вход по аккаунту


Cerebrospinal fluid polyamines Biochemical markers of malignant childhood brain tumors.

код для вставкиСкачать
Cerebrospinal Fluid Polyamines:
Biochemicd Markers of Malignant
Chddhood Brain Tumors
Peter C. Phillips, MD," Leon T. Kremzner, PhD,t and Darryl C. De Vivo, MDS
The clinical value of cerebrospinal fluid (CSF) polyamine determinations in childhood medulloblastoma has been
suggested. We performed 72 CSF polyamine determinations in 35 children with primary brain tumors. Spermine
values were normal and spermidine values were inconsistently elevated. CSF putrescine values, however, were consistently elevated in patients with histologically malignant brain tumors: medulloblastoma, epeodymoma, pineal germ
cell tumors, primitive neuroectodermal tumors, and brainstem gliomas. Children with supratentorial astrocytomas had
normal CSF polyamine values. CSF putrescine values were closely correlated with clinical state, with the highest
concentrations identified in patients with widely disseminated recurrent disease. We found CSF putrescine to be a
sensitive indicator of active disease in childhood malignant brain tumors. Further investigation is warranted into the
predictive value of CSF polyamines in determining tumor relapse before clinical or other diagnostic studies reveal
recurrent disease.
Phillips PC, Kremzner LT, De Vivo DC: Cerebrospinal fluid polyamines: biochemical markers of malignant
childhood brain tumors. Ann Neurol 19:360-364, 1986
The polyamines spermidine and spermine and the
diamine putrescine (hereafter referred to collectively
as PA) are low-molecular-weight aliphatic m i n e s
whose regulatory functions in cellular growth are incompletely understood [ 153. Although the association
of elevated urinary PA concentrations with certain disease states was reported in 1889 [25], the clinical value
of PA determinations in neoplastic disease has been
examined only recently. Studies by D. H. Russell and
others have supported the clinical value of serum and
urine PA determinations as a useful indicator of therapeutic response in patients with various systemic
malignancies [2, 18-21].
Cerebrospinal fluid (CSF) PA concentrations are
frequently elevated in the presence of primary central
nervous system (CNS) malignancy. Different CNS
tumors may demonstrate significant variations in CSF
PA concentrations or in the relative proportions of
putrescine, spermidine, and spermine. The high incidence of false-positive and false-negative CSF PA
values in adult malignant gliomas limits the usefulness
of the determination for predicting tumor recurrence
or for monitoring clinical status [3f. By contrast, the
characteristics of many childhood primary brain
tumors (proximity to the cerebral ventricles, rapid
growth rate, and tendency to disseminate throughout
the subarachnoid space) make these tumors ideal for
study of CSF PA concentrations.
Marton and associates [12f have reported an extensive longitudinal study of CSF PA values in childhood
medulloblastoma. Their findings suggest that CSF PA
determinations are sensitive markers of tumor activity
and of response to therapy. No false-positive CSF PA
values were detected in 210 determinations from 32
patients with medullobiastoma. Elevated CSF PA
values were the earliest indicators of recurrent tumor
in several patients, further underscoring the clinical
utility of these biological markers of malignancy.
We report the results of single and serial CSF PA
determinations in 35 children with primary brain
tumors. Our study population is broadly representative
of the most common types of childhood brain tumors.
Our findings confirm and extend previous studies advocating PA levels as useful CSF markers of malignant
brain tumors in childhood.
From the Departments of $Pediatrics, tRehabilitation Medicine, and
SNeurology,Division of Pediatric Neurology, the College of Physicians and Surgeons of Columbia University, The Neurological Institute of The Presbyterian Hospital in the City of New York, New
York. NY 10032.
Received Apr 16, 1985, and in revised form Aug
- 7. Accepted for
publication Aug 26, 1985.
Materials and Methods
Lumbar CSF samples were obtained from 35 children with
known or suspected primary brain tumors. PA control values
were determined in lumbar CSF specimens from 4 neurolog-
Address reprint requests
York, NY 1o032.
Dr De viva,
West 168th St, New
*Present address: Department of Neurology, Memorial SloanKettering Cancer Center, 1275 York Ave, New York, N Y 10021.
Table I . Cerebrospinal Fluid Polyamine Values in Children and Adults
CSF Polyamine (pmoVml)a
No. of
Mean Age,
yr (range)
Present studyb
Albright et al [ l ]
Marton et a1 [ 131
Marton et al El31
5.9 (2-17)
7.4 (1-16)
6.6 (2-9)
42 (19-65)
212 t 43
316 i 94
196.8 t 59.9
181 ? 53.5
116 t 40
189 t 60
136 t 46
153 t 49
"Values are mean
'Normal CSF cell count, protein, and glucose.
CSF = cerebrospinal fluid.
Table 2. Cerebrospinal Fluid Polyamine Values in Patients with Medulloblastoma
1. Tumor present ( n = 10)
2. Tumor absent
a. Consistently normal (n = 4 )
b. Transient elevation (n = 3)
815 t 334b
373 t 188b
250 i 104
435 i 287'
183 -+ 61
192 i 74
"Values are mean c SD.
Significantly different from control subjects: ' p < 0.001; ' p < 0.01.
cerebrospinal fluid; PA
ically normal children and 4 children with nonneoplastic
neurological diseases (Table 1). CSF samples were acidified
with 0.3 M perchloric acid, centrifuged at 30,000 g for 10
minutes, and the supernatants were frozen at -20°C.
Thawed samples were diluted with an equal volume of 12 N
hydrochloric acid and hydrolyzed under vacuum at 110°C for
20 to 24 hours, followed by drying under vacuum and resuspension in water. PA assays were performed with an automatic amino acid analyzer using HPAN-90 cationic resin
(Hamilton Co), column dimensions 0.5 X 7.0 cm. The initial
elutions were made with 0.35 sodium citrate (pH, 5.25) containing 296 1-propanol. At 30 minutes the sodium citrate
buffer was adjusted to 2.35 M. The pumping rate was 45 mV
minute and column temperature was maintained at 43.5"C.
PAS were quantitated with ophthaldehyde and fluorescence
was detected with a SpectroGlo Fluorometer (Gilson Corp)
[26]. The lower limit of assay sensitivity is 25 pmol.
Statistical analysis was performed using Student's t test
Table 1 indicates the population profile and CSF PA
values in control subjects. CSF spermine, inconsistently detected in both control and brain tumor populations, was excluded from data analysis. As no difference between the neurologically normal and the
nonneoplastic neurological disease control groups was
found, these CSF PA values were combined. A comparison of CSF PA values in control populations between the present study and previously reported studies in children and adults indicates good agreement
(Table 1).
Forty-four CSF PA determinations were performed
in 13 children with medulloblastoma (Table 2). Six
children had CSF PA determinations only while tumor
was present, 3 children had determinations only while
tumor was absent, and 4 children had determinations
during both clinical states. Putrescine and spermidine
values were significantly greater in all 10 children
(Group 1) with tumor present than those values in
controls ( p < 0.001). Twenty-three CSF PA determinations were performed during clinical remission in 7
children (Group 2) with medulloblastoma. N o child
had radiographic or CSF cytological evidence of tumor
recurrence. Two distinct patterns emerged from analysis of Group 2. Consistently normal CSF PA values
were recorded in 4 patients (Group 2a). Transient elevations in CSF putrescine values were recorded in 3
children (Group 2b) without evidence of hydrocephalus or recurrent tumor (Fig 1). Putrescine values returned to the normal range without intervening therapy. Spermidine values were normal during transient
putrescine elevations in 2 of the 3 patients in Group
2b. All patients with transiently elevated PA values
have remained in clinical remission for 26 to 40
months after the last PA determination.
CSF PA values in 14 children with histologically
malignant tumors other than medulloblastoma demonstrated elevated putrescine concentrations in all tumor
groups during active disease: primitive neuroectodermal tumors (PNET), p < 0.05; pineal germ cell
tumors, p < 0.01; ependymomas, p < 0.05; brainstem
gliomas, p < 0.05 (Fig 2). Elevated spermidine values
Phillips et al: CSF Polyamines and Childhood Tumors 361
lsl PUT
Fig 1 , Transient putrescine (PUT) elevation without evidence of
recurrent disease in 3 patients with medulloblastoma. spermidine
values were n o w 1 during putrescine elevation. All 3 patients
have remained in clinical remission 26 to 40 months after the
last cerebrospinalfiuid polyamine determination.
1300 -
1200 -
900 -
E" 1100'
300--200100 -
Fig 2. Mean cerebrospinalfiuid polyamine values (2SO) in
children with primary brain tumors measured during active disease. Dotted line indicates upper limit of the normal putrescine
range. Putrescine concentration (Put) was a more sensitive indicator of active tumor than spermidine (Spd)for all tumor
groups. Neither putrescine nor spermidine was a sensitive marker
for supratentorial astrocytomas. The values as indicated were
significantly different from control subjects: a = p < 0.001;
b = p < 0.01; c = p < 0.05. (PNET = primitive
neuroectodermal tumors.)
were associated with tumor presence only in patients
with PNET. One false-negative CSF putrescine and
spermidine value was noted in a child with PNET. No
PA values were obtained during clinical remission in
this group.
CSF PA determinations were performed in 8 children with low-grade supratentorial astrocytomas. Five
of these children had deep hemispheric tumors, and
the other 3 had tumors in the thalamus or third ventricle. N o significant elevations of the mean putrescine or
362 Annals of Neurology Vol 19 No 4 April 1986
Fig 3 . Mean cerebrospinalfiuidpolyamine values I SD) in
histologically malignant tumors obtained at initial presentation
(preoperative,postoperativelpreradiotherapy [RT)) or tumor recurrence (either local or disseminated). Putrescine concentration
(Put), greater than control values in all clinical states, was highest during tumor recurrence. Spemidine values (Spd) showed little variation among these four clinical states. The values as indicated were significantly d;fferentfrom control subjects: a = p <
0.05; b = p < 0.01.
spermidine values were detected in either subgroup
alone, although the mean putrescine value was elevated when the two groups were combined ( p < 0.05).
Combined PA values from histologically malignant tumors (medulloblastoma, PNET, pined region
germ cell tumors, ependymoma) were compared in
four clinical states: preoperative; postoperative/preradiotherapy; local tumor recurrence without CSF dissemination; local tumor recurrence with CSF dissemination (Fig 3). Recurrent tumors were documented
radiographically and all patients with CSF dissemination had abnormal CSF cytology or myelograms. CSF
samples from patients with recurrent tumor were obtained at the earliest clinical or radiographic evidence
of recurrence. Spermidine values showed little variation between clinical states. Compared to the control
group, putrescine values were significantly elevated in
patients at initial presentation (preoperative or postoperative/preradiotherapy, p < 0.03) and recurrent disease (local or widely disseminated, p < 0.01). Elevated
putrescine values were not statistically different between preoperative and postoperative/preradiotherapy
groups. Putrescine values in patients with recurrent
tumor were greater than those found at initial presentation (p < 0.05). The highest putrescine values were
found in patients with widely disseminated recurrent
The relationship between tumor cell growth, cell injury or death, and extracellular PA concentration is
complex. Intracellular PA concentrations are linked
directly to cell growth and proliferation in normal [IS]
and neoplastic [71 tissue. Although PA concentrations
are high in normal brain {8, 91, they are significantly
greater in histologically malignant brain tumors [ S ,
141. Hypercellular tumors (medulloblastoma, PNET,
glioblastoma) have higher PA concentrations than hypocellular tumors [ S , 141. Intracellular putrescine concentration most closely reflects the degree of malignancy, although both putrescine and spermidine
concentrations are increased in malignant brain tumors
[ S } . Harik and Sutton [ S } emphasize that putrescine
tissue levels directly reflect PA biosynthesis. Putrescine is the immediate product of the enzymatic decarboxylation of ornithine. The activity of ornithine decarboxylase, the rate-limiting enzyme in PA synthesis,
is greatest in highly malignant brain tumors {22, 231.
The effect of tumor growth and necrosis on CSF PA
concentration is not well understood; cellular reuptake
of PAS in extracellular fluid and variable clearance
rates of CSF PA when CSF dynamics are altered by
hydrocephalus may be further complicating factors.
The proposed mechanisms relating intracellular and
CSF PA concentrations during tumor proliferation and
death are controversial. Heby and Andersson 161 suggest that extracellular PA concentration reflects the
release of PA from dead or dying tissue. However, the
putrescine-spermidine ratio in normal cortex (0.14)
and malignant brain tumor (0.43) is reversed in the
CSF of normal subjects (1.2) and brain tumor patients
(1.4) [3, 5 ) . CSF PA values were not found to correlate with pathological or radiological evidence of necrosis in one study [ 3 ] . Russell proposed a model for
the interpretation of extracellular PA with respect to
cellular growth: extracellular putrescine is the major
PA associated with rapid cell growth, whereas spermidine values reflect spontaneous cell loss or cell destruction { 181. Evidence supporting this model has
been reported in studies of urinary PA concentrations
in cancer patients undergoing chemotherapy [Z, 191.
However, it remains essentially untested for CSF PA.
Clinical studies of CSF PA as biochemical markers
of primary or metastatic CNS malignancy have been
reported in adults {3, 4 , 10, 131 and children { I , 11,
12, 17, 24). Determination of CSF PA values has been
found to be insufficiently sensitive or specific to be of
clinical value for many brain tumors in adults 131.
However, the association of increased PA values with
hypercellular tumors and tumors in proximity to the
ventricular system { 3 ] , both features more typical of
childhood than adult brain tumors, suggests that CSF
PA values should be excellent markers of CNS malignancy in childhood. The positive correlation of CSF
PA values with tumor proximity to the ventricles is
best explained by the measured PA capillary permeability and diffusion coefficients in normal rat and cat
brains, respectively { 161. The putrescine permeability
coefficient was greater whereas the diffusion coefficient for putrescine was lower than respective values
for urea. These results suggest that putrescine released
by tumor will be cleared more readily by entering local
capillaries or nearby cells than by bulk transport of
extracellular fluid.
Studies of CSF PA values in childhood cancer identify patient age, presence of hydrocephalus, and CSF
sample site (lumbar or ventricular) as important considerations in the interpretation of CSF PA values.
Albright and co-workers { l }measured CSF PA concentrations in 7 6 children without neurological disease. They found the highest values in premature and
newborn infants, declining values during the first year,
and stable values after the first year. They also noted
elevated CSF PA values in children with hydrocephalus. However, in children with both hydrocephalus and
brain tumors, the PA values did not correlate with the
degree of ventriculomegaly, leading these authors to
conclude that CSF PA concentrations are more directly related to tumor type than is the degree of ventricular enlargement. PA concentrations are typically
higher in lumbar than in ventricular CSF. Albright and
associates [ l ]did not find significant differences in PA
concentrations between the lumbar and ventricular
CSF in children with brain tumors.
The data reported in this survey of childhood brain
tumors confirm the previously reported positive correlation of CSF PA values and clinical state in medulloblastoma [ 11, 12) and indicate that the clinical correlation of these markers extends to other histologically
malignant paraventricular tumors including PNET,
pineal germ cell tumors, ependymomas, and posterior
fossa malignant gliomas. The normal values of CSF PA
in low-grade astrocytomas suggest that CSF PA determinations in these tumors have little clinical importance. We found the putrescine level to be a more
sensitive tumor marker than the spermidine level for
all tumor groups. No patient had elevated spermidine
values who did not also have a raised putrescine concentration. CSF PA values were not significantly different from control values for low-grade supratentorial
astrocytomas located adjacent to the ventricles or distant from the ventricles. These results support the hypothesis that endogenous PA production is low in
these slowly growing tumors [Sl.
The ability to monitor treatment effect and accurately predict tumor recurrence before the manifestation of other clinical evidence using CSF PA
levels is directly dependent on the sampling frequency.
Using frequent CSF surveillance, Marton and COworkers { l l , 12) detected significant CSF putrescine
elevations in medulloblastoma patients before clinical
evidence of relapse. Although we were not able to
establish the predictive value of CSF PA determinations for tumor relapse before the occurrence of other
Phillips et al: CSF Polyamines and Childhood Tumors
clinical evidence in medulloblastoma or other childhood brain tumors, the number of serial determinations performed and the frequency of CSF PA determinations in these patients were less than those used in
the study by Marton and associates.
Marton and co-workers 1121 emphasized the absence of false-positive values in their series. The occurrence of multiple transient putrescine elevations with
or without elevated spermidine values represents
either the first reported false-positive PA values in
medulloblastoma or evidence of a successful host response to attempted tumor recurrence. These cases
further illustrate the importance of serial CSF PA determinations in individual patients. Until greater experience with CSF PAS is gained, caution should be exercised in the interpretation of elevated CSF PA values
without corroborative clinical, radiographic, or cytological evidence of tumor recurrence. The temporal
profile of CSF PA values, in fact, may prove to be
more important than single determinations in any
given patient.
Regular CSF surveillance in pediatric cancer patients
at high risk for CNS dissemination has become an
integral part of patient management. For brain tumor
patients, computed tomographic or magnetic resonance imaging scans at regular intervals may detect
tumor progression or recurrence before overt clinical
manifestations, yet the tumor burden may be considerable by the time radiographic evidence of recurrence
becomes apparent. Efforts must be intensified to improve the sensitivity of surveillance techniques, permitting the earlier detection and treatment of progressive or recurrent tumor. Our results indicate that
further evaluation of serial CSF PA determinations is
warranted in all histologically malignant brain tumors
of childhood.
1. Albright AL, Marton LJ, Lubich WP, Reigel DH: CSF polyamines in childhood. Arch Neurol 40:237-240, 1983
2. Durie BG, Salmon SE, Russell DH: Polyamines as markers of
response and disease activity in cancer chemotherapy. Cancer
Res 37:214-221, 1977
3. Fulton DS, Levin VA, Lubich WP, et al: Cerebrospinal fluid
polyamines in patients with glioblastoma multiforme and anaplastic astrocytoma. Cancer Res 40:3293-3296, 1980
4. Fulton DS, Martin LJ, Lubich WP, Wilson CB: Polyamine levels
in CSF from patients with pituitary tumors or non neoplastic
pituitary diseases. Arch Neurol 39:47-48, 1982
5 . Harik SI, Sutton CH: Putrescine as a biochemical marker of
malignant brain tumors. Cancer Res 39:5010-5015, 1979
364 Annals of Neurology Vol 19 No 4 April 1986
6. Heby 0,Anderson G: Tumor cell death: the probable cause of
increased polyamine levels in physiological fluids. Acta Pathol
Microbiol Scand [A]86:17-20, 1978
7. Heby 0, Marton LJ, Wilson CB, Martinez HM: Polyamine
metabolism in a rat tumor cell line: its relationship to the growrh
rate. J Cell Physiol 86:511-522, 1975
8. Kremzner LT: Metabolism of poiyamines in the nervous system.
Fed Proc 29:1583-1588, 1970
9. Kremzner LT: Polyamine metabolism in normal and neoplastic
neural tissue. In Russell D H (ed): Polyamines in Normal and
Neoplastic Growth. New York, Raven, 1973
10. Marton LJ: Polyamines and brain tumors. Natl Cancer Inst
Monogr 46:127-131, 1977
11. Marton LJ, Edwards MS, Levin VA, et al: Predictive value of
cerebrospinal fluid polyamines in medulloblastoma. Cancer Res
39:993-997, 1979
12. Marton LJ, Edwards MS, Levin VA, et al: CSF polyamines: a
new and important means of monitoring patients with medulloblastoma. Cancer 47:757-760, 1981
13. Marton LJ, Heby 0,Levin VA, et al: The relationship of polyamines in serological fluid to the presence of central nervous
system tumors. Cancer Res 36973-977, 1976
4. Moulinoux JP, Quemener V, LeCalve M, Darcel F: Polyamines
in human brain tumors. J Neuro-Oncol 2:153-158, 1984
5. P e g AE, McCann PP: Polyamine metabolism and function.
Am J Physiol 243:c212-221, 1982
6. Pierangelli E, Levin VA, Seidenfeld J, Marton LJ: Putrescine
diffusion in cat brain and capillary permeability in rat brain:
relation to CSF putrescine levels in brain tumor patients. Eur J
Cancer 17:143-147, 1981
17. Rennert OM, Lawson DL, Shukla JB, Miale TD: Cerebrospinal
fluid polyamine monitoring in central nervous system leukemia.
Clin Chim Acta 75:365-369, 1977
18. Russell DH: Clinical relevance of polyamines as biochemical
markers of tumor kinetics. Clin Chem 23~22-27, 1977
19. Russell DH, Durie BGM, Salmon SE: Polyamines as predictors
of success and failure in cancer chemotherapy. Lancet 2:797799, 1975
20. Russell D H , Levy CC, Schimpff SC, Hawk IA: Urinary polyamines in cancer patients. Cancer Res 31:1553-1558, 1971
2 1. Russell D H , Russell SD: Comparison of the relative usefulness
of serum, plasma and urine levels of polyamines as biochemical
markers of cancer. Clin Chem 212360-863, 1975
22. Scalabrino G , Modena D, Ferioli ME, et al: Degrees of malignancy in human primary central nervous system tumors: ornithine decarboxylase levels are better indicators than adenosyl
methionine decarboxylase levels. J Natl Cancer Inst 68~751754, 1982
23. Seldenfeld J, Marton LJ: Biochemical markers of central nervous system tumors measured in cerebrospinal fluid and their
potential use in diagnosis and patient management: a review. J
Natl Cancer Inst 63919-931, 1979
24. Smith BJ, Lee S, Sabia H , et al: Cerebrospinal fluid polyamines
in childhood leukemia. Ann Clin Lab Sci 14:225-231, 1984
25. Udranszky LV, Baumann E: Ueber das Vorkommen von
Diaminen, sogenannten Promainen, bei Cystinurie. 2 Physiol
Chem 13:562, 1889
26. Walters AS, Cote LJ, Perumal AS, Kremzner L T The effect of
polyamines on tyrosine hydroxylase and L-aromatic acid decarboxylase. Neurochem Res 7:977-985, 1982
Без категории
Размер файла
513 Кб
markers, malignant, childhood, brain, biochemical, tumors, fluid, cerebrospinal, polyamine
Пожаловаться на содержимое документа