The use of a scanning proton microprobe to observe anti-HIV drugs within cells

Life Sciences, Vol. 54, No. 21, pp. 160%1612, 1994 Copyright © 1994 Elsevier Science Lid Printed in the USA. All rights reserved

Pergamon

oo24-3206/94 $6.oo + .OO

THE USE OF A SCANNING PROTON MICROPROBE TO OBSERVE ANTI-HIV DRUGS WITHIN

CELLS

M° C h o l e w a 1, G . J . F . L e g g e 1, H° W e i g o l d 2, G. H o l a n 2, C . J . B i r c h 3 1Micro Analytical Research Centre (MARC), School of Physics, The University of Melbourne, Parkville, Vie. 3052, AUSTRALIA ZDivision of Chemicals and Polymers, CSIRO, Clayton, Vic.3168, AUSTRALIA aVirology Department, Fairfield Hospital, Fairfield, Vic.3078, AUSTRALIA (Received in final form March 11, 1994)

Summary A series of inorganic polyanions (viz. heteropolytungstates) has been shown to have antiviral activity but there was no evidence to indicate that the drugs reached their site of antiviral (HIV) activity intact. We have shown that with a scanning proton microprobe it is possible to analyse the metal content of individual cells (PBLs) treated with such a polyoxometalate drug and to determine the atomic ratio of the metals within the cell. This was found to be near that in the drug. The distribution of the metals (tungsten and cobalt) within the cell was measured and it was shown that both metals were located in the same region within the cell. These findings would suggest that the drug had entered the cells intact. Key Words: AIDS, inorganic drugs, proton microprobe, proton-induced X-ray emission

The drug under investigation in this study is one of a large number of polyanions (heteropolytungstates) that have been tested for their ability to inhibit the replication of HIV in a continous human T-lymphocyte line ( MT2 cells ) and in peripheral blood lymphocytes ( PBL cells ) from healthy blood donors. The mechanism of action of these compounds has been shown to be two fold, viz. the direct inhibition of the viral reverse transcriptase and interference with the adsorption/penetration of the virus into the target cells. Inouye et al. 1 found no direct correlation between the antiviral activity of heteropolytungstate drugs in MT4 cells and their ability to inhibit HIV-1 reverse transcriptase. It was observed by us 2 and others 1, that heteropolytungstate drugs of similar structure often show a large variation in antiviral activity in in vitro testing. Even the nature of the counter cation can have a significant effect on the drug efficacy. For example, exchanging the potassium for sodium on the polyanion can M a r i a n C h o l e w a , Micro Analytical Research Centre (MARC), School of Physics, The University of Melbourne, Parkville, Vic. 3052, Australia.

1608

Proton Microprobe in AIDS Research

Vol. 54, No. 21, 1994

markedly affect it's antiviral activity 2. A further interesting property of this family of drugs is that their anti-HIV activity expressed in vitro in MT2 cells is often much diminished when tested in PBL cells. The compound, I(lo[Co4(H20)2(PWg034)21" nil20 (n is ca. 25), which is the subject of this investigation, falls into this category. In the lack of evidence, it is tempting to rationalize many of these observations by assuming a sensitive relationship between the structure of the polyanion and the ability of the drug to penetrate cell walls. The aim of this study was therefore to establish a method for determining the quantitative distribution of metals in cells treated with a drug at non-toxic levels and, in particular, to establish whether the polyanion [Co4(H20)2(PWgOa4)2]1°- can enter PBL cells and whether if so, it remains intact within the cell? The subtoxic concentration of heteropolytungstate drugs varies, generally from about 5 to 200 #g of drug per mL of culture medium. The levels of subtoxic intracellular concentration of the signature elements are therefore in the parts per million region. Subtoxic is here defined as those concentrations of the drug below which drug-related morphological changes (observed by light microscopy) and cell death (estimated by counting viable cells in the presence of a vital stain) do not occur. By toxic we mean when more than 10% of cells are killed in the presence of drug and in the absence of virus. The advantage heteropolytungstates offer in these studies is that many of these compounds contain more than one metal not normally found in detectable amounts in individual cells. The detection within the cell of the metal atoms associated with the drug, together with the quantitative estimation of their ratio and distribution within the cell, can suggest an answer to the questions posed above. The compound of interest in this study contains two such signature elements, viz. tungsten and cobalt. If the ratio of these two elements within the cell is found to be the same as that in the drug and if their distribution within the cell is mapped and shown to be comparable, then presumably the drug is intact within the cell.

Methods Typical viable lymphocytes (MT2 and PBL) are about 10 #m in diameter and after plating out and freeze drying are about 12 #m in diameter and less than 1 #m in thickness. Therefore the mapping of elemental distributions within cells and the extraction of quantitative spectra requires a spatial resolution of about 1 #m, a high sensitivity to heavy metal elements and quantitative accuracy. The only instrument presently able to meet these requirements is the scanning proton microprobe - SPMP (or, more generally, ion microprobe) 3. This instrument (also sometimes known as a nuclear microprobe) utilises a beam of high energy ions of velocity comparable to that of the electrons in an electron microprobe; it is essentially nondestructive and particularly sensitive to the heavier elements, because of the very low bremsstrahlung background in its x-ray spectrum compared with that of the electron microprobe. The instrument at Melbourne was optimised for biological work 4 and, in order to maximise efficiency and to achieve quantitative accuracy in scanned data, it was necessary to develop a new technique for handling spectral data from scanning instruments 5. This technique, Total Quantitative Scanning Analysis (TQSA), is a procedure whereby all events (if required, from several detectors) are recorded without windowing during the scanning operation. Computer hardware keeps track of the energy (E) and beam position (x,y) for every event and the associated software constructs a three-dimensional (E,x,y) block of data from which the maps for all elements and spectra for all regions of interest can be extracted without further irradiation.

Vol. 54, No. 21, 1994

Proton Microprobe in AIDS Research

1609

This emphasis on efficiency is essential when trace elements are to be mapped in a very thin specimen. Although protons and electrons of comparable velocity have comparable ionizing efficiency, the much lower detection limits of the proton microprobe ( approximately 2 orders of magnitude for similar analyses ) are achieved by virtue of lower background radiation, not by higher yields of the characteristic x-rays. TQSA was also needed to define cell outlines and extract a quantitative spectrum for each cell. Thus, although TQSA offers benefits to all scanning instruments, it is essential to the work described here. Prior to analysis, peripheral blood lymphocytes (PBLs) are incubated in cell culture medium (RPMI-1640 containing 10% foetal calf serum and supplemented with interleukin2) containing subtoxic concentration (50 #g/mL) of drug. Drugs are added 24 hours in advance

'

l

'

l

'

l

'

1400

KI0[Co4(H20)2(PW90s4)2]*nH20

3 MeV

Cl

1200 I000 5q Z O O

H+

W

800 600

W PS K

400

Co

Si

x50~

200 ,

°o

2.0

4.0

6.0

ENERGY(keY)

8.0

~11 II1~12.0 10.0

Fig.1

A typical elemental spectrum from a single PBL exposed to 50 # g / m l of Klo[Co4(H20)2(PWg034)2]" nH~O drug. The spectrum, extracted from the data block, is that of the entire cell. The detected concentration of tungsten (,,~10 ppm) in the cell is well above the detection limit of the SPMP, but the level of the drug in the culture medium was a factor of 4 lower then the toxic {lethal) level. At the X-ray energy of 6 keV (horizontal axis) the scale is multiplied by a factor of 50 to make Co and W peaks more visible. The Fe peak at 6.4 keV is not from the drug.

1610

Proton Microprobe in AIDS Research

2A

6 .

2B

Vol. 54, No. 21, 1994

o

5

~o

,;

.

.

.

.

.

.

.

.

.

d

~o

Fig.2 Distribution of Phosphorus (A), Cobalt (B) and Tungsten (C) inside cells when exposed to a subtoxie level of Iflo[Co4(H20)2(PWgO34)2]'nH20in the culture medium. The colour scale of concentration is also shown. The apparent high concentrations are centered on the cell nuclei. Normalisation to the areal density (STIM) would not remove this effect. The original data, collected with 100 nm pixel resolution has here been smoothed with a gaussian function of 1 #m width (full width half maximum) similar to the resolution of the beam. The pixel size in these final images is also 1 #m. The cell on the left shows typical distribution of the drug elements.

Vol. 54, No. 21, 1994

Proton Microprobe in AIDS Research

1611

to give them maximal effect eg. if the drug is modified to an active form by the cell, it is necessary to allow some time for this to occur before infecting the cells. A good example is the nucleoside analogues. Probably for the inorganic polyanions this pre-incubation step is not necessary if the mechanism is inhibition of RT, but may be necessary if they inhibit at attachment. After 24 hours incubation, cells are harvested by centrifugation at 1500 rpm for 10 minutes, then resuspended in a small volume of 0.154 M a m m o n i u m acetate. For analysis, cells from this suspension are spotted onto duplicate grids and immediately snap frozen in isopentane cooled down to near freezing point by liquid nitrogen. The grids are transferred at liquid nitrogen temperature to a vacuum chamber and freeze dried at a pressure of 10 .6 torr as the cells slowly warm up to ambient temperature over a period of about 12 hours. The freeze dried cells are transferred to the specimen chamber of the SPMP and individually positioned and scanned with a 3 MeV proton beam of about 30 pA and 1 micron resolution. X-rays emitted by the cell are collected with a Si(Li) (energy dispersive) detector. Usually between 30 to 40 cells (including controls) were examined in a run with a given drug. Typically a time of 3 hours was necessary to collect sufficient statistics from a 20×20 # m 2 scan area which contained 1 or 2 cells. Analysis of data requires only a few minutes per cell. In order to minimize thermal damage to the cells, the beam is kept in steady motion and all data for the scanned cell are collected by TQSA. Although, as with electron probe irradiation, specimen shrinkage and elemental losses (generally only of light elements) can occur from ionization in the SPMP 6 (Scanning Proton Microprobe), losses of the elements of significance were negligible under the proton irradiation levels of this experiment, where even the shrinkage was small (2% as measured by STIM) and the mass loss not detectable. The data for each scan are sorted into a three-dimensional block, from which all elemental maps and spectra can be extracted. The ratios of signature elements (those not normally present in the cell) are compared with the same ratios measured for the pure compound, to see whether the compound breaks down or is stable within the cellular environment. The maps of the individual elements independently will show the distribution of the drug within the cell.

Results A typical elemental spectrum from one of these cells is shown in Fig.1. The heavy elements characteristic for this drug (K1o[Co4(H20)2(PWg034)2].nil20) are clearly measurable above the very low background characteristic of the P I X E (Proton Induced X-ray Emission) spectrum. The ability of the SPMP to map the elemental distributions of this drug is displayed in Fig.2. It shows the distribution of phosphorus (A), cobalt (B) and tungsten (C) in two PBLs within the scanned area (20 × 20 #m). The phosphorus map of Fig.2A is one of several elemental maps. Those for trace elements are naturally poor in statistics but, in order to correlate them with cell structure, they can be superimposed on other maps, such as this phosphorus map; this can be done very precisely, because all maps come from the same data set. The phosphorus map conveniently gives the outline of the cell.

Discussion The results obtained clearly indicate that the drug, even though of little activity against HIV1 in PBL cells, does enter these cells. Elemental levels were obtained from the areas under peaks in spectra such as that of Fig.1. Variability from cell to cell was a m a x i m u m of 50 %.

1612

Proton Microprobe in AIDS Research

Vol. 54, No. 21, 1994

The atomic ratio of the two elements ( W : Co ) in the cell is 12.2 within experimental error similar to that in the drug, v/z. 14.04. This, together with the observation illustrated in Fig. 2 that the tungsten and cobalt constituents of the drug are found in the same areas of the cell, suggests that the polyanion [Co4(H20)2(PWg034)2] 10- remains intact within the cell. The low antiviral activity of the drug tested in PBLs, compared to MT2 cells, does not therefore appear to be due to drug instability or inability to penetrate the cell membrane. In nearly all the treated PBL cells studied, the drug within the cells was strongly localised in the region where the P concentration also peaks. This may be the location of the cell nucleus. However, some cells had a different, more broad distribution of the drug within the cell. An example of each of these behaviours is observed in Fig.2. The cell on the left has the drug localised but in the cell on the right the drug can be seen to be spread more diffusely through the cell, and notably not concentrated within the cell nucleus, as judged from the P distribution map. Why a small minority of cells exhibit this bahaviour has not been investigated, but it may indicate that some cells are not fully viable. Their uptake of the drug appears to be less than that by the "normal" cells. While it was found that the techniques and instrumentation used in this study are appropriate for these and other investigations in which trace distributions of a drug within a cell may be identified by means of some characteristic marker element, it is imperative for the validity of any experimental findings that subtoxic levels of the drug be used. To extend this work to investigate more toxic drugs, the x-ray detection efficiency is being increased by an order of magnitude by instalation of a high efficiency Si(Li) detector.

Acknowledgements This work was supported by a CSIRO/University of Melbourne research grant and a Commonwealth Aids Research Grant.

References

1. Y. INOUYE, Y. T O K U T A K E , J. KUNIHARA, T. YOSHIDA, T. YAMASE, A. NAKATA, S. NAKAMURA, Chem. Pharm. Bull., 40 (1992) 805. 2. H. WEIGOLD, G. HOLAN, S.M. MARCUCCIO, C.J. BIRCH, I.D. GUST, Int. Patent Appl., PCT/AU/00280, 28 June, 1991. 3. J.A. COOKSON, A.T.G. FERGUSON, F.D. PILLING, J. of Radianal. Chem., 12 (1972) 39. 4. G.J.F. LEGGE, C.D. McKENZIE, A.P. MAZZOLINI, J. Microsc., 117 (1979) 185. 5. G.J.F. LEGGE, I. HAMMOND, J. Microsc., 117 (1979) 201. 6. M. CHOLEWA, G. BENCH, B.J. KIRBY, G.J.F. LEGGE, Nucl. Instr. and Meth., B54 (1991)101.