[CANCER RESEARCH55, 420-427, January 15, 19951
Cell Type-specific Fingerprinting of Meningioma and Meningeal Cells by Proton
Nuclear Magnetic Resonance Spectroscopy'
Catarina L. Florian,2 Nicholas E. Preece, Kishore K. Bhakoo, Stephen R. Williams, and Mark D. Noble
The Royal College of Surgeons Unit of Biophysics, Institute of Child Health, 30 Guilford Street, London WCIN JEH (C. L F., N. E. P., K. K. B., S. R. W.J; Department of
Neurobiology, Ludwig institute for Cancer Research. Courtauld Building, 91 Riding House Street, London WJP 8BT; and Departments of Biochemistry and Molecular Biology,
Anatomy and Developmental Biology, University College London. London (M. D. NJ, United Kingdom
ABSTRACT
We compared the properties of six human meningiomas with normal
rat meningeal cells using cell culture techniques, high resolution in vitro
‘H-NMR (nuclear magnetic resonance) spectroscopy,
graphic analysis. Cell cultures were immunocytochemicauy
and chromato
characterized
at all stages with specific antibodles Quantitative and qualitative metab
olite assessments
in cell extracts were obtained from ‘H-NMRspectra and
chromatographic analysis. Human meningioma cells expressed a charac
teristic spectrum of metabolltes including free amino acids, compounds
related to membrane phospholipid metabolism, energy metabolites, and
other intermediary products. These spectral characteristics, although dif
ferent in some respects, were strikingly similar to the ones of rat menin
geal cells. Particularly, several metabolites that allow discrimination
be
tween meningeal cells and other cell types of the central nervous system
were preserved in meningiomas. These similarities suggest that the regis
lation of intracellular
levels of such metaboiltes is so intrinsic to the
identity of cell type as to be conserved across species and through tram
formation. Additionally, human meningioma cultures expressed some
spectroscopic
characteristics
(7) have found that it is possible
that enabled them to be clearly distinguished
from primary rat meningeal cultures. Thus, human meninglomas may be
both specifically recognizable by ‘H-NMRspectroscopy and also distin
in vivo.
INTRODUCTION
The evolution of tumor classification has been closely correlated
with the development of methods for identifying the normal cells from
which tumors arise. For example, as the cellular specificity of labeling
with various antibodies has become more clearly defined, the use of
histological criteria for tumor classification has been gradually sup
plemented and, in some instances, replaced by the use of immunocy
tochemical analysis. In addition, the increase in our understanding of
the cellular lineages involved in normal development has led to the
recognition of categories of tumors that were previously not appreci
ated.
One of the tools for cell type recognition that has lately proved of
potential interest is ‘H-NMR3
spectroscopy. NMR analysis of extracts
obtained from cultured cells provides a reliable method to assess the
metabolites present in these cells (1). A major advantage of NMR
spectroscopy is that it is a noninvasive technique that can be applied
in vivo, enabling information to be obtained about the metabolic
composition of living tissue. Cerebral metabolites such as creatine and
An important
supported
by the Cancer
Research
Campaign
U.K.
and
3 The
whom
requests
abbreviations
for
reprints
used
are:
should
NMR,
be
nuclear
taming compounds; NAA, N-acetyl-aspartate;
MATERIALS
that
follows
from
our observations
that
AND
METhODS
Preparation and Analysis of Cell CUltUres
All cells were cultured in DMEM containing
BRL,
Paisley,
United
Kingdom),
FCS (Imperial Laboratories),
1 gfliter glucose (GIBCO
supplemented
with
10%
heat-inactivated
2 mM glutamine (Sigma Chemical Co., Poole,
United Kingdom), and 25 @g/m1
gentamicin (Flow Laboratories, Rickman
sworth, United Kingdom; DMEM-FCS). Cells were grown in NUNC tissue
culture flasks (GIBCO.BRL)
the Welcome
coated with 13 pg/ed poly-L-lysine (Sigma) and
maintained in a 37°Cincubator with humidified atmosphere containing 7.5%
addressed.
magentic
question
eventually be possible to use NMR for noninvasive diagnosis of
meningiomas and that cell type-specific metabolites are so integral to
the identity of a cell type that they are conserved across species and
through transformation.
Trust.
2 To
cere
mary rat meningeal cells. Thus, our results suggest that it might
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
was
unambiguously
different cell types express characteristic ‘H-NMRspectra relates to
the applicability of this technique to tumor diagnosis, i.e., whether
related tumor cells express similar spectra. To investigate this possi
biity, we compared ‘H-NMRspectra from six human meningioma
cell lines. We found that these spectra were not only similar to each
other but also shared characteristics with spectra obtained from pri
Received 8/11/94; accepted 11/11/94.
The costs of publication of this article were defrayed in part by the payment of page
work
to distinguish
bellar granule neurones, meningeal cells, cortical astrocytes, oligo
dendrocyte-type-2 astrocyte progenitor cells, and oligodendrocytes by
analysis of their ‘H-NMRspectra. Each cell population studied cx
pressed a characteristic distribution and amount of metabolites, in
cluding free amino acids and derivatives, compounds involved in
membrane biosynthesis and catabolism, and compounds related to
energy metabolism. Thus, different cell types could be recognized by
the application of ‘H-NMRspectroscopy.
guishable from normal rat meningeal tissue. Our results raise the eventual
possibility of using NMR In the nonlnvaslve diagnosis of brain tumors
I This
phosphocreatine, Cho and NAA can be detected routinely by ‘HNMR spectroscopy in vivo in animals and humans (2, 3). Additional
signals in normal brain, including those of glutamate, inositol, and
glucose, can be detected in animal studies or in humans with the latest
generation of clinical NMR scanners (4). Glutamine, alanine, lactate,
and GABA are present in lower amounts in normal brain and are
difficult to detect due to sensitivity and resolution problems but can be
detected with the aid of spectral editing techniques or in pathological
states when their concentration is elevated (5, 6). NMR spectra of
extracts obtained at high magnetic field strengths offer much im
proved sensitivity and resolution compared to spectra recorded in viva
and enable a wide range of hydrogen-containing metabolites to be
identified and quantified. Thus, such extracts can be used to detect
fine biochemical differences between various cell lines or types of
cells.
In examining purified populations of different cell types derived
from a variety of cellular lineages of the central nervous system, we
resonance;
Cho,
@2The
choline-con
immunocytochemical
characterization
of all cell cultures
was per
formed at each passage by plating cells in parallel onto poly-L-lysine-coated
GABA, @y-aminobutyricacid; GFAP, glial
fibrillary acidic protein; GalC, galactocerebroside; PCA, perchloric acid; TSP, 3-trimeth
ylsilyl-tetradeuterosodium propionate; 2D-COSY, homonuclear proton two-dimensional
correlated spectroscopy; PFG, pulsed field gradients; NAAG, N-acetyl-aspartyl-gluta
mate; PC, phosphoryl choline; GPC, glycerophosphoryl choline; OPA, o-phthaldialde
glass coverslips
hyde; BCA, bicinchoninic acid; PK, pyruvate kinase.
antibodies on mixed cultures of cerebral cortices.
and growing
them in the same conditions
as above. Antibodies
and staining methods described in the literature (8, 9) were used with minor
modifications.
The specificity of immunochemical
labeling was tested for all
420
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOUC PROFILES OF MENINGES AND MENINGIOMA CELLS
were accumulated and Fourier transformed. A line broadening of 0.2—0.3Hz
was applied for spectral processing.
2D-COSY was also used for identification of coupled resonances. Ab
solute-value and, where possible, phase-sensitive conventional COSY and
double quantum filtered COSY spectra were acquired, either using standard
phase cycling or with the application of PFG (20 0/cm; duration 2 ms, rise
and fall time of 100 @i.s)
to select coherence pathways (17). Typically, 192
transients were collected for each of 256 T, increments (repetition time, 1.2
Meningioma Cells. Six human meningioma cell lines, of which one was
derived from a recurrent meningioma,
were cultured. One of the cell lines
originated from a spinal meningioma. The composition of cultures was eval
uated by immunolabeling cells on coverslips with antibodies at 48—72h after
plating onto coverslips (initial density of 500 cells/coverslip). Cells were tested
for the presence of cytoplasmic intermediate filaments, such as glial fibrillary
acidic protein (jrotein
specific
to astrocytes
in mature brain; Refs. 10—12) and
vimentin (ii, 13), and also for fibronectin (12). Additionally, contamination by
other cell types of the nervous system was assessed by testing with monoclonal
s; spectral
supernatant,
Primary
diluted
Cells. Cell cultures were prepared as described pre
Metaboilte
viously (9). Leptomeninges of 1-day-old Wistar rats were removed and dis
sociated with collagenase and trypsin and then plated onto 175-cm2flasks and
grown in DMEM-FCS.
The cells formed a confluent monolayer within 5—6
days after plating, when they were passaged and replated at 1:4 dilution into
175-cm2flasks. Cells were plated in parallel onto coverslips (density of 3000
cells/coverslip).
When meningeal cells formed a confluent monolayer,
they
were harvestedfrom cultureflasks. The meningealcultureswere assessed for
purity by immunocytochemical labeling on coverslips with anti-Ran-2 (a
monoclonal antibody which recognizes an antigen present among others on the
surface of astrocytes and leptomeningeal cells; Ref. 16). All antibodies used in
the characterization of human meningioma cells were also used for rat
meningeal cells.
Hz). Water suppression
was achieved
by presatu
periods as in the one-dimen
Identification
and Quantification.
Metabolite peaks found in
the one-dimensional spectra were identified by: (a) their chemical shift and
coupling pattern as described in the literature (1, 18); (b) comparison with
spectra of metabolites in known concentrations obtained at the same pH and
spectroscopic conditions; and (c) two-dimensional spectroscopic methods.
The following metabolites were identified from their strongest and best
resolved resonances: g3-hydroxybutyrate.@yCH@
1.19 ppm (doublet); threonine
--yCH31.30 ppm (doublet); lactate [email protected] ppm (doublet); alanine @CH@
1.47
ppm (doublet); acetate -CH3 1.92 ppm (singlet); NAA -NCOCH3 2.02 ppm
(singlet); NAAG -NCOCH3 2.05 ppm (singlet); glutamate -yCH2 2.34 ppm
(triplet); succinate -CH2 2.41 ppm (singlet); glutamine --yCH2 2.44 ppm
(triplet); aspartate -CH2 2.56 and 2.75 ppm (double doublet); creatine -NCH3
3.04 ppm (singlet); taurine -SCH2 3.08 ppm (triplet) and -NCH2 3.42 ppm
(triplet); choline -N(CH3)33.21 ppm (singlet); PC -N(CH3)3 3.22 ppm (sin
Cell Harvesting, Extraction, and Preparation of Samples
glet); GPC -N(CH3)3 3.23 ppm (singlet); glycine -CH2 3.56 ppm (singlet); and
inositol(H2) 4.05 ppm (triplet). Metabolite amounts were calculated from their
intensities in 1H-NMRspectra by reference to the internal standard TSP after
baseline correction. The intensity of a given signal in the proton spectrum is
proportional to the amount of compound and number of protons contributing
to that signal.
All cells were continuously cultured until the amount necessary for high
resolution NMR spectroscopy was obtained (typically iO@-i0' cells for a
sample).
5000
sional spectra. Data were collected into 2K data points in F2, and 1K in Fl,
which was then zero-filled to 2K. Sine-bell and Gaussian weighting func
tions were applied during the double Fourier transform, and the COSY
plots were symmetrized after initial analysis.
1:10).
Meningeal
width,
ration of the HOD peak between acquisition
antibody A2B5 (a marker characteristic of neurones and cells of the oligoden
drocyte-type-2 astrocyte lineage (14); hybridoma supernatant, diluted 1:2), or
anti-GalC(a specific surface marker for oligodendrocytes; Ref. 15); hybridoma
They were harvested
at 95—100% confluence
and always
24 h after a
final medium change. Cells were detached from culture flasks by a 3-min rinse
with Ca2@/Mg@@-free DMEM (CMF-DMEM) containing 034 mM EDTA
(Sigma), followed by 3—5mm incubation with a solution of 300 units/mI
trypsin
(Sigma;
trypsin
inhibitor-DNase
bovine
pancreas,
(5200
type III) in CMF-DMEM.
units/mI
soybean
trypsin
HPLC Analysis of Metabolites
One ml soybean
inhibitor,
74 units/mi
Quantitative
determination
of amino acids, NAA, and NAAG was per
bovine pancreas DNase I, and 3 mglml BSA; Sigma) was added per 10 ml cell
suspension to prevent the cells from clumping and to stop the trypsinization
process. Cell suspensions were centrifuged at 500 X g for 5 mm, followed by
a 30-s spin at 700 x g. Cells were washed three times with 30 ml PBS, and the
pellet finally was extracted with 2 ml ice-cold PCA (12% v/v; Merck, Ltd.).
PCA extractswere sonicated(2 times 30 s, 4°C;
Soniprep150, MSEScientific
calculated
primary cultures of rat meningeal cells (replicate samples).
3 mm and then 150 pi mixture
formed by HPLC to complement the spectroscopic data. Amino acids were
analyzed by derivatization with OPA (19), with the concentrations being
from
external
standards.
A LiChroGraph
HPLC
system
(Merck
Hitachi, Darmstadt, Germany) equipped with a LiChrospher 100HP-18 sepa
ration column (5
@m;4 mm internal diameter, 25-cm length) was used. The
OPA derivatives were detected with a UV/VIS fluorometer(LiChroGraph
Instruments,
Crawley, United Kingdom) and then centrifuged
at 3000 X g at
F1050), characterized by wavelengths of 340 nm for excitation and 450 nm for
4°Cfor 20 mm.The pellets were keptat —20°C
for proteinquantification,and emission. Aliquots taken from NMR samples were diluted 5—200-foldwith
the supernatantswere furthertreated.At least two PCAextractswere prepared ultrapure water, and a 150—250-idsample was mixed with 350 p.1reagent
for each humanmeningiomacell line, and four extractswere obtainedfrom containing 0.8 mg/ml OPA (Sigma). The compounds were allowed to react for
The pH of the supernatantwas adjustedto slightly alkaline(pH > 7.5) and
then mixed with 5 mg Chelex-100 (Bio-Rad Laboratories, Richmond, CA) per
0.2 ml sample in order to remove multivalent ion contaminants. A subsequent
adjustment of pH to 8.5—8.9with 3 M KOH was followed by removal of the
precipitated KC104 by centrifugation at 3000 X g for 5 mm. The sample was
lyophilized and redissolved in 0.5—0.7ml D2O (Goss Scientific Instruments,
Ltd., Ingatestone, United Kingdom). As internal concentration and chemical
shift standard TSP (Goss Scientific Instruments, Ltd.) was added (10 @l
of 10
mM TSP in D20). Finally, the pH of the sample was adjusted to 8.9 with DC1
or NaOD (Aldrich, Gihingham, United Kingdom).
was injected
onto the column.
Elution
was
achieved at 40°Cwith a three-step linear gradient of methanol in phosphate
buffer (50 mM in ultrapure water, pH 6) starting at 20% methanol, increasing
to 35% methanol in 5 min, then decreasing to 30% methanol in 30 min, and
finishing at 35 mm with 20% methanol.
NAA and NAAG were detected with an HPLC System Gold (Beckman
Instruments, San Ramon, CA) equipped with a UV detector (wavelength, 210
nn@),and their concentrations were calculated from external standards. NMR
samples were diluted 1:10 with 5 mMH2SO4,and 50-id aliquots were injected
onto the column. NAA was analyzed by a method described previously (7)
using the same conditions. Detection for NAAG was carried out according to
Koller et al. (20) using an Machem SAX (46 mm internal diameter, 25-cm
‘H-NMRSpectroscopy
length)
All spectra were obtained at 26°C-30°C
on a Varian Unity-plus NMR
spectrometer
(Varian
Associates,
Inc.,
NMR
instruments,
Palo
Alto,
A second
radiofrequency
pulse
was applied
at the frequency
fitted with a 1-cm guard
column.
Protein Quantification
CA)
operating at a proton frequency ofSOOMHz. Single-pulse spectra (approaching
full relaxation) were acquired with 45°pulses applied every 5 s. The sample
spinning rate was 20 Hz, spectral width was 5000 Hz, and data size was 32K
points.
column
of the
water peak in the delay between acquisitions in order to suppress the residual
water (HOD) signal. For a satisfactory signal:noise ratio, 512 or 1024 scans
Protein quantification was performed on the final PCA precipitate by
complexing with BCA (21). The pellet was dissolved in 1 ml of 1 mM KOH
by incubation at 37°Cfor 30 min. BCA reagent (BCA kit, Pierce, Rockford,
CA) was furtheradded, and the absorbancewas detected after incubation
(wavelength,
256 sum; UV/VIS
Spectrophotometer
Ultrospecll),
and protein
was calculated by reference to BSA standards (BCA kit, Pierce).
421
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOLIC PROFILES OF MENJNGES AND MENINGIOMA CELLS
Statistics
cultured rat meningeal and human meningioma cells are shown in
Fig. 1. These high-field regions of spectra (0.5—4.5ppm) contain
Results are presented as mean ±SD. Immunocytochemical results are
signals from a variety of metabolites and present qualitative sim
expressed as a percentage of positive cells from the total number of cells. The
amounts of metabolites were referenced to the amount of protein in the ilarities and also qualitative and quantitative differences. Metabo
samples, as nmol/mg protein. The results from replicate samples of 2—6 lites detected and quantified include amino acids and related com
extracts for each of the six human meningioma
cell lines were averaged per
pounds; substances involved in membrane biosynthesis such as
each cell line. Data obtained per cell line were finally averaged per meningi
choline, Cho, and inositol; and intermediary metabolites such as
oma cells. Data from each of the four independent observations for meningeal succinate and lactate. Metabolite contents (nmol/mg protein) of the
cells were also averaged. Two sample Student's t tests were performed to
cultures examined are given in Tables 2 and 3. In Table 4, selected
determine
significant
differences
among group means, and Ps were quoted
without correction for multiple comparisons. The critical values for assessing metabolite ratios are presented for rat meningeal and human me
ningioma cells. These ratios were chosen because of their rele
significance levels were obtained after performing a Bonferroni correction for
multiple comparisons (22) independently for NMR and HPLC data sets (12 vance either to NMR in vivo (involving creatine, Cho, and inositol)
and 13 comparisons, respectively).These critical values were P < 0.005 for 5% or to previous work (7) identifying metabolite ratios which could
significance level (i.e., 0.05/12 or 0.05/13), and P < 0.001 for 1% significance
distinguish between preparations of rat meningeal cells and other
level (analogous).
primary cultures from rat brain (involving Cho, glutamate, aspar
tate, and creatine). Not all identifiable compounds in the ‘H-NMR
spectra could be accurately quantified because of incomplete res
RESULTS
olution in certain zones of the spectra (e.g., isoleucine, leucine, and
valine with signals between 0.9—1.05 ppm). In addition, signals
Immunocytochemical
and Cell Morphologj Characterization.
from some of the identifiable compounds were often too low for
All four preparations of primary rat meningeal cells contained flat,
bipolar or tripolar A2B5 and Ga1C cells. Positive immunocyto
accurate quantification.
chemical labeling was obtained for fibronectin and vimentin. There
Within each category of cells (rat normal and human tumor),
was over 95% positive labeling with Ran-2 antibody, showing the
spectra were qualitatively identical (i.e., all identifiable peaks were
purity of meningeal cultures relative to contamination with cells of the
present in all spectra from that category), with the exception of
nervous system. There were some GFAP@ astrocytes (4 ± 1%)
creatine, which was not detectable in spectra from one-half of the
scattered within the meningeal cultures.
meningioma cell lines (see below). Quantitatively, the spectra from
Consistent and reproducible results in the expression of specific
the meningioma cells lines displayed more variability between differ
antibodies were obtained in the six human meningioma cell lines
ent cell lines (Table 2), but excellent reproducibility was achieved in
(Table 1). The specificity of immunocytochemical labeling was con
replicate samples (i.e., preparations obtained from one cell line).
firmed by negative results in all controls. Neither anti-Ga1C nor the
There was very good reproducibility in replicate preparations from
monoclonal antibody A2B5 (which recognizes specific surface gan
normal rat meninges. Good agreement was obtained between quanti
gliosides) labeled any of the cells of the six human meningioma lines
fication from NMR spectra and by HPLC analysis in all samples. The
at either early or late passage. Fibronectin, an extracellular matrix
taurine content of human meningioma cell lines could not be reliably
glycoprotein with adhesive and opsonic properties (12), was widely
quantified
in the NMR spectra, although
it was present at levels
expressed by all six meningioma cell lines in all passages, the pro
detectable
by
HPLC.
The
greatest
variability
in absolute metabolite
portion of fibronectin@ cells in culture increasing with the passage
amounts was displayed by glutamate in human meningiomas,
but the
number. The expression of GFAP was either negligible in one cell line
or absent (in the other five cell lines). Vimentin was expressed by all data for glutamate from NMR spectra and by HPLC analysis were in
good agreement for individual cell lines.
six meningiomas (100% vimentin@ cells), irrespective of passage
The identification of coupled metabolites in single-pulse spectra
number or stage in culture.
was
confirmed and supplemented by identification of crosspeaks in
Metabolites Identified in ‘H-NMR Spectra and Comparison
2D-COSY spectra. Fig. 2 displays a 2D-PFG-COSY spectrumfrom
with HPLC Results. Representative
high-resolution
1H-NMR
one of the meningioma cell lines.
spectra obtained from extracts of acid-soluble metabolites from
Table 1 Immisnocytochemicalcharacteristics of the human meningioma cell lines
Results are calculated as a percentage of positive cells from the total number of cells. At least 200 cells were counted on each coverslip. At every passage (P), at least two coverslips
were prepared for each antibody, or combination of antibodies, for each cell line. Tests were carried out in the following combinations: A2B5 and aGalC; aFN; aGFAP; aVim. Figures
are given for the start and for the end passage number,
numbers.Passage
where these results differ for the two passage
numberHuman
meningioma
(%)IN1O67 cell line
Start
End
A2B5 (%)
aGalC' (%)
aFN (%)
aGFAP (%)
P6
P9
0
0
29.0±8.7 (P6)
82.0 ±8.9
0
100
P3
P6
0
0
26.7±3.4 (P3)
58.6 ±2.6
0
100
P3
P6
0
0
26.0 ±4.2 (P3)
0
100
0
100
(P9)lNlO69
(P6)IN1O81
(recurrent)
(P6)IN11I4
78.3±3.3
PS
P8
0
0
43.1±5.1(P5)
80.1±8.5
P3
P7
0
0
32.2±6.4 (P3)
67.0±5.6 (P7)
P3
P7
0
0
31.5 ±6.4 (P3)
79.5 ±14.9
(P8)IN1123
(P7)IN1239
(spinal)
(P7)a
aGalC,
anti-galactocerebroside;
aVim
aFN,
anti-fibronectin;
aGFAP,
anti-glial
fibrillary
acidic
protein;
aVim,
0.5 (P3)
2 ±03
0
100
100
anti-vimentin.
422
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOUC PROFILES OF MENINGES AND MENINGIOMA CELLS
Table2 Comparative
ofrat
compositionof metabolitesquantsfledfrom‘H-NMR
spectra
cellsMetabolite
meningeal and human meningioma
Spectraobtained
concentrations(nmol/mgprotein)are expressedas mean±SD.
RMm@iIn@@
cho
humanmeningioma
from replicate preparations for each of the independent observations for
referenceto cell lines and rat meningeal cells, respectively, were analyzed by
(a)
thenumber
TSP.P obtainedfromtwosampleStudent'st testaregivenwithoutcorrectionfor
Ac,
differentfrom
ofcomparisons. The results for normal cells were considered statistically
the results for tumor cells if P <level).Rat
0.05/12 (5% level) or P < 0.01/12 (1%
HumanMetabolites
ac
lines(nmol/mg
meningeal cells
PAlanineprotein)
0.004Choline
0.124GPC
0.526PC
<0.001Creatine
0.003Glutamate
<0.001Glycine
4.5
4.0
(n = 6)
15.6
0.77313.9±2.8
±3.8
Lac
@
meningioma cell
(n = 4)
3@5
3.0
2.0
1.5
1.0
pp@
4.4 ±0.7
14.0 ±3.6
17.6 ±4.8
114.1 ±8.0
7.3 ±5.8
298.4 ±43.9
0.033@3-Hydroxybutyrste
0.140Inositol
12.9 ±5.8
6.6 ±1.4
27.5 ±10.1
13.4 ±8.0
0.116Succinate
0.771Threonine
5.4 ±1.9
12.1 ±5.2
18.5 ±14.2
13.1 ±5.4
0.009Taurine
12.7 ±3.5
33.6 ±5.5
23.8 ±5.7
Not measurable
Human msnlnØoma (a)
Cho
2.0 ±0.4
7.5 ±2.2
16.3 ±3.6
27.1±5.1
6.2 ±4.8
9.1 ±4.5
<0.001
aThe
data
forcholine-containing
compounds
(Qio,
3.2ppm)
arethesum
ofthree
individual peaks: choline (3.21 ppm), PC (3.22 ppm), and GPC (3.23 ppm).
Table 3 Comparative composition
ratmeningea!of metabolites determined by HPLC analyses in
cellsMetabolite
and hwnan meningioma
theindependent concentrations
(nmol/mg protein) are expressed as mean ±SD of
observations for the rat meningeal cells preparations and for human menin
gioma
correctionfor
cell lines. P obtained from two sample Student's I test are given without
statisticallydifferent
the number of comparisons. The results for normal cells were considered
(1%level).Metabolites@'
from the results for tumor cells if P < 0.05/13 (5% level) or P < 0.01/13
lines(nmol/mg
protein)
4.1
4.0
3.5
3.0
2.5
2.0
1.8
1.0
[email protected],@a
ppm0.5
(C)
Human meningioma cell
(n = 6)
Alanine
Arginine
Asparagine
Aspartate
17.8 ±3.8
2.0 ±0.9
1.4 ±1.0
20.3 ±0.4
25.2 ±7.1
6.1 ±2.3
4.9 ±1.9
27.0 ±14.6
GABA
Glutamine
5.9 ±0.8
43.6 ±5.6
0.6 ±0.2
72.9 ±32.6
Glutamate
Hypotaurine
NAA
NAAG
Serine
Taurine
Tyrosine
as,
i@t meningeal cells
(n = 4)
105.4 ±2.6
19.4 ±2.9
Not detected
3.0 ±1.5
18.0 ±1.9
305.5 ±47.4
5.0 ±3.0
Not detected―
4.5 ±3.9
28.6 ±5.9
P
0.264
0.026
0.018
0.404
<0.001
0.145
<0.001
<0.001
0.474
0.018
38.4 ±4.2
18.0 ±2.8
<0.001
2.3 ±0.6
10.3 ±2.1
<0.001
a The amountof NAAwastoo lowto be detectedby HPLCinvestigationsin fiveof
the six meningioma cell lines we examined, but one meningioma cell line (IN1239)
contained a very low amount of NAA (2.9 ±1.7 nmol/mg protein).
v@LaoI.
Common Features of the Spectra of Rat Meningeal and Human
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
ppm
0.5
Fig. 1. Representative single-pulse high-field high resolution ‘H-NMRspectra from
rat meningeal and human meningioma cells. Spectra were obtained from PCA extracts
of primary cultures of rat meningeal cells (a) and cultures of human meningioma cell
lines (b and c) NMR spectroscopic analysis was performed at pH 8.9, with 512—1024
scans recorded on a spectrometer operating at the proton frequency of 500 MHz.
Typically, iO7@iO8
cells were obtained for one extract. The content in protein of samples
ranged from 5.1 to 8.2 mg for rat meningeal cell preparations and from 0.92 to 5.3 mg
for human meningioma cell extracts. Spectra referenced to TSP (0 ppm) from the
following cell lines are displayed: b = IN1067; c = 1N1114.The peak marked () was
anexogeneouscontaminant,probablyfromtheplasticwareusedinsamplepreparations
In spectra a and b, the height of the [ac doublet was truncated to 2/3 of the original
height in the spectrum. tie, isoleucine; Leu, leucine; Val, valine; (3-HB. @-hydroxybu
tyrate; Thr, threonine; Lac, lactate; Ala, alanine; Ace, acetate; Glu, glutamate; Suc,
succinate; GIn, glutamine; Asp, aspartate; Cr. creatine; Thu. taurine; 0,, choline; PC,
phosphorylcholine;GPC, glycerophosphoryl
choline;Cho,choline-containing
com
pounds; Gly, glycine; mo, inositol.
Meningioma Cell EXtraCts. In the light of existing data on metab
olites detectable by 1H-NMR in whole brain extracts and cell
preparations from either rat or human brain (5, 7, 23), we note that
spectra from both rat meningeal cells and human meningioma cell
lines were characterized by either very low or not detectable
signals for NAA (2.02 ppm, singlet), NAAG (2.05 ppm, singlet),
aspartate
(2.56 and 2.75 ppm, double
doublet),
and the neuroactive
amino acid GABA (3.0 ppm, triplet). NAA was not detectable by
HPLC (lower detection limit, 0.1 nmol) in any of the samples
examined, with the exception of one meningioma cell line that
contained a small amount of NAA. Low NAAG concentrations
were also determined by HPLC in preparations from both rat
meninges and human meningiomas, with no significant differences
between the two categories. Compounds such as tyrosine (3.05,
3.15, and 3.93 ppm, all double doublets) and hypotaurine (2.65 and
3.30 ppm, both triplets) were present in too low an amount to be
detected by NMR but were detectable by chromatographic analy
sis. Complementary determinations by HPLC showed statistically
423
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOUC PROFILES OF MENINGES AND MENINGIOMA CELLS
Table 4 Metabolite rajios in rat meningeal and human meningioma cells
Metabolite ratios (mean ±SD) were calculated from metabolite concentrations quasi
rifled from the ‘H-NMR
spectra of acid extracts of rat meningeal and human meningioma
cells. P values obtained from Two-sample Student's t test are given without correction for
the number of comparisons. The results for normal cells were considered statistically
different from the results for tumour cells if P < 0.05/12 (5% level) or P < 0.01/12 (1%
level) Abbreviations Cho choline-containing compounds.
meningeal cells
Metabolite ratiosRat
6)PCho/creatine1.12
(n = 4)Human
2.90.019creatine/inositol3.2
0.43<0.001aspartate/creatine1.2
3.30.024Cho/aspartate0.7
0.70.120Cho/glutamate0.12
0.050.217Cho/glycine1.2
1.60.277glycine/creatine0.78
meningioma
cell lines (n =
±0.185.5
±0.90.63
±0.35.6
±0.11.3
±0.020.1
±0.41.7
±0.235.9
±
±
±
±
±
±
±4.60.063
significant differences at the 1% level in the concentration of
amino acids such as GABA, hypotaurine, and tyrosine in extracts
of rat meningeal cells compared to human meningioma cells.
Spectra of preparations from rat meninges and human meningioma
cell lines had common features in the 1.0—1.5ppm region, displaying
the doublet signals of f3-hydroxybutyrate, threonine, lactate, and
alanine. The concentrations of none of these metabolites were signif
icantly different between rat normal and human tumor cells. Relative
to ‘H-NMRspectra in vivo, lactate signals were elevated in the spectra
obtained from cultured cells, probably due to anaerobic glycolysis
taking place in cells during the extraction process. Lactate content was
highly variable in the cell preparations (from 25 to 230 nmol/mg
protein), depending on the rapidity of cell extraction and on the
dimension of the cell pellet. The signals for alanine were relatively
high in spectra from preparations of both categories of cells compared
to the typical levels of alanine detected by spectroscopy in vivo in
normal human brain (5, 23).
In the methylene region of the spectra (2—3ppm), common signals
for both types of cells included those of glutamine and succinate. No
statistically significant differences were obtained in the concentrations
of these metabolites between the preparations of rat meningeal and
human meningioma cells. There was a large variation in the glutamine
concentration of tumor cells, ranging from 29.3 nmol/mg protein for
one cell line up to 105.0 nmollmg protein for another, but the aver
aged glutamine content in human meningiomas was not statistically
significantly different from the value for rat meningeal cells.
The methine region of the spectrum (3.2—4.0ppm) had some
common general characteristics for both types of cells examined,
including a relatively high signal for glycine and relatively low signal
for inositol. No significant difference in the inositol content deter
mined from the 1H-NMR spectra was obtained between normal and
transformed cells. We did not have a reliable independent assessment
for glycine and threonine content in samples because their separation
on the HPLC column was poor. Allowing for correction of the Ps for
the number of comparisons, the difference in threonine and glycine
contents of rat meningeal and human meningioma cells was not
significant at the 5% level.
Differential Characteristics in Spectra of Normal Rat Menhiges
and Human Meningiomas. There were also differences in the
1H-NMR spectra obtained from cell extracts of cultured rat meninges
and human meningiomas. For example, a prominent compound in
preparations of meningeal cells obtained from rat was taurine, a
metabolite which was either not detected by NMR in meningioma
extracts, or which gave very low signals in spectra. Quantification of
taurine in spectra obtained from extracts of human transformed cells
was not possible due to overlap of the creatine signal at 3.04 ppm with
one of the taurine peaks (3.08 ppm, triplet) and because of the
complexity of the ‘Hspectrum where the second taurine signal lies
(3.42 ppm, triplet). Taurine was among the most abundant free amino
acid detected by HPLC (Table 3) in rat meningeal cells and was
present in amounts significantly different (P < 0.001) from those
existent in human meningiomas.
Among other signals in the high field region, the singlets at 3.04
ppm and 3.94 ppm, which are assigned to creatine, were more prom
Fig. 2. Two-dimensional (pulsed-field gradient) corre
lated ‘H-NMR
spectrum of a human memngioma cell line.
The 2D-PFG-COSY spectrum was obtained from a PCA
extract of a human meningioma cell line (1N1067) on the
same spectrometer as above but using the application of
pulsed field gradients to select coherence pathways. For
each of the 256 T, increments 192 transients were collected.
The inset shows the whole 2D ‘H-NMRspectrum (0—
l0ppm), while the main figure represents an expansion of
the spectral region of interest (0.5—4.5ppm). The coupled
peaks with chemical shifts of 1.7 and 3.05 ppm were ten
tatively assigned to polyamines (PA) by running single
pluse ‘H-NMR
spectra of putrescine and spermidine and by
comparison of HPLC profiles ofsingle polyamines with the
HPLC profiles obtained from human meningioma cell line
extracts. Abbreviations same as for Fig 1; @-A1a,
f3-alanine.
n
424
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOLIC PROFILES OF MENINGES AND MENINGIOMA CELLS
agreement with these latter observations of commonalities among
meningiomas.
Our findings related to fibronectin and vimentin expression of
inent in ‘H-NMRspectra from normal rat meninges than in spectra
from human meningioma cell lines. The creatine peak in vitro repre
sents the sum of creatine and phosphocreatine present in the living
cells, since hydrolysis of phosphocreatine to creatine occurs during
the cell harvesting and extraction process. Spectra from preparations
of three meningioma cell lines were characterized by the absence of
creatine peaks; in two other cell lines, the creatine levels were below
the lowest amount in any of the meningeal preparations; in one cell
line, the creatine level appeared relatively “normal―
(16.9 nmol/mg
protein). However, overall there was a significant difference
(P = 0.002) in the creatine content of rat normal and human trans
formed cells. The creatine:inositol metabolite ratio was also signifi
cantly different at the 1% level in meningeal compared with menin
gioma cells (Tables 2 and 4).
The signals for glutamate were relatively higher in spectra of
human tumor cells than in those of normal rat cells. This was also
reflected, by HPLC, in the significantly elevated (P < 0.001) content
of glutamate in human meningioma lines (305.5 ±47.4 nmol/mg
protein) versus normal rat meningeal cells (105.4 ±2.6 nmol/mg
protein), and related to this, in a significantly different (P < 0.001)
glutamate:alanine ratio.
Peaks from Cho were prominent in spectra of both normal and
tumor cells. The signals in the 3.2—3.3ppm region are assigned to
choline itself (3.21 ppm, singlet), PC (3.23 ppm, singlet), and GPC
(3.24 ppm, singlet); these metabolites were identifiable in the
spectra of both categories of cells. Other metabolites, including
carnitine, ethanolamine, and phosphoethanolamine
might also have
contributed to signals in this region. A significant difference at the
5% level (P = 0.004) was detected in total Cho (sum of the three
peaks) in rat meningeal compared to human meningioma cell
extracts (Table 2).
The relative heights of individual peaks contributing to the Cho
signal in the two categories of cells were also different. There was a
significant increase (P < 0.001) in the amount of PC in human
meningiomas compared to rat meninges. A shift was observed in the
meningioma
cell lines were in concordance
with previous
studies
on
human intracranial meningiomas. It has been shown (12) that neo
plastic cells retain the intermediate filaments of their tissue of origin
and that virtually all meningeal neoplasms have a uniform distribution
of fibronectin filaments. Additionally, coexpression of vimentin and
desmosomal proteins and the association of this type of intermediate
filaments to desmosomal plaque proteins are a combination unique to
meningioma cells (13). The negative GFAP labeling obtained in the
meningioma
cells was in agreement
with their cell of origin
and their
usual localization outside brain tissue (10, 24).
Interpretation
of HPLC
ReSUlts and NMR
Findings.
Although
there were quantitative variations for certain metabolites between the
human meningioma cell lines, qualitatively identical ‘H-NMRspectra
were obtained from all six cell lines we examined. This fact suggests
a preservation of the characteristic metabolite profile within the type
of tumor, irrespective of its origin.
Spectra from preparations of rat meningeal cells displayed obvious
similarities to those we have reported previously from extracts of
primary cultures of rat meningeal cells (7). These results confirm the
reproducibility
of cell culture
techniques,
metabolite
extraction
pro
cedures, and NMR spectroscopy investigations. Although cell culture
procedures and methods of extract preparation could be a source of
variability
(1,26), in studies
such as ours, these potential
problems
can
be readily overcome if factors including composition of the culture
medium, degree of cell culture confluence, time of harvesting, and
method of extraction are carefully controlled.
We have shown previously that rat meningeal cells can be distin
guished from other primary cultures of central nervous system rat
cells (neurones, astrocytes, oligodendrocyte-type-2-astrocyte
progen
itors, and oligodendrocytes) on the basis of ‘H-NMRspectra of
extracts (7). The distinguishing features were the presence of high
amounts of succinate and @-hydroxybutyrate, the absence of NAA and
PC:GPCratio from below one in rat meningeal cells to greaterthan hypotaurine, and relatively low amounts (compared to oligodendro
cytes) of creatine. These characteristics were shared by all the human
one in the human meningioma cell lines. PC is synthesized in vivo in the
meningiomas we examined, with no significant differences in the
first step of phospholipid biosynthesis, whereas GPC is a phosphodiester
succinate and @3-hydroxybutyrate concentrations between rat menin
breakdown product of phospholipids. The predominance or otherwise of
one kind of phosphoesters in the composition of the Cho peak in these geal cells and human meningioma cell lines, and an even lower
creatine content in the meningioma preparations. In addition, the
spectra may indicate differences in phospholipid metabolism in menin
metabolite ratios Cho:aspartate and Cho:glutamate, which were used
gioma cells of human origin versus meningeal cells from rat.
previously to discriminate rat meningeal cells from rat astrocytes, do
not differ significantly between rat meningeal and human meningioma
cells (Table 4). The aspartate:creatine ratio, which was more than
DISCUSSION
10-fold higher in rat meningeal cells compared to cortical rat astro
We used cell culture techniques and investigations by ‘H-NMR cytes, was even higher in human meningioma cell lines, reflecting
their lower creatine content.
spectroscopy complemented with HPLC analysis in order to compare
Although the key features described above clearly relate human
metabolite profiles of normal rat meningeal cells and human menin
meningioma cells to their tissue of origin even when considered
gioma.
across these widely separated species, there were also a number of
Meningioma Cell Lines in the Light of Their Immunocytochem
characteristics which discriminated the transformed from normal
lcal Characteristics
and Cell Types of the Meninges. All cell types
of meningeal tissues covering the brain and spinal cord originate in cells. These statistically significant features of the human meningioma
the primitive meninx derived from the layer of mesenchymal tissue
metabolite profiles were reductions in creatine and taurine and in
that comes to surround the neural tube in the early development of the creases in Cho (attributable
mainly to increased PC) and glutamate.
embryo (24). Meningiomas, as primary tumors of the meninges com
Taurine is known to be much lower in human brain than in rat brain
posed of or differentiating towards arachnoidal cells (25), can involve
(5, 27); therefore, it is possible that the reduced taurine in the menin
dura mater, leptomeninges, or both and account for about 15% of all
gioma cell lines reflected a species difference rather than an effect of
human brain tumors. Tumors of the meninges display a high pheno
transformation. Although it is not yet known whether the dissimilar
typical variability in cell products, shape, pattern, and stroma, but
ities in other signals reflect differences across species or differences
equally, human meningiomas of various origins have similarities in between normal and transformed cells, arguments can be raised that at
their immunocytochemical characteristics, histological features, me
least some of these changes may be associated with transformation.
tabolism, behavior, etc. (11, 13, 25). The results we obtained were in
A feature of human meningioma cell lines that we examined was
425
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOUC PROFILES OF MENINGES AND MENINGIOMA CELLS
their high content of glutamate, consistently 2.5- to 3-fold higher than
rat meningeal cells. As previous studies (24) have demonstrated that
human meningiomas in culture express a relatively low uptake of
glutamate (lower even than normal human fibroblast cultures), the
question arises as to how the high glutamate content in meningioma
cells might be achieved. A variety of observations raise the possibility
that such an elevation would be a predicted consequence, both of
enzymatic alterations characteristic of brain tumors and of the high
alanine content of primary meningeal cells (7). We propose the
following hypothesis to explain the high glutamate levels observed in
our human meningioma cultures.
One of the enzymes involved in the glycolytic pathway in eukary
otic cells is PK, which catalyses the transfer of phosphate from
phosphoenolpyruvate
to ADP, yielding
free pyruvate
and ATP. PK is
“turned
off―when Al?, or other fuels including alanine, are present
in high concentration in the cell, and it “turns
on―when there is a
buildup of the preceding glycolytic intermediates, especially phos
phoenolpyruvate
(28).
It has been
shown
(24) that this enzyme
is
present both in normal brain and brain tumors mostly as the muscle
type (M), in the form of two subgroups: Ml, which is not inhibited by
alanine in its action; and M2, which is inhibited by alanine. The Ml
form is present in normal brain, while the M2 form is predominant in
tumors. These observations, considered in the light of the results of
our analyses of amino acids, especially alanine and glutamate, in rat
normal and human tumor meningeal cells, raise the possibility that
there might be a relationship between the relatively high alanine
content of both meninges and meningiomas and PK in these cells. It
might be that in normal meningeal cells, PK (the Ml form which is
not inhibited by alanine) will be active, while in meningioma cells, the
M2 form of PK which is inhibited by alanine is “turned
off.―Such an
inhibition
of PK would
lead to a lack of production
of pyruvate
in the
glycolytic process and accumulation of phosphoenolpyruvate. Alter
natively, pyruvate could be generated in the tumor cells via the
transamination
reaction from alanine and a-ketoglutarate,
yielding
glutamate in addition to pyruvate. The presence of this reaction would
account for the buildup of glutamate in meningioma cells without
uptake from the culture medium. This reversible transamination proc
ess could also replenish
alanine. Our results showing that human
meningiomas
display a constant glutamate:alanine
ratio, which is also
significantly different from the ratio for rat meningeal cells, are
consistent with this suggestion. Subsequently, when phosphoenol
pyruvate
has accumulated
in meningioma
cells in high amounts,
it will
“turn
on―the PK (M2 form), and glycolysis will follow its normal
path. This hypothesis is also consistent with observations that normal
allosteric factors regulating the glycolytic rate to that of the tricar
boxylic acid cycle are defective or altered in transformed cells (28).
Our results showed that the PC peak was predominant amongst
signals contributing to Cho in spectra from all six human meningioma
cell lines we examined, as opposed to a predominant GPC peak in the
3.2 ppm region of the ‘H-NMRspectra obtained from rat meningeal
cells. This finding was consistent with the idea that increased levels of
phosphomonoesters
(e.g., PC and phosphoethanolamine)
may be as
phodiesters) cannot be detected due to the decreased spectral resolu
tion achievable in NMR spectroscopy in vivo. 31P-NMR spectroscopy
has showed elevated phosphomonoester signals in untreated menin
giomas of human patients in vivo (31, 32), as well as in spectra from
PCA extractsof resected humanmeningiomas (32).
In the comparison
of NMR spectra
from extracts
to those obtained
in vivo, it is important to appreciate that metabolite levels may differ
because certain metabolite pools may be “NMR-invisible―
in vivo.
There is little evidence to suggest that most of the major metabolites
detected by ‘H-NMRspectroscopy are to any extent “invisible,―
in
contrast to the situation for 31P-NMR spectra in vivo in which ADP
and inorganic phosphate certainly contain NMR-invisible pools (33).
An exception
to this is glutamate,
which is about 80% NMR-visible
in
rat brain (34) in vivo or in guinea pig brain slices (35). It has been
suggested that the invisible pool is neurotransmitter glutamate (36), in
which case interpretation of our data from nonneuronal preparations
would be unaffected.
The ‘H-NMRprofiles we obtained from cell populations derived
from human meningiomas were very similar to the ‘H-NMRspectra
that have been reported (30) for extracts of meningiomas obtained
from human biopsy samples. These similarities included relatively
high signals for alanine and Cho, a reduced signal for creatine, and
low or not detectable signals for neuroactive amino acids and related
compounds such as GABA, NAA, and NAAG. The low or not
detectable levels of such compounds (which are present in analyses of
human central neural tissue) were consistent with the embryonic
derivation of the meninges and the location of meninges and
meningiomas outside of the central nervous system.
Metabolic regulation in cultured cells may diverge compared to the
situation in vivo. Furthermore, such changes may be different in
normal cells and in transformed cells. However, all of the features that
we observed in spectra from preparations of cultured meningioma
cells in vitro (see above) were in agreement with data reported (23,30)
by noninvasive NMR spectroscopy on meningiomas in vivo. Such a
consistency lends support to the ideas that: (a) the metabolic pattern
in vivo is generally preserved by cells in culture when culture condi
tions are carefully controlled; and (b) that meningioma cell lines may
offer a suitable model for studying at least some of the properties of
their parental tumors.
In conclusion, our investigations showed that human meningioma
cell lines derived from excised primary meningiomas expressed sim
ilar characteristic
metabolic
profiles,
even when isolated
from differ
ent patients. Furthermore, both rat meninges and human meningioma
cell lines shared a number of features, which makes them distinguish
able from other cells of the central nervous system. It is interesting to
note the conservation of metabolites such as @3-hydroxybutyrate,ala
nine, glutamine, succinate, inositol, threonine, and maybe glycine
across species and across transformation from normal to tumor tissue.
Additionally, we identified certain characteristics such as Cho, PCI
GPC, creatine,glutamate,and taurinewhich enabledthe spectrafrom
the human transformed cells to be distinguished from those of normal
rat meningeal cells. If these differences prove to be due to metabolic
alterations in transformed
cells rather than due to species differences,
sociated with intensified cell membrane synthesis and increased rate
of cell replication and have been, therefore, proposed as an indicator they might enable neoplastic and normal tissues to be distinguished
noninvasively by ‘H-NMRspectroscopy. Together, our findings raise
of cell proliferation (29).
the possibility of identifying meningiomas in vivo by performing
Our findings showing raised amounts of Cho in preparations from
noninvasive ‘H-NMRspectroscopy on patients.
human meningiomas compared to extracts from rat meningeal cells
were also in agreement with the outcome of studies performed on
human meningiomas in vivo and on human biopsy extracts. Clinical
ACKNOWLEDGMENTS
studies by ‘H-NMRspectroscopy on humans have detected an in
crease of Cho levels in meningiomas compared to normal contralat
We are indebtedto Dr. Sian Davies for help with the HPLCdeterminations
eral brain (30), but the differences in distribution of individual peaks
for NAA and NAAG and to Sati Sahota for assistance in the detection of amino
acids by HPLC. Drs. Albert Busza and Martin King are gratefully acknowl
contributing to the Cho signal (i.e., phosphomonoesters and phos
426
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
METABOLIC
edged for their
respectively.
advice
on software
packages
PROFILES
and statistical
OF MENINGES
CELLS
tion of subpicomole amounts of amino acids by precolumn fluorescence derivation
methods,
with o-phthaldialdehyde. Anal. Chem., 51: 1667—1674,1979.
20. Koller, K. J., Zaczek, R., and Coyle, J. T. N-Acetyl-aspartyl-glutainate: regional
levels in rat brain and the effects of brain lesions as determined by a new HPLC
REFERENCES
1. Sze, D. Y., and Jardetzky, 0. Determination of metabolite and nucleotide concentra
tions in proliferating lymphocytes by ‘H-NMR
of acid extracts. Biochim. Biophys.
Acta, 1054: 181—197,1990.
method. J. Neurochem., 43: 1136—1142, 1984.
21. Smith, P. K.. Krohn, R. I., Hermanson, 0. T., Mallia, A. K., Gartner, F. H.,
Provenzano, M. D., Fujimoto, E. K.. Goeke, N. M., Olson, B. J., and Kienk, D. C.
Measurement ofprotein using bicinchoninic acid. Anal. Biochem., 150: 76—85,1985.
22. Altman, D., Three or more independent groups of observations. In: Chapman and Hall
(eds.), Practical Statistics for Medical Research, pp. 205-218. Padstow, Cornwall: T.
2. Behar, K. L, Den Hollander, J. A., Stromski, M. E., Ogino, T., Shulman, R. 0.,
Petroff, 0. A. C., and Prichard, J. W. High resolution ‘HNMR study of cerebral
hypoxia in vivo. Proc. Natl. Acad. Sci. USA, 80: 4945—4948,1983.
3. Gadian, D. 0., Proctor, E., williams,
AND MENINGIOMA
S. R., Cady, E. B., and Gardiner, R. M.
Neurometabolic effects of an inborn error of amino acid metabolism demonstrated in
vivo by ‘H
NMR. Magn. Reson. Med., 3: 150—156,
1986.
4. Williams, S. R. In vivo proton spectroscopy: experimental aspects and potential. In:
M. Rudin and J. Seeig (eds.), NMR—BasicPrinciples and Progress: ln Vivo Mag
netic Resonance Spectroscopy, pp. 55—72.Heidelberg: Springer Verlag, 1992.
5. Rothman,D. L, Behar,K. L, Hetherington,H. P.,andShulman,R. 0. Homonuclear
‘Hdouble-resonance difference spectroscopy of the rat brain in vivo. Proc. Nail.
Acad. Sci. USA, 81: 6330—6334,1984.
6. Gadian,D.0. ProtonNMRstudiesofbrainmetabolism.Philos.Trans.R.Soc.Land.
A333: 561—570,1990.
7. Urenjak, J., Williams, S. R., Gadian, D. 0., and Noble, M. Proton nuclear magnetic
resonance spectroscopy unambiguously identifies different neural cell types.
J. Neurosci., 13: 981—989,1993.
J. Press(Padstow),Ltd.,1991.
23. Bruhn, H., Frahm, J., Gyngell, M. L., Merboldt, K. D., Hanicke, w., Sauter, R., and
Hamburger, C. Non-invasive differentiation of tumours with use of localised ‘H-MR
spectroscopy in vivo: initial experience in patients with cerebral tumours. Radiology,
172: 541—548,1989.
24. Kepes, J. J., Meningiomas-biology, pathology and differential diagnosis. In: Masson
Monographs in Diagnostic Pathology, pp. 3—20.New York: Masson Publishing,
1982.
25. Scheithauer, B. W. Tumors of the meninges: proposed modifications of the World
Health Organization classification. Acts Neuropathol., 80: 343—354,1990.
26. Patel, A. J., and Hunt, A. Concentration of free amino acids in primary cultures of
neurons and astrocytes. J. Neurochem., 44: 1816—1921,1985.
27. Frahm, J., Michaelis, T., Merboldt, K-D., Hanicke, W., Gyngell, M. L., and Bruhn, H.
On the N-acetyl methyl resonance in localized ‘H
NMR spectra of human brain in
vivo. NMR Biomed., 4: 201—204,1991.
8. Wolswijk, 0., and Noble, M. Identification of an adult-specific glial progenitor cell.
Development (Carob.), 105: 387—400, 1989.
28. Lehninger, A. L. Glycolysis. In: Biochemistry, pp. 417—441.New York: Worth
Publishers, Inc., 1976.
9. Noble, M., Fok-Seang, J., and Cohen, J. Ohs are a unique substrate for the in vitro
growth of central nervous system neurons. J. Neurosci., 4: 1892—1903,
1984.
10. Bignami, A., Eng@L F., Dahi, D., and Uyeda, C. T. Lacalization of the glial fibrillary
29. Ruiz-Cabello, J., and Cohen, J. Phospholipid metabolites as indicators of cancer cell
acidic protein in astrocytes by immunofluorescence. Brain Res., 43: 429-435, 1972.
11. Halliday, W. C., Yeger, H., Duwe, 0. F., and Phillips, M. J. Intermediate filaments
in meningiomas. J. Neuropathol. Exp. Neurol., 44: 617—623,1985.
12. Hitchcock, E., and Morris, C. Immunocytochemistry of intracranial meningiomas.
J. Neuro-Oncol., 5: 357—368,1987.
13. Schwechheimer, K., Kartenbeck, J., Moll, R., and Franke, W. W. Vimentin filament
desmosome cytoskeleton of diverse types of human meningiomas. Lab. Invest., 51:
548—591, 1984.
14. Drago, J., Reid, K. L, and Bartlett, P. F. Induction of the ganglioside marker A2B5
function. NMR Biomed., 5: 226-233, 1992.
30. Gill, S. S. Thomas, D. 0. T., van Bruggen, N., Gadian, D. 0., Peden, C. J., Bell, J. D.,
Cox, I. J., Menon, D. K., lies, R. A., Bryant, D. I., and Courts, 0. A. Proton NMR
spectroscopy of intracranial tumours: in vivo and in vitro studies. J. Comput. Assist.
Tomogr., 14: 497—504,1990.
31. Blankenhorn, M., Bachert, P., Ende, 0., Knopp, M. V., Hess, T., Tromier, V.,
Forsting, M., Kolem, H., Lorenz, W. J., and van Kaick, 0. Clinical phosphorus-CSI
in preoperative embolization therapy ofmeningioma. Proc. Soc. Magn. Reson. Med.,
1026, 1992.
32. Kalra, R., Wade, K. E., Cadoux-Hudson, T. A. D., Parkes, H. 0., McDonald, B., Esiri,
M.,andRadda,0. K.NMRstudyofphospholipidmetabolismofhumanmeningioma.
on cultured cerebellar neural cells by growth factors. Neurosci. Len., 107: 245—250,
Proc. Soc. Magn. Reson. Med., 628, 1993.
1989.
33. lIes, R. A., Stevens, A. N., Griffiths, J. R., and Moths, P. 0. Phosphorylation status
15. Raff, M. C., Mirsky, R., Fields, K. L, Lisak, R. P., Dorfman, S. H., Silberberg, D. H.,
Gregson, N. A., Liebowitz, S., and Kennedy, M. C. Galactocerebroside is a specific
cell surface antigenic marker for oligodendrocytes in culture. Nature (Land.), 274:
by 31P-NMR spectroscopy and its implications for metabolic control. Biochem. 3.,
229: 141—151,
1985.
34. de Graaf, A. A., and Bovee, W. M. M. 3. Improved quantification of in vivo ‘H-NMR
spectra by optimization of signal acquisition and incorporation of prior knowledge
into spectral fitting. Magn. Reson. Med., 15: 305—319,1990.
and peripheral nervous system. In: J. Brockes (ed), Neuroimmunology, pp. 141—177. 35. Kauppinen, R. A., and Williams, S. R. Nondestructive detection of glutamate by
New York: Plenum Publishing Corp., 1982.
‘H-NMRspectroscopy in cortical brain slices from the guinea pig: evidence for
17. Hurd, R. E. Gradient-enhanced spectroscopy. J. Magn. Reson., 87: 422-428, 1990
changes in detectability during severe anoxic insults. J. Neurochem., 57: 1136—1144,
18. Preece, N. B., Baker, D., Butter, C., Gadian, D. 0., and Urenjak, J. Experimental
1991.
allergic encephalomyelitis raises betaine levels in the spinal cord of strain 1 guinea
36. Kauppinen, R. A., Pirttila, T-R. M., Auriola, S. 0. K., and Williams, S. R. Compart
pigs. NMR Biomed., 6: 194-200., 1993.
mentation of cerebral glutamate in situ as detected by ‘H/'3C-NMR.Biochem.
19. Lindroth, P., and Mopper, K. High perfonnance liquid chromatographic determina
J.,298: 121—127,
1994.
813—816,
1978.
16. Mirsky, R. The use of antibodies to define and study major cell types in the central
427
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.
Cell Type-specific Fingerprinting of Meningioma and Meningeal
Cells by Proton Nuclear Magnetic Resonance Spectroscopy
Catarina L. Florian, Nicholas E. Preece, Kishore K. Bhakoo, et al.
Cancer Res 1995;55:420-427.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/55/2/420
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 16, 2017. © 1995 American Association for Cancer Research.