Anal. Chem. 1994,66, 3000-3004
Determination of Calcium Content in Individual Biological Cells by Inductively Coupled Plasma Atomic Emission Spectrometry Tsutomu Nomlru,. Satoshl Kaneco, Tomokaru Tanaka, Daisuke I to, and Hlroshl Kawaguchi Department of Materials Science and Engineering, Nagoya University, Nagoya 464-0 1 Japan Bert T. Vallee Center for Biochemical and Biophysical Sciences and Medicine, Harvard Medical School, Boston, Massachusetts 02 1 15
A cell analyzer is described which can continuously determine the calcium content in individual biological cells by inductively coupled plasma atomic emission spectrometry in real time. A cell suspension in ethanol (70% in water) is sprayed by a concentric nebulizer, and the dropletscontaining cells are dried ina drying chamber. The dried cells dispersed in air are directly introduced into the plasma to determine the calciumby emission spectrometry. Each flash derived from each cell in the plasma is converted into an electric pulse correspondingto the calcium content in the cell, and the heights of the pulses are measured by a pulse height analyzer. To calibrate this system, monodisperse calcium acetate aerosols are prepared by a vibrating orifice monodisperse aerosol generator. The determined calcium contents in several mammalian cell samples range from 0.057 to 0.27 pg. The detection limit of the present system is -0.01 pg of calcium in a cell.
Along with the development of sensitive analytical methods, it has been increasingly revealed that minor or trace elements (especially metal elements) in a biological cell play important roles in various biological processes including metabolism, cell growth, gene expression, and cancer Elemental composition in a biological cell has been conventionally determined by such sensitive spectroscopic analytical methods as atomic absorption spectrometry (AAS) and inductively coupled plasma atomic emission spectrometry (ICP-AES) after a large number of cells are digested with acids. Nevertheless, a resultant composition is an averaged value in those cells and no information on a distribution among the cells can be provided. An accurate absolute content of a given element in a cell substantially depends on the accuracy of the cell-counting technique. Recently flow cytometry has been developed which can simultaneously measure the cell size and the biological components of individual cells in the flow system by a laser light scattering method and laser fluorometry, * Author
to whom correspondence should be addressed (E-mail address,
[email protected];FAX, 8 1-52-789-3241). (1) Hay, R. W. Eio-inorganic Chemistry; Ellis Horwod Ltd.: Chichester, England, 1984. (2) Fiabane, A. M.; Williams, D.R. The Principles ojbio-inorganic Chemistry; The Chemical Society: London, 1977. (3) Riordan, J. F., Vallee, B. T., Eds. Meta[lobiochemistry; Merhods in Enzymology; Academic Press: San Diego, London, 1988: Vol. 158.
3000
Analytical Chemistry, Vol. 66, No. 19, October 1, 1994
re~pectively.~However, it does not have any method of measuring minor or trace elements in individual cells at present. Intracellular free calcium ions can be individually determined by fluorometry using a sensitivefluorescent reagent for calcium such as Fura-2, Quin-2, and A e q ~ o r i nbut , ~ the method cannot be applied to the determination of total calcium content in a cell. A quantitative imaging method for cultured cells with an ion microscope that can determine local intracellular concentrations of calcium and other elements in individual cells6 cannot be utilized for continuous determination of a large number of cells. In recent years, we have developed an airborne particle analyzer using ICP-AES7-9 or ICP mass spectrometry (ICPMS),'O in which air samples are directly introduced into the ICP and the elemental content in individual particles can be continuously determined by measuring the pulse height distribution of signals from the spectrometer. The analyzer can measure subpico- to femtogram amounts of elements in a particle, depending on the element. In this work, we attempted to apply the airborne particle analyzer using ICP-AES to the individual measurement of calcium content in dried biological cells in air. After a cell suspension was sprayed by a concentric pneumatic nebulizer, the liquid droplets, each of which was made to contain at most a single cell, were dried in a drying chamber. Individual dried cells in an air stream were introduced into the ICP, from which flashes consisting of atomic emission of elements in the individual cells were generated. After processing of signals from the ICP-AES, we obtained distributions of calcium content in individual cells for several mammalian cell samples. (4) (a) Landy, A. L., Auk, K. A., Bauer, K. D., Rabinovitch, P. S.,Eds. Clinicol Flow Cytometry; Ann. N.Y. Acad. Sci. 1993 677. (b) Gilman-Sachs, A. Anal. Chem. 1994, 66, 700A-707A. (5) Grynkiewcz, G.; Poenie, M.; Tsien, R. Y. J . Eiol. Chem. 1985, 260, 34403450. (6) Ausserer, W. A.; Ling, Y.;Chandra S.;Morrison, G. H. Anal. Chem. 1989, 61, 2690-2695.
(7) Kawaguchi, H.; Fukasawa, N.; Mizuike, A. Spectrochim. Acta 1986, 418, 1277-1286. (8) Kawaguchi, H.; Kamakura, K.; Maeda, E.; Mizuike, A. Eunreki Kagaku 1987, 36, 431435.
(9) Nomizu, T.; Nakashima, H.; Hotta, Y.;Tanaka, T.; Kawaguchi, H. Anal.
Sci. 1992, 8 , 527-531. (10) Nomizu, T.: Kaneco, S.;Tanaka, T. ; Kawaguchi, H. Anal. Sci. 1993, 9, 843-846.
0003-2700/94/0366-3000$04.50/0
0 1994 American Chemical Society
Monochromator PM
Preamp.
I
Personal computer
I
ribbon heater "'Am
A-n
air
Figure 1. Schematic diagram of the experimental system.
EXPERIMENTAL SECTION Instrumentsand Procedures. A schematic diagram of the experimental system used in this work is shown in Figure 1. A 1-mL cell suspension was injected into an ethanol (70% in water) stream through a rotary valve. The stream was fed to a concentric pneumatic nebulizer (Meinhard type) placed at the middle part of a drying chamber at a flow rate of 0.2 mL/min by a programmable HPLC pump (Model 576, GL Sciences Inc., Tokyo, Japan). Air (moisture was removed by a cold trap) was used to spray the cell suspension with the nebulizer and to dry the droplets at flow rates of 1.3 and 10 L/min, respectively. The outer wall of the drying chamber (Pyrex glass, 3-cm i.d.) was heated with a ribbon heater at -70 OC. As shown in the figure, two brass sieve plates (3mm holes were bored at pitches of 4 and 8 mm for the upper and lower ones, respectively) were incorporated into the chamber in order to minimize the turbulence of the air flow. To reduce the deposition of electrically charged cells on the wall of the chamber and feed tubes, the charge of the cells was neutralized by the irradiation of w a y emission from 3.7 X 106Bq (100pCi) of 241Am,which was attached to the inner wall of the chamber. The air stream containing dried cells was introduced into the ICP at a flow rate of 50 mL/min by using suction pressure of another concentric pneumatic nebulizer incorporated below the torch. A Seiko SPS-1 IOOH inductively coupled plasma atomic emission spectrometer (Tokyo, Japan) was used, except for the signal processing system. The ICP was operated at an incident radio frequency power of 1.1 kW (27.12 MHz) and at flow rates of 13, 1.1, and 0.87 L/min for the outer, intermediate, and carrier argon gas, respectively. The plasma was observed at a height of 15.1 mm above the load coil by the monochromater with a slit height of 4 mm, and an entrance and exit slit width of 20 and 5 pm, respectively. Signals from a photomultiplier generated by flashes of the Ca I1 393.4-nm line in the plasma were amplified and fed to a laboratorymade multichannel pulse height analyzer. Signal processing after the photomultiplier was the same as that of our previous system.' Cell Samples. Cell samples used were gifts from several researchers. A mouse fibroblast cell line, termed L929-ep0, was maintained in minimum essential medium (MEM; Gibco,
Grand Island, NY),supplemented with 10% fetal bovine serum (FBS) at 37 "Cin humidified atmospherecontaining 5% C02, and subcultured every 3-4 days. The cultured cells were detached from the wall of the culture bottle by trypsinization and suspended in the MEM for storage. A human pancreas cell line, termed SR-4-SH, was pooled and cultured in growth medium consisting of medium 199 (Hazleton, Lenexz, KS) buffered with 25 mM Hepes (Sigma, St. Louis, MO), 20% fetal calf serum (FCS; Biocell, Rancho Dominguez, CA), 90 pg/mL heparin (Sigma), and gentamycin (Gibco). A human umbilical vein endothelial cell line (HUVEC), which was isolated from two to four fresh human umbilical cord veins by collagenase treatment, was pooled and cultured in the same growth medium as that of a human pancreas cell line, except with addition of 20 pg/mL endothelial cell growth supplement (ECCG; Sigma). Before use, the cells were carefully washed by pipeting with a phosphate-buffered saline, PBS(-), which contained KCl, KH2P04, NaCl, and Na2HP04 in concentrations of 0.2, 0.2, 8.0, and 1.15 g/L, respectively. The calcium content of the saline was less than 0.7 pg/mL. After sedimentation of cells by gravity, the supernatant was removed by pipeting. The washing step was repeated at least five times. Finally, the buffered saline was pipeted off and then the cell sample was resuspended in ethanol (70% in water) at a cell number density of 106-107 cells/mL. The calcium content of the ethanol was less than 1 ng/mL. Calibration. In order to calibrate the measuring system, monodisperse aerosols were generated by a vibrating orifice monodisperse aerosol generator (Model 3450, TSI Inc., St. Paul, MN). A 20-pm-diameter orifice wasvibrated at a typical frequency (79.5 kHz). The dispersion and dilution gases (moisture-removed air) were used to dry the aerosols at flow rates of 1.5 and 30 L/min, respectively. To neutralize the electric charge of the particles, another 3.7 X 106 Bq (1OOpCi) of 241Amwas attached to the inner wall of the dispersion chamber. Calcium acetate solutions of 10, 3, and 1 pg of Ca/mL in 2-propanol (50% in water) were fed to the monodisperse aerosol generator at a flow rate of 0.17 mL/ min, which generated particles containing 0.36,O. 11,and 0.036 pg of calcium, respectively. Figure 2 shows pulse height spectra obtained by introducing the monodisperse aerosols into the ICP. The channel number of the pulse height analyzer is proportional to the logarithmic conversion of the emission intensity since a logarithmic amplifier is used to process signals, as shown in Figure 1. The channel number of each peak in Figure 2 was linearly correlated to the calcium content in each monodisperse aerosol particle, as shown in Figure 3. Signals at channels under 50 in Figure 2 were regarded as the background signals derived from the continuum emission of the ICP. RESULTS AND DISCUSSION A concentric pneumatic nebulizer used in ICP-AES normally generates droplets with a diameter ranging from 0.1 to 100 pm.' Since the size of mammalian cells ranges from 5 to 20 pm, the droplets may contain several cells or more if the cell number density of the cell suspension is sufficiently (1 1) Moore, G . L.Inrroducrion io Inductively Coupled PIosma Atomic Emission
Specfromerry; Elsevier: New York, 1989; Chapter 7.
AnalyticalChemistry, Vol. 66, No. 19, October 1, 1994
3001
*!
*.
100 150 200 250 Channel number
0
50
E[
: *
.! ..
w
0
E 2.0
. I v)
I
1
.*:*
a
w
!l E 2.0 3
8
I
L
'u 0
50
100 150 200
Channel number
1 250
Flgure 2. Pulse height spectra for calcium emission obtalned wlth monodisperse calcium acetate aerosols containing (a) 0.36,(b) 0.1 1, and (c) 0.036 pg of calcium.
0.001b
20
100 150 200 Channel number
25b
Flgure 3. Relation between the calcium content in a partlcle and the
channel number.
high. By diluting the cell number density in the suspension, however, most of the droplets can be controlled to contain a single or no cell. Droplets not containing cells did not generate emission at the wavelength of calcium (393.4 nm) as the calcium concentration in the medium was sufficiently low. When the spectrometer was set to a wavelength just away 3002
Analytlcal Chemistry, Vol. 66, No. 19, October 1, 1994
from the Ca I1 393.4-nm line, no signal arose and no background change occurred as the cell particle passed through the observation zone. The intensity of emission due to molecular bands by cell constituents was too low to enhance the background signals due to the continuum emission. In order to determine the true content of calcium in a cell, it would be better to introduce live cells suspended in a buffered saline into thedrying chamber. The cell samples in the present work, however, were fixed with ethanol (70%in water). The 70%ethanol does not lyse the membrane although it punctures some pinholes on the cell membrane. Because the cell fixed with it can sufficiently preserve the cell structure and components, it has been used as the most common fixing agent in flow cytometry. Even some leaching of the cell electrolyte does not affect the determination of the calcium content in the cell since most of the calcium is stored within organelles in the cell.12 The 70% ethanol was also convenient since it vaporized easily and generated no salt particle other than cells in the drying chamber. Without the drying process, droplets containing cells can easily deposit on the inner wall of feeding tubes and glassware. When the cell suspension was fed into a pneumatic nebulizer with a spray chamber that has been conventionally used for ICP-AES measurement of solution, all of the cells were removed in the chamber so that no signal for the cells was detected. Measurements of various cell samples including lymph, tissue, and cultured cells were tried. The lymph cells did not have enough calcium to be detected with the present system. Although rat or cat tissue cells could be easily obtained by mincing organ tissue such as liver, lung, and pancreas followed by filtering it with a 200-mesh nylon screen, they were accompanied by a large number of broken fragments which caused enhancement of background signals of ICP-AES. Consequently, cultured cells were used to confirm the feasibility of the present system because the cell size was relatively large and the amounts of broken fragments were extremely small (although careful washing was needed before use). The calibration using monodisperse aerosols is based on the assumption that a pulse signal derived from a cell has almost the same pulse height as that from an aerosol particle containing the same amounts of calcium. Traces a and b of Figure 4 show the signal profiles from the preamplifier produced with aerosol particles and biological cells, respectively. Since both the pulse signals have almost the same pulse width (1 ms), the calibration was considered to be feasible. More accurate calibration would be possible if standard cell samples were available. Figure 5 shows a micrograph of cultured mouse fibroblast cells, and the pulse height spectra obtained with them at different cell number densities are shown in Figure 6 . The scale of calcium content shown at the top of each graph in Figure 6 was obtained from the calibration curve shown in Figure 3. Signals seen at channels under 50 in Figure 6 are the same background signals as those in Figure 2. The channel number in Figure 6 corresponds to the calcium content in a single droplet generated by the nebulizer. If the cell number density of the cell suspension is so high that each droplet (12) Albertr,B.;Bray,D.;Lewis, J.;Raff,M.;RoberIs,K.; Watson, J.D. Molecular Biology ofrhe Cell, 3rd ed.; Garland Publishing: New York, 1994; p 508.
10 ms 0.4 V
I
Channel number
Figure 4. Output signals from the preamplifier produced by (a) monodisperse calcium acetate aerosols (calcium content 0.042 pg) and (b) mouse fibroblast cells. V-
0
50
100
200
150
,
250
Channel number ru,
1*5,, , , , y 1 , , , , ,,Oil , , , , , ,,;
I
I
\
I
Flgure 6. Pulse height spectra for calcium emission obtained with cultured mouse fibroblast cells at approximate cell number densities of (a) 8 X los, (b) 4 X los, and (c) 1 X los. Table 1. Determlned Calclum Content In Cultured Mammallan Cells
cell sample Flgure 5. Micrograph of cultured mouse fibroblast cells.
contains several cells, the cell number in a droplet can be reduced, and consequently, the peak of the pulse height distribution will be shifted to the lower channel with dilution of the cell suspension. When the cell number density is adjusted so that each droplet contains a single or no cell, further dilution of the cell suspension will lessen the number ratio of the droplet containing a single cell to that containing no cell. In this case, the height of the pulse signal should not change, but the count rate of the pulses having the same pulse height should decrease with dilution of the cell suspension. As shown in Figure 6, the peak of the pulse height distribution did not shift but the count rate of the peak was reduced with dilution of the cell suspension. Thus, it was concluded that each pulse signal should be derived from a single dried cell generated from a droplet containing a single cell. The line in the figure was obtained by smoothing the data based on the simple method of moving averages ( N = 13). The mean value of calcium content in individual cells was estimated to be 0.057 pg at the peak of pulse height distributions. Since the size of the cells ranged from 10 to 15 pm, the calcium concentration in the cell can be calculated to be in a range from 0.82 to 2.8 mM, assuming a spherical
mouse fibroblast cells human pancreas cells human endothelium cells
measured content (Pg/ cell) diameter mean std calcd cellular (pm) value dev concnO(mM) 10-15 15-20 15-20
0.057 0.16 0.27
0.029
0.04 0.04
0.82-2.8 0.94-2.2 1.6-3.7
Calculated from the cell diameter and the mean value of measured content for each cell sample by assuming a spherical cell.
cell. The calculated value is close to the intracellular total calcium concentration, 1-2 mM, in a typical mammalian cell reported in the literature (although the concentration of its intracellular free calcium ion is reported to be lV mM).I2 The result was further confirmed by the conventional determination. After the cell number density was measured with a Thoma glass counting chamber (0.1 mm deep, Erma Inc., Tokyo, Japan) under a microscope, the calcium concentration of a digested cell solution was determined by ICPAES with the result of 0.016 pg of Ca/cell. Although this result does not seem consistent with the above value, 0.057 pg of Ca/cell, the difference can be tolerated because the cell-counting technique used above generally gives an inaccurate value in the order of magnitude. Table 1 lists the mean values and the standard deviations of calcium content in individual cells as well as the calculated intracellular calcium concentration for three cultured mammalian cell samples. All of the calculated concentrationsareclose to the above literature AnalyticalChemistry, Vol. 66, No. 79, October 7, 7994
3003
values. The detection limit of calcium in the present system is practically -0.01 pg, considering the background signals at channels under 50 in Figure 6 . The comparison of the detection limits of the analyte in a particle by this system and the conventional ICP-AES measurement of solution was previously discussed.' CONCLUSION The present system using ICP-AES can continuously measure calcium content in individual mammalian cells in real time. In principle, it can not only measure thedistribution of calcium content of a large number of cells but can also show distributions of different groups of cells in a cell mixture without separation if there is a large difference of calcium content between thegroups (e.g., normal and carcinoma cells). However, the introduction efficiency of dried cells into the ICP is less than 0.1% and must be improved. The pneumatic nebulizer generates so large a number of droplets that a large volume of air, 10 L/min, was demanded to dry the droplets completely, while the air introduction into the ICP was limited to less than 50 mL/min. A vibrating nozzle4 adopted in flow cytometry may be preferable because the number of droplets generated by the system is much less than that of the pneumatic nebulizer and the droplet diameter can be controlled at a constant value. Extensive applications of the present system to other elements were limited by the insufficient sensitivity. Even
3004
Analytical Chemistry, Vol. 66, No. 19, October 1, 1994
sodium and potassium contents in a mammalian cell were not enough to be detected by the system. Although magnesium contents in a mammalian cell were expected to be determined, the measurement has not yet succeeded. If the cell size is as large as 50 pm, like planktons, various elements will be measured by the present system with a skillful drying technique. Utilization of ICPMS as a detection method may overcome the sensitivity problem for several elements although our airborne particle analyzer with ICPMS should be further improved. lo ACKNOWLEDGMENT
The authors acknowledge Prof. Izumi Nakashima and Mr. Toshio Hayakawa of the School of Medicine, Dr. Shigeo Yoshino of the School of Science, Prof. Hiroyuki Honda and Mr. Koji Tani of the School of Engineering of Nagoya University, and Prof. Sunao Yamazaki of the University of Tokyo for the biological cell samples. Support was provided to S.K. as a research fellow of the Japan Society for the Promotion of Science.
Received for review M a y 24, 1994. Accepted July 14, lQ94.' e Abstract published
in Advance ACS Abstracts, August IS, 1994.
