Download The Direct Mapping of the Uptake of Platinum

[CANCER RESEARCH 63, 1776 –1779, April 15, 2003]
Advances in Brief
The Direct Mapping of the Uptake of Platinum Anticancer Agents in Individual
Human Ovarian Adenocarcinoma Cells Using a Hard X-ray Microprobe1
Petr Ilinski, Barry Lai, Zhonghou Cai, Wenbing Yun, Daniel Legnini, Teresa Talarico, Marian Cholewa,
Lorraine K. Webster, Glen B. Deacon, Silvina Rainone, Don R. Phillips, and Anton P. J. Stampfl2
Experimental Facilities Division, Argonne National Laboratory, Argonne, Illinois 60439 [P. I., B. L., Z. C., W. Y., D. L., A. P. J. S.]; Department of Biochemistry, La Trobe
University, VIC 3083, Australia [T. T., D. R. P.]; Pharmacology and Developmental Therapeutics Unit, Peter MacCallum Cancer Institute, VIC 3002, Australia [T. T., L. K. W.,
S. R.]; School of Physics, University of Melbourne, VIC 3052, Australia [M. C.]; Department of Chemistry, Monash University, VIC 3168, Australia [G. B. D.]; and Physics
Division, Australian Nuclear Science and Technology Organisation, NSW 2234, Australia [A. P. J. S.]
Abstract
Uptake of platinum-based anticancer compounds into individual human ovarian andenocarcinoma cells was measured using an X-ray microprobe. The uptake of cisplatin, a platinum-based compound, in drugresistant cells is decreased by ⬃50% after 24 h, compared with the uptake
of the drug in nonresistant cells over the same time period. The Pt103
derivative of the drug, in contrast, showed an increased uptake by an
order of magnitude in resistant cells over the same time period. Increased
uptake appears to allow Pt103 to overcome the resistance mechanism
developed by the cell. This work additionally shows that the X-ray microprobe is able to directly quantify Pt drug uptake on a subcellular level and
can measure the mass of Pt down to a detectable limit of 20 attograms of
Pt (2 ⴛ 10ⴚ17 grams or 6 ⴛ 104 Pt atoms) in 1 s. Such exquisite elemental
sensitivity combined with high spatial resolution paves the way for quantitative submicron three-dimensional mapping of elemental distributions
within individual cells.
technical advances have been difficult to achieve because only trace
amounts of material are distributed within each cell, and no appropriate optical fluorescence marker appears to exist for small systems,
such as cisplatin or its derivatives. State-of-the-art hard XMPs that use
the brilliant X-rays produced by third generation synchrotron sources
are able to directly map subcellular distribution and uptake of mid and
high Z-materials with better than ppm3 sensitivity, as demonstrated in
a recent study of chromium uptake in individual V79 Chinese hamster
lung cells (4). This capability offers the unique possibility of yielding
the cellular uptake and intracellular distribution of potential anticancer
agents, which will provide new understanding, leading to more effective drug design. This study is to determine whether the XMP is
capable of quantifying drug uptake within individual adenocarcinoma
cells.
Materials and Methods
Introduction
Cisplatin is one of the few chemotherapeutic, platinum-based compounds in clinical use for the treatment of cancer and used primarily
for the treatment of testicular and ovarian tumors (1). Because tumors
slowly develop resistance to cisplatin (and to all anticancer drugs),
there is a need to find derivatives that continue to be active against
tumors once cisplatin resistance has developed. Derivative Pt103, or
Pt[N(p-HC6F4)CH2]2py2, was selected from an extensive series of
cisplatin analogues (2) on the basis of its good activity against
cisplatin-resistant tumors (3). Cellular studies of Pt103 were commenced several years ago in an attempt to define the chemical/
biochemical properties that led to this activity against cisplatin-resistant tumors.
A critical aspect of this work was to establish if the cellular uptake
and intracellular distribution of Pt103 were significantly different
from that of the parent drug cisplatin. Although cellular uptake into
tumor cells can be assessed by techniques, such as atomic absorption
spectroscopy, and the distribution between nuclear and mitochondrial
DNA can be determined using gene-specific DNA cross-linking assays or by quantitative PCRs, a major technical advance would be
made if both of these aspects of drug uptake could be determined
without the need for subcellular fractionation of the tumor cell. Such
Received 11/6/02; accepted 3/5/03.
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 the Australian Synchrotron Research Program, funded by the Commonwealth of Australia under the Major National Research Facilities Program, the U. S.
Department of Energy, under Contract W-31-109-Eng-38, and a grant from the Clive and
Vera Ramaciotti Foundation (T. T., M. C., and D. R. P.).
2
To whom requests for reprints should be addressed, at Argonne National Laboratory,
9700 South Cass Avenue, Building 401-B3210, Argonne, IL 60439. Phone: 1-630-2526275; Fax: 1-630-252-0161; E-mail: [email protected].
Human ovarian 2008 cells and cisplatin-resistant 2008/CDDP cells were
plated out into 250-ml tissue culture flasks and left to attach overnight. All
flasks were ⬃50 – 60% confluent before the addition of 10 ␮M cisplatin or
Pt103. Control cells were treated with an equal volume of saline (cisplatin
control) or ethanol (Pt103 control). The cells were lifted off the flasks using
Pronase and harvested at various times ranging from 0 to 24 h. Some of the
cultured cells were taken to perform a cell count using a hemocytometer and
determine viability using tryphan blue staining. Once harvested, the cells were
resuspended to a concentration of 1 ⫻ 106 cells/ml in 154 mM ammonium
acetate (pH 6.5); 2 ␮l of the suspension were applied to a nylon foil and
subsequently snap frozen in isopentane to near freezing point using liquid
nitrogen. Samples were then placed in a vacuum chamber and freeze dried at
10⫺6 Torr for ⬃12 h, then stored in a desiccator before analysis.
The experiment was performed using a hard X-ray scanning microprobe at
the 2-ID-D beamline at the APS (5). An APS undulator A generates the X-ray
beam, and a Kohzu double-crystal Si(111) monochromator produces monochromatic X-rays. A Fresnel zone plate focused the X-ray beam on the sample
to a spot size of 1 (h) ⫻ 0.2 (V) ␮m2. The flux passing through this point was
measured to be 1 ⫻ 1010 photons/s, using an ion chamber in place of the
sample (Fig. 1). The zone plate was located 71 meters from the undulator
source, whereas the sample was placed 146.5 mm from the zone plate. To
perform fluorescence microspectrometery, a specimen was mounted on a X-Y
translation stage and placed at right angles to the incident beam. An energydispersive Ge detector was used to monitor fluorescence from the sample that
was scanned across the focused X-ray spot. In this way, spatial maps of many
different elements were acquired simultaneously. An incident X-ray energy of
11.7 keV, just above the 2p3/2 absorption edge of Pt, was used to produce the
Pt L␣ emission lines. At this energy, X-rays penetrate biological cells without
significant absorption nor beam spreading; thus, no specimen thinning is
required, and fluorescence maps of individual elements represent a twodimensional projection of the volumetric distribution within the specimen.
3
The abbreviations used are: XMP, X-ray microprobe; ppm, parts per million; ROI,
region of interest; APS, Advanced Photon Source; SRM, standard reference material.
1776
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 2003 American Association for Cancer
Research.
DIRECT MAPPING OF PLATINUM UPTAKE IN INDIVIDUAL CELLS
Fig. 1. Schematic of the experimental setup.
X-rays were generated by the APS undulator
source, monochromatized by a pair of Si(111) crystals, and focused by a Fresnel zone plate onto the
sample. An order-sorting aperture was placed near
the sample to reduce background radiation. An
energy-dispersive Ge detector measured emission
from the sample.
Results
Elemental concentration within each cancer cell was obtained by
comparison of the characteristic X-ray emission intensities to those
from a suitable standard with uniform and known elemental concentration. Ten energy ranges spanning the characteristic X-ray lines of
elements corresponding to P, S, Cl, K, Ca, Mn, Fe, Cu, Zn, and Pt
were integrated in real time by software and stored as 10 ROI. The
collection time at each pixel varied from 2 to 30 s depending on the
total count rate of each detected signal. Part of the emission spectrum
from a cancer cell is shown in Fig. 2. As can be seen, the characteristic
X-ray line of interest, Pt L␣, overlaps with the Zn K␤ line, so that some
counts accumulated into the Pt L␣ ROI were contaminated by Zn K␤
emission. A corrected Pt count may be found by subtracting the
number of counts generated by Zn K␤ from the number of counts in
the Pt L␣ ROI. The number of Zn K␤ counts can be calculated from
the intensity of the Zn K␣ line by knowing the ratio between the K␣
and K␤ line intensities (6). The Zn K␣ ROI had, however, a contribution from the Cu K␤ line. The number of Cu K␤, counts is calculated
by using the Cu K␣, ROI, which does not overlap any other emission
line. Finally, an expression to calculate the corrected number of
counts in the Pt L␣ line, NPt, can be defined as:
NPt ⫽ NPt ROI ⫺ 0.176 ⫻ NZn K␣ ROI ⫹ 0.0037 ⫻ NCu K␣ ROI,
specimen and elemental concentration within a thin sample is defined
as (8):
I ⫽ I 0G ␶
r⫺1
r
␻ fW ␳ l,
(B)
where I0 is primary beam intensity, G is a geometrical factor, ␶ is the
photoelectric mass absorption coefficient, r is the absorption jump
ratio of the corresponding atomic shell, ␻ is fluorescence yield, f is the
fraction of the characteristic X-ray line within the ROI to the entire
emission series, W is the weight fraction of the element in the
specimen, ␳ is sample density, and l is sample thickness. Factor W␳l
in Equation B is the surface density of the element in a specimen,
which is defined below as D; therefore, when geometrical factors are
the same, the ratio of zinc and platinum emission can be written as:
(A)
where NPt ROI is the total number of counts in the Pt L␣ ROI,
NZn K␣ ROI is the number of counts in the Zn K␣ ROI, and NCu K␣ ROI
is the number of counts in the Cu K␣ ROI. Thus, if NPt ROI is ⬎20%
of NZn K␣ ROI and ⬎0.5% of NCu K␣ ROI, and with sufficient counting
statistics, a Pt signal can be reliably determined.
To calibrate the platinum concentration, the emission spectrum of
SRM 1833 for X-ray fluorescence spectroscopy supplied by the
National Bureau of Standards (7) was measured in the same geometrical configuration as for the cells. SRM 1833 does not contain
platinum; the platinum concentration in the specimen was defined by
comparison with the concentration of zinc in the standard. The relation between the intensity of characteristic X-ray emission from the
Fig. 2. A typical emission spectrum from a cancer cell.
1777
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 2003 American Association for Cancer
Research.
DIRECT MAPPING OF PLATINUM UPTAKE IN INDIVIDUAL CELLS
I Zn
IPt
冉
冉
I0 ␶
⫽
I0 ␶
r⫺1
r
冊
冊
Zn
r⫺1
r
␻f
␻f
DZn
.
DPt
(C)
Pt
The emission signal of the zinc K␣ line was measured to be 84,100
counts/sec, whereas the zinc surface density of SRM 1833 was 3.9
␮g/cm2. For the same primary beam intensity and platinum surface
density of 1 fg/␮m2, the platinum L␣ line signal, according to Equation C, will be equal to 116 counts/s. With typical backgrounds of
⬍10 counts/sec in the Pt L␣ ROI, the XMP provides extremely high
elemental sensitivity for platinum that can be detected down to a
surface density (minimum detectable concentration) as low as 0.1
fg/␮m2 (or 10⫺8 gram/cm2) in 1 s. For a normal cell thickness of
10 –20 ␮m, a surface density of 0.1 fg/␮m2 corresponds to a local
platinum concentration of 5–10 ppm.
To obtain a distribution of the elements in a cell, samples were
scanned within the plane using a step size of 1 ␮m in both the X and
Y directions. One of the resulting platinum distributions is shown in
Fig. 3. The total amount of platinum in the cell, given in units of fg,
can be obtained by summation of the amount of platinum defined in
each pixel:
M⫽
冘
j
Mj ⫽
冘
j
Ij
116
A j,
Fig. 4. Uptake of 10 ␮M cisplatin and Pt103 in human ovarian 2008 and cisplatinresistant 2008/CDDP ovarian cells.
(D)
and 53%. Most variation thus occurs between cells rather from the
uncertainty of individual measurements.
where Mj is the amount of platinum in a pixel j, Ij is the platinum
signal at this pixel, and Aj is the pixel area in units of ␮m2. The pixel
area was 1 ␮m2, which is defined by the step size of the scan. The
background was subtracted from the platinum distribution before
summation. The level of the background was defined by averaging the
pixel intensities along the cell boundary in Fig. 3. The amount of Pt
was calculated for each cell and then averaged for a set of cells. Each
set contained from three to four cells. Sources of error for the
individual estimates of Pt mainly occurred from measurement errors
of the Pt and Zn emission peaks and from the uncertainty in the Zn
surface concentration of the SRM 1833 standard. The SD for individual measurements of Pt varied between 1.5 and 12%. The SD of
platinum within batches of cells was also estimated to be between 3
Fig. 3. Gray-scale map of the distribution of platinum in human ovarian 2008 cells
treated with 10 ␮M Pt103 drug for 8 h.
Discussion
As shown in Fig. 4, drug uptake in cancer cells that are resistant to
cisplatin is reduced by ⬃50%, compared with uptake in normal cancer
cells over a 24-h period. On the other hand, resistant cells show an
increased uptake of the Pt103 derivative by up to an order of magnitude. The relative extent of uptake of Pt103 compared with cisplatin
into single ovarian cells parallels results reported previously using
atomic absorption spectroscopy on several million cells (3). The
enhanced uptake of Pt103 may allow the drug to overcome the
resistance mechanism induced by long-term exposure to cisplatin.
Interestingly, Pt103 uptake was found to be largest in resistant cells
over a 24-h period (Fig. 4), contrary to expectations and previous bulk
measurements (3); nonresistant cells would be expected to have the
highest uptake. Such a result highlights the difference between measurements performed on millions of cells as opposed to those performed on individual cells, as it appears from this work that measurable variation in uptake can indeed occur within cell batches.
This experiment successfully demonstrates the feasibility of using
the XMP for single cell analysis, even for clinical doses of anticancer
agents, which has not been possible previously with other microprobe
analytical techniques (9, 10). Using the results from this study, the
XMP technique provides a minimum detectable limit of 20 attograms
(2 ⫻ 10⫺17 grams or 6 ⫻ 104 atoms) for platinum in a beam spot of
1 (h) ⫻ 0.2 (V) ␮m2 obtained in 1 s. In this study, a step size of 1 ␮m
was chosen to map out the Pt distribution in the whole cell within a
reasonable data collection time. The XMP, however, is capable of
achieving a spatial resolution of ⬃150 nm (5) and may therefore be
used for subcellular imaging in future studies when cell sections are
used. For quantitative analysis, it is appropriate to use a minimum
analyzable limit, which is defined as a concentration that can be
measured with a SD of 10%, and that is at least six times higher than
the minimum detectable limit (8). A minimum analyzable limit that
1778
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 2003 American Association for Cancer
Research.
DIRECT MAPPING OF PLATINUM UPTAKE IN INDIVIDUAL CELLS
can be achieved by XMP is ⬃100 –1000 times lower than that currently measurable using state-of-the-art proton microprobes. Such
sensitivity, on the level of several ppm, enables quantification of
platinum and its spatial distribution directly in single cells. Sensitivity
is even higher for transition metals, such as Cr, Fe, Ni, Cu, and Zn, by
at least another order of magnitude. With additional improvements to
optics and X-ray source, the microprobe will enable accurate spatial
mapping of platinum and colocalization of other elements distributed
throughout a single tumor cell, e.g., by the reconstruction of XMP
maps from a series of cell sections. In the near future, the XMP will
be able to yield three-dimensional images of the distribution of
platinum-derived anticancer drugs throughout a single tumor cell.
Such spatial and elemental resolution will allow for direct quantification of drug distribution in the nuclear and mitochondrial compartments. And such subcellular information may 1 day allow the detailed
mechanisms of drug uptake to be studied directly.
References
1. Wong, E., and Giandomenico, C. M. Chem. Rev., 99: 2451–2466, 1999.
2. Webster, L. K., Deacon, G. B., Buxton, D. P., Hillcoat, B. L., James, A. M., Roos,
I. A. G., Thomson, R. J., Wakelin, L. P. G., and Williams T. L. J. Med. Chem., 35:
3349 –3353, 1992.
3. Talarico, T., Phillips, D. R., Deacon, G. B., Rainone, S, and Webster, L. K. Investig.
New Drugs, 17: 1–15, 1999.
4. Dillon, C. T., Lay, P. A., Kennedy, B. J., Stampfl, A. P. J., Cai, Z., Ilinski, P.,
Rodrigues, W., Legnini, D. G., Lai, B., and Maser, J. J. Biol. Inorg. Chem., 7:
640 – 645, 2002.
5. Cai, Z., Lai, B., Yun, W., Ilinski, P., Legnini, D., Maser, J., and Rodrigues, W. Am.
Inst. of Physics Conf. Proc., 507: 472– 477, 2000.
6. Salem, S. I., Panossian, S. L., and Krause, R. A. Atomic Data Nucl. Data Tables, 14:
91–109, 1974.
7. Pella, P. A., Newbury, D. E., Steel, E. B., and Blackburn, D. H. Anal. Chem., 58:
1133–1137, 1986.
8. Jenkins, R., Gould, R. W., and Gedcke, D. (eds.). Quantitative X-Ray Spectrometry.
New York: Marcel Dekker, 1995.
9. Moretto, P., Ortega, R., Llabador, Y., Smirnoff, M., and Bernard, J. Nucl. Instr. Meth.
Phys. Res., 104: 292–298, 1995.
10. Kirk, R. G., Gates, M. E., Chang, C-S., and Lee, P. Exp. Mol. Pathol., 63: 33– 40, 1995.
1779
Downloaded from cancerres.aacrjournals.org on June 14, 2017. © 2003 American Association for Cancer
Research.
The Direct Mapping of the Uptake of Platinum Anticancer
Agents in Individual Human Ovarian Adenocarcinoma Cells
Using a Hard X-ray Microprobe
Petr Ilinski, Barry Lai, Zhonghou Cai, et al.
Cancer Res 2003;63:1776-1779.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/63/8/1776
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 14, 2017. © 2003 American Association for Cancer
Research.