm
Breast, 1983
Other Symptoms
and Signs
6th nerve
No tumor
Purification and Biotinylation of IgG
In order to eliminate background staining from nonspecific
IgG present in tumors, we biotinylated IgG from patients
and control samples. This technique eliminates the need to
use a secondary goat anti-human IgG antibody, thus reducing
background staining. Serum from an anti-Ri-positive patient
was passed through a Protein A-Sepharose (Sigma Laboratories, S t . Louis, MO) column, and then washed extensively
with PBS; the bound IgG was eluted with 0.1 M sodium
citrate, p H 2.5, neutralized, and subsequently dialyzed
against PBS. The purified IgG in PBS as described above
was incubated with biotin N H S (Whydroxysuccinide) long
arm (Vector Laboratories) in an IgG-biotin ratio of 10:l wt/
wt and allowed to react at room temperature for 2 hours.
This material was dialyzed against PBS at 4°C overnight.
Partial PuriJScation of the Ri Antigen
The Ri antigen was partially purified from extracts of cortical
neurons using an immunoaffinity resin containing human
anti-Ri I g G (from Patient 5). Biotinylated IgG in PBS was
mixed with a slurry of streptavidin agarose (Sigma Laboratories) in a ratio of 4.6 mg of biotinylatcd IgGiml of swollen
resin and allowed to rock at 4°C for 12 hours. This mixture
was applied to a column and washed exhaustively first with
PBS 0.1% NP40, then with 0.1 M sodium citrate, pH 2.5,
Axillary, 1788
F,allopian tube, 1389
followed by 0.1% SDS, in order to remove any loosely
bound IgG, and was finally washed extensively with PBS
again. The contaminating human IgG washed off the column
was assayed by dot blot analysis on nitrocellulose paper with
12>I-ProteinA and found to be less than 0.2 pdrnl. Neuronal
protein extract containing 150 mg of protein was mixed with
3 ml of imrnunoaffinity resin, rocked at 4°C for 3 hours,
and then applied to the column. The column was washed
thoroughly with PBS (> 100 bed volumes) and eluted stepwise with 2% SDS; 10 fractions of 2.5 ml each were collected. Fractions containing the Ri antigen were assayed by
Western blor analysis and pooled. The Ri antigen purified
176-fold with a 26% recovery. Polyacrylamide SDS gel electrophoresis followed by silver staining revealed that the partially purified antigen was not homogeneous.
Clinical Findings
The clinical findings are summarized in Table 1. The
clinical findings of Patients 1 and 3 have been detailed
elsewhere E14, 151. All 8 women were ataxic and 7
had ocular motor abnormalities. The ataxia varied in
severity from mild difficulty with gait to severe truncai
ataxia preventing walking and appendicular ataxia limit1.uque et al: Anti-Ri and Parancoplastic Opsoclonus
Table I . Continzled
Clinical Course
Current Status
Frequent falls, gait difficulty
Chemotherapy (CAF), prednisone (improved)
Prednisone (no change)
Gait difficulties, frequent falls,
needs assistance
Gait difficulties (left more than
right), needs assistance
NED (March 1990)
Unsteady gait, MRI scan positive
First surgical resection: NED; Recurred 66 mo later 11988). Radiation therapy to chest wali,
chemotherapy (CAF), prednisone (improved)
Doing well, stable gait, occasional
Controlled save for gingival metastases
Tremor, abnormal handwriting,
gait difficulties, normal head
Lumpectomy, chemotherapy
(CAF): (no change)
Nursing home, paresthesias, loss
of balance, frequent falls, uses
wheelchair, tremor persists
Needed assistance
Mild gait ataxia, normal hcad
Normal mammogram, axillarv
lymph node: adenocarcinoma
Imbalance, gait difficulties, trutcal and appendicular ataxia, abnormal visual tracking and vestibulo-ocular reflexes
Wheelchair bound, unable to
walk, severe truncal and appendiculac ataxia, demented
Unsteadiness, gait difficulties
Surgical resection followed by radiation therapy and seed implant
Modified radical mastectomy
1983, negative lymph nodes,
prednisone (no response)
Gait difficulties, uses walker, dysarthric speech
NED (March 1990)
Ataxia improved; a myelopathy
developed 5 years after onset,
unable to walk
cyclophosphadmide, adriamycin, and Auorouracil; NED
no evidence of disease; MRI
ing use of the extremities. Although fine movements
of the hands were limited in 4 patients, all were able
to write legibly and to feed themselves. In 6 women,
the ocular motor abnormality was opsoclonus, a disorder of saccadic eye movements consisting of involuntary, arrhythmic, multidirectional, high-amplitude conjugate saccades, sometimes called saccadomania {l6J.
Patient 2 had nystagmus on right lateral and upward
gaze; bilateral palsies of the sixth cranial nerve appeared subsequently. Patient 6 was believed to have
opsoclonus on initial examination but subsequent neuro-ophthalmological evaluation revealed loss of smooth
pursuit movements, occasional upbeat nystagmus, and
episodes of ocular flutter (findings often confused at
the bedside with opsoclonus). Other less frequent
symptoms and signs included nausea, dizziness or vertigo, muscle weakness, dysarthria, dysphagia, and decreased hearing (see Table 1). Patient 8 developed a
myelopathy 5 years after the onset of opsoclonus. Myelogram and spinal computed tomography (CT) scan
had a normal appearance.
244 Annals of Neurology Vol 29 No 3 March 1091
Resting tremor, gait difficulties,
normal head MRI
Died April 1989, probably neurological causes; no autopsy
magnetic resonance imaging.
The onset of the neurological disorder was usually
rapid, with symptoms reaching their peak in 1 week to
4 months. Most patients stabilized; Patients 1, 3, and
4 improved substantially. All patients, however, continued to display some degree of neurological disability
1 to 12 years after the onset of their symptoms.
The CSF was abnormal in 3 of 6 women for whom
the information was available. All 3 had a mild lymphocytic pleocytosis (11 to 29 cells/mm3) and a mildly
elevated protein concentration (54 to 72 mg’dl). The
glucose concentration and cytological findings were
The neurological disorder was the presenting symptom in 4 of the 8 patients. Cancer was subsequently
discovered in 3. In the fourth, no cancer had been
discovered 21 months from the onset of neurological
symptoms. The other 4 patients developed neurological symptoms 1 month to 9 years after cancer had been
diagnosed and treated. Seven patients were alive at the
time of writing. Four patients received prednisone; in
2 there appeared to be some improvement, but 2 had
Fig 1 . Immunohistochemical reactioity of sera with human cerebral cortex. (A)Control using n o m d serum diluted 1500. ( B )
The serum, dihted 1:500,]>om an anti-Ki-positive patient reacts
with nuclei of cerebral cortex. giving the characteriiticbrown peroxidase reaction product. The cytoplasm also reacts slightly but the
nucleoli are spared. Glial cells do not react. (Counterstained with
Hematoxylin. X 250 befre 510% enlargement.)
n o response. Cyclophosphamide was used as chemotherapy for the breast cancer in 3 patients, 1 of whom
did not receive prednisone. The opsoclonus did not
Serum and CSF of all 8 patients contained a hightiter antibody called anti-Ri. During the time these patients were studied, we studied the serum and, in many
instances, the CSF of 8 7 other patients with breast
cancer, some of whom had neurological disability, but
none had the clinical syndrome found in these patients.
None possessed the anti-Ri antibody. Also during that
time, we studied the serum of 23 other patients with
opsoclonus, 10 of whom had cancer (small-cell lung
cancer in 3, neuroblastoma in 5 , uterine carcinoma in
1, and ovarian carcinoma in 1).None of the 23 patients
had detectable anti-Ri antibody.
Fig 2. Reactivity of affinity-purified anti-Ri IgG with human
cortex. Anti-Ri IgG bound to the 80- and SS-kd molecular mass
bands was eluted. (A)Immunohistoch~ii.rtuyperformed with elution of the SJ-kd band gives a reaction similar to the native antiRi serz47n. The eluatefrom the 80-kd band (not shown) was identical. ( x 100 before enlargement.) (Bj Same as A at higher magni&cation. ( x 250 before 223% enlargement.) (Cj No reaction is
detected with material eluted fmm an imevekznt area of the nitrocellulose strip. All countwstained with hematoxylin. ( x 100 before enlargement.)
Luque et al: Anti-Ri and Paraneoplastic Opsoclonus 245
Fig 3. Quantitative Western blot analysis. (A)Serial dilutions of
sera from 1500 to 1:20,000 and of CSF from 1 :10 to 1 :80 were
incubated with nitrocellulose strips containing the Ri antigen.
LneJ I through 5: Normal serum. Lanes 6 through 10: Anti-
Ri serum. ?'he area of nitrocellulose containing the Ri antigen
Detection and Characterization of Anti-Ri Antibody
Immunohistochemical study of serial dilutions of patients' serum and CSF revealed the presence of an antibody (anti-Ri) at high titer that reacted with CNS neuronal nuclei of all areas of the CNS examined (Fig 1).
Slight cytoplasmic staining of neurons was also seen
but glia did not react. Sections of cerebral cortex were
examined both before and after counterstaining with
hematoxylin. The antibody did not react with cells in
white matter, which only became identifiable after hematoxylin counterstaining. The antibody also did not
react with many cells in gray matter, which likewise
could be seen only after counterstaining. We concluded that the only reactive cells were neurons. The
serum titers varied from 1:5,000 to 1:320,000and CSF
titers, from 1:2,000 to 1:16,000. Buffer alone andor
control serum and CSF were negative for the antibody
at identical dilutions.
Biotinylated IgG [ 17) from anti-%-positive sera, but
not from control sera, reacted identically to the native
anti-Ri serum. Using the biotinylated IgG serum, no
reaction was found with frozen sections from non-CNS
tissues, including liver, lung, ovary, and endocrine tissues (thyroid, parathyroid, adrenal gland, testes, and
kidney) (data not shown).
appropriate bands on the Western blot. Immunohistochemistry performed with these affinity-purified
preparations reproduced the same staining pattern obtained with native serum (Fig 2B). Antibody bound to
an irrelevant area of the preparative nitrocellulose strip
used in the purification served as a negative control
(see Fig 2A).
Afinity PuriJication of Anti-Ri Antibody
In order to show that the antigen identified in the histochemical analysis was the same one identified by
Western blot, we affinity-purified the antibody that
recognized the 55- and 80-kd antigens by eluting the
246 Annals of Neurology Vol 29 No 3 March 1991
was cut out and the radioactivity determined using liquid scintillation spectroscopy. iB) The antibody reactivity is plotted against
the IgG concentration.
Western Blot Analysis
The sera and CSF of all 8 patients detected in a consistent and highly specific manner two protein antigens
with relative molecular masses of about 5 5 kd and 80
kd in immunoblots of cortical neuronal extracts (Fig 3;
see also Fig 4A). Binding of the anti-Ri IgG was detectable at 0.5 p,g/ml of IgG, whereas no reactivity of normal IgG was evident at a concentration of 50 Fg/ml.
Partial Purification of the Ri Antigen
To more stringently assay putative anti-Ri sera, the Ri
antigen was affinity-purified from human cortical neurons and used for Western blot analysis (Fig 4B). The
sera from all 8 patients recognized the purified Ri antigen. The serum from an additional patient with nonparaneoplastic opsoclonus, negative findings on immunohistochemical study, but a banding pattern by
Western blot consistent with anti-Ri failed to recognize
the purified Ri antigen (see Figs 4A and 4B).
Patients sera blotted with native anti-Ri serum
against extracts from non-CNS tissues, tumor, and tumor culture lines were negative for the Ri antigen except for a neuroblastoma line (SKN-S-M) in which an
Table 2. The Content of the R i Activity in Sera and CSF
Ri Activity ( c p d k g IgG)
Patient No.
CSFJSerum Ratio
82 5
The Selective Expression of the Ri Antigen in Tumor
Tissue from Anti-Ri-Positive Patients
The Ri antigen is extremely restricted in its expression.
Fig 4. (A)Western blots of cortical neuronal mtract show that
anti-Ri-containing sera identifis 55-kd and 80-kd molecular
m s s bands (Patients 1 through 8). Patient 9 shows similar but
dzfferentmolecular mass bands. Incubation in the absence of serum
(-) or with normal serum (N)or with Sera from patients with
neltroblastomas and opsoclonus (0) resulted in negative findings.
IB) Reactions of various sera with pur$ed R i antigen. Anti-Ricontaining sera (Patients 1 through 8) are positive. Sera from
Patient 9 did not recognize pur$ed Ri antigen.
antigen corresponding to the 80-kd band was identified; a medulloblastoma line had reactivity in the region
corresponding to the 55- but not the 80-kd band (data
not shown).
Synthesis of Antibody in CNS
Quantitation of the specific reactivity of the anti-Ri
antibody was carried out by serial dilution of serum
and CSF in a range where there was a proportional
linear relationship with the amount of IgG present in
the fluid (see Fig 2). Serum reactivity varied from 126
to 2,093 unitsiml in individual patients. Similarly, CSF
reactivity was 11 to 52 units/ml. The specific activity
( c p d k g of total IgG) was always higher in CSF than
in serum, with a CSF/serum ratio of 3 to 17 for the
55-kd protein (see Table 2).
The antigen was not detected in frozen sections of human tumors from patients who were anti-Ri negative
(breast {G patients}, small-cell lung cancer 12 patients},
non-small-cell lung carcinoma [ 3 patients}, pheochromocytoma [2 patients), neuroblastoma [3 patients],
neurofibroma [2 patients}, seminoma [2 patients)).
Frozen sections of the tumors of the anti-Ri-positive
patients were not available.
We examined paraffin-embedded tumor tissue from
4 anti-Ri-positive patients (1 from a lymph node containing metastatic breast cancer, 1 from an axillary
lymph node, 1 from adenocarcinoma of the fallopian
tube, and 1 from a primary breast adenocarcinoma)
by immunohistochemistry. Control human IgG from a
healthy volunteer showed no staining (Fig 5A). The
two specimens of tumor in lymph nodes of anti-Ripositive patients were positive for Ri antigen: 1 gave a
strong nucleolar staining (see Fig 5B), the other stained
nuclei but spared nucleoli and cytoplasm (Fig 5D). The
adenocarcinoma of the fallopian tubes showed cytoplasmic staining in some focal tumor cells. One specimen of breast cancer (Patient 3) was negative for Ri
antigen. Since we have observed that prolonged fixation of tissue in formalin destroys Ri antigenic reactivity, it is difficult to conclude that the Ri antigen is not
expressed in this specimen. The Ri antigen was not
detectable in either frozen or paraffin-embedded breast
tumors (n = 6) from patients whose serum was negative for anti-Ri antibody. Biotinylated IgG from 3
other anti-Ri-positive patients had identical reactions.
Affinity-purified biotinylated anti-Ri IgG from the
corresponding bands on nitrocellulose demonstrated
similar immunohistochemical staining on tumor tissues
as that described above.
The anti-Ri antibody as characterized in this report
identified a subset of patients with paraneoplastic ataxia
and usually opsoclonus (as defined clinically in the introduction), most of whom had an associated carcinoma of the breast. The antibody was not found in
serum of patients with paraneoplastic opsoclonus associated with neuroblastoma or small-cell lung cancer.
Luque et al: Anti-Ri and Paraneoplastic Opsoclonus 247
Fig 5. Expression of Ri antigen in tumor tissue from patients
with the Ri .yndrome. Immunohistochemist y with biotinylated
antibodies. (A)Breast cancer in lymph node from Patient 1 reacted
with n o w 1 IgG. (BI Same tumor reacted with anti-Rt Jerum.
There is strong reactivity in the nucleoli of the malignant cells.
The unstained cells are normal lymphocytes. iC) Breast cancer in
lymph node from a patient without anti-Ri syndrome reacted with
unti-Ri serum. There is no reaction. (0)Axillavy lymph node
from Putient 5 shows nuclear staining of a feu, malignant cells.
(All counrwstained with methylene blue, A, B x 250 befire
235% enlargement; C, 0 x 250 b4ore 225% enkwgement.j
Paraneoplastic opsoclonus, whether antibody positive or negative, is itself a subset of a more general
syndrome of opsoclonus caused by several different
CNS insults, including viral infections, brain tumors,
hyperosmolar syndromes, and metabolic abnormalities
116, 18, 191. No antibodies have been reported in
nonparaneoplastic opsoclonus, nor did we encounter
The presence of fallopian tube cancer in 1 patient
indicates that the antibody reaction is not unique to
breast cancer. The failure to find cancer in 1 anti-Ripositive patient can lead to one of at least three explanations: (1) A cancer is present but has not yet been
found. In many paraneoplastic syndromes it is not uncommon for months or years to elapse before the primary tumor is identified 1201, or for a very small cancer
to be encountered only at autopsy 121f. ( 2 ) The antibody reaction was elicited by a non-neoplastic antigen.
Only two-thirds of patients with Lambert-Eaton myasthenic syndrome have an underlying small-cell lung
cancer, although the antibody appears to be present in
both patients with and those without cancer 122). We
have encountered 2 patients with the anti-Hu antibody-associated paraneoplastic sensory neuronopathy-encephalomyelitis syndromes in whom no tumor
was encountered even at autopsy. Thus, antigenic insults other than tumor may be responsible for the antibody response and the clinical syndrome. (3) The
breast cancer elicited an immune response so powerful
that the tumor was destroyed. This happy speculation
has some' precedence. For years investigators have
commented on the indolent course of the cancer and
long survival of patients with paraneoplastic syndromes
C231, particularly in the opsoclonus-myoclonus syndrome associated with neuroblastoma [24}. A confounding factor is the likelihood that the appearance
248 Annals of Neurology Vol 29 No 3
March 1991
of a paraneoplastic syndrome leads to early diagnosis
of the cancer. However, Dalmau and associates [3}
recently showed that 15% of patients with small-cell
lung cancer harbor low titers of the anti-Hu antibody
usually associated with neurological disease. Those patients differ from the 85% of patients without the antibody in that they are more likely to have tumor limited
to the chest, despite the absence of a neurological disorder to call early attention to the tumor.
The clinical features of this syndrome, which includes the subacute onset of opsoclonus, ocular flutter,
nystagmus, or other eye movement disorders associated with truncal and sometimes appendicular ataxia,
cannot be discriminated from nonparaneoplastic opsoclonus. Likewise, the clinical course, which can be
marked by continuing neurological abnormalities, resolution of symptoms or remissions, and exacerbations,
does not differ from the nonparaneoplastic syndrome.
The CSF pleocytosis, which often resolves after several
weeks, indicates the presence of an inflammatory prvcess but that can also characterize CNS viral infections.
In one of our patients El41 (Patient 1). the magnetic
resonance image (MRI) scan revealed a small area of
hyperintensity in the dorsal midbrain on the T2weighted images. The MRI scan of Patient 3 was abnormal. MRI scans of 3 other patients were normal.
Other patients with nonparaneoplastic opsoclonus had
CT or MRI abnormalities in the pons {25} or the cerebellar vermis 1261.
The pathological lesion of paraneoplastic opsoclonus
is unknown. At the time of writing, only one of the
patients reported here had died, perhaps a testimony
to the indolence of the underlying cancer, but in other
patients with paraneoplastic opsoclonus, findings have
ranged from an entirely normal brain {27] to perivascular inflammatory infiltrates [2S] to a marked loss of
Purkinje cells of the cerebellum f291 or of olivary neurons of the brainstem {27). The lack of pathological
findings in some patients may explain why, in contrast
to paraneoplastic cerebellar degeneration, many patients with paraneoplastic opsoclonus undergo spontaneous remission 130, 311. The frequency of spontaneous remissions makes suspect reports of effective
treatment of opsoclonus with various drugs including
thiamine [32) and clonazepam [31, 331.
The presence of the Ri antigen identified by histochemical techniques in 3 of the 4 tumors available for
study suggests that it is the body’s immune response
to a tumor antigen that elicits the antibody response.
That 1 tumor (Patient 3) was negative for Ri antigen
is not surprising since the histochemical reaction is sensitive to prolonged formalin fixation and paraffin embedding; we have been unable to study frozen sections
of tumor from any of our patients. The 3 Ri positive
tumors strongly support, but do not prove, the hypothesis that the presence of anti-Ri antibody represents an
immune reaction to a tumor antigen. However, the
different histochemical staining of each of the 3 Ri
positive tumors is surprising. In 1 patient, the antigen
appeared to be restricted to the nucleolus, in another
it was nuclear but not nucleolar, and in the third it was
cytoplasmic. Only further study, preferably of freshfrozen tissue, will clarify the situation.
Fresh-frozen breast cancer from patients without
paraneoplastic opsoclonus is uniformly negative for the
Ri antigen. Thus, the situation with anti-Ri paraneoplastic opsoclonus is similar to that in anti-Yo-positive
paraneoplastic cerebellar degeneration where the antigen is present only in the tumors of those patients
who developed the antibody response 1173. The situation is different with the anti-Hu antibody and with
the Lambert-Eaton myasthenic syndrome antibody
where the antigen appears to be present in all smallcell lung cancers but only a small subset of patients
mount an antibody response 14, 341. Clinically anti-Ripositive patients differ from anti-Yo-positive patients
in that the latter do not have opsoclonus, usually have
more severe appendicular ataxia, are more likely to
have ovarian rather than breast cancer, and do not recover.
The role of the antibody in the pathogenesis of the
disease is not established. In the Lambert-Eaton myasthenic syndrome, plasmapheresis improves the clinical
symptomatology and infusion of patient’s IgG into experimental animals reproduces the neuromuscular abnormality, establishing the antibody as pathogenetic 14,
351. Such proof is lacking in paraneoplastic syndromes
such as opsoclonus-myoclonus that affect the CNS.
However, in the anti-Hu subacute sensory neuronopathy-encephalomyelitis syndrome, the antibody is
found in neurons of the CNS, suggesting a pathogenetic role [36}. Alternatively, the antibody could be a
marker simply indicating an immune response. To our
knowledge, no attempts have been made to produce
the opsoclonus syndrome in experimental animals by
transfer of anti-&-positive sera nor have any of our
patients been treated with plasmapheresis. The tendency for the disorder to spontaneously remit also
makes evaluation of plasmapheresis difficult.
Antibodies are being increasingly found in the serum
of patients with a variety of diseases, and neurological
paraneoplastic syndromes are no exception. Some
investigators have reported significant numbers of
normal persons who harbor antineuronal antibodies
137-391 and significant numbers of patients with cancer (particularly small-cell lung cancer) who also harbor
antineuronal antibodies that are unassociated with
paraneoplastic syndromes {3, 40, 41}. Using very high
concentrations of IgG (> 200 pg/ml), we often found
faint immunohistochemical staining of neurons and
one or more bands on Western blots in patients with
cancer without paraneoplastic syndromes or in patients
et al:
Anti-Ri and Paraneoplastic Opsoclonus 249
with cancer and neurological disease who may or may
not have a paraneoplastic syndrome. In a few instances,
a particular antibody is found at high titer in patients
with a definable clinical paraneoplastic syndrome usually associated with a limited subset of cancers. These
autoantibodies are not found in normal persons or in
persons with cancer who do not have the paraneoplastic syndrome. They are, as in this instance, also not
found in patients with the same paraneoplastic syndrome who do not have the specific cancer as, for example, in opsoclonus-myoclonus associated with neuroblastoma. Thus, only after an antibody has been
defined, and its relationship to a particular clinical syndrome and a particular cancer elucidated, can one accept it as an important marker for a specific disease.
A few semantic issues should be addressed. The first
is that of nomenclature. Lennon [42} suggested that
the current nomenclature for paraneoplastic antibodies
is inadequate. In the first clearly defined paraneoplastic
antibodies reported by Greenlee and Brashear [ 11 and
by Jaeckle and colleagues [43], in patients with paraneoplastic cerebellar degeneration, the term anti-Purkinje cell antibody was used and seemed appropriate.
However, a number of antibodies, including those we
have called anti-Yo, anti-Hu, and anti-Ri, react with
Purkinje cells and all may be associated with cerebellar
signs, although both the clinical situations and the underlying tumors differ with the individual antibodies.
Thus, some other term appears essential. When Graus
and associates discovered the antinuclear antibody associated with paraneoplastic sensory neuronopathy in
small-cell lung cancer, they named it anti-Hu after the
first two letters of the last name of the first patient they
studied [44}. We continued this practice with the antiY o and anti-Ri antibodies.
To determine if an antibody one encounters in a
patient is a previously identified one requires the use
of not only immunohistochemistry but also Western
blotting to determine the size of the antigen. The antiRi antibody appears histochemically identical to the
anti-Hu antibody. However, it is clearly different since
it identifies different antigens on Western blotting of
cortical neurons. Some of the difficulties distinguishing
among the various paraneoplastic syndromes producing cerebellar signs and associated with staining of Purkinje cells are easily resolved by the use of Western
For the clinician encountering a patient with a possible paraneoplastic syndrome, the identification of a
speclfic antibody may focus the search for an underlying neoplasm to one or only a few cancers and allow
an early diagnosis to be made, which may potentially
cure the patient. For the scientist, the presence of specific antigens shared uniquely between tumors and the
nervous system have many intriguing implications.
250 Annals of Neurology Vol 29 No 3 March 1991
Supported in part by National Institutes of Health grant NS26064
(Dr Posner) and American Cancer Society grant PDT 359 (Dr Furneaux).
Presented in part at the 114th Annual Meeting of the American
Neurological Association, New Orleans, LA, September 24-2 7
1. Greenlee JE, Brashear HR. Antibodies to cerebellar Purkinje
cells in patients with cerebellar degeneration and ovarian carcinoma. Ann Neurol 1983;14:609-613
2. Furneaux HM, Reich L, Posner JB. Autoantibody synthesis in
the central nervous system of patients with paraneoplastic syndromes. Neurology 1990;40:1085-1091
3. Dalmau J, Furneaux HM, Gralla RJ, er al. Detection of the
anti-Hu antibody in the serum of patients with small cell lung
quantitative Western blot analysis. Ann Neurol
4. Prior C, Lang B, Wray S, Newsom-Davis J. Action of LambertEaton myasthenic syndrome IgG at mouse motor nerve terminals. Ann Neurol 1985;17:587-592
5. Thirkill CE, Fitzgerald P, Sergott RC, et al. Cancer associated
retinopathy (CAR syndrome) with antibodies reacting with retinal optic-nerve and cancer cells. N Engl J Med 1989;
6. Anderson NE, Rosenblum MK, Posner JB. Paraneoplastic cerebellar degeneration: clinical-immunological correlations. Ann
Neurol 1988;24:559-567
7. Budde-Steffen C, Anderson NE, Graus P, et al. An anti-neuronal autoantibody in paraneoplastic opsoclonus. Ann Neurol
8. Luque A, Furneaux HM, Wray S, et al. Anti-Ri: an autoantibody
associated with paraneoplastic opsoclonus. Ann Neurol 1789;
26:178 (Abstract)
9. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase
complex (ABC) in immunoperoxidase techniques: a comparison
between ABC and unlabelled antibody (PAP) procedures. J Histochem Cytochem 1781;29:577-580
10. Blomstrand C, Hamberger A. Protein turnover in cell enriched
fractions from rabbit brain. J Neurochem 1969;16:1401-1407
11. Yaragihara T, Hamberger A. A method for separation of Purkinje cells and granular cell-enriched fractions from rabbit cerebellum. Brain Res 1973;59:445-448
12. Laemmli UK. Cieavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680-685
13. Towbin H , Staehelin T, Gordon J. Electrophoretic transfer of
proteins from polyacrylarnide gels to nitrocellulose sheets: procedures and some applications. Proc Natl Acad Sci USA 1979;
14. Case records of the Massachusetts General Hospital (Case 91988). N Engl J Med 1988;318:563-570
15. Herishanu Y,Apte R, Kupperman 0.Immunological abnormalities in opsoclonus cerebellopathy. Neuro-ophthamology 1985;
16. Leigh RJ, Zee DS. The neurology of eye movements. Philadelphia: FA Davis, 1983
17. Furneaux HM, Rosenblum MK, Dalmau J, et al. Selective expression of Purkinje cell antigens in tumor tissue from patients
with paraneoplastic cerebellar degeneration. N Engl J Med,
18. Digre KB. Opsoclonus in adults. Report of three cases and
review of the literature. Arch Neurol 1986;43:1165-1175
17. Noda S, Taka0 A, Itoh H, Umezaki H. Opsoclonus in hyperosmolar nonketotic coma. J Neurol Neurosurg Psychiatry 1985;
20. Croft PB, Wilkinson M. The course and prognosis in some types
of carcinomatous neuromyopathy. Brain 1969;92:1-8
2 1. Anderson NE, Budde-Steffen C, Wiley RG, et al. A variant of
the anti-Purkinje cell antibody in a patient with paraneoplastic
cerebellar degeneration. Neurology 1988;38:1018-1026
22. Lennon VA, Lambert EH. Autoantibodies bind solubilized calcium channel-o-Conotoxin complexes from small cell lung carcinoma: a diagnostic aid for Lambert-Eaton myasthenic syndrome.
Mayo Clin Proc 1989;64:1498-1504
73. Gulcher JR, Stefansson K. Paraneoplastic syndromes of the nervous system. Neuroimmune Dis 1988;8:295-3 14
24. Altman AJ, Baehnes RL.. Favorable prognosis for survival in
children with coincident opsoclonus myoclonus and neuroblastoma. Cancer 1976;37:846-852
25. Hattori T, Hirayarna K, Imai T, et al. Ponrine lesion in opsoclonus-myoclonus syndrome shown by MRI. J Neurol Neurosurg
Psychiatry 1988;5 1:1572-1575
26. Tuchman KF, Alvarez LA, Kantrowitz AB, et al. Opsoclonusmyoclonus syndrome correlation of radiographic and pathological observations. Neuroradiology 1789;31:250-252
27. Ridley A, Kennard C , Scholtz CL, et al. Omnipause neurons in
cwo cases 06 opsoclonus associated with oat cell carcinoma of
the lung. Brain 1987;110:1679-1709
28. Anderson NE, Budde-Steffen C, Rosenblum MK, et al. Opsoclonus, myoclonus, ataxia, and encephalopathy in adults with
cancer: a distinct paraneoplastic syndrome. Medicine 1788;
27. Ross AJ, Zeman W. Opsoclonus occult carcinoma and chemical
pathology m dentate nuclei. Arch Neurol 1967;17:546-55 1
30. Bellur SN. Opsoclonus: its clinical value. Neurology 1975;
31. Dropcho E, Payne R. Paraneoplastic opsoclonus-myoclonus. Association with medullary thyroid carcinoma and review of the
literature. .4rch Neurol (Chicago) 1986;43:410-4 15
32. Nausieda PA, Tanner CM, Weiner WJ. Opsoclonic cerebellopathy: a paraneoplastic syndrome responsive to thiamine. Arch
Neurol (Chcago) 1981;38:780-78 1
33. Leigh Rr. Zee DS. The diagnostic value of abnormal eye movements: a pathophysiological approach. John Hopkins Med J
34. Lennon CA, Lambert EH, Wittingham S, Fairbanks V. Autoimmunity with Lambert-Earon myasthenic syndrome. Muscle
Nerve 1982;5:521-525
35. O’Neill JH, Murray NMF, Newsom-Davis J. The LambertEaton myasthenic syndrome. A review of 50 cases. Brain
36. Dalmau J, Furneaux HM, Rosenblum RK, et al. Immunological
study of the nervous system and tumor of three patients with
paraneoplastic encephalomyelitis and small cell lung cancer.
Neurology 1790;40 (Supp 1):166 (Abstract)
37. Stefansson K, Marton LS, Dieperink ME. Circulating autoantibodies to the 200,000-dalton protein of neurofilaments in the
serum of healthy individuals, Science 1985;228:1117-1 117
38. Daar AS, F a b r e w . Organ-specific IgM autoantibodies to liver,
heart and brain in man: generalized occurrence and possible
functional significance in normal individuals, and studies in parients with multiple sclerosis. Clin Exp Immunol 1981;45:37-47
37. Plioplys AV, Thibault J, Bouchard JP, et al. Anti-CNS antibodies in neurological and psychiatric disorders J Neurol Neurosurg Psychiatry 1987;50:15 14-1521
40. Grisold W, Drlicek M, Liszka U, Popp W. Anti-Purkinje cell
antibodies are specific for small-cell lung cancer but not for
paraneoplastic neurological disorders. J Neurol 1989;236:64
41. Brashear HR, Greenlee JE, Jaeckle KA, Resew. Anricerebellar antibodies in neurologically normal patients with ovarian
neoplasms. Neurology 1989;39:1605-1609
42. Lennon VA. And-Purkmje cell cytoplasmic and neuronal nuclear antibodies aid diagnosis of paraneoplastic auroimmune
neurological disorders. J Neurol Neurosurg Psychiatry 1789;
52:1438-1433 (Letter to editor)
43. Jaeckle KA, Graus F, Houghton A, et al. Autoimmune response
of patients with paraneoplastic cerebellar degeneration to a Purkinje cell cytoplasmic protein antigen. Ann Neurol 1985;
44. Graus F, Elkon KB, Cordon-Cardo C, Posner JB. Sensory neuronopathy and small cell lung cancer: an antineuronal antibody
that also reacts with the tumor. Am J Med 1986;80:45-52
Luque et al: Anti-Ri and Paraneoplastic Opsoclonus
Без категории
Размер файла
1 184 Кб
associates, cancer, opsoclonus, anti, antibody, paraneoplastic, breast
Пожаловаться на содержимое документа