(CANCER RESEARCH 50, 527-532, February 1, 1990)
31
P Magnetic Resonance Spectroscopy of Human Colon Cancer1
John N. Kasimos,2 Thomas E. Merchant, Louis W. Gierke, and Thomas Glonek
Department of Pathology [J. N. K., L. W. G.] and the Magnetic Resonance Laboratory [T. E. M., T. GJ, Chicago College of OstéopathieMedicine; Pathologisch Instituât
[T. E. M.I, Rijksuniversiteit Utrecht, Utrecht, The Netherlands
ABSTRACT
Phosphatic metabolite profiles of 19 malignant and normal human
colon specimens were analyzed by techniques of perchloric acid extraction
and 31P magnetic resonance spectroscopy at 202.4 MHz. Thirty-one
individual phosphorus-containing intermediates of metabolism were iden
tified and quantified for statistical intergroup comparisons. Elevations in
relative concentrations of phosphorylethanolamine, IMP, NADP 2'-P,
an uncharacterized resonance at 3.72 A, glycerol 3-phosphorylcholine,
phosphorylated glycans and the nucleoside diphosphosugars were seen
in malignant tissues concurrently with reductions in relative concentra
tions of phosphorylcholine, phosphocreatine (PCr), and ATP. The malig
nant and normal tissue groups were further characterized and contrasted
by computing metabolic indices from spectral data. Significant elevations
in phosphomonoesters, glycerolphosphodiesters, the ratio of phosphorylethanolamine/phosphorylcholine, and phosphomonoesters/inorganic
orthophosphate were detected in malignant tissues along with significant
reductions in the ratios of PCr/inorganic orthophosphate, PCr/ATP, the
energy charge of the adenylate system and the tissue energy modulus.
These results revealed significant alterations in high energy metabolism,
low energy metabolism, and membrane metabolism characteristic of
malignant tissues. The reduction in high energy phosphates ATP and
PCr was balanced by the net increase in nucleoside diphosphosugar and
a shift in equilibrium to metabolism involving low energy phosphomon
oesters. The spectral data of the tumors, which were of epithelial origin,
demonstrated minor metabolites not previously detected in tissue extract
analysis of malignant tissues. Detection of these minor metabolites
represents an indirect measurement of phospholipid metabolism in ma
lignant tissues.
INTRODUCTION
tumor recurrence (1, 2) is an ineffective screening technique for
early carcinomatous lesions of the colon due to the high degree
of overlap with its appearance in malignant and benign abnor
malities in the colon and other organ systems (3, 4). Double
contrast barium enema cannot alone consistently provide defin
itive diagnosis (5) and fecal occult blood testing is fraught with
false negative and positive results and therefore must be com
plimented by additional tests (5, 6). Flexible proctosigmoidoscopy is an effective screening tool which can often detect lesions
at an early stage in asymptomatic patients having routine
physical examination (5, 7) and colonoscopy is the preferred
method for evaluating a positive fecal occult blood test (6).
Those patients who are at risk for colon cancer, however, are
often reluctant to undergo such procedures. Besides the psycho
logical challenge and mild discomfort of the procedures, colon
oscopy is an invasive procedure which bears the potential risk,
albeit small, of intestinal perforation (5).
The advancement of in vivo MRS may provide a safe, pain
less, and accurate potential screening procedure in which a
noninvasive in vivo study of the colon could identify colonie
lesions by obtaining and comparing phosphatic spectra of the
lesions with normal or control spectra.
High resolution ''P-magnetic resonance spectra of tumors
are obtained to study biochemical details of the pathophysiology
of malignancy. Malignant neoplasms are composed of cells
whose net cell survival is increased resulting in an accelerated
observed tumor doubling time (e.g., colon tumors) (8). In gen
eral, the more anaplastic the tumor, the more rapid its growth
and the greater its metabolic turnover (9). In this case the 31Pmagnetic resonance spectroscopic profile should hypothetically
demonstrate characteristic features reflective of an altered met
abolic rate. These variations will also generate altered concen
trations of high and low energy phosphate compounds com
pared with normal tissues (10).
Herein we report the findings of an ex vivo study in which
the phosphorus-containing
metabolites of normal human coIonic mucosa with fibromuscular wall are compared to the
phosphorus-containing metabolites in tumors of colonie epithe
lial origin using techniques of PCA extraction and "P-MRS.
Magnetic resonance spectroscopy has gained clinical rele
vance as a research and diagnostic tool and has the potential
ability to yield biochemical and pathophysiological information
in vivo. MRS3 study of ex vivo tissues in high resolution analysis
may have the ability to act as an analytic adjuvant to current
diagnostic techniques.
Current methods for definitive diagnosis of colonie abnor
malities incorporate invasive medical and surgical procedures.
In the majority of cases, the patient presents in a symptomatic
state with weight loss, hematochezia, melanotic stools, tenesmus, abdominal distension, anorexia, or anemia; in severe cases
the patient may present with spontaneous intestinal perforation
or fistula development. The circulating level of CEA while
useful as a postoperative prognostic indicator and monitor of
MATERIALS
Received 6/14/89; revised 10/20/89; accepted 10/25/89.
The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1Supported through intramural resources of the Chicago College of Ostéo
pathie Medicine.
2To whom requests for reprints should be addressed, at Department of
Pathology, CCOM-Olympia Fields OstéopathieMedical Center. 20201 South
Crawford, Olympia Fields, IL 60461.
'The abbreviations used are: MRS, magnetic resonance spectroscopy; CEA,
carcinoembryonic antigen; MR, magnetic resonance: PCA, perchloric acid; GPC,
glycerol 3-phosphorylcholine; PME, phosphomonoester; GPD glycerolphosphodiester; NMP, nucleoside monophosphates; PE, phosphorylethanolamine; PC,
phosphorylcholine; GPE, glycerol 3-phosphory lethanolamine; PCr, phosphocrea
tine; t»-GP,tt-glycerol phosphate; Glu 1-P, glucose 1-phosphate: GPG, glycerol
3-phosphorylglycerol; GPS, glycerol 3-phosphoryIserine; GPI, glycerol 3-phosphorylinositol; Pi, inorganic orthophosphate: PG. phosphorylated glycans; NS,
nucleoside diphosphosugar.
AND METHODS
Surgical Procedures. Human colon tissue specimens consisting of
malignant colonie epithelial neoplasms and noninvolved controls were
obtained from colectomy specimens of 10 patients undergoing surgery
for colon cancer previously diagnosed by endoscopie biopsy and radiographic studies. The colectomy specimens were surgically removed and
promptly submitted to the Department of Pathology in an unfixed state
within 10 min following excision. The excised colons were opened in a
linear fashion with an enterotome, washed free of fecal contaminants,
and the neoplasms identified. A sample of the neoplasm weighing
approximately 1.5 g was removed from an area which macroscopically
appeared to be the least necrotic. A second sample of noninvolved
colonie mucosa and fibromuscular wall of a similar weight was taken
as a control. Without delay, the tissue samples were separately im
mersed in liquid nitrogen for storage. The remaining surgical specimen
was then sectioned, sampled, paraffin-embedded, and microscopically
examined for anatomic pathological diagnosis after staining with hematoxylin & eosin.
527
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
JIP MRS OF COLON
The histopathological type of the 10 tumors analyzed in this study
was adenocarcinoma of colonie epithelial origin. By histopathology,
nine of the tumors were graded as well differentiated and one was
poorly differentiated. By Dukes' classification, six of the well-differen
tiated tumors were classified as Dukes' B2, one as Dukes' C, and two
as Dukes' D. The poorly differentiated tumor was classified as Dukes'
expression used in the calculation is computed from mole percentage
values and not from molarities. The mean values of the indices were
also compared with a two-tailed / test with significance at the P < 0.05
and P < 0.01 levels. For purposes of the statistical analyses, missing
values represent resonance signals lying below the limits of detection.
C. The nine normal controls analyzed were orderly colonie mucosa
with attached fibromuscular wall and serosa.
Chemical Procedures. Preparation of PCA extracts was performed
according to procedures previously described for phosphorus MR tissue
extract analysis (11).
3IP-MRS. "P-MR qualitative calibrations and analyses were per
formed according to procedures previously described for phosphorus
MR tissue extract analysis (12). The MR spectrometer employed was
a General Electric 500-NB equipped with deuterium stabilization,
variable temperature, and Fourier-transform capabilities and operating
at 202.4 MHz for "P-MRS. All 3'P experiments were conducted under
proton-decoupled conditions in a 10-mm probe. PCA extracts were
analyzed in a 0.5-ml microcell, where, under nonspinning conditions,
the linewidth of standard hydrogen broad-band decoupled trimethylphosphate in water is 0.7 Hz. The typical spectrometer conditions for
31Pexperiments were as follows: pulse sequence, 1 pulse; pulse width.
18 /¿s
(45°spin-flip angle); acquisition delay, 500 ^s; acquisition time,
RESULTS
Thirty-one individual phosphorus-containing
metabolites
were detected in 19 human colon tissue specimens analyzed by
"P-MR spectroscopy. Representative malignant and normal
colon tissue spectra contain (Fig. 1), from downfield to upfield,
two small uncharacterized signals at 4.63 and 4.55 6 on the
extreme downfield side of the spectrum which are followed
upfield by the phosphomonoester resonance band. The phosphomonoester band includes hexose 6-phosphate at 4.48 Ãánd
extends to but does not include Pi at 2.63 ó.The other 15
resolvable components of this band include six uncharacterized
resonances at 4.34, 4.02, 3.98, 3.81, 3.72, and 3.39 5 and nine
other previously characterized resonances. These characterized
resonances include «-GP at 4.29 e, the twin resonances of
fructose 1,6-diphosphate at 4.10 a and 4.05 á,0-glycerol phos
phate at 3.93 5, the substituted ethanol phosphates, PE at 3.85
5 and PC at 3.32 5 and the nucleoside monophosphates IMP
and AMP at 3.78 a and 3.74 ¿,respectively, and the 2' phos
1.64 s; sweep width, ±6024 Hz; number of acquisitions, 24,000. In
addition, a computer-generated filter time constant introducing 0.6 Hz
line broadening was applied as needed. To compensate for relative
saturation effects among various phosphorus signals detected in a single
"P-MR spectroscopic profile, the MR spectrum was standardized
against measured amounts of tissue-profile metabolites wherever these
were known. The procedures for this calibration, insuring that an
accurate quantitative measurement was obtained from the "P-MR
phate of NADP at 3.63 6. Pi at 2.63 & is followed by the
resonance signal of Glu 1-P at 2.05 ¿.The phosphodiester band
is found immediately upfield. This band contains GPG at 0.98
6, GPE at 0.85 a, GPS at 0.66 6, GPI at -0.10 ¿,GPC at -0.13
(5,and the broad resonance of PG at —¿0.70
5. The next detectable
signal upfield is that of PCr at -3.10 &followed by the bands
spectral profile, have been described (11). The chemical shifts reported
follow the convention of the International Union of Pure and Applied
Chemistry and are reported in field independent units of 6 relative to
the shift position of 85% phosphoric acid. The internal chemical-shift
reference was GPC, —¿0.13
0.
Verification of Phosphomonoesters and Diesters. Peak assignments
are based upon accurate measurements of the chemical shift of the
resonance which, under invariable experimental conditions, is repro
ducible with acceptable accuracy. Phosphomonoesters and diesters in
the aqueous "P-MR spectrum can be identified by adding a known
quantity of a pure phosphorus-containing compound and observing the
position of the resonance. Phosphomonoesters and diesters not previ
ously identified but detected in this study were verified by these methods
(12).
Data Analyses. Metabolite concentrations in relative phosphorus
mole percentages were computed for all detected resonances in the
analyzed colon specimens using the curve resolution software of the
spectrometer. Mean metabolite concentrations in relative mole per
centages of phosphorus were calculated for the malignant and normal
tissue groups. The two groups were compared at the level of the
individual metabolites by implementing a two-tailed t test to the mean
metabolite concentration for each group. Significance was determined
at the P < 0.05 and P< 0.01 levels. From the grouped metabolite data,
the following indices were calculated: PME, GPD, NMP, PE + PC,
PE/PC, GPE + GPC, GPE/GPC, (PE + PC)/(GPE + GPC), PME/
Pi, ATP/Pi, PCr/Pi, PCr/ATP, energy charge (13), phosphorylation
potential (14), and the energy modulus (high energy phosphates/low
energy phosphates) (15). Energy charge is the extent to which the
adenylate system is filled with high energy phosphate groups and in the
steady state living system is nominally 0.85. The phosphorylation
potential, 1",of a tissue is a measure of its high energy status in terms
MALIGNANT
NORMAL
PCr
GPC
JÃœA
of its ability to synthesize high energy phosphates, the bulk of which
are in the form of ATP and is therefore a measure of the potential of a
living system to carry out ATP-dependent processes (16). In calculating
the phosphorylation potential it is assumed that the total detectable
phosphorus was 45 mivi which is a number obtained from a variety of
tissues. The important feature of this ratio, however, is not the phos
phorus concentration but the molar relationship between ATP, ADP.
and Pi. A total phosphorus concentration must be assumed because the
-'V1—I—
-10
Fig. I. The resonance signals from downfield (left) to upfield (right) are as
follows: uncharacterized signals at 4.63 and 4.55 ó:hexose 6-phosphate at 4.48
a; six uncharacterized resonances at 4.34. 4.02. 3.98, 3.81, 3.72, and 3.39 6; aglycerol phosphate at 4.29 o: the twin resonances of fructose 1.6-diphosphate at
4.10 o and 4.05 6; /i-glyccrol phosphate at 3.93 ¿;PE at 3.85 i; inosine and
adenosine monophosphate at 3.78 and 3.74 ó;nicotine adenine dinucleotide
phosphate at 3.63 />;PC at 3.32 ¿;Pi at 2.63 A;glucose I-phosphate at 2.05 6;
glycerol phosphoglycerol at 0.98 ¿:glycerol 3-phosphorylethanolamine at 0.85 ó;
glycerol 3-phosphorylserine at 0.66 a; glycerol 3-phosphorylinositol at -0.10 6;
GPC at -0.13 a; PG at -0.70 a. The next resonance signals upfield are PCr at
—¿3.10
6 followed by the bands corresponding to the ionized end groups y at
-5.70 6 and fi at -6.02 o, of ATP and ADP. respectively. Further upfield are the
resonance bands of the a-esterified phosphates of ATP at —¿10.88
6; ADP at
—¿
10.48 ¿;the dinucleotides (/}A') (nicotine adenine dinucleotide and nicotine
adeninedinucleotidc phosphate)at -11.20 »andthe NS at -12.88 fi(e.g., uridine
diphosphoglucose). On the extreme upfield side of the spectra is the resonance
arising of the rf-phosphate of the nucleoside triphosphates. primarily ATP at
-21.33 S.
528
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
"P MRS OF COLON
corresponding to the ionized end groups y at -5.70 <5and ß
at
—¿6.02
&,of ATP and ADP, respectively. Further upfield are the
resonance bands of the a-esterified phosphates of ATP at
-10.88 o, ADP at -10.48 5, dinucleotides (NADP and
NADPH) at -11.20 a, and NS at -12.88 6 (e.g., uridine
diphosphoglucose). Finally, on the extreme upfield side of the
spectra is the resonance arising from the middle 0-phosphate
of the nucleoside triphosphates, primarily ATP at —¿21.33
o. In
general, the high energy phosphates resonate between -1.5 and
—¿25
5 and the low energy phosphates between 8 and —¿1.5
¿.
Because the 3'P-MR spectrum measures the total phosphorus
profile of a substance rather than a select chemical entity, the
spectra analyzed in this study revealed a number of phosphoruscontaining molecules that have not been previously identified
by other chemical procedures.
Qualitatively, observable differences can be seen between the
malignant and normal spectra. The malignant spectrum dem
onstrates numbers of readily resolvable resonance signals in the
region of 4 5, the PME region. In contrast, the relative absence
of the uncharacterized resonance at 4.55 6 and the resonances
of Glu 1-P and GPS in the normal spectra is discernible.
Additionally, the relatively more pronounced PG peak (Fig. 1)
is a feature of the malignant spectra not observed in the normal
spectra.
Relative metabolite concentrations in mole percentages of
phosphorus were calculated for all detectable metabolites of the
tissue spectra. Spectra were grouped according to histopathological diagnosis as malignant or normal. Mean metabolite
concentrations were computed for metabolites appearing in the
spectra of the two groups (Table 1). To determine if differences
existed between the mean concentrations of the metabolites in
the two tissue groups, a two-tailed t test was applied. Statisti
cally significant elevations were observed in the mean relative
concentrations of metabolites IMP, NADP 2'-P, GPE, and NS
at the P < 0.05 level and PE, uncharacterized at 3.72 5, and
PG at the P < 0.01 level. Statistically significant decreases in
malignant tissues compared to normal were seen in the metab
olites PC at the P < 0.05 level and in PCr and ATP at the P <
0.01 level.
The highly resolved spectra of this study permitted the cal
culation of fifteen applied and theoretical metabolic indices
(Table 2), and thus metabolic indices also were compared on a
statistical basis using the two-tailed t test. Significant differ
ences at the P < 0.05 level were found to exist in the elevation
of the PME/Pi index and the depression of the tissue energy
modulus in malignant tissues. Highly significant elevations (P
< 0.01) were found to exist in the PME, GPD, and PE/PC
indices, and significant depressions in the PCr/Pi, PCr/ATP
indices and the energy charge of the adenylate system in malig
nant tissues.
DISCUSSION
The analysis of phosphorylated intermediates of metabolism
can be divided into the interpretation of phosphorylated metab
olites related to the cell membrane and the evaluation of high
and low energy metabolism that describes energy-generating
pathways.
Phosphorylated Metabolites Related to Membranes. Phos
phorylated metabolic intermediates representing precursors and
products of membrane phospholipid metabolism are numerous
in this high-resolution study. The also represent a majority of
the significant findings. The precursor products of membrane
phospholipid metabolism can be further classified according to
Table 1 Phosphatic metabolic profile of human colon tissues
Phosphatic0
metaboliteUU"Hex shift
(6)4.63
64.55
ó4.48a4.34
±0.040.11
±0.100.180.55
±0.030.59
0.120.44
+
0.070.38
±
0.060.87
+
64.2964.1054.050.040.84
±
0.070.43
±
20.31±0.1
1,6-diPFru
20.45±0.1
±0.030.36
1,6-diPUUß-GPPEUIMPAMPUNADP2'-PUPCPiGlu-l-P*GPGGPEGPSGPI''GPCPGPCrATPA
64.02
0.050.57
±
0.030.29
±
0.031.43
+
63.98
31.68±0.1
0.321.76
+
63.9563.85
±0.272.13
0.777.99
+
0.376.05
+
0.51'1.11
+
0.431.83
+
63.81
0.193.79
+
¡3.78o3.7553.7263.63
±0.265.00
0.61r1.56
+
0.611.75
+
0.361.02
+
±0.222.45
0.20'1.18±0.24/1.50
+
0.342.87
±
0.601.24
+
63.39
63.32
±0.261.83
±0.612.72
\f23.16±0.3
62.6352.0550.9850.8560.666-0.106-0.135-0.70
±0.2620.50
2.490.61
±
±2.190.480.36
0.040.25
+
0.060.86±0.15/0.29
+
0.031.34
±
0.110.32
+
0.040.76
±
0.091.45
±
0.161.61
+
+0.2217.85
±0.165.53
1.29'3.64
±
5-3.106jrA1JMalignant*0.18
2.060.60
±
0.64'29.90
+
0.1614.85
+
17'7.27
±3.
2.804.44
+
0.992.05
±
1.253.07
±
0.415.74
+
0.623.11
+
+ 0.92Normalc0.26 ±O.S2/
°The abbreviations used are: U, uncharacterized; Hex 6-P, hexose 6-phosphate; Fru 1,6-diP, fructose 1,6-diphosphate; /J-GP, 0-glycerol phosphate; NADP
2'-P, the 2'-phosphate group of NADP.
"«=10.
cn = 9.
d Metabolite lying below levels of detection for normal tissue.
' Significant difference between mean values, P < 0.01.
^Significant difference between mean values, P < 0.05.
"ATP, «-10.885, ß
-21.33 6, y -5.705.
* ADP, a -10.486,/3-6.025.
6-PUa-GPFru
'DN,-11.25.
•¿'NS,
-11.5 5, -12.886.
Table 2 Phosphatic metabolic indices of human colon tissues
index"PMEGPDNMPPEPC
Metabolic
\.lbd2.37
±
1.673.81
±
0.70a1.80
+
±0.292.98
0.329.82
±
±0.468.77
0.662.36
+
PC)PE/PCGPEGPC
(PE +
0.345.67+
±
0.27''2.
±
1.222.95
0.320.90
+
GPC)GPE/GPCPEPC/GPEGPCPME/PiATP/PiPCr/PiPCr/ATPEnergy
(GPE +
±0.240.79
15
0.013.63
+
0.285.62
±
0.341.71
+
1.851.05
±
±0.11'0.47
0.240.29
±
0.060.03
+
0.070.17
+
0.04''0.36
+
0.010.11
+
0.05*0.78
+
±0.010.60
0.04''241.95
±
0.04178.10
+
chargePhos
58.910.48
+
56.340.87
+
potentModulusMalignant*30.38
+ 0.10Normal'23.01 + 0.12'
" The abbreviations used are: GPE/GPC, GPE to GPC; PEPC/GPEGPC;
PE
and PC to GPE and GPC; PME/Pi, PME to inorganic orthophosphate; ATP/
Pi; PCr/Pi, phosphocreatine to Pi; energy charge, energy charge of the adenylate
system, energy charge = ([ATP] + ([ADP] * 0.5))/((ATP] + [ADP] + [AMP]);
phos potent, phosphorylation potential, phosphorylation potential (O = [ATP]/
[ADP] [Pi]; tissue energy modulus (ratio of high to low energy phosphates).
"n = \0.
cn = 9.
d Significant difference between tissue groups, P < 0.01.
' Significant difference between tissue groups, P < 0.05.
the phosphorus polar head group linkage; phosphomonoesters
such as phosphorylethanolamine
or phosphodiesters such as
glycerol 3-phosphorylethanolamine.
The chemical shift effects
of the functional groups results in the phosphomonoesters
resonating between 8.0 to 1.2 6 and the phosphodiesters reso
nating between 1.2 to -1.0 o (15). In this study, the analyzed
phosphomonoesters of membrane phospholipid metabolism are
529
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
"P MRS OF COLON
predominantly PE which is elevated in malignant tissue and
PC which is depressed in malignant tissue. The elevation of PE
is consistent with previous findings by Merchant et al. (10)
where PE was significantly elevated in malignant breast tissues
compared to normal. The phospholipid analogue of PE, phosphorylethanolamine is a major component of the inner leaflet
(17) of the cell membrane and is responsible for alterations in
membrane shape with malignancy (18). PE has also been pos
tulated to be a false neurotransmitter with inhibitory action
(19). The increase in ethanolamine metabolism is further iden
tified through the ethanolamine phosphodiester GPE which is
significantly elevated in malignant tissues. This compound is
identified as a saponification product of the membrane phos
pholipid phosphatidylethanolamine
resulting from conditions
favoring the hydrolysis of fatty acid side chains. Saponification
represents cellular injury, degradation of cells, and an increase
in phospholipase activities (20).
The decrease in the relative concentration of PC is a phenom
enon expected in malignant tissue. A recent study by our group
reports an increase in the alkyl-phospholipid analogue of phosphatidylcholine, phosphatidylcholine plasmalogen.4 This alkylphospholipid has been shown to be released by malignant tissue
and is responsible for the activation of immune response cells
(21). PC characteristically appears in the outer leaflet of the
cell membrane as a residue of phosphatidylcholine and sphingomyelin (17). It has also been postulated that PC is a false
neutrotransmitter with excitatory action (19).
The remainder of the intermediaries corresponding to mem
brane metabolism are found in the phosphodiester region of
the spectrum. Aside from the significant elevation of GPE, the
appearance and elevation of the glycerol-linked precursor-prod
ucts GPC, GPS, and notably GPI, although not significant
except for the relative absence of GPI in normal tissue, poten
tially reveals the breakdown of membrane phospholipids in an
environment favoring hydrolysis by lysophospholipase activity.
The relative absence of GPI in normal tissues is remarkable.
The appearance of GPI, coupled with the elevation of PE may
be explained through the findings of Hefta et al. (22). In their
study, they described that CEA, a peptide and oncofetal antigen
expressing the dedifferentiation of a variety of carcinomas
including colorectal cancer, is anchored to a phosphorylated
inositol-glycan in the membrane through an ethanolamine link
age. This finding raises the possibility that in other tumors with
elevated levels of CEA, inositol may be concomitantly elevated.
Given the role of CEA in expressing the degree of tumor
dedifferentiation, further investigation may focus on the corre
lation of phosphorylated inositol levels with the degree of
differentiation. The role of inositol phosphates as second mes
sengers evoking a broad spectrum of intracellular reactions is
being defined (23).
Metabolic indices found to differ significantly between malig
nant and normal tissues include the PME and GPD indices in
which membrane precursor products play an important role.
Both indices are elevated in malignant tissues. The summation
of PC and PE concentrations and GPC and GPE concentra
tions, while individually elevated or depressed, are themselves
not significant. Therefore, although the components of the
PME band, PE and PC, do contribute, respectively, an impor
tant portion of the signals, they do not account for the signifi
cant elevation by themselves. The two residues of phosphatides
that give rise to asymmetric membrane characteristics, PC and
4T. E. Merchant, P. Meneses, L. W. Gierke. and T. Glonek, 31P-magnetic
resonance phospholipid profiles of neoplastic human breast tissues, submitted for
publication.
PE, act as apparent antagonists of each other. Their ratio, PE/
PC, is significantly elevated in malignant tissues. The elevation
of the glycerolphosphodiesters
in malignant tissue means is
accounted for by the significant elevation in GPE and the
singular appearance of GPI in malignant tissues. The phos
phorylated glycans, represented by a prominent broad band in
the phosphodiester region and whose relative concentration of
the tissues' total mole percentage of phosphorus is only ex
ceeded by the flux of the inorganic nutrient Pi, are a significant
expression of the malignant membrane's generation of catabolic
end products, destruction, and uncontrolled growth. These
phosphorylated substances may now have a more definable role
as indicators of tumor differentiation based upon the glycan
linked with the phosphatide residues discussed above. The
increased production of these products represents major com
promise of membrane function. The elevation of PG in malig
nant tissues may represent a significant expression of the ma
lignant cell membrane's generation of catabolic, anabolic, or
intermediate products. It cannot be definitely presumed that
these are products of neoplastic growth or end products of
tumor necrosis and inflammatory response. However, the tissue
utilized in our study was sampled from areas free of grossly
observable necrosis. The vast majority of the tissue processed
was composed predominantly of neoplastic cells and associated
stroma.
High Energy Metabolism. Intermediates responsible for high
energy metabolism are found principally from —¿1.5
to —¿24
b in
the phosphatic spectrum. Of these metabolites, the significant
diminution of PCr and ATP in malignant tissue was offset only
slightly by the significant increase in the NS in the malignant
tissue. The decrease in PCr in malignant tissue is an expression
of malignant tissues previously measured by Merchant et al.
(10). The diminution of PCr in malignant tissue, also repre
sented by the significant depression of the PCr/Pi index repre
sents the loss or expenditure of an energy generating source as
an alternative to, or coupled with, the formation of ATP. PCr
diminution may also represent an increase in the biosynthesis
of protein since PCr is one of the high energy metabolites
responsible for nucleoside triphosphate synthesis, the essential
high energy cofactors required for protein biosynthesis. The
relationship of PCr to Pi is manifested in the near 50% relative
reduction in the concentration of ATP in malignant tissue. The
reduction in the ATP/Pi index and ratio of PCr/ATP represents
the tissues' decreasing dependence on high energy sources of
metabolic energy.
The elevation of NS portrays an alternative source of high
energy metabolism whose relevance is seen in the role NS plays
in (a) energy generating pathways of the hexose monophosphate shunt, (b) as a carrier molecule in phospholipid biosyn
thesis, and (c) as a precursor in nucleotide biosynthesis.
The levels of ATP, representing 30% of the relative mole
percentage of phosphorus concentration in the noninvolved
tissue testifies that the handling of the surgical tissue specimens
was adequate for preservation. The noninvolved tissue has near
basal levels of relative ATP concentration which is evidence of
tissue preservation of metabolites (24-26). The PCr is also
preserved in the control and the Pi levels are approximately
equivalent in the noninvolved and malignant tissue. This ex
emplifies that the degradation in high energy metabolism is not
from handling and that even though the ATP levels are nearly
two times greater in noninvolved tissues, these differences are
real and do not represent artifactual changes.
The energy charge represents the ability of tissue metabolism
to keep the adenylate system phosphorylated. "P-MRS analysis
530
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
"P MRS OF COLON
shows that the colon tumors examined do not have this ability.
This defect would be exacerbated by anaerobic conditions. Since
the energy charge of healthy tissues is approximately 0.80 (27),
the energy charge of the normal tissues, 0.78, reflects timely
handling of the tissue specimens which prevents their degra
dation.
Low Energy Metabolism. The accumulation of low energy
phosphatic intermediates in malignant tissues is observed by
the relative increases seen in IMP, the unknown resonance at
3.72 ¿,NADP 2'-P, and the singular appearance in normal
tissue of minor metabolites at 4.55 o and Glu 1-P. These
intermediates probably represent an accumulation of precursors
to glycolysis, energy generating pathways, and lipid metabo
lism. While relative accumulation of these intermediaries may
measure a tendency or shift to glycolysis as an energy-generat
ing source, the accumulation might imply that a decrease in
glycolytic cofactors is present. One such factor would be oxygen.
Neoplastic cells are subjected to increasing hypoxia as the
tumor growth extends beyond necessary vascular supply. The
accumulation of these intermediates may characterize pathways
not discussed at present. Overall, the shift to these processes is
manifested by the increased PME index and PME/Pi ratio
which account for minor metabolic changes.
In order to compare the high energy and low energy processes
the spectral tissue energy modulus was computed and found to
be significantly different between malignant and normal tissue.
The spectral tissue energy modulus is a comprehensive indicator
of the energy status of malignant tissues because its value is
influenced by all measurable quantities in the 31P spectrum.
The depression of the modulus in malignant tissue typifies
malignancy. Its depression has also been measured in malignant
breast tissues using similar methods (10). The modulus value
of the normal control 0.87 ±0.12 is expected as healthy tissues
have a modulus near unity.
While the metabolic differences observed are significant be
tween the malignant tissues analyzed and their noninvolved
controls, several problems inherent to this study must be rec
ognized. Firstly, though the tissues were handled identically,
malignant cells are known to differ biochemically from normal
cells in respect to metabolic products, structural proteins, en
zymatic proteins, carbohydates, and lipid constituents. The
possibility exists that the differences observed are actually ar
tifacts secondary to processing effects on tissues of varying
constitution rather than true metabolic characteristics of the
malignant tissues. Secondly, the malignant neoplasms tend to
be composed of large amounts of fibrous connective tissue
admixed with tumor cells, and almost totally displace the
normal colonie mucosa and fibromuscular wall in involved
areas. Hence, the ratio of epithelial cells to fibrous stroma is
highly variable in the malignant specimens and quantitatively
may exceed that of normal controls of equivalent size and
weight. This variability is an obstacle in controlling for quantity
of epithelium versus stroma in different tumor specimens,
though the ratio is fairly constant in normal tissue. Neverthe
less, the fact that similar qualitative and quantitative changes
were identified in all tumor specimens analyzed, suggests that
these differences have little effect on the assay. Thirdly, it is
possible that the controls obtained from the noninvolved re
gions of the colon may actually contain focal areas of disease
and thus may not truly reflect the metabolic profile of normal
colon which contains no focal areas of disease. However, as it
is not feasible to obtain entirely healthy colon specimens from
human subjects, and since to our knowledge no previous 31PMR spectra of normal human colonie tissue have been docu
mented, the assumption can be made at this time that the
profiles obtained from these control specimens reflect the met
abolic constituents of normal colonie tissue. Despite the afore
mentioned limitations, the ability of "P-MR spectroscopy to
distinguish malignant colonie tissue from normal colonie tissue
remains demonstrated.
In summary, highly resolved 3IP-MR spectra can be consist
ently obtained from chemical tissue extracts of human colon
adenocarcinoma. Significant differences in high energy and low
energy metabolism and metabolism related to the cell mem
brane can be measured and quantified. The findings of this
study are consistent with spectroscopic data of other human
malignancies. This includes the decrease in the tissue energy
modulus and the elevation in the phosphomonoesters and diesters with malignancy. This study supports previous findings
related to membrane breakdown and alteration characteristic
of malignancy. The most notable findings are related to mem
brane metabolism in malignant tissue where a shift in equilib
rium is seen from choline to ethanolamine metabolism and the
appearance of minor phosphodiester resonances in the malig
nant tissue spectra.
REFERENCES
1. Evans, J. T., Mittleman, A.. Chu. M., and Holyoke E. D. Pre- and postop
erative uses of CEA. Cancer (Phila.). 42: 1419-1421, 1978.
2. Minton, J. P., and Martin, E. W., Jr. The use of serial CEA determinations
to predict recurrence of colon cancer and when to do a second-look operation.
Cancer (Phila.), 42: 1422-1427, 1978.
3. Dent, P. B., McCulloch, P. B., Wesley-James, O., Maclaren, R., Muirhead,
W., and Dunnett, C. W. Measurement of carcinoembryonic antigen in
patients with bronchogenic carcinoma. Cancer (Phila.), 42: \ 484-1491. 1978.
4. Goldenberg, D. M., Shankey, R. M., and Primus, F. J. Immunocytochemical
detection of carcinoembryonic antigen in conventional histopathology spec
imens. Cancer (Phila.), 42: 1546-1553, 1978.
5. Fleischer, D. E., Goldberg, S. B.. Browning, T. H., Cooper, J. N., Friedman,
E., Goldner, F. H., Keeffe, E. B., and Smith, L. E. Detection and surveillance
of colorectal cancer. J. Am. Med. Assoc., 267: 580-585, 1989.
6. Knight, K. K., Fidding, J. E., and Battista, R. N. Occult blood screening for
colorectal cancer. J. Am. Med. Assoc., 261: 586-593, 1989.
7. Selby, J. V., and Friedman, G. D. Sigmoidoscopy in the periodic health
examination of asymptomatic adults. J. Am. Med. Assoc., 261: 594-601,
1989.
8. Baserga, R. The cell cycle. N. Eng. J. Med., 304:453-459, 1981.
9. Robbins, S. L., Cotran, R. S., and Kumar, V. (eds.). Pathologic Basis of
Disease, pp. 218-221. Philadelphia: W. B. Saunders, 1984.
10. Merchant, T. E., Gierke, L. W., Meneses, P., and Glonek, T. P-31 MR
spectroscopic profiles of neoplastic human breast tissues. Cancer Res., 48:
5112-5118, 1988.
11. Barany, M., and Glonek, T. Phosphorus-31 nuclear magnetic resonance of
contractile systems. In: D. L. Frederiksen and L. W. Cunningham (eds.),
Part B of Methods in Enzymology, Vol. 85, pp. 624-676. New York:
Academic Press, Inc., 1982.
12. Greiner, J. V., Kopp, S. J., Sanders, D. R., and Glonek, T. Organophosphates
of the crystalline lens: a nuclear magnetic resonance spectroscopic study.
Invest. Ophthalmol. Vis. Sci.. 21: 700-713, 1981.
13. Lehninger, A. L. Biochemistry, p. 538. New York: Worth Publishers, 1975.
14. Lehninger, A. L. Biochemistry, p. 539. New York: Worth Publishers, 1975.
15. Glonek, T., and Kopp, S. J. Ex vivo P-31 NMR of lens, cornea, heart and
brain. Magn. Reson. Imag., 3: 359-376, 1985.
16. Greiner, J. V., Kopp, S. J., and Glonek, T. Phosphorus nuclear magnetic
resonance and ocular metabolism. Surv. Ophthalmol., 30: 189-202. 1985.
17. Rothman, J. E., and Lenard, J. Membrane asymmetry. Science (Wash. DC),
195: 743-753, 1977.
18. Cullis, P. R., Hope, M. J., de Kruijff, B.. Verkleij, A. J., and Tilcock, C. P.
S. Structural properties and functional roles of phospholipids in biological
membranes. In: J. F. Kuo (ed.), Phospholipids and Cellular Regulation, vol.
I, pp. 1-59. Boca Raton, FL: CRC press, 1985.
19. Pettegrew, J. W., Moossy, J., Withers, G., McKeag, D., and Panchalingham,
K. 31P nuclear magnetic resonance study of the brain in Alzheimer's disease.
J. Neuropathol. Exp. Neurol. 47: 235-248, 1988.
20. Robbins, S. L., Cotran, R. S.. and Kumar, V. (eds.). Pathologic Basis of
Disease, pp. 8-17. Philadelphia: W. B. Saunders, 1984.
21. Snyder, F., and Wood, R. The occurrence and metabolism of alkyl and alkI-enyl ethers of glycerol in transplantable rat and mouse tumors. Cancer
Res., 28: 972-978, 1968.
22. Hefta, S. A., Hefta, L. J. F., Lee. T. D., Paxton, R. J., and Shively. J. E.
Carcinoembryonic antigen is anchored to membranes by covalent attachment
to a glycosylphosphatidyl-inositol moiety: identification of the ethanolamine
531
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
"P MRS OF COLON
linkage site. Proc. Nati. Acad. Sci. USA, SS: 4648-4652, 1988.
23. Rhee. S. G., Sun, P-G., Ryu, S-H., and Lee, S. Y. Studies of inositol
phospholipid-specific phospholipase C. Science (Wash. DC), 244: 546-550,
,9g9
,,
'„_ „¿, , ,
,.
. i •¿
r i.
i •¿ u i-.
24. Burt,
C. T., Glonek,
T., and
Barany,
M. Analysis
of phosphatic
metabolites,
the intracellular
pH, and
the state
of adenosine
triphosphates
in intact
muscle
Cohen. M. M. P-31 nuclear magnetic resonance analysis of brain: II. Effects
°foxygen deprivation on isolated perfused and nonperfused rat brain. J.
Neu.ro- chcm- 4i: 1716-1731, 1984
26. Greiner. J. V., Kopp. S. J., and Glonek. T. Distribution of phosphatic
metabolites in the crystalline lens. Invest. Ophthalmol. Vis. Sci., 26: 135]42 1935 'D. E. Adenine nucleotides as sloichiometric coupling agents in
27 Atkmsorii
by phosphorus nuclear magnetic resonance. J. Biol. Chem., 251: 2584-2591,
1976.
25. Kopp, S. J., Krieglstein, J., Freidank, A., Rachman, A., Seibert, A., and
metabolism and as regulatory modifiers: the adenylate energy charge. In: H.
J. Vogel (ed.). Metabolic Regulations. Vol. 5. pp. 1-21. New York: Academic
Press, 1971.
532
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1990 American Association for Cancer Research.
31P Magnetic Resonance Spectroscopy of Human Colon Cancer
John N. Kasimos, Thomas E. Merchant, Louis W. Gierke, et al.
Cancer Res 1990;50:527-532.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/50/3/527
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 18, 2017. © 1990 American Association for Cancer Research.