Rhodamine-123 selectively reduces clonal growth of carcinoma cells in vitro.

Rhodamine-123, a cationic laser dye, markedly reduced the clonal growth of carcinoma cells but had little effect on nontumorigenic epithelial cells in vitro. This selective inhibitory effect of Rhodamine-123 on some carcinomas is unusual since known anticancer drugs, such as arabinosyl cytosine and methotrexate, have not been shown to exhibit such selectivity in vitro.

Anticarcinoma activity in vivo of rhodamine 123, a mitochondrial-specific dye.

Carcinoma cells and normal epithelial cells differ in the mitochondrial retention of a permeant cationic compound, rhodamine 123. The possibility of utilizing this difference in carcinoma chemotherapy was investigated. Rhodamine 123 exhibited anticarcinoma activity in mice, and this activity was potentiated by 2-deoxyglucose.

Reversal of resistance to rhodamine 123 in adriamycin-resistant Friend leukemia cells.

Pleiotropic resistance to rhodamine 123 (Rho-123) in Adriamycin (ADM)-resistant Friend leukemia cells was circumvented by cotreatment with 10 microM verapamil. Increased cytotoxicity corresponded to higher intracellular Rho-123 levels. The verapamil-induced increase of drug accumulation in resistant cells is accounted for at least in part by the blockage or slowing of Rho-123 efflux from these cells. In contrast, accumulation and consequent cytotoxicity of Rho-123 in sensitive cells are not increased by verapamil. Similar results were obtained when ADM was used in this cell system. These results suggest that the efflux system for Rho-123 and ADM in sensitive cells is either reduced or absent. Although Rho-123 accumulates specifically in mitochondria and ADM mainly in the nucleus, the loss of these two different classes of compounds from resistant cells appears to occur via a similar or common mechanism. The similarities in drug transport between Rho-123 and ADM may have important implications when applied to an in vivo environment.

Selective toxicity of rhodamine 123 in carcinoma cells in vitro.

The study of mitochondria in situ has recently been facilitated through the use of rhodamine 123, a mitochondrial-specific fluorescent dye. It has been found to be nontoxic when applied for short periods to a variety of cell types and has thus become an invaluable tool for examining mitochondrial morphology and function in the intact living cell. In this report, however, we demonstrate that with continuous exposure, rhodamine 123 selectively kills carcinoma as compared to normal epithelial cells grown in vitro. At doses of rhodamine 123 which were toxic to carcinoma cells, the conversion of mitochondrial-specific to cytoplasmic-nonspecific localization of the drug was observed prior to cell death. At 10 microgram/ml, greater than 50% cell death occurred within 7 days in all nine of the carcinoma cell types and lines of different origin studied, while six of six normal epithelial cell types and lines remained unaffected. Cotreating carcinoma cells with 2-deoxyglucose and rhodamine 123 enhanced the inhibition of growth by rhodamine 123 alone in clonogenic survival assays. The observation of the selective toxicity of rhodamine 123 appears to be unique in view of the absence of selective toxicity reported in vitro for the various antitumor agents currently in clinical use. Preliminary results with rhodamine 123 in animal tumor systems indicate antitumor activity for carcinomas.

Structural and functional effects of adriamycin on cardiac cells in vitro.

The effects of Adriamycin (ADR) on the heart, seen clinically as transient electrocardiographic changes and cardiomyopathy, have been simulated in an in vitro cardiac cell system. Structural and functional alterations in cultured heart cells can be dissociated based upon ADR dose and length of exposure. At high ADR doses (100 to 200 micrograms/ml), cessation of beating was rapid, and structural changes consistent with the in vivo cardiomyopathic picture (vacuolization and nucleolar fragmentation) were observed. At low ADR doses (0.1 to 0.5 micrograms/ml), arrhythmias were produced in the absence of ultrastructural changes (within 48 hr); the incidence and severity of the arrhythmias were demonstrated to be dose dependent. Continued treatment of cultures at low dose levels for sustained periods of time (up to 17 days) resulted in a striking loss of muscle fiber without concomitant vacuolization and nucleolar fragmentation. An intermediate ADR dose of 10 micrograms/ml for 1 hr exposure caused vacuolization and cessation of beating, with lysis of cells within 72 hr. The parallel between the effects of ADR on in vitro cardiac cell structure and function with those seen in vivo suggests that this simple system may have value in studies directed towards the mechanism of ADR-induced cardiac toxicity and in the screening of anthracycline analogs for their potential effects on the heart.

Cross-resistance to rhodamine 123 in Adriamycin- and daunorubicin-resistant Friend leukemia cell variants.

Cross-resistance to rhodamine 123 (Rho-123) has been found in Adriamycin (ADM)-resistant and daunorubicin (DNR)-resistant Friend leukemia cell variants. Cytotoxicity in sensitive cells correlates with the intracellular amount of Rho-123, as determined by high-pressure liquid chromatography. Differential resistance coincides with Rho-123 accumulation which in sensitive cells was 20-fold higher than in resistant cells after 180 min of treatment. Sodium azide, which has been shown to inhibit ADM efflux and consequently increase drug accumulation in ADM-resistant cells, did not inhibit Rho-123 efflux. The difference in Rho-123 accumulation between sensitive and resistant cells correlates with cytotoxicity, which is in contrast to what has been found in these cells when treated with either ADM or DNR. Moreover, in contrast to the known effects of ADM and DNR on macromolecular synthesis, Rho-123 in sensitive cells was found to inhibit protein synthesis but had no effect on DNA or RNA synthesis. At Rho-123 doses which inhibited protein synthesis, drug localization changed from mitochondrial specific to generalized cytoplasmic. This effect was never achieved in resistant cells, even with prolonged drug exposure. The relevance of these findings is that different mechanisms of resistance to different drug types can be identified in the same cells even though similar resistance occurs. The similarity in resistance need not share a common mechanism. Although the drugs are effluxed more efficiently in resistant cells, the mechanisms for resistance in each case seem to differ. In the case of ADM and DNR, it appears to be multifactorial, whereas with Rho-123, total intracellular accumulation seems to be most important.

Structural requirements of simple organic cations for recognition by multidrug-resistant cells.

We previously noted that a wide variety of drugs which are recognized by multidrug-resistant cells (MDR+) are positively charged. However, it remains unclear why and how such a large number of structurally different compounds can be distinguished by MDR+ cells. The majority of the diverse compounds subject to MDR are complex and thereby complicate definitive structure/function characterization of the P-glycoprotein-mediated MDR mechanism. Using a series of simple aromatic (alkypyridiniums) and nonaromatic (alkylguanidiniums) organic cations differing in their lipophilicity by stepwise additions of single alkyl carbons, we demonstrate by growth inhibition studies that a single aromatic moiety and a critical degree of lipophilicity (log P > -1) are required for recognition of these simple organic cations by MDR+ cells. Thus, MDR+ cells are not cross-resistant to the nonaromatic guanidiniums but do show cross-resistance to those aromatic pyridiniums with chain lengths > four. Resistance ratios, as determined by comparison of 50% inhibitory doses in MDR- versus MDR+ cells, increase as a function of increasing chain lengths of these latter simple aromatic compounds. Resistance to pyridinium analogues in MDR+ cells is reversible by co-treatment with nontoxic doses of verapamil. Preliminary uptake data with radioactive analogues further implicate the MDR mechanism of lowered drug accumulation in accounting for resistance to the pyridinium homologues. Utilization of these simple organic cations provides a rational basis for better defining the physical chemical properties of more complex compounds processed by the MDR mechanism and suggests a strategy for designing chemotherapeutic agents with reduced susceptibility to MDR.

Relationship of multidrug resistance to rhodamine-123 selectivity between carcinoma and normal epithelial cells: taxol and vinblastine modulate drug efflux.

Preferential retention and cytotoxicity of Rhodamine-123 (Rho-123) was originally reported in a number of carcinoma cell types isolated from a variety of tissues as compared to normal epithelial cells from a limited number of other tissues. In the present study, we have examined Rho-123 selectivity in normal and tumor cell lines isolated from the same tissue source, i.e., human breast. We found that: (a) in matched pairs of normal and carcinoma breast cells, Rho-123 displays no preferential retention in either cell type; (b) there is no preferential toxicity in carcinoma as compared to normal breast cells; in fact, one of the carcinoma cell lines (MDA-MB231) shows moderate resistance to this dye; (c) all of the human breast cell lines do not express P-glycoprotein-mediated multidrug resistance; (d) the normal monkey kidney epithelial cell line CV-1, which was originally used as a model to demonstrate the relative resistance of normal epithelial cells to this drug, is found to express high levels of the mdr-1 gene, is resistant to other multidrug-resistant drugs (taxol and vinblastine), and its resistance to Rho-123 as well as decreased Rho-123 retention can be reversed by verapamil; and (e) taxol and vinblastine are found to block increased Rho-123 efflux in CV-1 cells. Thus, overall the data suggest that preferential retention and cytotoxicity of Rho-123 in carcinoma versus normal epithelial cells is related to the differential expression of the mdr-1 gene.

3'-Deamino-3'-morpholino-13-deoxo-10-hydroxycarminomycin conquers multidrug resistance by rapid influx following higher frequency of formation of DNA single- and double-strand breaks.

The mechanism of action of 3'-deamino-3'-morpholino-13-deoxo-10-hydroxycarminomycin (MX2) was examined in a human leukemia cell line (K562) and its Adriamycin (ADM)-resistant subline (K562/ADM). ADM and MX2 showed an equivalent antitumor effect against K562. K562/ADM was highly resistant to ADM. In cellular pharmacokinetic studies, MX2 showed faster and greater influx than did ADM in both K562 and K562/ADM. The efflux of ADM was rapid in K562/ADM but not in K562. On the other hand, the efflux of MX2 was rapid in both cell lines. The formation of DNA single-strand breaks and double-strand breaks by ADM was significantly lower in K562/ADM than K562. On the other hand, formation of those breaks by MX2 was not decreased. Although some of the DNA breaks induced by MX2 were resealed, there was no difference in the degree of resealing in K562 and K562/ADM cells. On the other hand, most of the small number of DNA breaks in K562/ADM induced by ADM were resealed. The topoisomerase II activity in K562 and K562/ADM was not significantly different. It is concluded that MX2 conquers multidrug resistance by rapid influx following a higher frequency of formation of DNA single- and double-strand breaks in K562/ADM cells.

Accumulation of simple organic cations correlates with differential cytotoxicity in multidrug-resistant and -sensitive human and rodent cells.

Structure/functional studies previously reported showed that in a series of simple organic cations in which the charge is delocalized, an aromatic ring and a minimal degree of lipophilicity (log P > -1) were required for recognition by murine cells which express P-glycoprotein (p-gp)-mediated multidrug resistance (MDR). In the present report we find that 3H-octylpyridinium, the simple aromatic cation which has been shown to be preferentially toxic to MDR- as compared to MDR+ cells, accumulates 4.7-fold greater in the MDR- cell line. In contrast, we find that 3H-guanidinium which displays no selective toxicity between MDR+ and MDR- cells, shows no significant uptake differences between these two cell types. We also present data which demonstrate that other organic cations which contain aromatic rings, a minimal degree of lipophilicity (log P> -1) and carry a delocalized (Rho 123) or shielded (triphenylmethyl phosphonium) positive charge, also accumulate to a greater degree in MDR- vs MDR+ cells. Additionally, we find that human cells which express p-gp MDR, have similar requirements for recognition of these simple compounds. In fact, the sensitivity profiles of these compounds closely correlate between murine and human cell lines. It was also found that none of the series of simple organic compounds tested showed modulatory activity in MDR+ cells, as assayed by monitoring retention of Rho 123. Thus, the requirements for MDR recognition vs those for MDR modulation are clearly distinguished with these simple structured compounds. In comparison, the calcium channel antagonist, verapamil, and a calcium channel agonist, Bay K 8644, both showed modulatory activity by increasing Rho 123 retention in MDR+ cells, further supporting the interpretation that verapamil's modulation of MDR is unrelated to its action on calcium flux. Overall, the data presented here add further information for defining the structural requirements of compounds for their recognition by, or modulation of, human cells expressing p-gp-mediated MDR.

Membrane potential differences between adriamycin-sensitive and -resistant cells as measured by flow cytometry.

Using the fluorescent membrane potential probe, 3,3'-dihexyl-oxacarbocyanine (DiOC6(3], we found a 4-fold higher uptake in Adriamycin (ADM)-sensitive versus -resistant Friend leukemia cells (FLC). When sensitive cells were treated in the presence of high potassium (120 mM K+), there was a greater than 80% reduction of DiOC6(3) uptake. Using carbonylcyanide 4-trifluoromethoxy-phenylhydrazone (FCCP), a specific inhibitor of mitochondrial membrane potential, DiOC6(3) accumulation was reduced by less than 30% in these cells. Both results support the conclusion that a greater uptake of DiOC6(3) in ADM-sensitive than in -resistant cells indicates an increased plasma transmembrane potential. Since electronegative plasma membrane potentials are a driving force for the transport of lipophilic positively-charged compounds, differences in membrane potentials between sensitive and multiple drug resistant (MDR) tumor cells could have an important influence on drug accumulation and cytotoxicity. The drugs which our ADM-resistant FLC display multiple drug resistance to are positively charged. In MDR FLC, the calcium channel antagonist, verapamil, has been shown to block the efflux of Rhodamine 123 (Rho 123) and other positively-charged compounds. Since DiOC6(3) is also positively-charged, we used verapamil to investigate its effects on drug uptake. In MDR FLC, verapamil increased DiOC6(3) accumulation by 1.9-fold, whereas in sensitive cells it was increased 1.5-fold. In contrast, verapamil increased the levels of Rho 123 in resistant cells 7.8-fold but lowered them in sensitive cells 1.5-fold. The minimal loss of DiOC6(3) from both sensitive and MDR cells and the above results can best be interpreted as indicating that DiOC6(3) is not transported by the efflux "pump" system but that verapamil induces a plasma membrane potential increase in sensitive and resistant cells that DiOC6(3) is sensitive to. On the other hand, since Rho 123 did appear to be actively effluxed from these resistant cells, the enhancement of this compound by verapamil was more likely due to inhibition of the MDR "pump." How, or whether, plasma membrane potentials and the MDR efflux "pump" are related remains to be investigated. In the resistant cells, verapamil also induced an increase (13-fold) in the accumulation of the electrically neutral fluorescent probe for calcium, INDO-1/AM. However, verapamil had no effect on the efflux of this compound, which was equivalent in both resistant and sensitive cells. Thus, a new effect of verapamil on drug accumulation in MDR cells is identified here.

Rho(0) tumor cells: a model for studying whether mitochondria are targets for rhodamine 123, doxorubicin, and other drugs.

A human osteosarcoma cell line devoid of mitochondrial DNA (rho(0)) and its wild-type parental cell counterpart (wt) are presented as a model to investigate drug targeting. By virtue of the absence of mitochondrial DNA, rho(0) cells cannot perform electron transport or oxidative phosphorylation. Since most of the drugs studied are transported by the efflux pumping systems controlled by the MDR1 and MRP1 genes, both cell lines were examined for the expression of these genes, and it was found that no MDR1 and only low amounts of MRP1 were expressed. Growth inhibition experiments indicated that doxorubicin (Dox), vinblastine, and paclitaxel were equitoxic in these cell lines. On the other hand, the IC(50) for rhodamine 123 (Rho 123) in rho(0) cells was 50 times higher than in wt cells. This result correlates with a lower accumulation of Rho 123 in rho(0) cells as measured by fluorescence microscopy and flow cytometry (3 times less than in wt cells). In contrast, when stained with Dox, both cell types accumulated similar amounts. Surprisingly, in these non-P-glycoprotein expressing cells, verapamil increased both Dox and Rho 123 retention. Overall, these data suggest that: (i) functional mitochondria do not appear to be targets for the growth inhibitory activities of Dox, paclitaxel, or vinblastine; (ii) for lipophilic cations like Rho 123, however, normal functioning mitochondria and maintenance of a normal mitochondrial membrane potential (Deltapsi(mt)) appear to play a critical role in the intracellular accumulation and subsequent cytotoxicities of these compounds; and (iii) verapamil increases drug accumulation in non-P-glycoprotein expressing cell lines, most likely by direct action on Deltapsi(mt) for Rho 123 and safranin O, and on heretofore unidentified plasma membrane transporters, as well as via interaction with low levels of MRP1, for Dox. These results should be considered when Rho 123 and verapamil are used to detect P-glycoprotein.

Relevance of the chemical charge of rhodamine dyes to multiple drug resistance.

Previously, we have shown that multiple drug resistant (MDR) Friend leukemia cells (FLC) are cross-resistant to the positively-charged dye, Rhodamine 123 (Rho 123), and that this resistance can be reversed by verapamil (VER). In the present study we used two zwitterionic rhodamine analogs, Rhodamine 116 and Rhodamine 110, and another positively-charged analog, Rhodamine 6G, to determine whether drug accumulation, resistance and modulation were affected by changes in the charge of these compounds. While there was no differential sensitivity between sensitive and resistant FLC to zwitterionic rhodamines, there was marked differential toxicity between these cell types for the positively-charged analogs. The IC50 values were 1000- and 100-fold greater in resistant than in sensitive cells for Rho 123 and Rho 6G respectively. Intracellular drug accumulation was significantly higher in sensitive as compared to resistant cells for both Rho 123 and Rho 6G, but little difference in drug uptake between these two cell types was observed for Rho 110 and Rho 116. It was also found that the intracellular to extracellular ratio of the positively-charged compounds was greater than unity in both sensitive and resistant cells whereas for the zwitterionic analogs this ratio was less than 1. Furthermore, this ratio of drug uptake was found to be significantly higher for Rho 6G than for Rho 123, which correlated with the high oil:water partition coefficient of Rho 6G (115.6). In MDR cells, verapamil increased Rho 123 and Rho 6G accumulation by 9.4- and 8.6-fold respectively. In addition, IC50 values in resistant cells were reduced greater than 100-fold for Rho 6G and greater than 1000-fold for Rho 123 in the presence of 10 micrograms/ml of verapamil. In contrast, less than 2-fold reduction of IC50 values for both of the zwitterionic analogs could be obtained under the same conditions. These results indicate that the chemical charge of rhodamines plays an important role in their differential accumulation, cytotoxicity and sensitivity to modulators such as verapamil, in sensitive and multi-drug resistant cells. The data also suggest that increased lipophilicity of the positively-charged rhodamines may increase their ability to accumulate in, and subsequently kill, MDR cells.

Correlation of a unique 220-kDa protein with vitamin D sensitivity in glioma cells.

We have investigated the antitumor and apoptotic effects of 1, 25-dihydroxyvitamin D(3) (VD(3)) in glioma cell lines and in primary cultures derived from surgical specimens from patients. Our results showed that certain glioma cells underwent apoptosis, whereas others were resistant. In an attempt to search for parameters that dictate VD(3) sensitivity, we discovered a unique 220-kDa protein in glioma cells that were sensitive to VD(3). This protein was not a classical vitamin D receptor (VDR), but was recognized by two different anti-VDR monoclonal antibodies. Furthermore, the level of the 220-kDa protein was inversely correlated with the IC(50) of VD(3) in these glioma cells. This 220-kDa protein was also present in frozen brain tumor samples, and the level of expression appeared to correlate with their corresponding primary cultures. Thus, our findings suggest that this 220-kDa protein may play an important role in determining VD(3) sensitivity in malignant glioma.

Potentiation of adriamycin accumulation and effectiveness in adriamycin-resistant cells by aclacinomycin A.

Variants of Friend leukemia cells (FLC) selected for resistance to either adriamycin (ADM), daunorubicin (DNR) or aclacinomycin A (ACM) by step-wise exposure to each drug, were found to be cross-resistant to ADM and DNR but not to ACM. In addition, an epithelial cell line isolated from normal monkey kidney (CV-1) was found to be intrinsically resistant to ADM and DNR but not to ACM. In contrast, a human breast carcinoma cell line (MCF-7) was found to be sensitive to all three compounds. In these latter cell lines as well as in the FLC variants, lowered intracellular amounts of ADM and DNR correlated with resistance, but ACM levels were the same in sensitive and resistant cells. When cells with either acquired or intrinsic resistance were treated with ACM in combination with ADM or DNR, significant increases in the intracellular amounts of these latter compounds were found. Increased drug accumulation in resistant cells treated this way was accompanied by increased cytotoxicity. When resistant cells were exposed to ACM in combination with other anthracyclines, similar results were obtained. In comparison, these phenomena were not observed when either one of the sensitive cell types (parental FLC and MCF-7) were treated similarly. Since ADM and DNR resistant cells are sensitive to ACM and their resistance circumvented by ACM, this drug may have important clinical applications when used in combination with other anthracyclines.

Comparison of topoisomerase I and II expression in primary brain tumor and lung cancer.

Topoisomerase I (TOP I) and II (TOP II) activities and their corresponding levels were analyzed in 27 primary brain tumors and 32 lung cancers (28 NSCLC, 4 SCLC). The TOP I and II activities in primary brain tumors varied from 500-2,000 units/mg and 100-3,000 units/mg respectively. Their corresponding levels varied from <0.01-3.30 (TOP I) and 0.24-8.30 (TOP II) arbitrary units. In lung cancer, the TOP I and II activities ranged from 1,000-4,000 and 500-4,000 units/mg respectively with their levels ranging from 0.30-61.60 and 0.2-14.2 arbitrary units respectively. These parameters were compared in both tumors using the Wilcoxon rank sums test, the difference were statically significance for all four parameters with a p<0.0001 for TOP I and II activities and TOP I levels and p<0.09 for TOP II levels. Using linear regression analysis, there was no correlation between TOP I and II activities and their corresponding levels in primary brain tumor. However, in lung cancer, the relationship between TOP I activities and levels were linear with r(2)=0.2 and p<0.0094, but not for TOP II activity and their levels. There was no relationship between TOP I and TOP II levels in the same tumor for both types of cancer. The future clinical implication of these findings are discussed.

Effect of verapamil on rhodamine 123 mitochondrial damage in adriamycin resistant cells.

Mitochondrial damage was found in Friend leukemia cells treated with rhodamine 123 (Rho 123). In contrast, when cells resistant to the drug were similarly treated, mitochondria were unaffected. These results correlated with higher levels of Rho 123 in sensitive as compared to resistant cells. However, when resistant cells were co-treated with verapamil, intracellular Rho 123 levels reached those of sensitive cells. At these levels mitochondrial damage and subsequent cytotoxicity in resistant cells were the same as in sensitive cells. These data suggest that differences in Rho 123 mitochondrial damage and subsequent cytotoxicity in sensitive and resistant cells result entirely from increased intracellular drug levels and not from differences in mitochondrial sensitivity or other mechanisms.

Comparison of annamycin to adriamycin in cardiac and MDR tumor cell systems.

Based on the response of a wide variety of tumors to the anthracycline, Adriamycin, numerous studies have been initiated to find an even more effective analog. In this pursuit two of the obstacles that have been necessary to overcome are a unique dose dependent Adriamycin-induced cardiotoxicity reported in patients treated with this chemotherapeutic agent as well as p-gp-mediated multi drug resistance (MDR) which has been found in tumor cells exposed to Adriamycin in vitro and in vivo as well as in human tumor samples. Using an in vitro cardiac cell system and MDR+ and MDR- Friend leukemia cell lines we find that a relatively new anthracycline, Annamycin, has reduced cardiotoxic activity but is more effective in inhibiting the growth of MDR+ cells than Adriamycin. The reduced cardiotoxicity of Annamycin is approximately 10 fold lower than Adriamycin whereas the increased efficacy against the MDR+ Friend leukemia tumor cell line is about 2 fold. The observation that Adriamycin preferentially accumulates in cardiac-muscle (CM) but not in cardiac non-muscle (NM) cells while Annamycin accumulates equally in both, may explain in part the reduced cardiotoxicity of Annamycin. Moreover, the cytosolic accumulation of Annamycin vs the nuclear localization of Adriamycin suggests a different target site for each drug.

Cytotoxic and photodynamic effects of Photofrin on sensitive and multi-drug-resistant Friend leukaemia cells.

To study cross-resistance to Photofrin (PF) photosensitization, a Friend leukaemia cell line (ADM-RFLC) with a high level of multi-drug resistance (MDR) and the parental sensitive cell line (FLC) have been used. PF uptake measured by HPLC shows a similar intracellular drug accumulation in both cell lines. The ID50s for cell growth inhibition by PF are also similar after exposure in the dark in the two cell lines, while after illumination they are slightly lower in ADM-RFLC than in FLC cells. Moreover, verapamil, known to reverse the MDR phenotype induced by P-glycoprotein over-expression (the drug efflux mechanism), affects equally ADM-RFLC and FLC cells sensitivity to PF. In addition, photodynamic treatment with PF did not reverse the resistance to rhodamine 123 and aclarubicin, but partly reverses resistance of ADM-RFLC cells to antitubulin drugs such as vinblastine or vincristine. These latter results could have clinical application in the treatment of tumours expressing the MDR phenotype.

Selective killing of carcinoma cells "in vitro" by lipophilic-cationic compounds: a cellular basis.

Lipophilic positively-charged compounds are facilitated across biological membranes by the transmembrane potential of intact cells. One such compound, rhodamine 123, has recently been shown to be selectively toxic toward a variety of transformed (carcinoma), epithelial cells in vitro (Lampidis et al., 1982; Bernal et al., 1982; Lampidis et al., 1983). A mechanism that could account for the selectivity of this agent would be a difference in the plasma membrane potential between normal and carcinoma cells. We report here that a significantly higher transmembrane potential has been found in a pair of carcinoma (83 mV for human breast and -99 mV for human cervix) as compared to normal (-56 mV for marsupial kidney and -48 mV for monkey kidney) epithelial cell lines. We also identified 3 other positively-charged lipophilic compounds, safranin 0, rhodamine 6G and tetraphenylphosphonium chloride (TPP+), which show selective toxicity toward carcinoma cells in vitro, while an uncharged lipophilic analog, rhodamine 116, does not. These data suggest that the higher plasma membrane potential of carcinoma cells may in part contribute to the preferential accumulation and selective toxicity of the lipophilic cationic compounds we have examined. An extension of this concept to an in vivo environment could lead to a class of cationic compounds which selectively exploit differences between normal and carcinoma cells.