[Cancer Research 60, 5179-5186, September 15, 2000]
© 2000 American Association for Cancer Research
Experimental Therapeutics |
Noninvasive Real-Time Monitoring of Intracellular Cancer Cell Metabolism and Response to Lonidamine Treatment Using Diffusion Weighted Proton Magnetic Resonance Spectroscopy1
Yael Mardor,
Ofer Kaplan,
Marina Sterin,
Jesus Ruiz-Cabello,
Elissa Ash,
Yiftach Roth,
Israel Ringel and
Jack S. Cohen2
Advanced Technology Center, Sheba Medical Center, Tel Hashomer 52621, Israel [Y. M., Y. R., J. S. C.]; The Hebrew University, Jerusalem 91120, Israel [M. S., I. R.]; Complutense University, Madrid 28040, Spain [J. R-C.]; Tufts University, Boston, Massachusetts 02111 [E. A.]; and Tel-Aviv Medical Center, Tel-Aviv 64239, Israel [O. K.]
 |
ABSTRACT
|
|---|
We have used diffusion-weighted proton magnetic resonance spectroscopy
(DWMRS) to noninvasively selectively observe only the intracellular
metabolites of breast cancer and melanoma cell lines in
vitro in real time. Breast cancer cell lines representing
different stages in breast cancer progression were chosen for study.
Intracellular biochemical profiles of six cell lines perfused in
alginate beads were obtained. Spectral differences between groups of
cell lines, including choline, lactate, and threonine peaks, were
investigated. We also monitored response to the antineoplastic agent,
lonidamine (LND), as a function of time and drug concentration in
perfused cancer cells. Previous studies reported that this drug induced
intracellular acidification and lactate accumulation. Diffusion
weighted proton spectra demonstrated a 2- to 9-fold increase in the
intracellular lactate signal as a response to LND treatment in several
cancer cell lines. These results are consistent with the hypothesis
that the principal mechanism of LND in some cancer cells is marked
inhibition of lactate transport. Moreover, we have shown that there is
a factor of two to three between the response of melanoma cells and
that of some types of breast cancer cells. The higher sensitivity of
the melanoma cells, as predicted by proton DWMRS, was correlated with
changes in water-suppressed magnetic resonance spectra and confirmed by
a biological assay. This study demonstrates the feasibility of using
DWMRS for monitoring intracellular metabolism and for studying the
effects and mechanisms of action of anticancer drugs. We believe that
this method can be used for noninvasive clinical applications, such as
the differentiation between benign and malignant tissue, real-time
monitoring of response to therapy, dose response, and toxicity effects.
|
INTRODUCTION
|
|---|
Development of effective therapies against malignant diseases and
monitoring their effect in a noninvasive manner is one of the most
important challenges to biomedical research.
MRS3
is a very useful experimental approach to this end because it enables
noninvasive continuous monitoring of biochemical processes.
31P and 13C nuclei have
been used in the majority of MRS studies of the metabolism of perfused
cells (1)
. Proton is the most sensitive MR nucleus. It is
intrinsically 14 times more sensitive than 31P,
thus enabling shorter acquisition times and higher spatial resolution
in vivo. Moreover, proton MRS offers the advantage that
numerous natural biological compounds, as well as drugs and their
metabolites, can be detected. However, the presence of signals
originating from extracellular compounds and the immense water signal,
as well as many overlapping proton signals, have hindered the
application of proton MRS in studies of biological samples. Some proton
MRS studies have been reported in vivo (2, 3, 4)
;
however, these have not yet attained translation to routine clinical
application. In this work, we present the first application of DW
proton MRS (5)
to selectively observe only the
intracellular metabolites of several cell lines in vitro. We
anticipate that this method will significantly improve the application
of proton MRS for in vivo and clinical applications. This
method is based on differences in motional properties of the components
attributable to the higher intracellular protein concentration and to
the consequent restriction of molecular motion within cells.
Intracellular components have a lower ADC than extracellular components
and free water. We address the signal overlap problem by working with a
high magnetic field strength (600 MHz), which results in greater
chemical shift dispersion and by using curve fitting for the data
analysis.
Many breast cancer tumors appear to be initially responsive to
endocrine manipulation, and/or are sensitive to cytotoxic chemotherapy,
but later progress to a more malignant phenotype, characterized by
invasive/metastatic foci that are resistant to both endocrine
manipulation and chemotherapeutic intervention (6)
. To
study the progression from estrogen-dependent growth to
estrogen-independent and antiestrogen-resistant tumors, a series of
MCF7 variants has been isolated, which together represent the most
critical stages in breast cancer progression (6, 7, 8)
. The
cell lines studied in this work are all variants that have been
selected and extensively characterized, and exhibit specific phenotypic
changes that reflect critical characteristics of the progression in
therapy from the hormone-sensitive to -insensitive phenotype. The
characteristics of the cells chosen for study are detailed in Table 1
(8)
.
Developing new, effective anticancer drugs and optimizing the use of
presently available drugs requires understanding of their mechanism of
action. LND is an anticancer drug derived from indazole-3-carboxylic
acid, which was first introduced as an antispermatogenic and
embryotoxic agent (9)
. Clinical studies have shown that a
wide range of cancer types respond to the drug (10)
.
Previous studies, based on measurements of the extracellular milieu
alone (9
, 11)
, suggested that LND inhibits glycolysis,
thus inhibiting production of lactate (the end product of glycolysis).
Previous 31P and 13C MRS
studies (12
, 13)
concluded that LND induced intracellular
acidification and lactate accumulation. These results supported the
hypothesis that the main effects of LND are not through inhibition of
glycolysis, but are based on profound intracellular acidification and
inhibition of lactate efflux. This proposed mechanism led us to choose
LND as an example for these studies. Observing the intracellular
lactate signal with proton DWMRS enabled us to study the mechanism of
action of this anticancer drug directly. Moreover, it enabled us to
determine the metabolic response to therapy of different types of
cancer cells.
In this study, we have investigated potential biochemical differences
among human breast cancer cell lines and a murine melanoma cell line by
comparison of their intracellular proton spectra. We have also
investigated the effects of LND on the metabolism of several of these
cell lines, as a function of time and drug concentration. To confirm
the drug sensitivity of these cells, as detected by the DWMRS method,
we correlated it with other standard MRS and with non-MR studies.
|
MATERIALS AND METHODS
|
|---|
Cell Culture.
MDA-MB231, MDA-MB435, and MCF7 human breast cancer cells were
routinely maintained in DMEM and supplemented with
penicillin-streptomycin, L-glutamine (2 mM),
and 5% FCS. MCF7/MIII cells, an estrogen-independent variant of
MCF7 cells, and MCF7/LCC2 cells, an antiestrogen-resistant variant of
MCF7/MIII, were maintained in DMEM without phenol red (which has an
estrogenic-like activity; Ref. 14
) and supplemented with
penicillin-streptomycin, L-glutamine (2 mM),
and 5% charcoal-stripped calf serum. F10–9 murine melanoma cells were
routinely maintained in DMEM and supplemented with
penicillin-streptomycin, L-glutamine (2 mM),
and 10% FCS. All cell cultures were maintained at 37°C in a
humidified 5% CO2/95% air atmosphere incubator.
Cell Preparation for Proton MRS Studies.
Cells were grown in flasks to 90% confluence, harvested with 0.05%
trypsin, centrifuged at 4°C at 1000 x g
for 5 min, and washed twice with the growth medium. Cell pellet
(0.5–0.6 ml containing 
5 x 107 cells) was mixed with an equal volume of
liquid alginate (2%) in PBS. Beads were prepared by dripping the
mixture from a capillary into a calcium chloride solution. This caused
immediate gelling, and the microcapsules were then washed three times
with fresh medium. The capsules were kept at 4°C for a few hours and
then perfused with growth medium at 37°C for at least 30 min before
the experiment started.
LND Treatment.
LND experiments were performed with low glucose DMEM (5.5
mM glucose), similar to glucose physiological
concentration. LND was dissolved in DMSO and PEG200 (1 g of LND/10 ml
DMSO/25 ml PEG) before adding to the perfusion solution; the final LND
concentration used in the perfusion solution was 200 µg/ml unless
otherwise mentioned. In control experiments, we used the same volume of
DMSO and PEG as in the LND solutions.
Evaluation of Cellular Response to LND Treatment with a Cell
Counting Method.
Cells were grown in flasks to 90% confluence, and a pellet of
107 cells was harvested as described above. The
cells were resuspended in 2 ml of the growth medium with the addition
of LND at two concentrations: 400 and 600 µg/ml LND was dissolved in
DMSO and PEG200 before it was mixed in the medium as described above.
In control experiments, we used the same volume of DMSO and PEG as in
the LND solutions. The cell suspension was incubated for 2 h
during which the suspension was mixed every 10 min. The cells were
rinsed twice with PBS and then resuspended in 2 ml of PBS. To count the
percentage of dead cells, 20 µl of cell suspension were mixed with 20
µl of trypan blue (0.2%) and placed for counting under the
microscope.
Cell Perfusion.
The perfusion system used was described previously (15)
.
The inflow tube was 0.5 mm inner diameter, and was placed near the
bottom of a Wilmad 8-mm MR tube. The outflow was directed into openings
in a plastic insert, and then into an outflow tube. The gel beads,
which filled the MR tube, were concentrated at the bottom of the tube
by the plastic insert. Perfusion rate (0.9 ml/min) was maintained
using a peristaltic pump. Before the experiment, the system was washed
with 70% ethanol and then with the growth medium. The cells were
perfused continuously with appropriate fresh media at 37°C. The
medium was bubbled gently with a mixture of 5%
CO2/95% O2 to ensure
sufficient oxygenation and to maintain a physiological pH.
DW MR Method.
Stejskal and Tanner (16)
showed that application of a pair
of pulsed magnetic field gradients sensitizes spin echo MR experiments
to diffusion. Tanner (17)
subsequently showed that the
stimulated echo obtained in a three 90° rf pulse experiment is useful
when strong diffusion weighting is required with minimal loss of echo
time. To further suppress the water signal, a CHESS sequence was added,
which consists of a selective 90° rf pulse followed by gradient
dephasing, before the first and third rf pulses (18)
. A
schematic representation of the pulsed gradient stimulated echo
sequence used in this study is shown in Fig. 1
.
In the pulsed gradient method, the normalized intensity of the water
signal is given by:
where, I and I0 are
signal intensities in the presence and absence of diffusion sensitizing
gradients respectively, 
is the gyromagnetic ratio of the nuclei,
g and 
are gradient strength and duration, respectively,
(
- /3) is the effective diffusion time,
D is the molecular diffusion coefficient, and b
is the diffusion weighting factor, which is expressed in units of
s/cm2. By varying g, , and/or ,
a diffusion curve can be obtained (5)
, and from the echo
amplitude dependence on b, one can calculate the ADCs (Eq.
A). In biological systems, there are usually several spin populations
with different ADCs, so that the signal attenuation is not
monoexponential. In the simplest case of a two-compartment system
(intra- and extracellular), with very slow exchange between them, the
attenuation of the water signal should be a biexponential function of
b. The lower D value originates from the
intracellular molecules, and the higher D value arises from
the sum of the intra- and extracellular molecules.
Presaturation MR Method.
In the case of non-DW proton MR spectroscopy, water suppression is
essential to observe the much smaller metabolite signals. A common
approach is to presaturate the water signal using a selective inversion
pulse followed by a delay chosen so that the water is passing through a
null when the localization part of the sequence is started. We used
this sequence to obtain water- suppressed spectra that represent a
convolution of intra- and extracellular metabolites. In subsequent
reference to water-suppressed spectra, we mean spectra obtained using
the presaturation method.
MR Experiments.
The experiments were carried out on a Bruker 600 DMX spectrometer,
using a txi8mm probe with one gradient in the z direction with a
maximum intensity of 50 gauss/cm. Proton DW and water-suppressed MR
spectra were continuously recorded in an alternating manner. All
experiments were performed at 37°C. Intracellular spectra were
acquired with echo time TE = 11 ms,
repetition time TR = 2.2 s, diffusion
gradient intensity of 17.5 gauss/cm, duration of 5 ms, and diffusion
time of 127 ms. For the water suppression CHESS sequence, we used 8-ms
single lobe sinc rf pulses.
Data Analysis.
For well-resolved peaks, quantification of the spectra was obtained by
integrating over the peak of interest (after a baseline correction),
using the program MestRe-C, version 1.5.1 (written by Departamento de
Quimica Organica, Facultad de Quimica, Universidad de Santiago de
Compostela, Santiago, Spain) on a Pentium III personal computer.
In more complicated cases, such as the lactate and threonine peaks, the
region of interest was fitted to a function consisting of a sum of
gaussian curves (again, after a baseline correction), using the program
Physics Analysis Workstation (PAW version 2.09/18, CERN Program Library
Q121, Cern, Geneva 1995) on a Pentium III personal computer. Resonance
assignments are based on the abundant literature of extracts and
suspensions of different tumor cell lines and other tissues (1
, 19)
. Some assignments were confirmed by perfusing with high
concentrations of specific metabolites and observing the rise in the
corresponding intra/extracellular peaks.
|
RESULTS
|
|---|
Differentiation between Intra- and Extracellular Signals via DWMRS.
The logarithmic normalized water signal intensity as a function of the
squared gradient intensity for cells embedded in alginate beads is
shown as the open triangles in Fig. 2
(5)
. The gradient duration and separation time were kept
fixed. Therefore, the squared gradient intensity is proportional to the
DW factor b (Eq. A). The curve is clearly biexponential,
where the slow component represents the intracellular region. To prove
that the slow component is indeed intracellular, we repeated the
experiment with beads and medium only (see filled triangles in Fig. 2
),
and no slow component was found. An additional test was performed by
changing the extracellular medium to a buffer without any metabolites
(PBS). The intracellular spectrum did not change (Fig. 3A)
. By contrast, some of the peaks in the water-suppressed
spectrum (consisting of both intra- and extracellular molecules)
decreased or vanished (Fig. 3B)
. The latter are correlated
with the peaks in the extracellular spectrum, representing the
metabolic profile of the perfusion medium (Fig. 3C)
.
View larger version (41K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 3. Intra/extracellular separation using DWMRS.
A, the top plot shows intracellular
spectra (proton DWMRS with a gradient intensity of 17.5 gauss/cm) of
perfused MDA-MB231 cells. B, middle plot
shows the intra- and extracellular spectra (water-suppressed proton
MRS) of the same cells. The spectra in A and
B were measured alternately, so that the first spectrum
in both plots was acquired while perfusing with the growth medium. The
second spectrum and on were taken 5 min apart, after the perfusion was
switched to PBS (a buffer without any metabolites). C,
the bottom plot is a water-suppressed spectrum of the
growth medium, representing the extracellular spectrum.
|
|
Signal:Noise Ratio.
Comparing DWMRS with the water suppression approach (Fig. 3, A and B)
shows a significantly greater noise
level in the DWMRS spectra (Fig. 3A)
, although both
measurements were acquired alternately with the same number of scans.
This results from several factors. The stimulated echo sequence used in
the DWMRS method loses a factor of 2 in signal relative to the spin
echo sequence used in the presaturation method. Diffusion weighting
also reduces the resonance intensities of the intracellular species
(the slow component in Fig. 2
). Finally, the intra- and extracellular
metabolites have a higher absolute concentration than the intracellular
alone.
Comparison of DWMR Cell Spectra.
We measured intracellular proton spectra of five human breast cancer
cell lines (see Table 1
; Ref. 8
) and one mouse melanoma
cell line, using the DWMRS method. The intracellular spectra were
obtained by choosing a gradient value in the region of the slow slope
of the diffusion curve (Fig. 2)
, where the MR signal obtained
originates from slow moving, i.e., intracellular, molecules.
Between three and seven spectra were recorded for each cell line, at
least 1 month apart. Fig. 4
shows an example of one spectrum taken for each cell line. By looking
at the shapes of these spectra, it seems reasonable to divide them into
three groups: the first group consists of the top three spectra, MCF7,
MDA-MB231, and MDA-MB435; the second group consists of the next two
spectra, MCF7/LCC2 and MCF7/MIII; and the third is the melanoma cell
line. The differences between the cell lines are described in terms of
ratios between areas under peaks at given regions. We define the
choline, lactate/threonine, and leucine/isoleucine regions as regions
a, b, and c, respectively (Fig. 4)
. The ratio between the integrated
signal intensities of regions b and a is >1.0 in the first group,
<1.0 in the second group, and close to 1.0 in the third. The same is
true for the ratio between regions b and c. The lactate/threonine
region contains several overlapping peaks. The area under the lactate
(b1) and threonine (b2) subpeaks was calculated using curve fitting. It
can be seen that in the first group of cells, the threonine subpeak
dominates the lactate/threonine region, whereas in the second group,
the lactate dominates. The melanoma cell line has similar contributions
from both metabolites. The calculated ratios discussed above are
displayed in Fig. 5
. The ratios presented in the figure were calculated per spectrum and
then averaged over several measurements per cell line. The error bars
are the SDs of these measurements. Fig. 5
, A–D
shows in a quantitative manner the qualitative difference seen in Fig. 4
. Fig. 5
, A–C shows the average ratios a:b,
c:b, and a:c. Fig. 5
>D shows the ratio of b2:b1. In all plots
of Fig. 5
, the dashed line represents the average value of the first
group of cell lines: MCF7, MDA-MB231, and MDA-MB435. The dotted line
represents the average value of the second group: MCF7/LCC2 and
MCF7/MIII. The lines demonstrate the difference in the metabolite
ratios for the two groups of cell lines. Fig. 5
,
A–C suggests that the difference between the two
groups of cell lines originates from the lactate/threonine region
(region b). To determine which metabolite it is that contributes to
this difference, we calculated the ratio of the subpeaks in this
region, the lactate signal (b1) and the threonine signal (b2), to the
choline (a) and leucine/isoleucine (c) regions, and to each other.
These calculated ratios (Fig. 5
, E–H) suggests
that the main component that contributes to the difference between the
cell groups is threonine.
View larger version (25K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 4. A Qualitative comparison between cell lines. Intracellular
proton DWMR spectra of perfused human breast cancer and murine melanoma
cell lines at 37°C. Data were acquired with the following parameters:
TE/TR = 11/2200 ms,
diffusion gradient intensity and duration of 17.5 gauss/cm and 5 ms,
respectively, and diffusion time of 127 ms and 256 scans, respectively.
The choline, lactate/threonine, and leucine/isoleucine regions are
defined as regions a, b, and c
respectively. The lactate and threonine peaks within region
b are defined as b1 and
b2, respectively.
|
|
View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 5. A Comparison between peak ratios of five breast cancer
cell lines. Intracellular proton DWMR spectra of five breast cancer
cell lines are compared. The data are presented as ratios of regions of
peaks (regions a, b, b1, b2, and c are
defined in Fig. 4
). Between three and seven intracellular proton
spectra were obtained for each cell line. The results presented are
averaged over the measured spectra. Bars, SD. ----, the
average value of the first group of cell lines: MCF7, MDA-MB231, and
MDA-MB435. ····, the average value of the second group:
MCF7/LCC2 and MCF7/MIII.
|
|
Ischemia.
The effect of ischemia on the metabolism of MDA-MB435 breast cancer
cells was simulated by stopping the perfusion for a certain period of
time. The first spectrum was taken while perfusing at a flow rate of
0.9 ml/min. The next three spectra were taken while the perfusion
was off, 5 min apart. The perfusion was turned back on, and the last
three spectra were acquired. The main response to ischemic stress
observed using proton DWMRS was a rise in the intracellular lactate
signal. Fig. 6
shows the lactate signal intensity (calculated using curve fitting and
normalized to the choline signal) as function of time. It can be seen
that the lactate rise, after stopping the perfusion, was steeper than
the recovery, after perfusion was resumed. It can also be seen that
about 50 min after the perfusion was turned back on, the lactate signal
restored to its original value. The rise and the recovery of the
lactate signal were fitted to an exponential function:
p + qert. The
2 of the fit was 0.1.
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 6. Cellular response to ischemic stress. The intracellular
normalized lactate signal intensity of MDA-MB435 cells is plotted as a
function of time. The perfusion was stopped at time
t = 0 and turned back on 20 min later.
The lactate signal rise (----) and recovery (····) were fitted
separately to the following exponential function: p + qert.
|
|
Drug Response, Comparison between Cell Lines.
The effect of the antineoplastic agent LND on the metabolism of
perfused cancer cells, MCF7, MDA-MB231 and MCF7/MIII human breast
cancer cells, and F10–9 murine melanoma cells, was studied by DWMRS.
LND was dissolved in the perfusion medium as described in the
"Materials and Methods" section.
DW spectra demonstrated a 2- to 9-fold increase in the intracellular
lactate signal as response to LND treatment (Figs. 7
and 8)
. A moderate decrease of the lactate signal in the perfusate was also
observed in the water-suppressed spectra of the effluent solution. No
change was detected in the lactate signal while perfusing with the
control medium. Similar results were obtained with all four cancer cell
lines.
View larger version (32K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 7. Cellular response to LND treatment. Intracellular proton
DWMR spectra of perfused MCF7 cells, at 37°C. Each spectrum
represents 20 min of perfusion. The first spectrum represents cells
perfused with control medium, then the perfusion solution was replaced
by medium with LND. Note the marked elevation of lactate signal at 1.33
ppm.
|
|
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 8. Cellular response to LND treatment. The response to LND
treatment is defined as the increase in the intracellular lactate
signal intensity (calculated using curve fitting and normalized to the
choline signal) 50–60 min after adding LND to the perfusion medium.
Two to three experiments were performed for each cell line, at least 1
month apart. Note the higher response of the murine melanoma cell line,
F10–9.
|
|
The response of F10–9 to LND was significantly stronger than that of
the breast cancer cell lines (Fig. 8)
, suggesting increased sensitivity
to the drug. To test this hypothesis, we compared the DW spectra
(intracellular only) of the cells with the water-suppressed spectra
(intra- and extracellular) before and after treatment. The peaks chosen
for comparison were the DMSO and the choline peaks. DMSO was dissolved
in the perfusion medium in a known concentration and crossed the
membrane freely. Choline was present in the perfusion medium only in
insignificant quantities; therefore, the choline signal in the
water-suppressed spectra represented the intracellular choline only,
whereas the DMSO water-suppressed signal represented the sum of the
intra- and extracellular signals. We have compared the intracellular
choline:DMSO peak ratio before and after LND treatment and found that
this ratio did not change. Therefore, a decrease of the choline:DMSO
peak ratio in the water-suppressed spectra in response to LND treatment
implies a decrease in the total number of intact cells, or in other
words, an indication of cell death. The change in the choline:DMSO
peaks for three cell lines, MCF7, MDA-MB231, and F10–9, is shown in
Fig. 9
. The open triangles represent the DW data, serving as a
control, and the filled triangles represent the water-suppressed
spectra. The melanoma water-suppressed choline:DMSO ratio was
significantly decreased, whereas the breast cancer cells showed no
significant change.
To confirm the higher sensitivity of the murine melanoma cells to LND
treatment, three cell lines were tested via a biological assay
(described in the "Materials and Methods" section): MCF7,
MDA-MB231, and F10–9. The cells were treated for 2 h at two LND
concentrations: 400 and 600 µg/ml. Control experiments were performed
with medium containing DMSO and PEG200 in the same volume as in the LND
experiments. The percentage of dead cells in the LND treated samples
relative to control is presented in Fig. 10
. It can be seen that F10–9 cells show the highest sensitivity to LND
treatment.
Drug Response, Concentration Study.
To test the effect of LND concentration on the metabolism of cancer
cell lines, we perfused F10–9 murine melanoma cells with increasing
concentrations of LND. Spectra were acquired 30 min after each LND
increase. Fig. 11
shows the intensity of the lactate peak (arbitrary units) as a function
of LND concentration (in µg per ml). It is shown that the response to
the drug increases up to a certain level where is plateaus. This can be
explained either by a saturation of the cell response or by a
saturation of the medium in which the drug is dissolved.
View larger version (18K):
[in this window]
[in a new window]
[Download PPT slide]
|
Fig. 11. LND concentration dependence. The increase of the
intracellular lactate signal (normalized to the choline signal) of
F10–9 murine melanoma cells is plotted as a function of LND
concentration. The data were fitted to a sigmoidal function. See text
for details.
|
|
To find the maximum value of LND response and the midpoint of the
curve, we fitted the data to the following function, derived from the
Henderson-Hasselbach equation (20)
, describing the
transformation from one equilibrium state to the other:
where R is the lactate response (the factor by which
the lactate signal increased), Rmax
and Rmin are the maximal and minimal
response values, C is the LND concentration,
Cmid is the midpoint of the curve, and
a is a free parameter that determines the slope. The results
of the fit are: Rmax = 4.6 ± 0.3, Rmin = 0.6 ± 0.2, and
Cmid = 195.4 ± 14.4 µg/m.
The maximal response to LND was a 7.7-fold increase (4.6/0.6) of
the lactate signal intensity, and the midpoint of the response was an
LND concentration of 195.4 µg/ml.
|
DISCUSSION
|
|---|
We applied the method outlined by van Zijl et al.
(5)
and showed that we are able to obtain a complete
intra- and extracellular signal separation by using proton DWMRS on
perfused cells with a 600-MHz MR spectrometer. We used the method to
compare spectra of different cancer cell lines and to monitor their
response to therapy and stress.
The development of estrogen independence, antiestrogen resistance, and
increased metastatic potential are among the most important phenotypic
changes associated with malignant progression in breast cancer
(6
, 21)
. It has been shown previously using
31P MRS (8)
that relative to the
parental MCF7 cells, the baseline spectra for both MCF7/MIII and
MCF7/LCC2 cells exhibit 50% higher UDPG levels. These
differences are associated with perturbations in the
glucose/glycoprotein/proteoglycan metabolism and with the acquisition
of an intermediate estrogen-independent and
estrogen-responsive/ER-positive phenotype. The authors conclude that
this may reflect a fundamental difference in how the intermediate
phenotype of the MCF7/MIII and MCF7/LCC2 cells use glucose metabolites
relative to ER-positive, E2-responsive cells (e.g., MCF7).
We also observed qualitative differences in the proton DWMR spectra
between two equivalent groups of breast cancer cell lines (Figs. 4
and 5)
. However, the differences we observed after final analysis were
mainly related to the threonine signal.
Some of the average differences in the calculated ratios between the
two groups of breast cancer cell lines were very large. For example,
the ratio of the lactate region:choline region shown in Fig. 5A
was about an order of magnitude larger in the MCF7,
MDA-MB231, and MDA-MB435 group than in the MCF7/MIII, MCF7/LCC2 group.
Still, the metabolic profile of cells can be easily affected by growth
conditions and preparations for MRS studies (22
, 23)
. As a
result, we also observed deviations between different measurements of
the same cell line, leading to large SDs of the average calculated
ratios (Fig. 5)
. Therefore, we conclude that for the purpose of
differentiating between cell lines, we were able to show the general
trend, but not to prove that there is a statistically significant
difference between cell lines in vitro. However, non-DW
proton MRS studies in vivo have shown significant
differences between peak ratios of different stages of malignancies in
the brain (24, 25, 26)
. Moreover, in a comparison between
in vitro and in vivo proton MR spectra of
squamous cell carcinoma, it was observed that the ratio of
choline:creatine in vivo was consistently larger than the
in vitro ratio (27)
. These results suggest that
DWMRS might be a better tool for in vivo studies than for
in vitro studies in this respect.
In contrast to the comparison of cellular spectra, the cell response to
LND treatment, as measured by proton DWMRS, is reproducible and
statistically significant. All of the cell lines tested responded to
LND treatment by a marked increase of the intracellular lactate signal
and a moderate decrease in the extracellular lactate signal. Thus,
real-time monitoring of intracellular metabolic response to LND
treatment was consistent with the LND mechanism of action as described
by Ben-Horin et al. (12)
. Moreover, we
have shown that there is a factor of 2–3 between the response (the
increase in intracellular lactate) of F10–9 melanoma cells and that of
breast cancer cells (Fig. 8)
. The higher sensitivity of the melanoma
cells, as predicted by proton DWMRS, was correlated with a decrease in
the choline/DMSO peak in the water-suppressed MR spectra (Fig. 9)
and
confirmed by a biological assay (Fig. 10)
.
Proton DWMRS was also used to monitor the intracellular metabolic
response to the ischemic stress induced by stopping the perfusion for a
limited period of time, and to monitor the metabolic recovery after the
perfusion was turned back on. The exponential behavior, in response to
ischemic stress, was previously shown by Knop et al.
(28)
for ATP and Pi in phosphorus MR spectra of perfused
cells at 25°C. The authors were not able to repeat the measurement at
37°C because the response at higher temperature is faster and there
were too few data points to adequately fit an exponential function.
This is probably attributable to the long acquisition times needed in
phosphorus MRS (each spectrum was obtained in 22 min). This example
demonstrates the advantage of using proton MRS in terms of time
resolution. The proton DW spectra (8 scans) were obtained in 40 s
each (Fig. 6)
. The short acquisition time enabled us to measure the
time dependence of the response to ischemic stress at 37°C with good
temporal resolution.
This study demonstrates the feasibility of using DW proton MRS
for monitoring intracellular metabolism noninvasively and in real time.
We have shown that this method can be used for following the metabolic
response of cells to stress as a function of time and to follow their
recovery. We have also shown that DWMRS is a very useful tool to study
the effects and mechanism of action of an anticancer drug on perfused
cells, as a function of time and drug concentration. Finally, we have
shown that DWMRS was able to predict that one type of cell is more
sensitive to a certain anticancer treatment than others. In comparison
with other MR techniques, the proton DWMRS method has distinct
advantages in that it is noninvasive (no contrast agents used),
information can be obtained for proton-containing molecules (most drugs
and metabolites), the MR signal is strong (allowing shorter acquisition
times and better spatial resolution than 31P
MRS), and complete intra- and extracellular separation of signals is
attainable (compared to standard proton MRS).
We anticipate that DWMRS will provide a significant improvement
over standard proton MRS, especially in the case of cancer. Because of
the heterogeneous nature of tumors, standard proton MRS averages over
different types of tissues, such as viable tissue, necrosis, and edema.
The extracellular volume fraction in tumors has been shown to vary
between 30% in viable tissue to 90–100% in necrotic areas
(29, 30, 31)
. By using DWMRS to suppress the signals from
nonviable extracellular matter, enhanced sensitivity for the viable
tissue can be obtained. The results presented in this work demonstrate
the potential of using DWMRS for noninvasive in vivo and
clinical applications, such as real-time differentiation between benign
and malignant tissue, detection of residual malignancy after treatment
(surgery, laser ablation, cryogenic ablation, radiation), monitoring of
cellular metabolic response to therapy, dose response studies, and
investigation of metabolic effects in tissues susceptible to drug
toxicity. However, applying the DWMRS method clinically will not
be a trivial task. There are several problems that have to be overcome,
such as the fact that magnetic field gradients on clinical MRI machines
are normally limited in intensity. Therefore, to obtain high
b values, it will be necessary to use longer gradient
duration times and longer diffusion times, allowing more exchange
between intra- and extracellular species (32
, 33)
. Still,
the exchange between intra- and extracellular metabolites is slower
than that of water, thus the separation between inta- and extracellular
metabolites should be better than that of water. Other technical
problems concerning clinical DWMRS are that the gradients might not
have strong enough shielding, the magnetic field homogeneity in
vivo might be problematic, and DWMRS is sensitive to motion
artifacts. We are presently carrying out in vivo DWMR
spectroscopy and imaging animal studies and intend to extend to
clinical studies in the near future.
|
ACKNOWLEDGMENTS
|
|---|
We thank Dr. Peter van Zijl for helpful technical advice in
carrying out the DWMRS method, Dr. Gil Navon for consultations on LND
experiments, Dr. Hugo Gottlieb for assistance in carrying out the NMR
experiments, and Dr. Eva Meirovitz for providing us with machine time
at Bar-Ilan University. We thank Dr. Robert Clarke of Georgetown
Medical School (Washington, D.C.) for generously providing us with the
breast cancer cells and Drs. Yoel Kloog and Amiram Raz of Tel-Aviv
University for generously providing us with the melanoma cells. We
thank Dr. Walter Frosecchi of F. Angelini Institute (Rome, Italy), for
kindly supplying us with LND. We thank Sharona Salomon for performing
excellent laboratory work. This research was performed in the Advanced
Technology Center, Sheba Medical Center (Tel-Hashomer, Israel). We
thank the Head of the Center, Dr. Arie Orenstein, for his generous help
and encouragement.
|
FOOTNOTES
|
|---|
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.
1 Supported by United States Army Grant
DAMD-17-94-J-4032 and Grant BIO97-0543 (to J. R-C.). 
2 To whom requests for reprints should be
addressed, at Advanced Technology Center, Sheba Medical Center,
Tel-Aviv University, Sackler School of Medicine, Tel Hashomer 52621,
Israel.
3 The abbreviations used are: MRS, magnetic
resonance spectroscopy; MR, magnetic resonance; DW, diffusion-weighted;
LND, lonidamine; DWMRS, DW proton MRS; PEG, polyethylene glycol; rf,
radiofrequency; CHESS, chemical shift selective saturation; ADC,
apparent diffusion constant; ER, estrogen receptor; TAM, tamoxifen;
UDPG, uridine di-phospho-glucose.
Received 2/23/00.
Accepted 7/20/00.
|
REFERENCES
|
|---|
-
Kaplan, O., van Zijl, P. C. M., and Cohen, J. S. NMR studies of metabolism of cells and perfused organs. In: J. Seelig and M. Rudin (eds.), In Vivo Magnetic Resonance Spectroscopy III: In Vivo MR Spectroscopy: Potential, and Limitations, Vol. 28, Heidelberg: Springer-Verlag, pp. 4–52, 1992.
-
Garcia-Segura J. M., Sanchez-Chapado M., Ibarburen C., Viano J., Angulo J. C., Gonzalez J., Rodriguez-Vallejo J. M. In vivo proton magnetic resonance spectroscopy of diseased prostate: spectroscopic features of malignant versus benign pathology. Magn. Reson. Imaging, 17: 755-765, 1999.[Medline]
-
Kim, J. K., Kim, D. Y., Lee, Y. H., Sung, N. K., Chung, D, S., Kim, O. D., and Kim, K. B. In vivo differential diagnosis of prostate cancer and benign prostatic hyperplasia: localized proton magnetic resonance spectroscopy using external-body surface coil. Magn. Reson. Imaging, 16: 1281–1288, 1998.
-
Nofray J. F., Tomita T., Byrd S. E., Ross B. D., Berger P. A., Miller R. S. Clinical impact of MR spectroscopy when MR imaging is indeterminate for pediatric brain tumors. Am. J. Roentgenol., 173: 119-125, 1999.[Abstract/Free Full Text]
-
van Zijl P. C. M., Moonen C. T. W., Faustino P. J., Pekar J., Kaplan O., Cohen J. S. Complete separation of intracellular, and extracellular information in NMR spectra by diffusion weighting. Proc. Natl. Acad. Sci. USA, 88: 3228-3232, 1991.[Abstract/Free Full Text]
-
Clarke R., Dickson R. B., Brunner N. The process of malignant progression in human breast cancer. Ann. Oncol., 1: 401-407, 1990.[Abstract/Free Full Text]
-
Leonessa, F., Boulay, V., Wright, A., Thompson, E. W, Brunner, N., and Clarke, R. The biology of breast tumor progression: acquisition of hormone-independence and resistance to cytotoxic drugs. Acta Oncol., 31: 115–123, 1991.
-
Ruiz-Cabello J., Berghmans K., Kaplan O., Lippman M. E., Clarke R., Cohen J. S. Hormone dependence of breast cancer cells and the effects of tamoxifen and estrogen: 31P NMR studies. Breast Cancer Res. Treat., 33: 209-217, 1994.
-
Silvestrini B., Palazzo G., De Gregorio M. Lonidamine and related compounds. Prog. Med. Chem., 21: 111-135, 1984.
-
Putzuolo S., Taddei I. New perspectives in anticancer therapy: A review of lonidamine. Drugs Today, 25: 111-135, 1989.
-
Floridi A., Paggi M. G., Marcante M. L., Silvestrini B., Caputo A., De Martino C. Lonidamine, a selective inhibitor of aerobic glycolysis in tumor cells. J. Natl. Cancer Inst., 66: 497-499, 1981.
-
Ben-Horin H., Tassini M., Vivi A., Navon G., Kaplan O. Mechanism of action of the antineoplastic drug lonidamine: 31P and 13C nuclear magnetic resonance studies. Cancer Res., 55: 2814-2821, 1995.[Abstract/Free Full Text]
-
Vivi A., Tassini M., Ben-Horin H., Navon G., Kaplan O. Comparison of action of the anti-neoplastic drug lonidamine on drug-sensitive and drug-resistant human breast cancer cell: 31P and 13C nuclear magnetic resonance studies. Breast Cancer Res. Treat., 43: 15-25, 1997.[Medline]
-
Levenson A. S., Jordan V. C. MCF-7: the first hormone-responsive breast cancer cell line. Cancer Res., 57: 3071-3078, 1997.[Free Full Text]
-
Daly P. F., Lyon R. C., Faustino P. J., Cohen J. S. Phospholipid metabolism in cancer cells monitored by 31P NMR spectroscopy. J. Biol. Chem., 262: 14875-14878, 1987.[Abstract/Free Full Text]
-
Stejskal E. O., Tanner J. E. Spin diffusion: Spin echoes in the presence of a time-dependent field gradient. J. Chem. Phys., 42: 288-292, 1965.
-
Tanner J. E. Use of the Stimulated Echo in NMR Diffusion Studies. J. Chem. Phys., 52: 2523-2526, 1970.
-
Moonen C., van Zijl P. Highly effective water suppression for in vivo proton NMR spectroscopy. J. Magn. Reson., 88: 28-41, 1990.
-
Robinson S. P., Barton S. J., McSheehy P. M., Griffiths J. R. Nuclear magnetic resonance spectroscopy of cancer. Br. J. Radiol., 70: S60-S69, 1997.
-
Edsall J. T., Wyman J. Biophysical ChemistryVol Academic Press 1. New York 1958.
-
Clarke R., Skaar T., Baumann K., Leonessa F., Lippman J. M., Thompson E. W., Freter C., Brunner N. Hormonal carcinogenesis in breast cancer: cellular and molecular studies of malignant progression. Breast Cancer Res. Treat., 31: 237-248, 1994.[Medline]
-
Shedd S. F., Lutz N. W., Hull W. E. The influence of medium formulation on phosphomonoester and UDP-hexose levels in cultured human colon tumor cells as observed by 31P NMR spectroscopy. NMR Biomed., 6: 254-263, 1993.[Medline]
-
Franks S. E., Kuessel A. C., Lutz N. W., Hull W. E. 31P MRS of human tumor cells: effects of culture media and conditions on phospholipid metabolite concentrations. Anticancer Res., 16: 1365-1374, 1996.[Medline]
-
Norfray J. F., Darling C., Byrd S., Ross B. D., Schwalm C., Miller R., Tomita T. Short TE proton MRS, and neurofibromatosis type 1 intracranial lesions. J. Comput. Assisted Tomogr., 23: 994-1003, 1999.[Medline]
-
Meyerand E. M., Pipas M. J., Mamourian A., Tosteson T. D., Dunn J. F. Classification of biopsy-confirmed brain tumors using single-voxel MR spectroscopy. Am. J. Neuroradiol., 20: 117-123, 1999.[Abstract/Free Full Text]
-
Shino A., Nakasu S., Matsuda M., Handa J., Morikawa S., Inubushi T. Noninvasive evaluation of the malignant potential of intracranial meningiomas performed using proton magnetic resonance spectroscopy. J. Neurosurg., 91: 928-934, 1999.[Medline]
-
Mukherji S. K., Schiro S., Castillo M., Kwock L., Muller K. E., Blackstock W. Proton MR spectroscopy of squamous cell carcinoma of the extracranial head and neck: in vitro and in vivo studies. Am. J. Neuroradiol., 18: 1057-1072, 1997.[Abstract]
-
Knop R. H., Chen C. W., Mitchell J. B., Russo A., McPherson S., Cohen J. S. Metabolic studies of mammalian cells by 31P-NMR using a continuous perfusion technique. Biochim. Biophys. Acta, 804: 275-284, 1984.[Medline]
-
Paran Y., Edgar R., Grobgeld D., Bendel P., Degani H. Mapping the intracellular and extracellular apparent diffusion coefficient and volume fraction of MCF7 human breast cancer. Proc. Intl. Soc. Magn. Reson. Med., 8: 767 2000.
-
Stubbs M., Bhujwalla Z. M., Tozer G. M., Rodrigues L. M., Maxwell R. J., Morgan R., Howe F. A., Griffiths J. R. An assessment of 31P MRS as a method of measuring pH in rat tumors. NMR Biomed, 5: 351-359, 1992.[Medline]
-
Bhujwalla Z. M., McCoy C. L., Glickson J. D., Gillies R. J., Stubbs M. Estimations of intra- and extracellular volume and pH by 31P magnetic resonance spectroscopy: effect of therapy on RIF-1 tumors. Br. J. Cancer, 78: 606-611, 1998.[Medline]
-
Assaf Y., Cohen Y. Non-mono-exponential attenuation of water and N-acetyl aspartate signals due to diffusion in brain tissue. J. Magn. Reson., 131: 69-85, 1998.[Medline]
-
Pfeuffer J., Flogel U., Leibfritz D. Monitoring of cell volume and water exchange time in perfused cells by diffusion-weighted 1H NMR spectroscopy. NMR Biomed., 11: 11-18, 1998.[Medline]
This article has been cited by other articles:
|
|
|
|
 
J. Fang, Q. J. Quinones, T. L. Holman, M. J. Morowitz, Q. Wang, H. Zhao, F. Sivo, J. M. Maris, and M. L. Wahl
The H+-Linked Monocarboxylate Transporter (MCT1/SLC16A1): A Potential Therapeutic Target for High-Risk Neuroblastoma
Mol. Pharmacol.,
December 1, 2006;
70(6):
2108 - 2115.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. Papaldo, M. Lopez, E. Cortesi, E. Cammilluzzi, M. Antimi, E. Terzoli, G. Lepidini, P. Vici, C. Barone, G. Ferretti, et al.
Addition of Either Lonidamine or Granulocyte Colony-Stimulating Factor Does Not Improve Survival in Early Breast Cancer Patients Treated With High-Dose Epirubicin and Cyclophosphamide
J. Clin. Oncol.,
September 15, 2003;
21(18):
3462 - 3468.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
ABSTRACT
INTRODUCTION
) and control gel (
). The squared gradient
intensity is proportional to the diffusion factor b
because 
