Tetraphenylphosphonium chloride induced

Int J Cancer: 67,72-79 (1996) 1996 Wiley-Liss, Inc Publication of the International Union Against Cancer Publication de I'Union Internationale Contre le Cancer TETRAPHENYLPHOSPHONIUM CHLORIDE INDUCED MR-VISIBLE LIPID ACCUMULATION IN A MALIGNANT HUMAN BREAST CELL LINE Edward J DELIKATNY~J, Sandrine K ROMAN',Rebecca HANCOCK', Thomas M JEITNER3s5,Catherine M LANDER', Darryl C RIDE OUT^.^ and Carolyn E MOUNTFORD' Departments of 'Cancer Medicine and 3Patholonv,Universityof Sydney, Sydney, NSW, 2006, Australia and 4Scripps Clinic, La Jolla, CA, USA The effect of the cationic lipophilic phosphonium salt tetraphenylphosphonium chloride (TPP) on a human malignant breast cell line, DU4475, was monitored with proton nuclear magnetic resonance (IH MRS) TPP caused a dose- and timedependent increase in resonances arising from MR-visible lipid as measured by the CH2/CH3ratio in the I-dimensional 'H MR spectrum Two-dimensional MRS identified increases in the glycerophosphocholine/Iysine cross-peak ratio and corresponding decreases in the phosphocholine/lysine ratio in a dosedependent fashion in TPP-treated cells Lipid metabolic changes are discussed in the light of other MR experiments, and the data indicate that accumulation of MR-visible lipids may arise from the rearrangement of phospholipids accompanying mitochondrial destruction or from the catabolism of phospholipids associated with early events in the cytotoxic process o 1996 Wilty-Liss,h c Current chemotherapy of human neoplasia is limited because the toxicity mechanisms of the drugs act with equal efficacy and on the same targets in normal and malignant cells Consequently, the development of a new class of antineoplastic drugs, which has a unique mechanism of cancer cell selectivity resulting in enhanced and selective toxicity both in vitro and in vivo, would be of great importance Cationic lipophilic compounds exhibit anti-tumour activity in vitro as they selectively inhibit the proliferation of a variety of cultured carcinoma cells relative to untransformed cells (Davis et al., 1985; Herr et al., 1987; Rideout et al., 1989, 1990, 1994) The most common examples of this class of anti-neoplastics are tetraphenylphosphonium chloride (TPP) and the fluorescent dye rhodamine 123, both of which have been extensively used as indicators of mitochondrial membrane potential (Chen, 1988) TPP has subsequently been used as a prototype for the synthesis of a number of selective anti-neoplastic agents, which in certain combinations can covalently self-assemble in situ and act synergistically (Rideout et al., 1989, 1990) The combination of cancer cell selectivity, which is largely independent of proliferative rates, and the ability for synergistic self-assembly potentially make these drugs very effective anti-tumour agents In addition, there is a possibility of less significant side effects, including the decreased risk of secondary tumour development arising from the nuclear mutations induced by classical chemotherapeutic agents (Rideout et al., 1994) TPP-based cationic lipophilic salts have been shown to have selective anti-neoplastic activity in pancreatic carcinoma and Ehrlich lettre ascites (ELA) cells when compared with untransformed monkey kidney fibroblast cells (Rideout et al., 1989, 1990) In addition, other lipophilic cations, including rhodamine 123, have been shown to have anti-tumour activity in mice when administered on a multidose schedule (Herr et al., 1987; Chen, 1988) The cytotoxic mechanism of cationic lipophilic compounds involves the inhibition of mitochondrial function via destruction of the mitochondrial membrane potential The specificity of these agents for malignant cells has been attributed to the high mitochondrial and plasma membrane potentials which characterise carcinoma cells and effect an increase in drug accumulation and retention in the mitochondria (Davis et al., 1985; Rideout et al., 1989) TPP is known to accumulate selectively in the mitochondria of breast carcinoma cells (MCF-7) and has been shown to inhibit oxidative phosphoylation in rat liver mitochondria (Davis et al., 1985; Lampidis et al., 1985) The relationship between selectivity and lipophilicity of lipophilic phosphonium cations has been explained in terms of diffusion rates across membranes and is a key factor in determining the degree of selective anti-neoplastic activity (Rideout et al., 1989) Proton magnetic resonance spectroscopy ('H MRS) has been successfully employed to monitor biochemical changes of intact cells at different stages of neoplastic development (Mackinnon et al., 1994) and activation (Dingley et al., 1992), as well as in response to drug treatment (Guidoni et al., 1994; Pilatus et al., 1994; Ross et al., 1994) and changes in culture conditions (Callies et af., 1993; Delikatny et af., 1996) Of particular interest have been the changes recorded in the mobile lipid region of the spectra, and it has been suggested that this represents a characteristic of the cells in response to stress (Delikatny et al., 1996), necrosis (Kuesel et al., 1994a, b ) or the changes accompanying neoplastic development and proliferation (Mackinnon et al., 1994) Further resolution of these changes has been accomplished with 2-dimensional (2D) MRS, which allows the assignment of MR frequencies to specific chemical moieties (May et al., 1986; Lean et ul., 1992) The aim of the present study was to investigate the effect of TPP on MR-visible metabolites, in particular lipid and lipid metabolites, to further characterise the specific response of these cells to selective anti-neoplastic agents MATERIAL AND METHODS Cell culture DU4475 is a poorly differentiated human breast ductal carcinoma cell line derived from an S.C metastatic nodule Stock cultures of DU4475 cells were grown as suspension culture in RPMI-1640 medium buffered with 0.11% (w/v) sodium bicarbonate and 0.02 M HEPES and supplemented with 10% (v/v) FCS (Cytosystems, Castle Hill, Australia; batch 81012053), 0.03% (w/v) L-glutamine, 250 units/l human insulin and 0.1% (v/v) gentamycin Growth-inhibition assays DU4475 cells were seeded (180 pI at 2.5 x 105/ml) in 96-well plates, and TPP was added hr later TPP was prepared fresh prior to each experiment as a 10 mM stock *To whom correspondence and reprint requests should be sent, at Department of Cancer Medicine, Blackburn Building, D06, University of Sydney, Sydney, NSW, 2006, Australia Fax: (61) 2-351-4105 'Present address: Department of Pathology, Division of Experimental Pathology,The Albany Medical College, Albany, NY, 12208, USA 6Present address: San Diego Regional Cancer Center, 11099 North Torrey Pines Road, Suite 250, La Jolla CA 92037, USA Received: October 6,1995 and in revised form February 3,1996 TPP EFFECTS ON 'H MR SPECTRA solution in PBS (140 mM NaCI, 0.35 mM KH2PO4, 0.35 mM Na2HP04) and then diluted in the same buffer TPP doses were added as 20 pl aliquots per well to a final concentration of 0.25-100 pM, and control wells each received 20 pl of the drug vehicle (PBS) At the end of 48 hr, cells were counted on a haemocytometer and their membrane integrity measured by the exclusion of Trypan blue, present at a final concentration of 17% (v/v) At least 200 cells were counted per treatment in each experiment In companion studies, pCi of [methyL3H] thymidine (specific activity 20 Ciimmol; ICN Seven Hills, Australia) was added to each well during the last hr of the drug treatment period Cells were subsequently harvested using a Titertek (ICN) multiple cell harvester, washed, hypotonically disrupted and precipitated onto glass fibre discs Air-dried discs were suspended in scintillation fluid (Canberra-Packard, Five Dock, Australia) and emissions counted using a LKB Wallac (Bromma, Sweden) 1217 Rackbeta liquid scintillation counter 73 H3N+ CH2-CH2- cross-peak on the corresponding side of the diagonal at 1.67,3.00 ppm In 2D spectra, this cross-peak is prominent, well removed from the diagonal and relatively easy to integrate This cross-peak may contain contributions from other polyamines, such as arginine and spermidine (Cerdan et aL, 1988; Sze and Jardetsky, 1990; Kuesel et aL, 1994a, b), and, in whole cells, a variable contribution from lysine residues in proteins and peptides with differing motional properties However, the lysine cross-peak appears to be of reasonably constant intensity in a number of cultured cell models (Mackinnon et al., 1994; Delikatny et al., 1996) and in human biopsy systems (Mackinnon et al., 1995) RESULTS TPP causes a concentration-dependent increase in MR-visible lipid The addition of TPP to DU4475 cells causes a dosedependent loss of membrane integrity as measured by Trypan blue exclusion with a concomitant decrease in thymidine incorporation (Fig 1) The cell density after TPP treatment MR spectroscopy MR spectra were acquired on a Bruker (Karlsruhe, Ger- was either lower or equal to the seeding density (2.5 x lo5/ many) AM360 wide bore MR spectrometer equipped with a ml), reflecting cytotoxic and cytostatic concentrations, respec5-mm dedicated proton probehead Samples were spun at 20 tively The ICso for the depletion of cells excluding Trypan blue Hz and the temperature was regulated at 37°C using a Bruker was 1.3 pM TPP (n = 5) Using this TPP concentration range, it was not possible to calculate a true ICsofor the attenuation VTlOOO temperature regulation unit of thymidine incorporation (Fig 1) However, thymidine Sample preparation DU4475 ce:lls were seeded (135 ml at incorporation was more sensitive to the actions of TPP than 2.5 x 105/ml)into 175-cm2culture flasks and TPP (or PBS as a Trypan blue exclusion control) added hr later to a concentration of 0.25-5 pM For Subsequently, the DU4475 cells were treated with TPP at kinetic studies, cells were treated at a constant concentration concentrations up to pM and the 1D and 2D MR profiles of of pM Cells were subsequently processed for MRS at the treated cells documented The 1D proton MR spectra of various times after drug addition in the kinetic studies and at TPP-treated and control DU4475 cells are shown in Figure 48 hr for the dose-response studies No medium changes were The spectrum of untreated DU4475 cells displays a number of made during the drug treatment period An average of x lo7 resonances arising mainly from amino acids/proteins, with cells were used for every experiment Cells were collected by some resonances which also can be assigned to cholinecentrifugation (400g), resuspended and washed in ml PBS/ containing metabolites, creatines, carbohydrates, lactate and D a total of times before finally being suspended in 400 p1 HEPES buffer (Table I) The 1D proton MR spectra of the PBS/D20in a 5-mm MR tube for MR measurements DU4475 cell line treated with and pM TPP, respectively, ID ' H spectra One-dimensional spectra were obtained at are shown in Figure 2b and c These spectra are chacacterised 37°C with gated pre-saturation of the residual water signal A by an intense methylene resonance at 1.3 pprn and correspondrelaxation delay of sec was used and 128 8K transients were ing increases in other lipid resonances, including the methyl acquired A Hz line-broadening was applied before Fourier resonance at 0.9 ppm and the olefinic resonance at 5.2 pprn transformation to increase apparent signal/noise, and spectra (Table I) The increase in MR-visible lipid, as measured by the were baseline-corrected using a 4th-order polynomial Changes ratio of the methylene/methyl resonances (CH2/CH3) at in the lipid resonance were quantified by changes in peak 1.3/0.9 ppm, respectively, is shown in Figure At TPP concentrations 20.5 pM, a prominent increase in ithis ratio height ratios in 1D spectra ' H spectra Magnitude-mode correlated spectroscopy was observed in a concentration-dependent fashion ( p < 0.01 (COSY) spectra consisting of 200 free induction decays (FIDs) of 32 transients ( < hr) were collected with gated water pre-saturation at 37°C A spectral width in the F2 domain of lo4 I 2,874 Hz (8 ppm) was used, and 2K time domain points in t2 (sequential quadrature detection) were acquired (acquisition lo4 -C time = 356 msec) The initial evolution delay between the pulses, t l , was 1.005 msec with an increment, At,, of 348 psec a During a 1-sec relaxation delay, continuous-wave irradiation lo4 (23 dB below 0.2 W) was used to pre-saturate the water resonance One-dimensional spectra were acquired before and x after the COSY spectrum to check on sample stability lo4 Two-dimensional spectra were zero-filled to 1K complex Q V points in tl, and 2D Fourier transformation was performed with a sine-bell window function in t l and a LorentzianGaussian window (Lorentzian line-broadening parameter [LB] = -30, Lorentzian-Gaussian lineshape transformation parameter [GB] = 0.12) in t2 (Delikatny et al., 1991) Changes in 2D spectra were measured by integration of cross-peak volumes in 2D COSY spectra Cross-peak volumes were FIGURE1- Dose-response cuwe of DU4475 cells treated with measured on either side of the diagonal in unsymmetrised TPP for 48 hr as measured by Trypan blue exclusion and spectra and calculated relative to the lysineipolyamine 3H-thyrnidineuptake Trypan blue exclusion, 3H-thymidine z : v 74 DELIKATNY ETAL TABLE I - ASSIGNMENT OF MAJOR RESONANCES IN 1D PROTON MR SPECTRA OF DU4415 HUMAN BREAST CELLS’ -CHp Chemical shift inom) Lipid’ 0.89 1.28 1.58 2.03 2.24 2.88 Molecule Fatty acyl chain Fatty acyl chain Fatty acyl chain Fatty acyl chain Fatty acyl chain Fatty acyl chain 5.32 Fatty acyl chain Amino acids 0.9-1.0 Neutral amino acids (Val, leu, ile) Threonine 1.33 Alanine1.49 Lysine/ polyamines 1.71 2.17 3.02 porn , FIGURE - One-dimensional proton MR spectra of the human breast cell line DU4475 showing control and treatment with TPP for 48 hr (a) Control, (b) ,uM TPP, (c) FM TPP Spectra (128 8K accumulations) were collected at 37°C at 360 MHz with the sample spinning at 20 Hz A line broadening of Hz was applied to each spectrum before Fourier transformation 3.78 Other 1.33 2.08 2.98 3.02 Time dependence of TPP-induced changes The dependence of the increases in MR-visible lipid on length of exposure to pM TPP (or PBS as control) was investigated, and the results are shown in Figure Increases in the CH2/CH3 ratio were evident after as little as hr of TPP treatment The value of the CH2/CH3 ratio continued to increase, almost linearly (R = 0.91, linear regression analysis), for up to 99 hr of exposure In control samples, to which an equal volume of PBS had been added, the CHz/CH3 ratio was relatively high immediately after drug addition and declined gradually to 99 hr (R = 0.87) CHr- CH3-CHCH34HH3N+-CHZ-CH*CH2Glutamate/glutamine CH-CH, CHqCOO-j-NH3+ Lysine H3N+-CH*-CH2CH2Alanine /glu /gln a-protons CH3-CHCH3-COON-CH2-CHZ-SO3 CH3-N-CHz- 3.02 Phosphocreatine CH3-N-CH2- 3.21 3.21 3.21 -N+(CH3)3 3.23 Choline Phosphocholine Glycerophosphocholine HEPES 3.25 3.25 Inositol Taurine HO-CH2CHL-N-CH-CH(0H)H3N+-CHz-CHz 3.43 Taurine H3N +-CH2-CH2 3.87 HEPES 7.91 at [TPP] 0.5 pM, Student’s t test) The increase in methylene signal was also significant when measured relative to the HEPES resonance at 3.9 ppm or to the lysine/lipid peak at 1.7/1.6 ppm (data not shown) CH3-CHrCH2-CH’ kM In particular, cross-peaks arising from protons on unsaturated acyl chains (C and D) and from the glycerol moiety of the head group (G and G’) were greatly enhanced Spectral differences in TPP-treated DU4475 cells were also observed in cross-peak ratios arising from the -CH2-CH2moieties of choline (cho), phosphocholine (PC) and glycero- TF'P EFFECTS ON 'HMR SPECTRA 0 B 20 I 40 60 80 I 100 Time (hours) FIGURE - Dose-dependent changes in methylene/methyl (CHJ CH3; 1.3/0.9 ppm) ratios from 1D proton MR spectra of DU4475 cells treated with UM TPP Data show the results of independent experiments performed on different days; pM TPP 0control phosphocholine (GPC) at 3.51, 4.05 ppm; 3.56, 4.17 pprn and 3.67,4.33 ppm, respectively, measured relative to the H3N+C H - C H r cross-peak of lysine and polyamines at 1.67,3.00 ppm The integrals of the PC/lysine cross-peaks gradually decreased with increasing TPP (p < 0.01 at [TPP] 0.5 pM), whereas those of GPC/lysine increased in drug-treated spectra ( p < 0.01 at [TPP] pM), as s.hown in Figure 6b and c The cho/lys integral ratio remained relatively constant throughout the concentration range studied (Fig 6u) These measured ratio differences are in accordance with the visual changes observed in cho, PC and GPC: cross-peaks in 2D COSY spectra, validating the use of the lysine polyamine cross-peak as an internal reference Current studies in our laboratory also indicate that TPP treatment of the HBL-100-transformed human breast cell line (up to 100 pM) does not cause significant changes in the lysine polyamine cross-peak relative to the cross-peak from an external standard of 10 mM p-aminobenzoic acid (data not shown) DISCUSSION TPP-induced changes to proton Mi? spectra The advantage of MR studies is that they enable the measurement of global changes in cell metabolism and structure non-invasively and so allow an assessment of the alterations accompanying drug treatinent in intact cells Specifically, MRS allows the measurement of increased mobile MRS lipids accompanying TPP treatment The MR-visible lipid typically represents a small fraction of the total lipid content of the cell, and the observed changes may represent a redistribution of existing lipids (see below) rather than a biochemical modification As such, isolation of the mobile lipid fraction from the bulk of non-MR-visible lipids is difficult and proton MR remains one of the only methods capable of observing such cellular changes In the current study, we have investigated the effects of TPP on the metabolism of DU4475 cells in terms of lH MR-visible metabolites as a first step to determining those changes which reflect the selective cytotoxic action of this drug on neoplastic cells In DU4475 cells, TPP was found to induce cytotoxicity in a dose-dependent manner at micromolar concentrations In companion studies, it was also found that TPP attenuated thymidine incorporation to a greater extent than Trypan blue exclusion The fact that TPP attenuates DNA synthesis at lower concentrations than the destruction of membrane integrity may reflect the greater energy requirements needed for cell division This is consistent with the direct effects of TPP, whereby the loss of ATP production due to the inhibition of mitocliondrial function causes an 75 inhibition of DNA synthesis However, it is also possible that the salvage of purines may be particularly sensitive to the actions of TPP The major changes measured by 'H MRS included a 3-fold increase in the methylene/methyl ratio at 1.3/0.9 pprn as well as increases in other lipid resonances and significant changes in 2D cross-peak ratios assigned to PC and GPC relative to lysine (and polyamines) The increase in MR-visible lipid signal was dose-dependent and began at concentrations below the IC,,, and kinetic studies demonstrated that the MR modifications began as early as hr after TPP treatment These results show that MRS monitors early events in the cytotoxic process occurring at low TPP concentrations It is of interest that high CH2/CH3 ratios were observed in both control and treated cells examined early after seeding, indicating the presence of substantial amounts of mobile lipid when the cells were in lag phase (Fig 4) The reasons for high CH2/CH3 ratios immediately after subculturing are uncertain but may represent the uptake of lipids from lipoproteins present in the serum of fresh medium Most resonances in proton spectra of whole cells contain contributions from several metabolites (Tables I, I]), which makes the choice of a denominator for ratio calculations difficult It is important to point out that a lH MR spectrum arising solely from lipid would display a constant CH2/CH3 ratio of approximately 7.3 (based on the palmitate molecule C16:0, which has 22 protons that resonate at 1.3 ppm and protons at 0.9 pprn) The existence of substantial amino acid methyl resonances at 0.9 ppm, even in the absence of mobile lipids, as shown in Figures 2a and 5a, means that increases in mobile lipid can still be measured using the CH2/CH3 ratio At high mobile lipid concentrations, this ratio will level off at a value dependent on the baseline concentration of amino acid methyl groups, which may explain the shape of the curve in Figure Increased mobile lipid occurs in a number of vstems The modifications in lH MR spectrum accompanying TPP treatment are consistent with changes associated with cellular stress and cytotoxicity measured by others For example, a correlation between necrosis and increased proton lipid methylene MR signal per unit weight of tumour tissue has been demonstrated in human brain biopsies (Kuesel et al., 1994a, ) In addition, increased CH2/CH3 ratios have been reported in HeLa cells treated with the glycolysis inhibitor londamine (Guidoni et al., 1994) Qualitative increases in the resonance at 1.3 ppm have also been reported in perfused MCF-7 breast cells treated with the differentiating drug phenyl acetate (Pilatus et al., 1994) and in implanted rat glioma tumour cells upon treatment with the anti-viral agent ganciclovir (Ross et al., 1994) Whereas in this study, where 2D MRS was used to show clear increases in mobile lipid, it is uncertain from the data presented by Pilatus et al (1994) and Ross et al (1994) whether the reported changes arise from an increase in lipid methylene resonances or from the methyl groups of lactate and threonine Mobile lipid accumulation measured by proton MR can also be induced as a function of culture conditions A qualitative increase in the methylene resonance has been observed after addition of exogenous fatty acids to the medium of AgX63.653 myeloma cells (Callies et al., 1993), and increases in the CH2/CH3ratio have been reported in L cells under conditions of high density, serum deprivation or low pH (Delikatny et af., 1996) Qualitative accumulation of mobile lipid has also been observed in CCRF-CEM cells made resistant to vinblastine (Holmes et al., 1989) These studies have also used 2D MRS to confirm the lipid nature of the increased 1.3 ppm resonance Taken together, these data indicate that increased mobile lipid in malignant cells may belie changes in tumour cells arising 76 DELIKATNYETAL -1 -2 -3 -4 -5 -PPm I PPm , , I , , , -1 -2 -3 -4 -5 B -PPm FIGURE - Representative uns mmetrised 2D COSY spectra of (a) control DU4475 cells and (b) DU4475 cells treated with pM TPP for 48 hr Spectra of x 10Y cells were collected at 37°C at 360 MHz with the sample spinning at 20 Hz Sine bell and Lmentzian-Gaussian (LB = -30; GB = 0.12) window functions were applied in the t l and t2 domains, respectively before Fourier transformation Relevant cross-peak assignments are labelled TPP EFFECTS ON 'H MR SPECTRA 77 average inter- and intramolecular dipolar interactions Phospholipids in bilayers not generally experience enough motional narrowing to be observed using high-resolution MR techniques (Davis, 1983; Bloom and Thewalt, 1995) In addition, the cross-peak arising from the CH2-CH2 moiety of GPC is indistinguishable from the cross-peak arising from the head group of phosphatidylcholine within the resolution obtainable in whole cells Thus, the increase in the GPC/lysine ratio indicates an increased mobility of the CH2-CH2 residue of either GPC or the head group of phosphatidylcholline This increased mobility could result from either rearrangement of existing molecules into more mobile forms or increases in 0.1 J++ concentration arising from chemicalmodification or both.d 10 Therefore, one possible explanation for the observed GPC/ [TPP] p M lysine ratio increases involves the rearrangement of membrane phosphatidylcholine into rapidly tumbling, and hence MRvisible, vesicular structures as a result of destruction of the mitochondrial membrane with TPP treatment This hypothesis is supported by ultrastructural studies in which TPP damaged mitochondrial cristae and caused pronounced membrane blebbing in FaDu hypopharyngeal squamous carcinoma cells (Rideout et al., 1994) However, it is also possible that TP;P induces catabolism of phosphatidylcholine by phospholipases A1 and A2 This would result in increased concentrations alf GPC, a small and relatively mobile metabolite, while simultaneously releasing fatty acyl chains to be incorporated into neutral lipids This would concur with reports showing increased concentrations of cytoplasmic lipid droplets in hypoxic (Freitas et al., 1990) and necrotic (Kuesel et al., 1994b) tumours and would indicate that the TPP-induced accretion of molbile lipids represents a more general phenomenon associated with cell stress and/or cell death Decreases observed in the PC/lysine cross-peak volume ratio imply decreased concentration or mobility in the CH2CH2 moiety of the PC molecule Since PC is a soluble, and hence relatively mobile, membrane precursor, this decrease can be correlated with a reduced PC concentration One possible mechanism would involve inhibition of the ATPdependent enzyme choline kinase, resulting in reduced PC levels Previous studies using 31P MR have correlated decreases in PC with a reduction in membrane synthesis arising 10 from decreased activity of choline kinase (Daly et al., 1990; [TPPI PM Kuesel et al., 1990) It is also possible that TPP causes attenuation of the activity of phospholipase C, which would FIGURE - Cross-peak volume ratios from the CH2-CH2 cross- also cause decreased PC levels via a catabolic mechanism peaks of (a) choline, (b) phosphocholine and (c) glycerophosphoIn addition to mitochondrial accumulation as a result of choline relative to the H3N+-CH2 CH2 cross-peak of lysine and polyamines at 1.67,3.00 ppm as a function of 48-hr TPP treatment membrane potentials, the selectivity of specific cationic lipophilic phosphonium compounds for carcinoma cells has been from cell stress or cell death associated with drug treatment In explained in terms of diffusion rates across membranes In fact, intracellular lipid accumulation, or fatty change (Robbins particular, a correlation has been noted between drug lipophilicet al., 1984), has been shown histopathologically to occur in ity (as measured by the octanol water partition coefficient) and cells adjacent to those that have undergone necrosis More- drug selectivity In addition, 3H TPP has been :shown to over, fatty change is considered to be an indicator of non-lethal concentrate intracellularly 77-fold in PaCa-2 human pancreinjury and sometimes the harbinger of cell death (Robbins et atic carcinoma cells within hr of treatment (RidelDut et al., al., 1984) 1989) This points to an uptake mechanism involving diffusion across the membrane and pre-concentration in the cytoplasm Metabolite levelsfvom spectroscopy The observations made here with 2D NMR allow a more followed by localisation in the mitochondria This localisation detailed analysis of the potential changes underlying the is accompanied by a rearrangement or chemical reorganisation toxicity of TPP Changes observed in phospholipid metabolites of lipids leading to a more mobile lipid fraction that can be suggest that TPP-induced increases in mobile lipid arise from observed with high-resolution 'H MRS These observations either mitochondrial membrane disruption or intracellular suggest the following hypothesis for the cytotoxic actions of neutral lipid accumulation accclmpanying lipid catabolism TPP Once inside the cell, TPP either alters the activity of the Significant increases in the GPCilysine cross-peak ratio were enzymes involved in phospholipid metabolism, which results in observed in TPP-treated cells at concentrations 0.5 pM The the generation of cytoplasmic neutral lipid at the expense of visibility of a high-resolution MR signal depends in part on the choline-bearing phospholipids, or causes an accumulation of presence of isotropic molecular motion sufficiently rapid to phospholipids resulting from the physical destruction of the DELIKATNY ETAL 78 TABLE I1 - ASSIGNMENT OF MAJOR CROSS-PEAKS IN 2D PROTON MR SPECTRA OF DU4475 HUMAN BREAST CELLS Molecules Lipid Fatty acyl chains Glycerol backbone Abbreviation A B C D E F G (vicinal) G' (geminal) Arachadonic acid Amino acids Alanine Aspartate I Asparagine Ala AspIAsn Glutamate/Glutaminel GluIGln Glutamate Isoleucine Glu Ile Leucine Lysine/polyamines2 Threonine Tyrosine Valine Phospholipid metabolism Choline Leu !iE % Cho Ethanolamine GlyceroPC Eth GPC Phosphocholine PC Phosphoethanolamine PE Other HEPES HEPES HEPES Hypotaurine Inositol Lactate Taurine Pyrimidine diphosphate Inos Lac Tau CDP UDP Ribose Unassigned Polyamine? Coupling partners (bold face) Cross- eak coord &pm) -(CH2),-CHZ-CH -CH=CH-CHz-CHz-CHz-CHz-CH