вход по аккаунту


Detection of 14-3-3 protein in the cerebrospinal fluid supports the diagnosis of Creutzfeldt-Jakob disease.

код для вставкиСкачать
Detection of 14-3-3 Protein in the
Cerebrospinal Fluid Supports the Diagnosis
of Creutzfeldt-Jakob Disease
Inga Zerr, MD," Monika Bodemer, BA,* Olaf Gefeller, PhD,t Markus Otto, MD,* Sigrid Poser, MD,*
Jens Wiltfang, MD,$ Otto Windl, PhD,$ Hans A. Kretzschmar, MD,$ and Thomas Weber, MD"
The analysis of 14-3-3 protein in cerebrospinal fluid (CSF) was shown to be highly sensitive and specific for the diagnosis of Creutzfeldt-Jakob disease (CJD). However, the predictive value of this test in the clinical diagnosis of, and its
relation to, sporadic, genetic, and iatrogenic CJD cases have yet to be established. CSF samples of suspect CJD cases seen
in the prospective German surveillance study were tested for the presence of 14-3-3 protein by using a modified western
blot (WB) technique. W B detected 14-3-3 protein in 95.4% of definite and 92.8% of probable cases. In two patients
classified initially as not having CJD the test was positive, and both were later proved to have definite CJD. The positive
predictive value is 94.7% and the negative predictive value is 92.4%. False-positive results in a single CSF analysis were
seen in patients with herpes simplex encephalitis, hypoxic brain damage, atypical encephalitis, intracerebral metastases of
a bronchial carcinoma, metabolic encephalopathy, and progressive dementia of unknown cause. WB analysis for 14-3-3
protein was positive in only 5 of 10 cases of familial forms of spongiform encephalopathies. CSF analysis for 14-3-3
protein should thus be performed in any case suspect for CJD.
Zerr I, Bodemer M, Gefeller 0, Otto M, Poser S, Wiltfang J, Windl 0, Kretzschmar HA, Weber T.
Detection of 14-3-3 protein in the cerebrospinal fluid supports the diagnosis
of Creutzfeldt-Jakob disease. Ann Neurol 1998;43:32-40
Recently, the 14-3-3 protein has been described as a
useful cerebrospinal fluid (CSF) marker for the in vivo
diagnosis of Creutzfeldt-Jakob disease (CJD).
14-3-3 proteins are a group of highly conserved proteins involved in the regulation of protein phosphorylation and the mitogen-activated protein kinase pathway. The y isoform is thought to be specific for
No comparison of results of 14-3-3
protein tests in the CSF, in relation to sporadic, genetic, and iatrogenic CJD cases, is available.
Negative results of 14-3-3 protein analysis in the
CSF in neuropathologically confirmed CJD cases suggested questions about the predictive value of this
novel test."' False-positive results were reported in
patients with herpes simplex encephalitis (HSE), multiinfarct dementia, acute infarction, stroke, subarachnoidal hemorrhage, and viral encephalitis and in a patient who died of carcinomatous meningitis with a
primary small cell carcinoma of the
and negative predictive values of a test provide useful
information to support or to exclude the clinical
diagnosis of a disease. They are most important
for the clinician to establish a clinical and laboratorysupported diagnosis. CJD, Gerstmann-StrausslerScheinker syndrome (GSS), and fatal familial insomnia
(FFI) belong to the group of transmissible spongiform
encephalopathies (TSEs), such as scrapie in sheep and
goats and bovine spongiform encephalopathy in cattle.'
A definite diagnosis of TSE can be achieved only at the
time of autopsy.' By current clinical diagnostic criteria,
CJD cases are misdiagnosed, especially those with atypical clinical symptoms and those cases of new variant
CJD (vCJD)."-" At present, detection of typical periodic sharp and slow wave complexes (PSWCs) is the
only laboratory test required to substantiate a clinical
This abnormality is seen in only 60% of
sporadic CJD and is absent in V C J D . ' ~ , ' ~ , 'Further*
more, PSWCs are by no means specific for the diagnosis of CJD but may be seen in cases of intoxication
with lithium, phencyclidine, tricyclic antidepressants,
and barbiturates, in metabolic and anoxic encephalopathies, and in viral encephalitides such as HSE.'5-20
From *Neurologische Klinik und Poliklinik, TAbteilung Medizinische Statistik, SPsychiatrische Klinik und Poliklinik, and SInstitut fur Neuropathologie, Georg-August-Universitat, Gortingen; and
"Neurologische Klinik, Marienkrankenhaus, Hamburg, Germany.
Address correspondence to Dr Zerr, Neurologische Klinik und
PoliMinik, Georg-August-Universitat Gottingen, Robert-Koch-Str
40, D-37075 Goctingen, Germany.
Received Apr 16, 1997, and in revised form Jul 25. Accepted for
publication Aug 23, 1997.
Copyright 0 1998 by the American Neurological Association
The detection of pathological changes associated
with TSE a n d the pathological isoform of the prionrelated protein (PrP) permits a definite diagnosis post
mortem. Unfortunately, the pathological isoform PrP’“
cannot be detected in CSF, serum, o r blood by currently available methods i n cases of sporadic CJD o r
GSS.21’22Detection of PrPS‘ by tonsillar biopsy in a
case of v C J D is n o t practicable as a routine clinical
test. 2 3 Besides complications of bleeding and inflammation, nearly 16% of patients with CJD have had
tonsillectomies. The clinical value of CSF analysis i n
the differential diagnosis of CJD has already been demonstrated by significantly elevated levels of neuronspecific enolase (NSE), S100, a n d T protein i n CJD
cases compared with controls, reflecting acute neuronal
damage a n d astrocytic changes. 12,24-28
So far, no data have been reported about the stability of 14-3-3 protein in CSF. Suboptimal CSF storage
conditions have been discussed as causative of negative
results in autopsy-proved CJD and vCJD.~In addition
to assessing these parameters, we used different strategies in the evaluation of western blots (WBs) to define
the reliability a n d validity of WB analysis of 14-3-3
protein in CSF for the differential diagnosis of CJD.
We describe the predictive values of detecting 14-3-3
protein in CSF for the clinical diagnosis of CJD. In
addition, we defined molecularly the PrP gene (PRNP)
in o u r patients a n d correlated the CFS findings with
the known mutations a n d polymorphism at codon
Subjects and Methods
In the German national CJD surveillance study, 484 patients
with clinically suspected CJD were seen, in a prospective
case-control study, from June 1993 to December 1996. Suspect cases were reported to the CJD unit by the participating
physicians. After informed consent was given by the relatives,
the patients were visited by a research physician. Each underwent a detailed neurological examination. Clinical data
based on a semistructured questionnaire were obtained from
the relatives, usually the spouse. Data regarding the technical
investigations were extracted from hospital records. Electroencephalographic (EEG) records, computed tomographic
scans, and, if available, magnetic resonance imaging scans
were taken to the CJD unit and evaluated by one of us
(S.P.),using previously established riter ria.^' CSF samples
were obtained by the visiting research physician and immediately frozen on dry ice, as previously de~cribed.”,’~
All patients with PrPS‘, detected immunohistochemically
in brain tissue after hydrolytic autoclaving, were classified as
having “definite” CJD.8 Clinical diagnosis was made according to previously established criteria based on clinical signs
and symptoms and EEG
Patients with rapidly progressive dementia of less than 2 years’ duration,
PSWCs, and two of the following four findings: myoclonus,
visual andfor cerebellar symptoms, pyramidal and/or extra-
pyramidal signs, and akinetic mutism, were classified as
“probable” cases.3”’-’ Those fulfilling the above criteria
without PSWCs were classified as “possible.” Patients who
did not fulfill the clinical criteria outlined before were classified as “other” and served as internal controls within the
framework of the CJD surveillance study. During follow-up,
patients were reclassified in accordance with the development
of their disease, with the diagnostic findings reported by the
notifying physicians, and with the findings by autopsy. Cases
showing clinical improvement were classified as “other,” even
if no precise diagnosis could be made.
External Controls
CSF samples from 60 patients seen in the Department of
Neurology at the Georg-August-Universitiit Gottingen served
as additional external controls.
14-3-3 Protein Immunoassay
The 14-3-3 protein immunoassay in CSF was performed as
previously described.2 Detection of the bound polyclonal antibody to the p isoform of 14-3-3 protein (Santa Cruz Biotech, Santa Cruz, CA) was performed by using the enhanced
chemilurninescence (ECL) detection kit according to the
manufacturer’s instructions (Arnersham Buchler, Germany).
Molecular weight markers and a positive control from a case
with confirmed CJD were run on every gel.
CSF samples from 2 patients with definite CJD were used
to test the stability of 14-3-3 protein.
Genetic Analysis
The analysis of the PRNP was performed on genomic DNA
isolated from blood according to standard procedure^.^^ The
coding region of PRNP was amplified by a polymerase chain
reaction assay with the primers 895W and 896W.33 The amplification product was screened by using the single-strand
conformational polymorphism technique for potential point
mutation^.'^ The coding region was reamplified in four overlapping fragments, which were analyzed in parallel with
DNA samples of patients with known mutations.35 In addition, codon 129 of the PRNP was examined by digestion
with the restriction endonuclease NspI. The final sequence
confirmation was obtained by direct solid-phase sequencing
of the complete coding region of PRNP on an automated
system (Model 4000L; LI-COR, Lincoln, NE).
Western Blot Evaluation
Evaluation by WB was performed by two investigators (M.B.
and I.Z.) blinded to diagnosis. To test the reliability of the
assay, samples of definite, probable, and other cases were
evaluated in two different experiments. Each gel was analyzed independently twice without knowing the previous results. The WBs were assessed by using the categories positive,
weakly positive, and negative.
Several different strategies for combining the information
of repetitive WB analyses, assessed independently by two raters, to yield a final test result are possible. We considered the
following four strategies in detail:
1 . Maximal avoidance of false-positives. The W B test is defined as positive only if both raters on all occasions agree
unanimously on the presence of 14-3-3 protein.
Zerr et al: Detection of 14-3-3 Protein in CSF Suggests CJD
2. Specificity increasing. In addition to those situations defined as positive by strategy 1 , contradictory or ambiguous assessments are also defined as positive test results if
the majority of ratings indicate the presence of 14-3-3
3. Sensitivity increasing. To increase the sensitivity of the
diagnostic classificarion rule further, contradictory or ambiguous situations are defined as positive test results not
only in the case of a majority of positive ratings but also
in the case of an equilibrium of positive and negative
4.Maximal avoidance of false-negatives. The WB test result
is defined as negative only if both raters on all occasions
unanimously agree on the absence of 14-3-3 protein.
StatisticaL Analysis
The statistical description, regarding the diagnostic validity
and reliability of the WB test of 14-3-3 protein in CSF for
detecting CJD, comprised the estimation of standard measures of test validity (sensitivity, specificity, and predictive
values) and test reliability (K value). To indicate the influence of sampling variability, the point estimates are accompanied by their 95% confidence intervals (CIS),which have
been calculated as exact intervals from the corresponding binomial distributions. Descriptive information on the distribution of continuous variables is provided by giving selected
quantiles of the observed data.
WE3 analysis of 14-3-3 protein was performed on CSF
samples from 289 of 484 suspect CJD cases seen from
June 1993 through December 1996 (examples are
given in Fig 1). Among these 484 patients referred to
the CJD unit, CSF specimens were available from 46%
of the definite cases, from 74% of probable, from 51%
of possible, and from 61% of cases classified as “no
CJD.” The demographic data of the patients are shown
in Table 1. Of the patients with CSF available for 143-3 analysis, there was no difference compared with
those with CSF that did not undergo analysis. Most of
the patients with CJD were sporadic cases (134 definite
and probable cases). The distribution of the polymorphism at codon 129 showed an overrepresentation of
homozygosity for methionine in definite and probable
cases (Table 2). Five cases with a mutation at codon
178 (D178N), 2 cases with a mutation at codon 200
(E200K), 2 cases with a mutation at codon 102
(PIOIL), and 1 case with a mutation at codon 210
(V21OI) were identified in the study. Four patients
with the D178N mutation were homozygous for methionine at codon 129, and 1 patient was heterozygous
with methionine and valine at this position. All patients
with the D178N mutation showed the FFI genotype.
One patient with iatrogenic CJD 10 years after dura mater transplant was heterozygous at codon 129.
Methodo Logical Issues
CSF samples from 2 patients with definite CJD were
used to test the stability of 14-3-3 protein. In the first
sample, 14-3-3 protein was clearly detectable initially
and showed a strong band; the sample of the second
patient showed only a weak band. Both samples were
stored at 22”C, 4”C, and -20°C and were measured
on days 2, 4, 6, 8, and 12. The aliquots stored at
-20°C were freeze-thawed six times. All samples tested
were clearly 14-3-3 protein-positive, showing that protein 14-3-3 is stable under various conditions. Only in
the second sample, with an initially weak band, was
there a slight decrease of intensity on day 12 (data not
shown). CSF samples from 202 patients were tested for
detection of p 130/ 131 by using two-dimensional gel
electrophoresis and for 14-3-3 protein by using the
WB technique. Results are given in Table 3. The 14-
Fig 1. Detection of 14-3-3 protein
in cerebrospinal fluid (CSF) in patients with Creunfeldt-jakob ditease
(CJD) and in controls. Lanes 1
through 4 show CSFfrom patients
with definite CJD; lanes 5 through
8 show CSFfiom controls (4 cases
with Alzheimev j disease).
Annals of Neurology
Vol 43
No 1
January 1998
Table 1. Ane at Onset and Sex Distribution of Cases by Diagnostic CLassi$cation (n
Age Distribution
Definite (n = 65)
Probable (n = 69)
Possible (n = 40)
Other (n = 104)
Genetic" (n = 10)
Iatrogenicb (n = I )
Male : Female
"Five cases with the D178N mutation, 2 cases with the E200K mutation, 2 cases with the P102L mutation, and 1 case with the V2101
case of iatrogenic
Creutzfeldt-Jakob disease 10 years after dura mater transplant.
Table 2. Codon 129 Polymorphism
Definite (n = 65)
Probable (n = 69)
Possible (n = 40)
Other (n = 104)
Genetic (n = 10)
Iatrogenic (n = 1)
MIM, % (n)
84.6 (44)
80.0 (32)
36.4 (8)
47.4 (27)
80.0 (8)
MIV, Yo (n)
5.8 (3)
7.5 (3)
40.9 (9)
38.6 (22)
20.0 (2)
VIV, % (n)
9.6 (5)
12.5 (5)
22.7 (5)
14.0 (8)
"Number of cases analyzed.
M = methionine: V
3-3 protein was detected in all samples that were positive for p130/131. In addition, 23 samples negative
for p130/13 1 were 14-3-3 protein-positive, indicating
the higher sensitivity of this procedure. To assess the
reliability of the 14-3-3 protein determination, two
evaluators (M.B. and 1.2.) assessed the WB independently (n = 575). For this assessment, samples from
the CJD surveillance study and the external controls
were combined. There was a high level of agreement
between the two evaluators, leading to a K value of
0.87 (95% CI, 0.83-0.91), and 93.1% of decisions
were concordant based on the rating scale negative,
weakly positive, and positive.
Different strategies in the evaluation of the gels were
considered, to assess the validity of the WB determination (Table 4). If maximal attention is directed to the
avoidance of false-positives (strategy l ) , a high price,
with respect to a substantial reduction in sensitivity
(69.4%), must be paid. In contrast, the maximal avoidTable 3. Comparison of Two-Dimensional Gel Electrophoresis
f . r Proteins I30/131 and the 14-3-3 Protein Immimoassay
f n = 202)
ance of false-negatives (strategy 4) has a less pronounced effect on the specificity, which decreases to
85.6%. Nevertheless, better overall results regarding
both measures of test validity are obtained simultaneously with the less extreme strategies, 2 and 3 . In
particular, the sensitivity increasing strategy 3, in which
ambiguous or contradictory assessments of the WB results are resolved by defining the final classification results as 14-3-3 protein-positive, achieved simultaneously a high sensitivity of 94% and a comparable
specificity of 93.3%. All further data on the properties
of the WB test thus refer to the application of strategy
3 for the evaluation of 14-3-3 protein presence.
Sensitivig, Spec;fcig, and Predictive Values of
14-3-3 Protein WB
14-3-3 Protein was detected in 95.4% (95% CI, 87.199.0) of pathologically proven definite cases (Table 5).
Based on definite and probable cases, the sensitivity
was slightly lower (94.0%; 95% CI, 88.6-97.4).
14-3-3 Protein was detected in 7 patients with suspect
CJD who were classified clinically as having other diseases. One of these patients most likely suffered from
CJD and developed typical clinical signs and symptoms, and a typical EEG, but could not be classified as
probable, because the rapid clinical deterioration to
akinetic mutism prevented a formal testing of demen-
Zerr et al: Detection of 14-3-3 Protein in CSF Suggests CJD
Table 4. Sensitivig, SpeciJcity, and Predictive Values, Using Diferent Interpretation StrategieJ
Strategy 4
Strategy 1
Definite and probable
Positive predictive value
Definite, probable, and
other" cases
Negative predictive value
Definite, probable, and
other" cases
Avoidance of
Yo (95% CI)
Strategy 2
Strategy 3
Yo (95% CI)
Yo (95% CI)
Avoidance of
Yo (95% CI)
(58.8-8 1.3)
(50.9-8 1.4)
(56.1-8 5.4)
(77.3-9 1.7)
(94.2-1 00)
"Cases presented initially to the surveillance study as suspected
Creutzfeldt-Jakob disease with other diagnosis
tia. One patient suffered from bronchial carcinoma
with intracerebral metastasis, and 1 patient died after
barbiturate intoxication. One patient had a disease duration of more than 5 years and developed progressive
dementia and extrapyramidal symptoms. This patient
is still alive, but the clinical course does not yet permit
a definite diagnosis. In these 4 cases, the determination
of 14-3-3 protein was done retrospectively, and, unfortunately, no CSF could be examined at follow-up.
Three other patients were initially 14-3-3 protein-positive. Because of the clinical classification of no CJD, a
second lumbar puncture was performed 2 weeks later.
All these patients showed clinical improvement of their
symptoms and were 14-3-3 protein-negative on the
second lumbar puncture. A more likely diagnosis in
these cases was atypical encephalitis (n = 1) or hypoxic
brain damage (n = 2).
14-3-3 Protein was detected in 2 patients with HSE
and in none with other neurological and psychiatric
diseases (see Table 5).
The positive predictive value for detection of 14-3-3
protein in sporadic cases, based on definite and probable cases, is 94.7% (95% CI, 89.5-97.9); the negative
predictive value is 92.4% (95% CI, 85.5-96.7), based
on cases seen in the study as suspected CJD cases with
another diagnosis at follow-up. As mentioned above, 7
patients who were false-positive were classified as no
CJD on the clinical examination. The predictive value
of 14-3-3 detection in CSF is higher in patients with
probable and possible CJD.
14-3-3 Protein was detected in both cases with the
36 Annals of Neurology Vol 43 No 1 January 1998
at follow-up (internal controls).
E200K mutation and in the case with the V210I mutation. In 5 patients with the D178N mutation, the test
was positive only in 1 case, as well as in 1 of both cases
with the P1O2L mutation (Table 6). The detection of
14-3-3 protein does not depend on the polymorphism at
codon 129. Sensitivity and specificity of 14-53 protein
WB for genetic cases were only 50% each. The patient
with iatrogenic CJD presented clinically as a probable
case and was 14-3-3 protein-positive.
Disease Duration a n d T i m e to Spinal Tap
The mean disease duration of patients with definite
sporadic CJD and the detection of 14-3-3 protein in
CSF was 6 months (range, 2.1-26.4 months). The
mean time from onset of symptoms to spinal tap was
3.9 months (range, 1.4-18.6 months). Six patients
were presented initially to the surveillance study as suspect cases and were classified as other, due to insuficient clinical signs and symptoms and/or ambiguous
EEG findings. Of these, 2 were later reclassified as definite and 4 as possible CJD, after clinical deterioration.
Five of these patients were clearly 14-3-3 proteinpositive in the CSF at entry into the study. This suggests that 14-3-3 protein may be detectable quite early
in the course of the disease.
The inconsistency of the detection of 14-3-3 protein
in the CSF of several genetic cases prompted us to analyze the relation between the presence or the absence
of 14-3-3 protein, and the disease duration and the
time point of the spinal tap, respectively (see Table 6).
In the D178N patient positive for 14-3-3 protein, the
Table 5. Detection of 14-3-3 Protein in Cerebrospinal Fluid
in Patients with Creutzfeldt-Jakob Disease and in Controls,
Based on the Sensitivity Increasing SpateD 3
Creutzfeldt-Jakob disease (CJD)
Other (internal controls)
Neurological (external) controls
"One patient with typical clinical symptoms and a typical electroencephalogram could not be classified as probable due to a very
rapid clinical progression to akinetic mutism and no formal testing
for dementia; one patient each, with bronchial carcinoma and intracerehral metastasis, metabolic encephalopathy after barbiturate intoxication, and progressive dementia and extrapyramidal symptoms
of unknown cause. In these cases, the determination of 14-3-3 was
done retrospectively; no cerebrospinal fluid could be examined at
follow-up. Three other patients were initially 14-3-3 positive. Because of the clinical classification as no CJD, a second lumbar puncture was performed 2 weeks later. All these patients showed clinical
improvement of their symptoms and were 14-3-3 negative on the
second lumbar puncture. The most likely diagnosis of these patients
was atypical encephalitis (n = 1) or hypoxic brain damage (n = 2).
Both cases with herpes simplex encephalitis were positive for
14-3-3, The diagnoses of the remaining 14-3-3 negative neurological controls (n = 58) are as follows: multiple sclerosis (n = 9), no
neurological disease (n = 7), optic neuritis (n = 4), Bell's palsy
(n = 4),SDAT (n = 2), viral encephalitis (n = 3 ) , depression (n =
3), borreliosis (n = 2), essential myoclonus (n = 2), epilepsy
(n = 2), vertigo (n = 2), and 1 case each of systemic lupus erythematosus, postinfectious encephalitis, tuberculous meningitis,
Guillain-Barrt syndrome, myopathies, spastic spinal paralysis, progressive spinal muscle atrophy, multiinfarct dementia, progressive
dementia without neurological symptoms, stroke, thrombosis of cerebral sinus, transient amnestic episode, meningiosis carcinomatosa,
bronchial carcinoma without intracerebral metastasis, hydrocephalus, brain atrophy, aneurysm, and meningioma.
spinal tap was performed very early after development
of the first symptoms (2 months; disease duration, 9
months), whereas the cases that were 14-3-3 proteinnegative were seen in the study 6 months after onset
(range, 4-10 months) with a mean disease duration of
12.5 months (range, 6-22 months). The disease duration in the patient with the P102L mutation who was
positive for 14-3-3 protein did not differ from that in
the case that was 14-3-3 protein-negative. The patient
was seen in the study 7 months after clinical onset of
disease (disease duration, 9 months); in the patient
who was 14-3-3 protein-negative, the spinal tap was
performed 6 months after the first symptoms (disease
duration, 15 months).
Detection of 14-3-3 Protein and Follow- Up
The results of 14-3-3 protein determination in CSF in
suspected sporadic CJD cases classified initially as possible or probable CJD, and the subsequent clinical
courses, are shown in Figure 2. Samples from 3 pa-
tients classified as having probable CJD, and shown
at autopsy to have Alzheimer's disease, were clearly
14-3-3 protein-negative on WB. Twelve patients were
classified initially as possible CJD cases and were later
reclassified as no CJD, after clinical improvement (n =
9) or because of other diagnoses made at autopsy (n =
3). All these patients were 14-3-3 protein-negative.
Twenty-five patients who were 14-3-3 protein-positive
were classified initially as possible CJD and were reclassified as definite (n = 19) or probable (n = 6 ) after
autopsy or clinical deterioration and appearance of typical EEG. Two cases classified initially as no CJD, and
later, at the time of autopsy, shown to be definite, were
clearly 14-3-3 protein-positive in CSF (results not
shown in Fig 2). These findings on probable and possible cases reflect the high predictive value of 14-3-3 protein detection in CSF in an appropriate clinical setting.
Clinical diagnosis of probable or possible CJD, in patients in our prospective series who were positive for
14-3-3 protein in CSF (60 of 63 patients), was confirmed by neuropathological examination in all cases.
The positive predictive value (probability that the patient has CJD if 14-3-3 protein, by WB, is positive)
in this subgroup is therefore very high. Based on all
cases examined, the positive predictive value is lower
(94.7%), because we identified 7 false-positive cases in
a single CSF test. All these false-positive cases were patients who did not even fulfill the criteria for possible
CJD. The second lumbar puncture, performed 2 weeks
later, was 14-3-3 protein-negative in these patients.
The negative predictive value (the probability that a
patient truly does not have CJD if 14-3-3 protein, by
WB, is negative) is 92.4%, based on the whole collective. In cases where clinical criteria for CJD were fulfilled (probable and possible cases), all patients who
were later reclassified as no CJD, due to clinical improvement of their illness or due to neuropathological
diagnosis, were 14-3-3 protein-negative. Positive and
negative 14-3-3 protein CSF findings were therefore
highly predictive for the final diagnosis in probable or
possible CJD cases. In contrast, we identified 3 confirmed CJD cases who were 14-3-3 protein-negative.
In these patients, only PSWCs on EEG corroborated
the clinical diagnosis of CJD. Absence of 14-3-3 protein in the CSF by WB analysis cannot ultimately rule
out the diagnosis of CJD but makes it unlikely. WB
analysis of 14-3-3 protein is a highly sensitive test to
support the clinical diagnosis of CJD, and it provides
valuable and specific information. ',' In the setting of
our CJD surveillance study, it has a sensitivity of
94.0% with a specificity of 93.3%. Detection of
14-3-3 protein in the CSF predicts the diagnosis
of CJD more specifically than any other clinical test,
and it adds useful information especially in CJD cases
Zerr et al: Detection of 14-3-3 Protein in CSF Suggests CJD
Table 6. Disease Duration, Codon 129 Polymorphism, and CSF Findings in Patients with PRNP Mutations
Time until
Codon 129
D 178N
P 102L
v 2 101
Spinal Tap”
“In months.
CSF = cerebrospinal fluid; PRNP
prion-related protein gene; M = methionine; V
Fig 2. Initial clinical diagnosis of suspected Creutzfeldtqakob
disease (CYD) cases (only sporadic cases), detection of 14-3-3
in cerebrospinul Juid (classiflcation of test results based on
strategy 3), and Jnal diagnosis. Numbers in parentheses refer
to the percentages of test results positive for 14-3-3 in the corresponding subpopulations. None of the group classifled initially as probable was classifled finally as possible.
without PSWCs during the patient’s lifetime, a condition that may occur in 40% of cases, even on repetitive
EEG recording. 13336 Nevertheless, the detection of
PSWCs provides important information and completes
the clinical diagnosis, because few CJD cases are
14-3-3 protein-negative. WB analysis of the CSF for
detection of 14-3-3 protein is more sensitive than twodimensional gel electrophoresis for detection of pl30/
131 proteins.l.2.1 1,30,37,38
Based on the validity and reliability of the 14-3-3
protein WB, which was performed twice in each case,
four strategies can be devised (see Table 2). Using
strategy 3, optimal results yielding a balance between
sensitivity and specificity were obtained for our data
set. The overall accuracy of the test under these conditions is 93.7%. It should be recognized that all these
Annals of Neurology
Vol 43
No 1 January 1998
estimates of diagnostic performance could not yet be
validated on an independent data set. Thus, as our
classification strategy has been selected by maximizing
sensitivity and specificity, our values may be slightly
optimistic and may need to be confirmed by external
validation. 14-3-3 protein is believed to indicate acute
neuronal loss and acute brain damage. 1,324,1 Therefore, it is not surprising to find 14-3-3 protein in the
CSF of patients with diseases accompanied by acute
neuronal damage such as ischemic stroke, subarachnoidal hemorrhage, and CJD. In our study, we detected
14-3-3 protein in patients with hypoxic brain damage,
paraneoplastic encephalopathy, HSE, and metabolic
brain damage after barbiturate intoxication. Detection
of 14-3-3 protein in the CSF of such patients limits
the value of this marker in clinical screening and emphasizes the need to use 14-3-3 protein as a marker for
CJD only in an appropriate clinical setting in cases fulfilling the clinical criteria for CJD.9p12Encephalitis,
hypoxic brain damage, infarction, or brain tumor, as a
cause of rapid progressive dementia, should be excluded by appropriate investigations, including CSF
analysis and brain imaging. Based on our findings,
blind testing of the samples, without clinical or paraclinical evidence of signs and symptoms suggestive of
CJD, cannot be recommended.
It is intriguing that 4 of 5 patients with the FFI genotype and l of 2 patients with the P102L mutation
(GSS genotype) were 14-3-3 protein-negative, whereas
patients with familial CJD genotype and the E200K or
the V 2101 mutation were clearly 14-3-3 protein-positive. The detection of 14-3-3 protein does not depend
on the polymorphism at codon 129, the disease duration, the disease progression, or the time of spinal tap
after beginning of the symptoms. This underscores the
need to combine biochemical tests of CSF with genetic
analysis of the PRNP in suspect cases of CJD.39 Although the detection of 14-3-3 protein in CSF is
thought to reflect acute brain damage, the presence of
14-3-3 protein in 1 patient with D178N mutation and
in 1 patient with P102L mutation cannot be explained
by a more rapid disease progression, because the overall
disease duration does not differ in genetic cases that are
positive or negative for 14-3-3 protein. The number of
patients with proven PRNP mutations, seen thus far, is
too small to draw any definite conclusions. It is tempting to speculate, however, that these differences may be
due to ( 1 ) different localization of the lesions or (2) rate
of neuronal loss. Another cause of false-negative findings
may be suboptimal sample storage‘; but this does not
explain the results of the present study, because all samples were stored on dry ice immediately after the lumbar
WB analysis of 14-3-3 protein in CSF is of particular value because it allows accurate diagnosis at a very
early stage of the disease, long before typical clinical
symptoms such as severe dementia and myoclonus are
evident, as can be shown in some cases in our study.
These findings indicate that 14-3-3 protein may be
present in CSF at a very early stage of the disease.40
These data support preliminary results of an animal
study. 1,40
WB analysis of CSF for 14-3-3 protein should be
performed in every suspect case of CJD, to validate the
clinical diagnosis. Although detection of 14-3-3 protein
in CSF is not appropriate as a screening test, we believe
that in patients with clinically suspected CJD who fulfill the diagnostic criteria but without typical PSWCs,
clinical diagnosis of CJD can be substantiated by the
detection of 14-3-3 protein in CSF. This test helps to
identify true CJD cases in the group of those with possible CJD and to distinguish CJD from atypical Alzheimer’s disease with myoclonus and PSWCs. Biochemical markers in the CSF, such as NSE, pl30/ 13 1 ,
and 14-3-3 protein, should be included together with
the currently used EEG and clinical criteria for the diagnosis of CJD.1’2’10-13In cases that do not meet the
clinical classification criteria but are 14-3-3 proteinpositive, a second lumbar puncture may help to distinguish between patients in the early stages of CJD and
patients with hypoxic brain damage or atypical encephalitis. Further data are required ( 1 ) to determine the
optimal time point for a second lumbar puncture, and
(2) to validate this assumption.
We believe that the detection of 14-3-3 protein in the
CSF is a more sensitive test than EEG, to support the
clinical diagnosis of CJD, and it should be included in
the evaluation of suspect cases of sporadic CJD. Because
incidence figures are calculated based on neuropathologically confirmed cases and cases classified clinically as
probable based on typical EEG recordings, which are
found in only 60 to 70% of sporadic
a better
monitoring of incidence figures may be achieved by including CSF markers in the diagnostic criteria.
We have shown that 14-34 protein, as detected by an
antiserum to the p isoform of 14-3-3 protein, is a stable
marker. Given the stability of 14-3-3 protein in the
CSF, the failure to detect a positive reaction by WB is
either due to a very low level of 14-3-3 protein in the
CSF or to the absence of the protein. A further refinement of the accuracy of 14-3-3 protein analysis may be
achieved by using a quantitative test such as an enzymelinked immunosorbent assay, using monoclonal antibodies to the presumed brain-specific subunits y and 5 of
the 14-3-3 protein.41 A comparative analysis of these
isoforms may help to distinguish between CJD and
other TSEs, and other conditions causing the release of
14-53 protein such as HSE, hypoxic brain damage,
metastatic carcinoma, and barbiturate into~ication.~
Our preliminary data suggest adding the term
laboratory-supportedCJD to the currently used diagnostic criteria of CJD.9-12,28 By extending our previous
work based on the analysis of cases seen prospectively
in the German case-control study, we now provide
further evidence for the inclusion of biochemical CSF
parameters into the currently used clinical diagnostic
criteria of CJD.2,1’”2’27228Sensitivity, specificity, and
positive and negative predictive values of 14-3-3 protein detection in CSF surpass the diagnostic accuracy
of the currently used EEG criteria, to achieve a probable diagnosis of CJD. However, this test must be further evaluated in patients with dementia and other
neurodegenerative disorders. We emphasize that only
those patients who meet the clinical criteria for at least
“possible CJD” and who are 14-3-3 protein-positive
should be considered as having laboratory-supported
CJD. Until there are means to identify reliably the
pathological agent of TSEs in clinical samples, the detection of 14-3-3 protein in CSF should be regarded as
a characteristic and highly suggestive, but nonspecific,
marker for CJD that always must be interpreted within
the context of the clinical differential diagnosis.
This study was supported by a grant from the Bundesgesundheitsministerium fur Gesundheit to H.A.K. and to S.P. (BMG
We gratefully acknowledge the work of the site visiting physicians
Maria Lantsch, Kati Weidehaas, Anke Otto, and Christian Riedemann. We thank Katharina Skworc and Stefanie Arendt for genetic
analysis and Bernhard Steinhoff for the EEG analysis. We also
thank all physicians notifying suspect cases to the German CJD surveillance unit, for providing pertinent clinical, neuroradiological,
and biochemical data, as well as for their help in obtaining CSF
specimens. Finally, we thank Asta Medica, Frankfurt, Germany, for
its financial support.
1. Hsich G, Kenney K, Gibbs CJ, et al. The 14-3-3 brain protein
in cerebrospinal fluid as a marker for transmissible spongiform
encephalopathies. N Engl J Med 1996;335:924-930
2. Zerr I, Bodemer M, Weber T. The 14-3-3 brain protein and
Zerr et al: Detection of 14-3-3 Protein in CSF Suggests CJD 39
transmissible spongiform encephalopathy. N Engl J Med 1997;
336:874 (Letter)
Aitken A. 14-3-3 proteins on the MAP. Trends Biochem Sci
Aicken A, Collinge DB, van Heusden BP, et al. 14-3-3
Proteins: a highly conserved, widespread family of eukaryotic
proteins. Trends Biochem Sci 1992;17:498-501
Moussavian M, Potolicchio S, Jones R. The 14-3-3 brain protein and transmissible spongiform encephalopathy. N Engl
J Med 1997;336:873-874
Will RG, Zeidler M, Brown P, et al. Cerebrospinal fluid test
for new-variant Creutzfeldt-Jakob disease. Lancet 1996;348:955
Prusiner SB, Hsiao KK. Human prion diseases. Ann Neurol
Kretzschmar HA, Ironside J , DeArmond S, Tateishi J. Diagnostic criteria for sporadic Creutzfeldt-Jakob disease. Arch Neurol
Masters CL, Harris JO, Gajdusek DC, et al. Creutzfeldt-Jakob
disease: patterns of worldwide occurrence and the significance of
familial and sporadic clustering. Ann Neurol 1979;5:177-188
Will RG, Ironside JW, Zeidler M, et al. A new variant of
Creutzfeldt-Jakob disease in the UK. Lancet 1996;347:921-925
Zerr I, Bodemer M, Otto M, et al. Diagnosis of CreutzfeldtJakob disease by two dimensional gel electrophoresis of cerebrospinal fluid. Lancet 1996;348:846-849
Zerr I, Bodemer M, Racker S, et al. Cerebrospinal fluid concentration of neuron-specific enolase in diagnosis of CreutzfeldtJakob disease. Lancet 1995;345:1609-1610
Steinhoff BJ, Racker S, Herrendorf G, et al. Accuracy and reliability of periodic sharp wave complexes in Creutzfeldt-Jakob
disease. Arch Neurol 1996;53:162-165
Chazot G, Broussolle E, Lapras CI, et al. New variant of
Creutzfeldt-Jakob disease in a 26-year old French man. Lancet
1996;347:1181 (Lerter)
Kuroiwa Y, Celesia GG, Chung HD. Periodic EEG discharges
and status spongiosus of the cerebral cortex in anoxic
encephalopathy: a necropsy case report. J Neurol Neurosurg
Psychiatry 1982;45:740-742
Kuroda Y, Ikeda A, Kurohara K, et al. Occurrence of paroxysmal synchronous EEG discharges in subcortical arteriosclerotic
encephalopathy (Binswanger’s disease). Intern Med 1993;32:
Smith SJ, Kocen RS. A Creutzfeldt-Jakob-like syndrome due to
lithium toxicity. J Neurol Neurosurg Psychiatry 1988;51:120123
Koponen H, Honkonen S, Partanen J, Riekkmen PJ. Epileptic
attack, delirium, and a Creutzfeldt-Jakob-like syndrome during
mianserin treatment. Neuropsychobiology 1990;23: 164-1 68
Takahashi M, Kubota F, Nishi Y,Miyanaga K. Persistent synchronous periodic discharges caused by anoxic encephalopathy
due to cardiopulmonary arrest. Clin Electroencephalogr 1993;
24: 166-172
Brick JF, Brick JE, Morgan JJ, Gutierrez AR. EEG and pathologic findings in patients undergoing brain biopsy for suspected
encephalitis. Eleccroencephalogr Clin Neurophysiol 1990;38:
Tagliavini F, Prelli F, Porro M, et al. A soluble form of prion
protein in human cerebrospinal fluid: implications for prionrelated encephalopathies. Biochem Biophys Res Commun
Kawashima T, Furukawa H, Doh-ura K, Iwaki T . Diagnosis of
new variant Creutzfeldt-Jakob disease by tonsil biopsy. Lancet
1997;350:68-69 (Letter)
40 Annals of Neurology
Vol 43
No 1 January 1998
23. Hill AF, Zeidler M, Ironside J, Collinge I. Diagnosis of new
variant Creutzfeldt-Jakob disease by tonsil biopsy. Lancet 1997;
349:99-100 (Letter)
24. Jimi T, Wakayama Y, Shibuya S, et al. High levels of nervous
system-specific proteins in cerebrospinal fluid in patients with
early stage Creutzfeldt-Jakob disease. Clin Chim Acta 1992;
25. Mokuno K, Kato K, Kawai K, et al. Neuron specific enolase
and S-100 protein levels in CSF of patients wich various neurological diseases. J Neurol Sci 1983;60:443-45 1
26. O’Brien C. Mad cow disease: scant data cause widespread concern. Science 1996;271:1798
27. Otto M, Wiltfang J, Tumani H, et al. Elevated levels of tauprotein in CSF of patients with Creutzfeldt-Jakob disease. Neurosci Lett 1997;225:210-212
28. Otto M, Stein H , Szudra A, et al. Elevated level of SlOO protein concentration in CSF of patients with Creutzfeldt-Jakob
disease. Akt Neurol 1996;23:71 (Abstract)
29. Finkenstaedt M, Szudra A, Zerr I, et al. Magnetic resonance
imaging of Creutzfeldt-Jakob disease. Radiology 1996;199:793798
30. Freed E, McCormick F, Ruggieri R. Proteins of the 14-3-3
family associate with Raf and contribute to its activation. Cold
Spring Harb Symp Quant Biol 1994;59:187-193
31. Will RG. Slow virus infection of the central nervous system. In:
Hayward RD, Quinn NP, Wade JPH, Shorvon SD, eds. Current medical literature. Neurology 1991;7:67-71
32. Sambrook J, Fritsch EF, Maniatis T. Molecular cloning: a laboratory manual, vol 2. 2nd ed. Cold Spring Harbor NY: Cold
Spring Harbor Laboratory Press, 1989
33. Nicholl D, Windl 0, de Silva R, et al. Inherited CreutzfeldcJakob disease in a British family associated with a novel 144
base pair insertion of the prion protein gene. J Neurol Neurosurg Psychiatry 1995;58:65-69
34. Orita M, Iwahana H, Kanazawa H, et al. Detection of polymorphism of human DNA by gel electrophoresis as singlestrand conformation polymorphism. Proc Natl Acad Sci USA
35. Windl 0, Dempscer M, Estibeiro JP, et al. Genetic basis of
Creutzfeldt-Jakob disease in the Unired Kingdom: a systematic
analysis of predisposing mutations and allelic variation in the
PRNP gene. Hum Genet 1996;98:259-264
36. Brown P, Gibbs CJ Jr, Rodgers-Johnson P, et al. Human spongiform encephalopathy: the National Institutes of Health series
of 300 cases of experimentally transmitted disease. Ann Neurol
1994;35:5 13-529
37. Chen Z, Fu H, Liu D, et al. A NaCI-regulated plant gene encoding a brain protein homology that activates ADP ribosyltransferase and inhibits protein kinase C. Plant J 1994;6:729-740
38. Tanji M, Horwitz R, Rosenfeld G, Waymire JG. Activation of
protein kinase C by purified bovine brain 14-3-3: comparison
wich tyrosine hydroxylase activation. J Neurochem 1994;63:
39. Mascrianni JA, Iannicola C, Myers RM, et al. Mutation of the
prion protein gene at codon 208 in familial Creuczfeldt-Jakob
disease. Neurology 1996;47:1305-1312
40. Kenney K, Harrington MG, Gibbs CJJ. The 14-3-3 brain protein and transmissible spongiform encephalopathy. N Engl
J Med 1997;336:874-875
41. Martin H , Rostas J, Patel Y, Aitken A. Subcellular localization
of 14-3-3 isoforms in rat brain using specific antibodies. J Neurochem 1994;63:2259-2265
Без категории
Размер файла
942 Кб
detection, jakos, support, protein, disease, diagnosis, fluid, cerebrospinal, creutzfeldt
Пожаловаться на содержимое документа