- Email: [email protected]
Nuclear microscopy of single whole cultured cells: preparation and analysis of human Chang liver cells
.-EB
c2sJ !kj
ELSEVIER
Beam Interactions with Materials 8 Atoms Nuclear Instruments
and Methods
in Physics Research B 130 (1997) 35 l-357
Nuclear microscopy of single whole cultured cells: preparation and analysis of human Chang liver cells P.S.P. Thong ‘,*, F. Watt ‘, R. Paramanantham b, B.H. Bay b, K.H. Sit b a Nuclear Microscopy L,aborutory. Department of Physics. National University of Singapore, Lower Kent Ridge Road, Singapore 119260, Singapore b Department
qf Anatomy, Faculty ctfMedicine, National Uniuersity of Singapore. Lower Kent Ridge Road, Singpapore I1 9260, Singapore
Abstract Nuclear microscopy is a powerful tool for the measurement of elemental concentrations in single cells. Six methods involving the use of various fixing agents, rinsing agents and drying methods were tried in the preparation of cultured human Chang liver cells for nuclear microscopy and the suitability of each method was evaluated by monitoring the K/Na ratios and shapes of individual cells. The K/Na ratio is a commonly used criteria for the ionic integrity of cells; K/Na ratios well above 1 indicates minimal perturbation of the intracellular ionic composition. Non-stimulated human Chang liver ceils in a resting state are usually polygonal in shape and flattened in firm anchorage to the substrate, while dividing or stimulated cells appear rounded. Therefore the shapes of the cells can be used as an indicator of whether the cells are in a resting or stimulated state. It is not desirable for cells to be in a stimulated state since then the effects of other external stimuli cannot be observed independently. Of the six methods tested, chemical fixation, as expected, was considered non-ideal for the preparation of human cultured Chang liver cells. Ice-cold 1.50 mM sucrose was found to be the most suitable rinsing solution for the preparation of cultured human Chang liver cells. Both freeze-drying and air-drying were used as drying methods and cells processed by either method were found to have K/Na ratios well above 1. Hence both drying methods were found to be suitable although membrane blotting followed by air-drying was preferred as excess rinsing solution can be very quickly removed during the blotting process. The K/Na ratios of cells on the same target holder but from different regions were found to be dependent on the local cell density. Cells which are locally dense-packed were found to have a much higher K/Na ratio than cells in a less dense region.
1. Introduction Nuclear
Microscopy
(NM)
is a collective
term
to refer to a group of microanalytical techniques based on the use of a focused beam of MeV particles. Increasingly, NM is being recognised as a powerful tool for the analysis of both tissue sections used
* Corresponding [email protected]
author.
Fax:
+ 65-777-6126;
email:
and single whole cells in biomedical research. The strengths of the NM include a lower bremsstrahlung background cont~bution to the X-ray energy spectrum, a greater range of the projectile beam in the target and a much lesser degree of broadening as the beam traverses the sample than when compared to the electron microscope. The use of cell cultures in biomedical research has gained popularity in the past decade. Its favourabiIity lies in the fact that firstly, experimental conditions can be controlled and secondly, the effects of
0~68-583X/97/$17.~ 0 1997 Elsevier Science B.V. Ail rights reserved. PII SOi68-583X(97)00225-5
VIII. BIOLOGY
P.S. P. Thong et at. / Nucl Instr. and Meth in Phy.
352
external factors, toxic agents and stimuli can be studied even in single cells [l]. The crucial step in the meaningful analysis of biomedical materials lies in preserving the ionic integrity of the samples. In the past two decades, electron microscopists have accumulated a wealth of information on the preparation of cells for X-ray microanalysis. Warley [21 pointed out that whereas fixation methods were popular before the early SO’s, the trend is toward cryopreparation in the 90’s since cryotechniques often yield the best results with regards to preserving the in situ ionic concentrations. This criterion is especially important where analysis of diffusible ions is involved. In choosing a workable preparation procedure for the nuclear microscopic analysis of cells, the same considerations are applicable as for electron microscope X-ray microanalysis, viz. the minimal disturbance of the ionic concentrations arising from preparation. Hence the nuclear microscopist can adopt procedures for microanalysis generally recommended by the electron microscopist. Nevertheless, as no two cell types behave in the exactly the same way, no single preparatory method can be hailed as the ideal method for all cell types. In the present work, we experimented with various processes for the preparation of cultured human Chang liver cells for single whole cell analysis using the NUS Nuclear Microscope [3], A total of six preparation methods, involving the use of different fixatives, washing mediums and drying methods were tried. A summary is outlined in Table 1. The suitability of each method
Table 1 Methods used for the preparation
(3)
Methanol in graded concentrations 4% fo~aldehyde in phosphate buffered saline 5% glutaraldehyde in 0.1 M phosphate buffer
(4) (5) (6)
1
was evaluated by looking at the K/Na ratios and shapes of individual cells to find the most suitable method for processing the cells. NM applications using cultured human Chang liver cells have included studies in apoptosis [4,5].
I.1. K/Na
ratio
The intracellular K/Na ratio is a commonly used indicator of the ionic integrity of the cell [6]. K/Na ratios well above 1 indicates minimal perturbation of the intracellular ionic composition while K/Na ratios below 1 is indicative of at least partial damage to the cell. Here we show that the K/Na ratio of individual cells is both dependent on the cell preparation method used as we11 as on the local cell density.
1.2. Celt shapes Non-stimulated Chang liver cells in a resting state are usually polygonal and flat. Cell rounding occurs either during mitosis or in the presence of an external stimulus [7]. Thus the shapes of the cells can be used as an indicator of whether the cells are in a resting or active state. It is not desirable for cells to be in a stimulated state since then the effects of other external stimuli cannot be observed independently. We show in this paper that cells exhibit rounding up due to perturbation by certain process chemicals.
of cultured human Chang liver cells for nuclear microscopy
Fixing agent (1) (2)
Rex 8 130 I19971 351-357
Rinsing agent
reagent type I
Drying method (no further drying required) freeze-drying
I
freeze-drying
I50 mM ammonium acetate (pH 7.0) ice-cold 150 mM sucrose ice-cold 150 mM sucrose
freeze-drying
reagent type
freeze-drying membrane blotting followed by air-drying
P.S.P. Thong et cd./ Nucl. fnsrr. und Meth. in Phys. Res. 6 I30 (1997)
2. Materials
and methods
351-357
353
solution adjusted to pH 7.0 (method 4 in Table 11, or (b) ice-coid 150 mM sucrose solution (methods 5,6).
2.1. Monolayer culture 2.4. Dr?,ing methods American Type Culture Collection human Chang liver cells (ATCC CCL 13) were cultured as a monolayer on six Pioloform (Agar Scientific)-coated target holders as described by Sit et al. [8]. (a) 20 ml of a cell suspension in pure fetal bovine serum (Cytosystems) was dropped onto the Pioloform coated target holders in a 70 ml sterile container. A cell density of 3 X lo3 cells on the 5 mm diameter hole of the target holder was found to be suitable for single cell analysis using the nuclear microscope, (b) The container was capped and incubated for one hour at 37°C. (c> At the end of the one hour incubation, the serum-pulsed cells were gently flooded with 10 ml of DMEM (Dulbecco’s Modified Eagle’s Medium, Sigma) supplemented with 10% fetal bovine serum. (d) The cufture was then placed in a 37°C incubator and was ready for processing after 24 h.
Samples fixed in methanol (method 11 required no further drying since methanol both fixes and dehydrates. Samples fixed in fo~aidehyde and glutaraldehye (methods 2,3) as well as those rinsed in ammonium acetate and sucrose (methods 45) were freeze-dried as follows: (i) After rinsing, the edges of the target holders were touched with a piece of filter paper. (ii> The samples were placed in a small volume of liquid nitrogen (- 170°C) in a Dewar flask and freeze-dried in a freezer at - 20°C. The sixth remaining sucrose-rinsed sample (method 6) was blotted with a piece of low protein binding polysulfone membrane (BioTrace, Celman), The cells were then air-dried on a laminar flow bench and place in a desiccator until analysis by nuclear microscopy. 2.5. Nuclear micrasco~~c analysts
2.2. Fixation The six target holders with the cells on the Pioloform film were removed from the growth medium and 3 were placed in 70 ml beakers containing 10 ml of the following fixatives: (1) Graded concentrations of 50%, 70% and 100% methanol (ho-analysi grade, Merck); (2) 4% formaldehyde in phosphate buffered saline “A” (Pro-analysi grade, Merck); (3) 5% glutaraldehyde in 0.1 M phosphate buffer (Pro-analysi grade, Merck). Fixation was for 5 min at room temperature (24°C). 2.3. Rinsing The fo~aldehyde and glutaraldehyde fixed samples were rinsed in 18 MfZ cm resistivity ASTM (American Society of Testing and Materials Standard) Type I reagent water (MilliQ, Millipore). Excess growth medium from the three unfixed samples was drained by touching the target holder with a piece of filter paper, after which the cells were dipped in either (a) 150 mM ammonium acetate
Analysis of the cells was done using the National University of Singapore Nuclear Microscope facility operating with a 2 MeV proton beam focused to a 1 p,m spot size [3]. Particle Induced X-Ray Emission (PIXE), Rutherford Backscattering Spectrometry (RBS) and Scanning Transmission Ion Microscopy @TIM) can be simultaneousiy employed to extract complementary information on the sample. PIXE provides information on the concentration and distribution of elements at the part per million (ppm> level; RBS on the matrix composition, thickness and density used in the normalisation of elemental concentrations; and off-axis STIM facilitates sample positioning and imaging. In order to maximise the efficiency of X-ray detection, the XYZ sample manipulator was set at 45” to the beam axis. A 400 p.m perspex filter with a I mm central hole was used with the 60 mm* Si(Li) X-ray detector (Link Systems) to optimise the system for the detection of trace elements from iron upwards in the Periodic Table 193. Single cell data were recorded either by scanning over the cell only, or by scanning a large area and using listmode data handling techniques VIII. BIOLOGY
354
[lo] to extract data off-line data analysis.
P.S. P. Thong
from
et al. / Nucl.
individual
cells
Instr.
und Meth.
during
in Phys.
Res. B 130 (1997)
351-357
be due to the direct use of liquid nitrogen. A more satisfactory result may have been obtained by freezing in liquid nitrogen cooled isopentane [l 11.
3. Results and discussion
3.2. Density-dependent
3.1. Process-dependent
The elemental concentrations of cells were also found to depend on the local cell density. Cells which had been rinsed in sucrose and air-dried were used in this comparison. The last two columns in Table 2 show the elemental concentrations of cultured human Chang liver cells grown on the same target holder (and therefore processed under identical conditions) but from two different regions. It can be seen that cells from a region in which the cells were densely packed (see Fig. lb) contained much higher potassium levels and lower sodium levels compared to those from a region where the cells were less densely packed (see Fig. la), resulting in a good K/Na ratio of about 15 for ceiis in the denser region compared with about 6 in the other. This is consistent with the hypothesis that there are K channels in the basolateral cell membranes which are open to potassium diffusion where the cells are separate from its neighbours, but not so in dense-packed cells 161. However, notwithstanding the good K/Na ratio, there are some problems associated with working with dense packed cells. Firstly, it is difficult to determine the boundaries of individual cells. Secondly, the cells tend to overlap each other in a dense-packed region and may no longer be a monolayer locally. Hence a region for single cell analysis is usually chosen such that individual cell boundaries are well defined. Regions with moderate cell densities in which individual cells can be identified and still exhibit a K/Na ratio well above 1 (Fig. la) thus offer a good compromise.
cell elemental profiles
Table 2 shows the average elemental concentrations of cultured human Chang liver cells which have been subjected to the preparation methods as listed in Table 1. Methanol-fixed cells (method 1) have a high level of chlorine (40 mg/g> and an average K/Na ratio of around 1. Glutaraldehydefixed cells suffered from magnesium and arsenic (not shown in Table 2) contamination, probably from the buffer solution, and had an average K/Na ratio of less than 1. The formaldehyde-fixed cells showed a good K/Na ratio of about 12. Hence fo~aldehydefixing yielded the best results among the three fixing agents tried. The average elemental concentrations of cells which had been rinsed in 150 mM ammonium acetate and freeze-dried (method 4) were calculated from a spectrum extracted from a total of about 1.5 cells in a 300 pm X 300 p_m scan using listmode data handling techniques. The cells were found to have high levels of sodium and chlorine, but low potassium levels, resulting in a K/Na ratio of less than 1. Ammonium acetate was therefore not considered a suitable rinsing solution for cultured human Chang liver cells. Ice-cold 150 mM sucrose, however, was found to be suitable as a rinsing solution. The average elemental concentrations of cells rinsed in sucrose and freeze-dried (method 5) or air-dried after blotting with pofysulfone membrane (method 6) show that cells dried by either process contained high potassium and low sodium levels, giving K/Na ratios well above 1 (6 and 12 for freeze-dried and air-dried cells respectively). Hence the cultured human Chang liver cells appeared to tolerate air-drying after the extracellular medium had been removed. The airdried cells contained much higher chlorine concentrations and marginally higher concentrations of phosphorus, sulphur and potassium compared to freeze-dried cells. The greater loss of P, S, Cl and K in freeze-dried cells compared to air-dried cells could
cell elemental profiles
3.3. Cell shapes as an indicator tion
of chemical
agita-
Fig. 1 shows the off-axis STIM maps of human Chang liver cells which had been processed with (a) and (b) 150 mM sucrose, (c) methanol, (d) formaldehyde, (e> glutaraldehyde, and (f) 150 mM ammonium acetate. From these maps, it can be seen that the cells exhibited rounding up to varying degrees depending on the prep~ation process. Cells which
concentrations
(parts per million,
dry weight)
-
Drying method:
12.82 (0.83)
75 (8) < 30 h
204 (25)
I. I6
Fe
Zn
K/Na
h Limit
of detection.
a Average elemental
concentration
extracted
204 (36)
4637 (180)
Ca
16274 (3042)
(I 756)
from total spectrum
of 15 cells
0.55 fO.10)
227 (48)
195 (65)
98 (12)
I16 (104) < 30 h
404 (38)
I599 (487)
SO934
(459)
I 1740
23357 (7405)
(0.09)
19719 837
0.57
121
135
120
971
11168
50074
88%
5479 (946) 7843 (757)
12325 (69)
K
ratio
2992Of1286) 64478 (5972)
NH,-acetate Freeze-drying
22034
13057 (757)
cu
Reagent type I Freeze-drying
8844 ( 1224)
I
-
(4)n=l5”
(n = 5)
14567 (1183)
i 1.92(0.87)
15.65 (I.711
31s (114) 156 (96) 5.79 (0.45)
I38 (38)
146 (88) 72 (42)
< 30 h
61 (64)
338 (197) 306(118) 71 (37)
70 (28)
33322 (4805)
471 (146)
27550 (2083)
6 I869 (4485)
22302 (403) (2269)
12160 (2282)
I 1575
1378 (234)
9002 (474)
2296 (157) 41471(1953)
23087 (1517)
1305 (238)
5812 (1078)
3980 (372)
(b) Dense
01=
IO)
(a) Moderate
B~otting/air-dying
I50 mM sucrose
(6)
n is the number of individual
17188 (1765)
199s (92)
2315 (163)
Freeze-drying
I50 mM sucrose
(5) II = 4
human Chnng liver cells subjected IO various methods of preparation.
CJlutaraldehyde
(3) n = 5
of cultured
30317 (I 199)
43835 (18272)
19655 (541)
P
1717 (48)
3983 (128)
S
I878 (89)
Mg
Freeze-drying
Reagent type
Formaldehyde
(2) )I= 4
Cl
20000 (5240)
Nil
Cell density:
-
Rinsing agent:
Methanol
II = 5
(I)
agent:
Method:
Fixing
cells analvsed and the standard errors are shown in oarentheses.
Average elemental
Table 2
P.S.P. Thong et uI./Nucl.
356
Instr. and Merh. in Phys. Res. B 130 (19971 35/-357
had been rinsed in sucrose (Fig. la,b) show the least sign of rounding with shrinkage, most remaining polygonal with about 30-40 p,rn diagonal lengths. The occasional round cells seen may be in mitotic phases. Cells fixed in glutaraldehyde (Fig. le) also appeared polygonal though some rounded cells can be seen. In contrast, cells which had been processed in methanol (Fig. lc), fo~aldehyde (Fig. Id), and ammonium acetate (Fig. lf> all appeared to have rounded up to about 25 pm, 20 Frn and 30 km in diameters respectively. In this respect, sucrose rinsing was found to have caused the least shrinkage to the cells and thus is the preferred medium for the preparation of cultured cells.
4. Conclusion Six methods involving the use of various fixing agents, rinsing agents and drying methods were tried in the prep~ation of cultured Chang liver cells for nuclear microscopy. The suitability of each method was evaluated by looking at the K/Na ratios and shapes of individual cells. A summary of the six preparation methods and their pros and/or cons is given in Table 3. Among cells which have been fixed, formaldehyde-fixed cells showed the most favourable elemental profiles. However they showed a great degree of rounding up, which may be an indicator of chemi-
Fig. 1. Off-axis STIM maps of (a) sucrose-rinsed ceils; (b) dense-packed sucrose-rinsed cells; (c) methanol fixed cells; (d) fomaldehyde fixed cells; (e> glutaraldehyde fixed cells and (f) ammonium acetate-rinsed cells.
Table 3 A summ~
of the six p~p~ation
methods and their pros and/or
cons
Fixing agent
Rinsing agent
Drying method
Pros
Cons
Recommended?
(I)
Methanol
-
Formaldehyde
Reagent type I
Simple, no drying required Ionic integrity preserved
(3)
Glu&~ldehyde
Reagent type
(41
-
(5)
-
(6)
-
150 mM NH,-acetate Ice-cold 150 mM sucrose Ice-cold 150 mM sucrose
ionic integrity not preserved Cells appear agitated Ionic integrity not preserved Ionic integrity not preserved Sucrose crystals may obscure cells -
No
(2)
(no further drying required) Freeze-drying
1
Freeze-drying Freeze-drying Freeze-drying Membrane blotting followed by air-drying
Ionic integrity preserved Ionic integrity perserved; excess rinsing solution removed
No No No Yes Yes; most suitable
P.S.P. Thong et al./ Nucl. Instr. and Meth. in Phys. Res. B 130 (1997) 351-357
cal perturbation. Hence chemical fixation was considered non-ideal for the preparation of cultured cells. Two commonly recommended rinsing solutions were tried, viz. sucrose and ammonium acetate. Cells which have been rinsed in 150 mM ammonium acetate were found to have K/Na ratios of less than 1. Cells which have been rinsed in 150 mM sucrose, however, were found to have K/Na ratios well above 1. Moreover, most of the cells remained substrate-anchored, exhibiting the least perturbation due to the preparative process. Hence ice-cold 150 mM sucrose was found to be the most suitable rinsing solution for the preparation of cultured Chang liver cells. In addition, to avoid sucrose crystal formation, which makes identi~cation of cells and their subsequent analysis difficult, rapid removal of the excess rinsing solution by blotting with a piece of biomembrane was considered an important step. Both freeze-drying and air-drying were tried as drying methods. Cells processed by both methods were found to have K/Na ratios well above 1. Hence both drying methods were found to be suitable but air-drying following membrane blotting was preferred as excess rinsing solution is removed very quickly during the blotting process. Overall, method 6 in Table 1 is considered the most suitable. The K/Na ratios of cells on the same target holder but from different regions were found to be dependent on the local cell density. Cells which are locally dense-packed were found to have a much higher K/Na ratio than cells in a less dense region. The concentrations of phosphorus and sulphur were
357
also higher in the dense-packed cells, showing that cells are less vulnerable to elemental redistribution in a region of high cell density. However, analysis of single cells is often difficult in regions of high cell densities since the cells tend to overlap and cease to be a monolayer locally. Hence regions of moderate cell densities from which individual cells can be identified and yet have K/Na ratios well above 1 are ideal for single cell analysis.
References [I] R. Wroblewski, J. Wroblewski, Scan. Microsc. Suppl. 8 t I9941 149. 121 A. WarIey, Scan. Microsc. Suppl. 8 (1994) 129. [3] F. Watt, 1. Orlic, K.K. Loh, C.H. Sow, P. Thong, S.C. Liew, T. Osipowicz. T.F. Choo, S.M. Tang, Nucl. Instr. and Meth. B 85 (1994) 708. [4] K.H. Sit, R. P_anantham, B.H. Bay, H.L. Chan, K.P. Wong, P. Thong, F. Watt, The Anatom. Rec. 245 (1996) I. [51 K.H. Sit, R. Paramanantham, B.H. Bay, K.P. Wong, P. Thong, F. Watt, Experientia 52 (1996) 778. 161 K. Zierold, in: X-ray Micro~~ysis in Biology: Experimental Techniques and Applications, eds. DC. Sigee, A.J. Morgan, A.T. Sumner and A. Warley (Cambridge University Press, Cambridge, 1993) pp. 101-I 16. 171 K.H. Sit, B.H. Bay, K.P. Wong, Japan. J. Physiol. 42 (1992) 355. [S] K.H. Sit, B.H. Bay, R. Paramanantham, P. Thong, F. Watt, J. Tiss. Cult. Meth 15 (1993) 199. 191 F. Watt, Nucl. Instr. and Meth. B 104 (1995) 276. 1101 G.W. Grime, M.A. Dawson, Nucl. Instr. and Meth. B 89 (1994) 223. [I 11 S. Van Lierde, W. Maenhaut, J. De Reuck, R.D. Vis, Nucl. Instr. and Meth. B 104 (1995) 328.
VIII. BIOLOGY
c2sJ !kj
ELSEVIER
Beam Interactions with Materials 8 Atoms Nuclear Instruments
and Methods
in Physics Research B 130 (1997) 35 l-357
Nuclear microscopy of single whole cultured cells: preparation and analysis of human Chang liver cells P.S.P. Thong ‘,*, F. Watt ‘, R. Paramanantham b, B.H. Bay b, K.H. Sit b a Nuclear Microscopy L,aborutory. Department of Physics. National University of Singapore, Lower Kent Ridge Road, Singapore 119260, Singapore b Department
qf Anatomy, Faculty ctfMedicine, National Uniuersity of Singapore. Lower Kent Ridge Road, Singpapore I1 9260, Singapore
Abstract Nuclear microscopy is a powerful tool for the measurement of elemental concentrations in single cells. Six methods involving the use of various fixing agents, rinsing agents and drying methods were tried in the preparation of cultured human Chang liver cells for nuclear microscopy and the suitability of each method was evaluated by monitoring the K/Na ratios and shapes of individual cells. The K/Na ratio is a commonly used criteria for the ionic integrity of cells; K/Na ratios well above 1 indicates minimal perturbation of the intracellular ionic composition. Non-stimulated human Chang liver ceils in a resting state are usually polygonal in shape and flattened in firm anchorage to the substrate, while dividing or stimulated cells appear rounded. Therefore the shapes of the cells can be used as an indicator of whether the cells are in a resting or stimulated state. It is not desirable for cells to be in a stimulated state since then the effects of other external stimuli cannot be observed independently. Of the six methods tested, chemical fixation, as expected, was considered non-ideal for the preparation of human cultured Chang liver cells. Ice-cold 1.50 mM sucrose was found to be the most suitable rinsing solution for the preparation of cultured human Chang liver cells. Both freeze-drying and air-drying were used as drying methods and cells processed by either method were found to have K/Na ratios well above 1. Hence both drying methods were found to be suitable although membrane blotting followed by air-drying was preferred as excess rinsing solution can be very quickly removed during the blotting process. The K/Na ratios of cells on the same target holder but from different regions were found to be dependent on the local cell density. Cells which are locally dense-packed were found to have a much higher K/Na ratio than cells in a less dense region.
1. Introduction Nuclear
Microscopy
(NM)
is a collective
term
to refer to a group of microanalytical techniques based on the use of a focused beam of MeV particles. Increasingly, NM is being recognised as a powerful tool for the analysis of both tissue sections used
* Corresponding [email protected]
author.
Fax:
+ 65-777-6126;
email:
and single whole cells in biomedical research. The strengths of the NM include a lower bremsstrahlung background cont~bution to the X-ray energy spectrum, a greater range of the projectile beam in the target and a much lesser degree of broadening as the beam traverses the sample than when compared to the electron microscope. The use of cell cultures in biomedical research has gained popularity in the past decade. Its favourabiIity lies in the fact that firstly, experimental conditions can be controlled and secondly, the effects of
0~68-583X/97/$17.~ 0 1997 Elsevier Science B.V. Ail rights reserved. PII SOi68-583X(97)00225-5
VIII. BIOLOGY
P.S. P. Thong et at. / Nucl Instr. and Meth in Phy.
352
external factors, toxic agents and stimuli can be studied even in single cells [l]. The crucial step in the meaningful analysis of biomedical materials lies in preserving the ionic integrity of the samples. In the past two decades, electron microscopists have accumulated a wealth of information on the preparation of cells for X-ray microanalysis. Warley [21 pointed out that whereas fixation methods were popular before the early SO’s, the trend is toward cryopreparation in the 90’s since cryotechniques often yield the best results with regards to preserving the in situ ionic concentrations. This criterion is especially important where analysis of diffusible ions is involved. In choosing a workable preparation procedure for the nuclear microscopic analysis of cells, the same considerations are applicable as for electron microscope X-ray microanalysis, viz. the minimal disturbance of the ionic concentrations arising from preparation. Hence the nuclear microscopist can adopt procedures for microanalysis generally recommended by the electron microscopist. Nevertheless, as no two cell types behave in the exactly the same way, no single preparatory method can be hailed as the ideal method for all cell types. In the present work, we experimented with various processes for the preparation of cultured human Chang liver cells for single whole cell analysis using the NUS Nuclear Microscope [3], A total of six preparation methods, involving the use of different fixatives, washing mediums and drying methods were tried. A summary is outlined in Table 1. The suitability of each method
Table 1 Methods used for the preparation
(3)
Methanol in graded concentrations 4% fo~aldehyde in phosphate buffered saline 5% glutaraldehyde in 0.1 M phosphate buffer
(4) (5) (6)
1
was evaluated by looking at the K/Na ratios and shapes of individual cells to find the most suitable method for processing the cells. NM applications using cultured human Chang liver cells have included studies in apoptosis [4,5].
I.1. K/Na
ratio
The intracellular K/Na ratio is a commonly used indicator of the ionic integrity of the cell [6]. K/Na ratios well above 1 indicates minimal perturbation of the intracellular ionic composition while K/Na ratios below 1 is indicative of at least partial damage to the cell. Here we show that the K/Na ratio of individual cells is both dependent on the cell preparation method used as we11 as on the local cell density.
1.2. Celt shapes Non-stimulated Chang liver cells in a resting state are usually polygonal and flat. Cell rounding occurs either during mitosis or in the presence of an external stimulus [7]. Thus the shapes of the cells can be used as an indicator of whether the cells are in a resting or active state. It is not desirable for cells to be in a stimulated state since then the effects of other external stimuli cannot be observed independently. We show in this paper that cells exhibit rounding up due to perturbation by certain process chemicals.
of cultured human Chang liver cells for nuclear microscopy
Fixing agent (1) (2)
Rex 8 130 I19971 351-357
Rinsing agent
reagent type I
Drying method (no further drying required) freeze-drying
I
freeze-drying
I50 mM ammonium acetate (pH 7.0) ice-cold 150 mM sucrose ice-cold 150 mM sucrose
freeze-drying
reagent type
freeze-drying membrane blotting followed by air-drying
P.S.P. Thong et cd./ Nucl. fnsrr. und Meth. in Phys. Res. 6 I30 (1997)
2. Materials
and methods
351-357
353
solution adjusted to pH 7.0 (method 4 in Table 11, or (b) ice-coid 150 mM sucrose solution (methods 5,6).
2.1. Monolayer culture 2.4. Dr?,ing methods American Type Culture Collection human Chang liver cells (ATCC CCL 13) were cultured as a monolayer on six Pioloform (Agar Scientific)-coated target holders as described by Sit et al. [8]. (a) 20 ml of a cell suspension in pure fetal bovine serum (Cytosystems) was dropped onto the Pioloform coated target holders in a 70 ml sterile container. A cell density of 3 X lo3 cells on the 5 mm diameter hole of the target holder was found to be suitable for single cell analysis using the nuclear microscope, (b) The container was capped and incubated for one hour at 37°C. (c> At the end of the one hour incubation, the serum-pulsed cells were gently flooded with 10 ml of DMEM (Dulbecco’s Modified Eagle’s Medium, Sigma) supplemented with 10% fetal bovine serum. (d) The cufture was then placed in a 37°C incubator and was ready for processing after 24 h.
Samples fixed in methanol (method 11 required no further drying since methanol both fixes and dehydrates. Samples fixed in fo~aidehyde and glutaraldehye (methods 2,3) as well as those rinsed in ammonium acetate and sucrose (methods 45) were freeze-dried as follows: (i) After rinsing, the edges of the target holders were touched with a piece of filter paper. (ii> The samples were placed in a small volume of liquid nitrogen (- 170°C) in a Dewar flask and freeze-dried in a freezer at - 20°C. The sixth remaining sucrose-rinsed sample (method 6) was blotted with a piece of low protein binding polysulfone membrane (BioTrace, Celman), The cells were then air-dried on a laminar flow bench and place in a desiccator until analysis by nuclear microscopy. 2.5. Nuclear micrasco~~c analysts
2.2. Fixation The six target holders with the cells on the Pioloform film were removed from the growth medium and 3 were placed in 70 ml beakers containing 10 ml of the following fixatives: (1) Graded concentrations of 50%, 70% and 100% methanol (ho-analysi grade, Merck); (2) 4% formaldehyde in phosphate buffered saline “A” (Pro-analysi grade, Merck); (3) 5% glutaraldehyde in 0.1 M phosphate buffer (Pro-analysi grade, Merck). Fixation was for 5 min at room temperature (24°C). 2.3. Rinsing The fo~aldehyde and glutaraldehyde fixed samples were rinsed in 18 MfZ cm resistivity ASTM (American Society of Testing and Materials Standard) Type I reagent water (MilliQ, Millipore). Excess growth medium from the three unfixed samples was drained by touching the target holder with a piece of filter paper, after which the cells were dipped in either (a) 150 mM ammonium acetate
Analysis of the cells was done using the National University of Singapore Nuclear Microscope facility operating with a 2 MeV proton beam focused to a 1 p,m spot size [3]. Particle Induced X-Ray Emission (PIXE), Rutherford Backscattering Spectrometry (RBS) and Scanning Transmission Ion Microscopy @TIM) can be simultaneousiy employed to extract complementary information on the sample. PIXE provides information on the concentration and distribution of elements at the part per million (ppm> level; RBS on the matrix composition, thickness and density used in the normalisation of elemental concentrations; and off-axis STIM facilitates sample positioning and imaging. In order to maximise the efficiency of X-ray detection, the XYZ sample manipulator was set at 45” to the beam axis. A 400 p.m perspex filter with a I mm central hole was used with the 60 mm* Si(Li) X-ray detector (Link Systems) to optimise the system for the detection of trace elements from iron upwards in the Periodic Table 193. Single cell data were recorded either by scanning over the cell only, or by scanning a large area and using listmode data handling techniques VIII. BIOLOGY
354
[lo] to extract data off-line data analysis.
P.S. P. Thong
from
et al. / Nucl.
individual
cells
Instr.
und Meth.
during
in Phys.
Res. B 130 (1997)
351-357
be due to the direct use of liquid nitrogen. A more satisfactory result may have been obtained by freezing in liquid nitrogen cooled isopentane [l 11.
3. Results and discussion
3.2. Density-dependent
3.1. Process-dependent
The elemental concentrations of cells were also found to depend on the local cell density. Cells which had been rinsed in sucrose and air-dried were used in this comparison. The last two columns in Table 2 show the elemental concentrations of cultured human Chang liver cells grown on the same target holder (and therefore processed under identical conditions) but from two different regions. It can be seen that cells from a region in which the cells were densely packed (see Fig. lb) contained much higher potassium levels and lower sodium levels compared to those from a region where the cells were less densely packed (see Fig. la), resulting in a good K/Na ratio of about 15 for ceiis in the denser region compared with about 6 in the other. This is consistent with the hypothesis that there are K channels in the basolateral cell membranes which are open to potassium diffusion where the cells are separate from its neighbours, but not so in dense-packed cells 161. However, notwithstanding the good K/Na ratio, there are some problems associated with working with dense packed cells. Firstly, it is difficult to determine the boundaries of individual cells. Secondly, the cells tend to overlap each other in a dense-packed region and may no longer be a monolayer locally. Hence a region for single cell analysis is usually chosen such that individual cell boundaries are well defined. Regions with moderate cell densities in which individual cells can be identified and still exhibit a K/Na ratio well above 1 (Fig. la) thus offer a good compromise.
cell elemental profiles
Table 2 shows the average elemental concentrations of cultured human Chang liver cells which have been subjected to the preparation methods as listed in Table 1. Methanol-fixed cells (method 1) have a high level of chlorine (40 mg/g> and an average K/Na ratio of around 1. Glutaraldehydefixed cells suffered from magnesium and arsenic (not shown in Table 2) contamination, probably from the buffer solution, and had an average K/Na ratio of less than 1. The formaldehyde-fixed cells showed a good K/Na ratio of about 12. Hence fo~aldehydefixing yielded the best results among the three fixing agents tried. The average elemental concentrations of cells which had been rinsed in 150 mM ammonium acetate and freeze-dried (method 4) were calculated from a spectrum extracted from a total of about 1.5 cells in a 300 pm X 300 p_m scan using listmode data handling techniques. The cells were found to have high levels of sodium and chlorine, but low potassium levels, resulting in a K/Na ratio of less than 1. Ammonium acetate was therefore not considered a suitable rinsing solution for cultured human Chang liver cells. Ice-cold 150 mM sucrose, however, was found to be suitable as a rinsing solution. The average elemental concentrations of cells rinsed in sucrose and freeze-dried (method 5) or air-dried after blotting with pofysulfone membrane (method 6) show that cells dried by either process contained high potassium and low sodium levels, giving K/Na ratios well above 1 (6 and 12 for freeze-dried and air-dried cells respectively). Hence the cultured human Chang liver cells appeared to tolerate air-drying after the extracellular medium had been removed. The airdried cells contained much higher chlorine concentrations and marginally higher concentrations of phosphorus, sulphur and potassium compared to freeze-dried cells. The greater loss of P, S, Cl and K in freeze-dried cells compared to air-dried cells could
cell elemental profiles
3.3. Cell shapes as an indicator tion
of chemical
agita-
Fig. 1 shows the off-axis STIM maps of human Chang liver cells which had been processed with (a) and (b) 150 mM sucrose, (c) methanol, (d) formaldehyde, (e> glutaraldehyde, and (f) 150 mM ammonium acetate. From these maps, it can be seen that the cells exhibited rounding up to varying degrees depending on the prep~ation process. Cells which
concentrations
(parts per million,
dry weight)
-
Drying method:
12.82 (0.83)
75 (8) < 30 h
204 (25)
I. I6
Fe
Zn
K/Na
h Limit
of detection.
a Average elemental
concentration
extracted
204 (36)
4637 (180)
Ca
16274 (3042)
(I 756)
from total spectrum
of 15 cells
0.55 fO.10)
227 (48)
195 (65)
98 (12)
I16 (104) < 30 h
404 (38)
I599 (487)
SO934
(459)
I 1740
23357 (7405)
(0.09)
19719 837
0.57
121
135
120
971
11168
50074
88%
5479 (946) 7843 (757)
12325 (69)
K
ratio
2992Of1286) 64478 (5972)
NH,-acetate Freeze-drying
22034
13057 (757)
cu
Reagent type I Freeze-drying
8844 ( 1224)
I
-
(4)n=l5”
(n = 5)
14567 (1183)
i 1.92(0.87)
15.65 (I.711
31s (114) 156 (96) 5.79 (0.45)
I38 (38)
146 (88) 72 (42)
< 30 h
61 (64)
338 (197) 306(118) 71 (37)
70 (28)
33322 (4805)
471 (146)
27550 (2083)
6 I869 (4485)
22302 (403) (2269)
12160 (2282)
I 1575
1378 (234)
9002 (474)
2296 (157) 41471(1953)
23087 (1517)
1305 (238)
5812 (1078)
3980 (372)
(b) Dense
01=
IO)
(a) Moderate
B~otting/air-dying
I50 mM sucrose
(6)
n is the number of individual
17188 (1765)
199s (92)
2315 (163)
Freeze-drying
I50 mM sucrose
(5) II = 4
human Chnng liver cells subjected IO various methods of preparation.
CJlutaraldehyde
(3) n = 5
of cultured
30317 (I 199)
43835 (18272)
19655 (541)
P
1717 (48)
3983 (128)
S
I878 (89)
Mg
Freeze-drying
Reagent type
Formaldehyde
(2) )I= 4
Cl
20000 (5240)
Nil
Cell density:
-
Rinsing agent:
Methanol
II = 5
(I)
agent:
Method:
Fixing
cells analvsed and the standard errors are shown in oarentheses.
Average elemental
Table 2
P.S.P. Thong et uI./Nucl.
356
Instr. and Merh. in Phys. Res. B 130 (19971 35/-357
had been rinsed in sucrose (Fig. la,b) show the least sign of rounding with shrinkage, most remaining polygonal with about 30-40 p,rn diagonal lengths. The occasional round cells seen may be in mitotic phases. Cells fixed in glutaraldehyde (Fig. le) also appeared polygonal though some rounded cells can be seen. In contrast, cells which had been processed in methanol (Fig. lc), fo~aldehyde (Fig. Id), and ammonium acetate (Fig. lf> all appeared to have rounded up to about 25 pm, 20 Frn and 30 km in diameters respectively. In this respect, sucrose rinsing was found to have caused the least shrinkage to the cells and thus is the preferred medium for the preparation of cultured cells.
4. Conclusion Six methods involving the use of various fixing agents, rinsing agents and drying methods were tried in the prep~ation of cultured Chang liver cells for nuclear microscopy. The suitability of each method was evaluated by looking at the K/Na ratios and shapes of individual cells. A summary of the six preparation methods and their pros and/or cons is given in Table 3. Among cells which have been fixed, formaldehyde-fixed cells showed the most favourable elemental profiles. However they showed a great degree of rounding up, which may be an indicator of chemi-
Fig. 1. Off-axis STIM maps of (a) sucrose-rinsed ceils; (b) dense-packed sucrose-rinsed cells; (c) methanol fixed cells; (d) fomaldehyde fixed cells; (e> glutaraldehyde fixed cells and (f) ammonium acetate-rinsed cells.
Table 3 A summ~
of the six p~p~ation
methods and their pros and/or
cons
Fixing agent
Rinsing agent
Drying method
Pros
Cons
Recommended?
(I)
Methanol
-
Formaldehyde
Reagent type I
Simple, no drying required Ionic integrity preserved
(3)
Glu&~ldehyde
Reagent type
(41
-
(5)
-
(6)
-
150 mM NH,-acetate Ice-cold 150 mM sucrose Ice-cold 150 mM sucrose
ionic integrity not preserved Cells appear agitated Ionic integrity not preserved Ionic integrity not preserved Sucrose crystals may obscure cells -
No
(2)
(no further drying required) Freeze-drying
1
Freeze-drying Freeze-drying Freeze-drying Membrane blotting followed by air-drying
Ionic integrity preserved Ionic integrity perserved; excess rinsing solution removed
No No No Yes Yes; most suitable
P.S.P. Thong et al./ Nucl. Instr. and Meth. in Phys. Res. B 130 (1997) 351-357
cal perturbation. Hence chemical fixation was considered non-ideal for the preparation of cultured cells. Two commonly recommended rinsing solutions were tried, viz. sucrose and ammonium acetate. Cells which have been rinsed in 150 mM ammonium acetate were found to have K/Na ratios of less than 1. Cells which have been rinsed in 150 mM sucrose, however, were found to have K/Na ratios well above 1. Moreover, most of the cells remained substrate-anchored, exhibiting the least perturbation due to the preparative process. Hence ice-cold 150 mM sucrose was found to be the most suitable rinsing solution for the preparation of cultured Chang liver cells. In addition, to avoid sucrose crystal formation, which makes identi~cation of cells and their subsequent analysis difficult, rapid removal of the excess rinsing solution by blotting with a piece of biomembrane was considered an important step. Both freeze-drying and air-drying were tried as drying methods. Cells processed by both methods were found to have K/Na ratios well above 1. Hence both drying methods were found to be suitable but air-drying following membrane blotting was preferred as excess rinsing solution is removed very quickly during the blotting process. Overall, method 6 in Table 1 is considered the most suitable. The K/Na ratios of cells on the same target holder but from different regions were found to be dependent on the local cell density. Cells which are locally dense-packed were found to have a much higher K/Na ratio than cells in a less dense region. The concentrations of phosphorus and sulphur were
357
also higher in the dense-packed cells, showing that cells are less vulnerable to elemental redistribution in a region of high cell density. However, analysis of single cells is often difficult in regions of high cell densities since the cells tend to overlap and cease to be a monolayer locally. Hence regions of moderate cell densities from which individual cells can be identified and yet have K/Na ratios well above 1 are ideal for single cell analysis.
References [I] R. Wroblewski, J. Wroblewski, Scan. Microsc. Suppl. 8 t I9941 149. 121 A. WarIey, Scan. Microsc. Suppl. 8 (1994) 129. [3] F. Watt, 1. Orlic, K.K. Loh, C.H. Sow, P. Thong, S.C. Liew, T. Osipowicz. T.F. Choo, S.M. Tang, Nucl. Instr. and Meth. B 85 (1994) 708. [4] K.H. Sit, R. P_anantham, B.H. Bay, H.L. Chan, K.P. Wong, P. Thong, F. Watt, The Anatom. Rec. 245 (1996) I. [51 K.H. Sit, R. Paramanantham, B.H. Bay, K.P. Wong, P. Thong, F. Watt, Experientia 52 (1996) 778. 161 K. Zierold, in: X-ray Micro~~ysis in Biology: Experimental Techniques and Applications, eds. DC. Sigee, A.J. Morgan, A.T. Sumner and A. Warley (Cambridge University Press, Cambridge, 1993) pp. 101-I 16. 171 K.H. Sit, B.H. Bay, K.P. Wong, Japan. J. Physiol. 42 (1992) 355. [S] K.H. Sit, B.H. Bay, R. Paramanantham, P. Thong, F. Watt, J. Tiss. Cult. Meth 15 (1993) 199. 191 F. Watt, Nucl. Instr. and Meth. B 104 (1995) 276. 1101 G.W. Grime, M.A. Dawson, Nucl. Instr. and Meth. B 89 (1994) 223. [I 11 S. Van Lierde, W. Maenhaut, J. De Reuck, R.D. Vis, Nucl. Instr. and Meth. B 104 (1995) 328.
VIII. BIOLOGY