| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Advances in Brief |
Departments of Medicine [R. P. K., D. P. P., M. L. W., P. C., J. K.], Radiology [E. X. W., P. O. A., J. K.], and Pharmacology [R. P. K.], Columbia University, New York, New York 10032, and Department Pathology, University of Minnesota, Minneapolis, Minnesota [M. S.]
 |
ABSTRACT |
|---|
 Top ABSTRACT IntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
|
Introduction |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
Because of the biological importance of sodium, its relative abundance, and its sensitivity, sodium NMR is a particularly useful tool for the study of pathophysiological processes. The large transmembrane ionic gradient causes the much lower IC concentration ([Na]i) to be highly dependent on active processes (Na+/K+-ATPase pump) and thus responsive to metabolic suppression. With special relevance to the study of neoplasms, [Na]i is correlated with the proliferation rate of nonneoplastic and malignant cell populations (5) , presumably because of the role of Na+ influx in initiating movement through the cell cycle. Na+ flux is mediated by transmembrane exchangers for both Ca2+ and H+ (6 , 7) . Thus, not only is there probable elevation of [Na]i because of the toxic effects of chemotherapy, but the altered values of baseline [Na]i, if detectable, could also have diagnostic significance.
Measurement of sodium content clinically has typically been done using SQ NMR techniques (8 , 9) . A significant disadvantage of SQ NMR is the relatively larger abundance of EC versus IC [Na]. SQ NMR requires paramagnetic shift reagents to discern Nai. Because of their relative membrane impermeability, they selectively shift the MR spectrum of the EC sodium. However, shift reagents would have severe disadvantages, including toxicity, in clinical use. An alternative MR approach to measure IC sodium content is based on the interaction of sodium polyanions and their resultant effects on nuclear spin transitions. Spin 3/2 nuclei (such as Na and K) have a nonvanishing quadrupole moment, allowing interaction with electrostatic field gradients (10) . In certain complex environments, such as those occurring in IC space, MQ spin transitions occur that can be detected by specific pulse sequences called MQ filters (11, 12, 13, 14) . The EC sodium diffuses relatively freely, with a small but presumed constant component electrostatically bound to plasma proteins and the surface of cell membranes. Thus, the presence of an MQ signal can be used to identify populations of sodium nuclei by their molecular environment and can be used to detect [Na]i changes (15, 16, 17, 18, 19, 20, 21, 22) .
In this study, we capture this utility of MQ sodium MRI, the ability to identify sodium by its microenvironment, but use instead a SQ pulse sequence. The higher signal-to-noise of SQ sodium MRI is a distinct clinical advantage. The sodium subpopulation weighting is accomplished by applying an IR pulse sequence to null the signal from sodium nuclei with long longitudinal relaxation times (T1), as found in free solution or plasma. The image signal is then weighted toward bound sodium, which is primarily IC. The IR technique is commonly used in proton imaging to distinguish fat and water. This study was designed to test the adequacy of such a weighting scheme and whether images so derived would have the hypothesized diagnostic utility.
|
Materials and Methods |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
Flow Cytometry and Fluorescent Markers for Apoptosis.
Fluorescent end-labeling techniques for flow cytometry were as
described (23)
. Briefly, cells were fixed in ice-cold
methanol free formaldehyde (1% in PBS), resuspended in 70% ethanol,
and stored at -20°C (24)
. Kit directions were followed
(APO-BrdU; Phoenix Flow Systems, San Diego, CA). For Annexin labeling,
cells were treated, pelleted, and exposed to Annexin V/FITC conjugated
antibody. (Chemicon International; Ref. 25
). Trypsin,
which cleaves the antibody binding site, was avoided. Flow cytometry
was performed on a Becton/Dickinson FACStar II (APO-BrdU) or exCaliber
flow system (Annexin) at the Columbia University Cancer Center and
analyzed using Winmidi Software.
Ionic Activity Measurements with Ratiometric Dyes.
For fluorescent ion measurements (free [Na]i
and [Ca]i), cells were plated in 96-well, low
autofluorescence culture plates (Costar). After treatments, wells were
washed twice with PBS and covered for 1 h with 125 µl of dye
loading buffer: (a) for [Ca]i
measurements, 50 µg Fura II/acetoxymethylester (26)
was
added with 30 µl pluronic to 6 ml of clear DMEM; (b) for
[Na]i measurements, Fura II/AM was replaced by
100 µg of SBFI acetoxymethylester (27
, 28)
and loading
continued for 3 h (dyes from Molecular Probes, Eugene, OR). Wells
were then washed twice with PBS (plus 1.0 mmol/l
Ca2+, 0.6 mmol/l Mg2+, and
2 mg/ml dextrose).
For [Na]i, an in situ calibration technique was used whereby cells were permeabilized to Na+ with monensin (20 µM) and gramicidin (40 µM); [Na]o was varied in EC solutions while adjusting [K]o, [Cl]o, and isothionate as described (27 , 28) . For [Ca]i measurements, one row of wells was exposed to ionomycin (10 µM) and one to ionomycin plus EGTA (20 mM) to derive Rmax and Rmin, (Kd was assumed to be 371 nM). The average control concentration for the five experiments was 126 nM. We displayed the percentage of change in control concentration calculated from the ratio measurements (26) . For ion measurements, plates were run on a Fluoroskan II fluorescent plate reader (Titertek). Ratios were derived by dividing fluorescent intensity at 345 nm excitation by that at 390 nm, both with a 508-nm emission filter. For all plates, dye-free wells were examined for both control and drug-treated cells and served as background values. For [Ca]i experiments, exposure to 6 mM MnCl2, which quenches dye, was an added control (26) . Averaged background values were subtracted from each well in the protocol grouping prior to deriving the ratio. Thapsigargin was added as a positive control for [Ca]i elevation and generated rapid elevations in excess of 1 µM.
MRI and Analysis.
All imaging experiments were performed on a high field (4.23 Tesla)
whole body MRI system at the Columbia University Hatch NMR Center. A
small quadrature birdcage radiofrequency coil (50-mm inside diameter;
Morris, Inc.) and a high-strength gradient insert coil (30 mT/m; Bruker
model G-33) were used in this study. Acquisition parameters were:
TR, 100 ms; TE, 5.6 ms; FOV, 40 mm;
slice thickness, 2.5 mm; inversion time, 25 ms; and flip angle, 90°.
Acquisition matrix was 64 x 64 x 8. Both inversion and
excitation pulses were nonselective.
Two reference phantoms of 200 mmol/l NaCl with either 30% (P1) or 40% Ficoll (P2) were glued to a plastic animal platform. The anesthetized mouse was aligned in a deep grove with the tumor positioned through a small elliptical opening in the plastic holder, thus maintaining the same relative position to the phantoms from experiment to experiment. The 40% phantom is the brightest phantom on the IR image, and the 30% phantom is the brightest on the SQ image. IR images and SQ images were normalized to the brightest phantom within each slice.
Analysis of tumor intensity change was measured at the brightest region using line profiles or region-of-interest resident software. This was generally at the center for uniformly bright images or on the largest portion of the annulus. To confirm tumor and kidney positions, paired proton and sodium quadrature birdcage coils, identical in dimension, were used.
Administration of Chemotherapy.
Drug was dissolved in 100–150 µl of vehicle and slowly injected into
a femoral vein under microscopic observation and while the mouse was
anesthetized with ketamine and xylazine. We injected 10 mg/ml Taxotere
(i.e., 
1.0 mg) or 2 mg/ml VP-16 (0.3 mg). The reported
toxic mouse dose for Taxotere is 4.5 mg. The mouse doses equivalent to
human doses were calculated from standard equations that convert weight
to surface area for small mammals. Because of problems with viscosity
for the mouse i.v. injections, the VP-16 dose was about one-third the
calculated equivalent human dose. All animals were used in accordance
with Institutional Animal Care and Use Committee and Columbia
University rules for humane treatment of animals.
Tumor Postmortem.
Formalin-fixed, paraffin-embedded sections of the explanted neoplasms
were cut at 4 µm and stained with H&E or deparaffined and
fluorescently end-labeled (100-µl aliquots of APO-BrdU kit reagents
applied directly to slides in a moist environment). The number of
mitotic figures was enumerated in up to 20 HPFs (x400) in the rim of
viable cancer cells for each specimen. Care was taken not to include
central areas with cellular debris or foci with back-to-back apoptotic
cells. In some specimens, <20 HPFs were available for evaluation. The
mitotic index represents the average number of mitotic figures per HPF
(29
, 30)
|
Results |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
The ability to weight for different populations of sodium nuclei is
easily demonstrated using a cell-free system, e.g.,
phantoms. The phantoms were tubes of NaCl solution. Some contained Agar
or Ficoll to simulate the increased viscosity and electrostatic binding
of sodium, which occurs to some extent in the EC space but more
extensively in the IC space. With appropriate choice of inversion time,
depending on the concentration of Agar or Ficoll, the signal from a
particular test phantom can be totally suppressed. We imaged two
phantoms (with and without 4% Agar) at nine different inversion times,
and in the absence of an inversion pulse, and plotted the intensity for
each image as a separate point. Complete suppression was obtained at
inversion times of 19 and 31 ms, respectively (Fig. 1)
. The straight-line is a fit of the
theoretical relationship between image intensity,
T1 and inversion time (assuming
T1 values of 27 and 43 ms). This
exercise demonstrates how one can totally suppress the signal from a
particular homogeneous population of sodium nuclei based on relaxation
times.
|
, insets). Furthermore, the mouse
kidneys, which appeared very bright on the SQ MR image with no
inversion pulse, were also totally suppressed on the IR image (Fig. 2
, compare C and
D). This inversion pulse, therefore, was quite effective at
nulling EC bulk fluid.
|
with traditional (no inversion pulse) SQ
sodium images (Fig. 2, A and C)
for solid tumors
formed from two human cell lines (DU145, Fig. 2, A and B
; PC3, Fig. 2, C and D
). Thus, 25 ms
is the optimal inversion time for both suppressing EC sodium nuclei as
well as relatively enhancing the image of solid tumors at the magnetic
field strength (4.23 Tesla) used in our experiments. Note that the
tumor is also bright on the traditional SQ image (see
"Discussion").
Intracellular Ion Elevation during Drug-induced Apoptosis in
Culture.
The dissipation of the transmembrane sodium gradient during cell
necrosis has been well studied. Less well studied are changes in ionic
gradients during apoptosis. To examine the relationship between sodium
elevation and the extent of apoptotic cell progression in a controlled
setting, cultured PC3 (and in some cases DU 145) cells were exposed to
the two drugs used in the in vivo experiments,
i.e., Taxotere (which disrupts microtubule assembly), and
VP-16 (a topoisomerase inhibitor). Apoptosis was assessed with flow
cytometry using both fluorescent end-labeling of DNA fragments and
Annexin V binding (Fig. 3, A and B)
. [Na]i elevation was examined
using the fluorescent ratiometric dye, SBFI/AM (Fig. 3
C).
Because Ca2+ is coupled to
Na+ by various membrane exchangers and is linked
to cell death and apoptosis, free concentration
([Ca]i) was also measured using a different
ratiometric dye, Fura II/AM (Fig. 3
D). Because
[Ca]i elevation would be expected to accompany
[Na]i elevation, this additional measurement
served as an independent confirmation of ionic response.
|
A). In PC3
cells, the earliest Annexin V response was 5 h (VP-16), but
it required an additional 24 h (VP-16) or 48 h (Taxotere) to
fully develop at the drug concentrations used (Fig. 3
B). The
ionic elevations (approximated at 10–20 mM for
[Na]i and 150 µM for
[Ca]i) were already significant as early as
2–6 h and peaked at 6–24 h (Fig. 3, C and D)
.
Thus, drug-induced ionic elevation preceded the development of
apoptotic markers, suggesting that the ionic responses are not a
trivial manifestation of the loss of membrane integrity associate with
cell death. The rapid ionic responses further indicated that we should
set the earliest in vivo measurement time to no later than
24 h and motivated future experiments to identify the early
apoptotic mechanisms involved.
Increased Intensity of IR Sodium Image during Drug Effects
in Vivo.
The effect of chemotherapy on the IR sodium image intensity of
in vivo tumors was consistent with the culture results,
i.e., IR sodium image intensity increased after in
vivo drug administration. This is suggestive of increased
[Na]i. Control and posttherapy images were
acquired in 16 mice. Tumors in two additional mice had images
reacquired after control saline injections. Taken as a whole, the
tumors exhibited increased intensity (36% ± 4%) after chemotherapy
(P < 0.001; n = 16), a response
significantly different (P < 0.002) than that after
saline injections (-5% ± 4%; n = 4; not significant
versus control at either 24 or 48 h). Taken separately,
the change for each drug on PC3 cells was significant; Taxotere induced
a 32% increase (P < 0.001; n = 5;
Fig. 4
, top left) and VP-16
induced a 38% change (P < 0.0005; n =
7; Fig. 4
, bottom left). The responses grouped by drug were
not significantly different from each other (P >
0.15). The average signal at 48 h was still elevated compared with
control (28%, P < 0.05; n = 11) for
the PC3 cell tumors. Four experiments using a second human prostate
cancer cell line (DU145; two each with Taxotere and VP-16) also showed
significant elevation at 24 h taken as a group (34%;
P < 0.05; n = 4). An independent
measure of drug effect, reduction in mitotic index, is plotted in the
right panel and is shown to correlate with image enhancement
(see below).
|
In HPFs of the surviving rims of these tumors (Fig. 5)
, one can see that the untreated tumor
(Fig. 5
A) has more mitotic figures than the treated tumor
(Fig. 5
B). Because a reduction in proliferation,
attributable to cell cycle arrest, is a useful indicator of
antineoplastic efficacy (29
, 30)
, the incidence of mitotic
figures was measured for up to 20 HPFs in the surviving rim of each
tumor. Mitotic figures were reduced significantly (P <
0.0001) in the treated (n = 7) tumors [2.24 ±
0.163 (SE) per field; m = 127 fields]
versus untreated (n = 3) tumors (8.57 ± 0.84 per field; m = 47). Importantly, there was,
among all treated tumors, a significant inverse correlation
(P < 0.02) between the reduction in mitotic figures
and the relative enhancement in the IR sodium image (Fig. 4
,
right panel).
|
Pre- and postdrug IR images are shown along with the control SQ image
for a tumor with significant central necrosis. Tumors with significant
central necrosis had a characteristic dark-center/bright-annulus
presentation on the inversion image (Fig. 6, A and B)
. These
tumors were nevertheless uniformly bright on the SQ image (no
inversion; Fig. 6
C). Note that the effect of drug exposure
(enhanced image intensity) is clearly visible comparing the two IR
images but is confined to the enhanced intensity of the annulus. The
dark center on the IR image was unchanged after drug administration.
The histology of the tumor (Fig. 6
D) shows a significant
central area with nonviable cells and cellular debris and a narrow,
peripheral rim of viable cells (between is a region of back-to-back
apoptosis).
|
). A broad band of
densely apoptotic cells is seen below the surviving rim, and the
nonviable center is comparatively smaller than in Q (Fig. 7
F, compare with Fig. 6
D). The border between the
surviving rim, and densely apoptotic cells are shown at high power in
the inset (Fig. 7
F). Furthermore, on the images there is a
clear increase in the diameter of the area of brightness on the IR
image, which occurred between 24 and 48 h after treatment. We
examined this region of recent IR brightening under higher power.
Evidence of a well-defined rim of apoptotic cells just under the
surviving rim was visible in adjacent tissue sections stained either
with H&E (Fig. 7
D) or immunofluorescent end-labeling (Fig. 7
E). We examined the fluorescent staining; the cells just
under the surviving rim were brightest, and the cells closer to the
small nonviable center were less bright. Presumably, the most brightly
stained cells most recently underwent apoptosis, and the less bright
cells had more time to lose their fragmented DNA. Thus, cells with no
nuclear staining on H&E micrographs were present only in the small
center of tumor M (300 µm). In contrast, tumor Q showed a
large center of cells without nuclei. On the basis of the image
broadening and immunofluorescence, a plausible interpretation is that a
portion of the inner surviving rim was converted to a zone of
back-to-back apoptotic cells between 24 and 48 h. Thereafter, but
not yet at the time of tumor removal, cellular debris would have
accumulated in the center, as seen with Q.
|
|
Discussion |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
and 4)
,
various antineoplastics (e.g., colchicine, lonidamine, and
VP-16) have been reported to elevate [Na]i
and/or [Ca]i (37, 38, 39, 40)
.
For Taxotere, the in vivo injection doses we used for mice
were comparable with human clinical doses. Furthermore, human plasma
levels of Taxotere after a single loading dose (41)
remain
above the levels (10 nmol/l) used in our tissue culture studies for
24 h. Thus, we expect that our tissue culture results (elevation of
both [Na]i and [Ca]i
and significant apoptosis) would apply during clinical in
vivo tumor studies. The VP-16 tissue culture levels we used were
comparable with average human plasma levels during the first 4–6 h
after a standard dose (100 mg/m2; Ref.
42
).
On the basis of our in vivo and tissue culture studies and in vitro literature reports, the most straightforward explanation for the increased image intensity after chemotherapy is increased [Na]i. We have, however, considered several alternate explanations, which although less likely, still must be investigated: (a) decreased IC T1 (or change in EC T1); (b) increased EC binding sites; and (c) increased EC [Na]o. Explanation c seems unlikely, because even in excitable tissue, where there is a high density of sodium channels and Na+/K+-ATPase pump sites, there is no evidence of bulk changes in [Na]o beyond the unstirred layers around channel pores. For explanation b, one may speculate whether drug-induced processes could alter production of metalloproteinases, affecting the density of EC sodium binding sites. During hypoxia and Na+/K+-ATPase pump blockade in the heart, multiple quantum sodium MRI studies showed that the contribution of EC sodium to the increase in sodium image intensity is negligible. As to potential explanation a, a reduction in IC T1 (or increase in EC T1) could also contribute to increased image intensity, and one could postulate that large changes in [Na]i by themselves could alter IC T1. However, during the intervention of hypoxia or ouabain poisoning, which can dramatically increase [Na]i, IC T1 values were reported stable (43) . Likewise, IC T2 values are stable during pump blockade in the heart (21) . Thus, although such alternate hypothesis motivate additional studies, it seems unlikely that such results could alter the diagnostic potential of this measurement.
Absolute quantitation, although a long-term goal of this approach, is confounded by various issues, not the least of which is tissue, cellular, and subcellular compartmentalization. For example, diffusion on a molecular level could result in an exchange of sodium between molecular domains with different T1 values during the recovery from the inversion pulse. This would be the case if the diffusion relaxation time constant were comparable with inversion time. As has been shown in submicron vacuoles (44) , this exchange process leads to an averaging of the T1 values and interferes with the ability to distinguish sodium inside and outside the vacuole membrane. However, the size of tumor cells and the limited unidirectional exchange through the polarized cell membranes make substantial exchange between IC and EC compartments during an inversion pulse unlikely (45 , 46) . Furthermore, measured values of T1 in mammalian IC and EC spaces show primarily single exponential relaxation kinetics. The EC T1 value is longer, consistent with our hypothesis. Sodium localized to regions of the interstitial spaces, where there are more numerous binding sites, do have short T1s (47) , but (as above) there is no evidence that they or T1s of the IC contribute a changing component to the IR image (43 , 48) .
The correlation between apoptosis and IR image intensity is apparent in
comparing the histology and images. Tumors with significant fluid
spaces or large centers with cellular debris had dark centers on the IR
weighted images. In this study, the only tumor with no IR signal had a
narrow rim of surviving cells, a large nonviable center with
significant fluid, but no significant interface with back-to-back
apoptosis. (No chemotherapy was administered for this tumor.) Tumors
with bright annuli had sizeable rims of back-to-back apoptosis. Tumors
with bright centers had back-to-back apoptosis extending into the
central region with minimal nonviable tissue. Although we could not do
a time study of the histological parameters, the results are
consistent with the notion that the brightness comes from dense regions
of apoptosis and/or early apoptotic processes in the surviving rim. On
the basis of the spatial resolution of the image (1 mm within each
slice), the enhanced intensity could be from a spatial broadening of
the apoptotic rim, an increased density of apoptotic cells (including
early commitment to apoptosis), or both.
Apoptosis is a dynamic process, which ends with complete fragmentation
and loss of cell nuclei and most of the other cellular macromolecules.
Hence, the absence of a weighted IC sodium MRI signal in late apoptosis
can be explained by the loss of polyanions, because the IC space
equilibrates its chemical composition with the EC fluid. This
prediction is consistent with our finding (Fig. 4)
that the average IR
intensity increase is smaller at 48 than at 24 h. A more detailed
quantitative analysis is required to correlate specific cell
populations with image responses and specific ionic responses with
molecular events induced by the chemotherapy.
Because proton MRI, which uses the hydrogen nucleus, is best suited for structural studies and angiography, we have chosen to develop clinical diagnostic approaches using sodium MRI because of the major involvement of sodium in important dynamic processes in the cell, one of which, illustrated here, is the interaction of chemotherapy, cellular apoptosis, and ions. The ease with which sodium images can be coregistered with proton high resolution images is a further advantage of this technique, as is the fact that it does not require costly or potentially toxic reagents, as do traditional SQ sodium MRI or the new techniques used in gene marker or receptor imaging.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported in part by a grant from
Rhone-Poulenc Pharmaceuticals, Inc. and the Rose M. Badgeley Residuary
Charitable Trust. 
2 To whom requests for reprints should be
addressed, at Department Pharmacology, PH-7-West Room 318, 630 West
168th Street, Columbia University, New York, NY 10032. Phone:
(212) 305-8370; Fax: (212) 305-8780; E-mail: rpk1{at}Columbia.edu
3 The abbreviations used are: MRI, magnetic
resonance imaging; SQ, single quantum; MQ, multiple quantum; EC,
extracellular; IC, intracellular; IR, inversion recovery; VP-16,
etoposide; HPF, high power field.
Received 10/11/99; revised 3/13/00; accepted 3/23/00.
|
REFERENCES |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Petrovsky, E. Schellenberger, L. Josephson, R. Weissleder, and A. Bogdanov Jr. Near-Infrared Fluorescent Imaging of Tumor Apoptosis Cancer Res., April 15, 2003; 63(8): 1936 - 1942. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
; 1
M NaCl plus 4% agarose, •) were examined during SQ
acquisition and for nine different IR times (horizontal
axis; Tinv). SQ is equivalent to
zero inversion time. The lines are the theoretical relationships
assuming T1s of 27.4 and 43.3 ms (inversion
times of 19 and 31 ms). Intensity of IR signal varies, such as
ABS (1 -
2exp(-Tinv/T1)) .
Inset, a cystic tumor (fluid measured by needle
aspiration) imaged with a 25-ms inversion pulse sequence (right
inset) and normal SQ sequence (left inset). The
cystic tumor was completely eliminated by inversion pulse. The mouse
body contour is delineated in red.
) and Taxotere (10
nM; •) is plotted. One-dimensional annexin histograms
(FITC intensity) for live cells are shown for control (bottom
inset) and after 24 h VP-16 (top inset).
C, [Na]i elevation in PC3 cells adherent
to 96-well plates is measured with a fluorescent ratiometric dye,
indicating an elevation of 10–20 mM (from the in
situ calibration technique), which starts as early as 2–6 h
