The multilayered postconfluent cell culture as a model for drug screening.

New drug development requires simple in vitro models that resemble the in vivo situation more in order to select active drugs against solid tumours and to decrease the use of experimental animals. In this paper, we review the characteristics and scope of a relatively simple cell-culture system with a three-dimensional organisation pattern - the multilayered postconfluent cell culture model. Solid tumour cell lines from diverse origins when grown in V-bottomed microtiter plates reach confluence in 3-5 days and then start to form multilayers. The initial exponential growth of the culture is followed by a plateau phase when cells reach confluence. This produces changes in the morphology of the cells. For some cell lines, it is possible to observe cell differentiation. A substantial advantage of the system is the use of the sulforodamine B (SRB) assay to determine relative cell growth or viability, which allows semiautomation of the experiments. Several experiments were performed to assess the differences and similarities between cells cultured as monolayers and multilayers, and eventually, compared with the results for solid tumours and some other models such as spheroids. Cell-cycle analysis for multilayers showed a lower S-phase arrest, which is accompanied by a decrease in the expression of cell-cycle-related proteins and a decrease in cellular nucleotide pools. Gene and protein expression of topoisomerase I, topoisomerase II and thymidylate synthase expression were lower for multilayers, but no substantial changes were observed for the expression of DT-diaphorase. P53 expression increased. Multilayer cultures present distinctive properties for drug transport across the membrane, drug accumulation and retention. In fact, the transport of antifolates across the membrane, accumulation of topotecan and gemcitabine-triphosphate are reduced in multilayers when compared with monolayers, which may be related to a decrease in drug penetration to the inner regions of the multilayers. Alteration of these pharmacodynamic parameters is directly related to a decrease in drug activity. The most powerful application of multilayers is in the assessment of cytotoxicity. Solid tumour cell lines from different origins have been treated with several conventional and investigational anticancer drugs. The data show that multilayers are more resistant to the drugs than the corresponding monolayers, but there are substantial differences between the drugs depending on culture conditions, e.g. the difference was rather small for a drug such as cisplatin, miltefosine and EO9, a drug, which is activated under hypoxic conditions. Gemcitabine was active against ovarian cancer but not against colon cancer, resembling the in vivo situation. This observation was not evident with monolayer experiments. Another interesting application is the possibility to perform drug combination studies. The combination of gemcitabine and cisplatin proved to produce selective cell kill in H322 cells (non-small cell lung cancer cell line). Neither of the drugs was independently able to produce similar effects. In summary, multilayer cultures are relatively simple three-dimensional systems to study the effect of microenvironmental conditions on anticancer drug activity. The model might serve as a base for a more rigorous secondary in vitro screening.

Determinants of activity of the antifolate thymidylate synthase inhibitors Tomudex (ZD1694) and GW1843U89 against mono- and multilayered colon cancer cell lines under folate-restricted conditions.

The cytotoxicity and metabolic effects of two thymidylate synthase (TS) inhibitors, Tomudex (Raltitrexed, ZD1694) and GW1843U89, were studied in WiDr colon cancer cells under four different growth conditions: as standard monolayers and as postconfluent multilayers grown under either high (WiDr, 8.8 microM folic acid) or low (WiDr/F, 1 nM leucovorin) folate conditions. Both GW1843U89 and ZD1694 were 13-15-fold more active against WiDr/F than WiDr cells when cultured as monolayers (IC50s in WiDr/F cells were 0.22 and 0.39 nM, respectively). WiDr cells were markedly less sensitive to the drugs when grown as multilayers (4-15-fold), in contrast to the WiDr/F cells, which were equally sensitive. However, total growth inhibition could not be achieved in WiDr multilayers (concentration causing total growth inhibition > 10,000 nM), whereas in WiDr/F multilayers, it could be achieved at 0.42 nM ZD1694 and 150 nM GW1843U89. Growth conditions markedly affected the TS levels when using different enzyme assays. At nonsaturating substrate concentrations, the catalytic activity of TS was similar in mono- and multilayers grown under high folate conditions but lower in multilayers at saturating concentrations. In cells grown under low folate conditions, TS catalytic activity was 3-6-fold lower in multilayers than in monolayers. This was consistent with a decrease in the number of S-phase cells in multilayers. Western blotting revealed less pronounced (2-3-fold) differences in the TS protein content. Exposure of the cells for 24 h to the drugs increased the TS levels by 4-fold. Because this increase in TS levels might explain the decrease in sensitivity to the TS inhibitors, we measured TS inhibition (TSI) by the drugs in intact cells using the TS in situ assay. GW1843U89 was more active than ZD1694. However, after 4 h of exposure in WiDr/F mono- and multilayers, TSI was in the same range for both drugs [50% TSI (TSI50), 0.5-1.7 nM]. In WiDr cells, the TSI50 for ZD1694, but not GW1843U89, was 10 times higher in the multilayers as compared to the monolayers. Despite the increase in TS protein levels, the extent of TSI was similar or even more pronounced in both cell lines grown as either multi- or monolayers. Because the cells were grown under depleted and folate-rich conditions that may affect folate uptake, we measured folate transport using methotrexate (MTX) as the reference drug for the activity of the reduced folate carrier. MTX uptake was 4-fold lower in multilayers compared to monolayers in both WiDr and WiDr/F cells. Uptake of MTX was 5-fold more effective in WiDr/F cells than in WiDr cells in both mono-and multilayers. In conclusion, the resistance of WiDr multilayers to the novel antifolates ZD1694 and GW1843U89 may be due to the high folate medium concentrations, which may be responsible for impaired drug uptake along with less effective TSI. In contrast, WiDr/F monolayers and multilayers were very sensitive to these antifolates. These effects of folate homeostasis may explain some of the variable results seen in treatment of solid tumors with new antifolate TS inhibitors.

Selective cell kill of the combination of gemcitabine and cisplatin in multilayered postconfluent tumor cell cultures.

Both gemcitabine (2',2'-difluorodeoxycytidine, dFdC) and cisplatin (cis-diammine-dichloroplatinum) have significant anticancer activity against ovarian, head and neck, and non-small cell lung cancer (NSCLC). dFdC can be incorporated into DNA and RNA, and inhibit DNA repair, while cisplatin can form Pt-DNA adducts. We previously observed schedule-dependent synergism of the combination of dFdC and cisplatin in monolayer cell cultures. We now evaluated whether the combination would also enable selective cell kill in multilayered postconfluent cell cultures, since each compound showed variable activity in multilayered cells. The combination was tested in multilayered cultures from cell lines with a different histological origin: the human head and neck squamous cell carcinoma cell line UMSCC-22B (22B), the human NSCLC cell line H322, and ADDP, a cisplatin-resistant variant of the human ovarian cancer cell line A2780. Sensitivity of the multilayered cells was dependent on exposure duration and sequence of the drug combinations, which were added in a constant molar ratio (dFdC:cisplatin 1:100), with a total exposure time of 96 h. The type of interaction was related to the degree of resistance of the cell lines to either dFdC or cisplatin. Thus, the very sensitive 22B cells only showed an additive effect when cells were preincubated for 24 h with dFdC prior to exposure to the combination. In contrast, in the resistant ADDP and H322 cells, synergism was the most common profile (three out of four schedules tested). This is of special relevance when we take into account that these drugs only show cytostatic effects when administered alone, whereas the combination produced cytotoxic cell killing. In conclusion, combining dFdC with cisplatin can be at least additive, but synergistic in multilayered postconfluent cells resistant to dFdC and cisplatin.

Postconfluent multilayered cell line cultures for selective screening of gemcitabine.

The in vitro cytotoxicity of gemcitabine (dFdC) was tested in ovarian and colon cancer cell lines grown as monolayers and three-dimensional multilayered cell cultures. In our model, dFdC showed slight selectivity in cytotoxicity against ovarian over colon cancer cells, when cell lines were grown as monolayers. However, when cell lines were grown as multilayers, this selectivity was accentuated: A2780 multilayers were 14 times less sensitive than monolayers, but the colon cancer cell lines were more than 1000 times more resistant than their corresponding monolayers. The accumulation of the active metabolite, dFdCTP, after 24 h exposure to 1 microM dFdC varied between 1100 and 1900 pmol/10(6) cells in monolayers. This was 5 times lower in multilayers compared with monolayers of all four cell lines, which can, in part, explain the lower sensitivity of the multilayers. In addition, it appears that the amount of the active metabolite retained is more important than the amount accumulated initially, since the differences between the ovarian and the colon cancer cell lines were more evident in retention experiments. Exposure to dFdC caused a 2-3-fold increase in the levels of several nucleotides, except for the CTP pools in the colon cancer lines, which were reduced by 3-fold at the highest dFdC concentration (10 microM). The findings with the multilayer model are in better agreement with in vivo activity in ovarian cancer and colon cancer than those with the monolayer system. This indicates the potential of the multilayer system to be a better predictive model than the conventionally used monolayer cultures.

Preclinical combination therapy with gemcitabine and mechanisms of resistance.

Cisplatin and gemcitabine both have activity in solid tumors, such as non-small cell lung, ovarian, and head and neck cancers. These drugs have the desired features needed to obtain synergistic activity, different side effect profiles, and mechanisms of action. Cisplatin acts by forming DNA-DNA cross-links (both intrastrand and interstrand) and DNA-protein cross-links; resistance to cisplatin is thought to be due to excision repair of the affected DNA. Gemcitabine acts by its incorporation into nucleic acids, leading to masked chain termination. By combining gemcitabine with cisplatin, it might be possible to achieve a better therapeutic effect than either drug alone and to bypass resistance to one or both drugs. Acquired resistance to gemcitabine was associated with a deoxycytidine kinase deficiency in vitro, but this was difficult to achieve in vivo. Proper scheduling may overcome intrinsic and transient resistance due to physiologic circumstances or aberrant biochemical properties. Preclinical in vitro and in vivo combination studies with cisplatin showed schedule- and model-dependent synergistic and additive effects between cisplatin and gemcitabine. Incorporation of gemcitabine into DNA might facilitate cisplatin-DNA adduct formation. Combining gemcitabine and cisplatin inhibited the DNA excision-repair process more than gemcitabine alone. This implies that if gemcitabine nucleotide is incorporated into the DNA strand, the action of the proofreading exonucleases is less efficient. In addition, both deoxyribonucleotide and ribonucleotide pools, essential for good functioning of DNA repair, are seriously depleted by gemcitabine. It is concluded that combining gemcitabine with cisplatin can be at least additive providing the right schedule is chosen, giving the best balance between acceptable toxicity and an enhanced antitumor activity.

Development, pharmacology, role of DT-diaphorase and prospects of the indoloquinone EO9.

1. The indoloquinone EO9 (3-hydroxymethyl-5-aziridinyl-1-methyl-2- (1H-indole-4,7-dione)-propenol; E85/053; NSC 382,459) is a synthetic bioreductive alkylating agent that is structurally related to mitomycin C (MMC). 2. EO9 does, however, show a different mechanism of action and a broader antitumour profile than MMC. It is also a more potent cytotoxic agent in vitro than MMC, probably because of its impressive efficient activation by reductive enzymes, particularly DT-Diaphorase. This enzyme is elevated in several tumours compared to normal tissues. 3. The preferential cytotoxicity of EO9 under hypoxic conditions makes it an interesting compound to combine with radiation. 4. In preclinical and the Phase I clinical studies, no myelosuppression was observed but reversible proteinuria was dose-limiting. Phase II clinical studies were started in the summer of 1994.

DT-diaphorase activity in normal and neoplastic human tissues; an indicator for sensitivity to bioreductive agents?

DT-diaphorase (DTD) is an important enzyme for the bioreductive activation of the new alkylating indoloquinone EO9. In preclinical studies, EO9 has shown selective anti-tumour activity against solid tumours and under hypoxic conditions. The levels of three reductive enzymes have been determined in three types of human solid tumours, together with corresponding normal tissues and normal liver. DTD enzyme activities were measured in tumour extracts using 2,6-dichlorophenolindophenol (DCPIP) and NADH as substrates; cytochrome P450 reductase or cytochrome b5 reductase activities were assessed with cytochrome c and NADPH or NADH respectively. DTD activity was highest in non-small-cell lung (NSCLC)-tumours (mean 123 nmol DCPIP min-1 mg-1), followed by colon carcinoma (mean 75 nmol min-1 mg-1) and squamous cell carcinoma of the head and neck (6-fold lower than NSCLC). DTD activity was very low in normal liver and normal lung (4-6 nmol min-1 mg-1), while the levels in normal colon mucosa or normal mucosa of the head and neck region were in the same range as the corresponding tumours. The levels of the two other reductive enzymes, cytochrome P450 reductase (CP450R) and cytochrome b5 reductase (Cb5R), were 5 to 25-fold lower than those of DTD in all the tissues, except for normal liver, in which DTD was 2 to 4-fold lower. The degree of variation found for DTD (range 4-250 nmol min-1 mg-1), was not observed for these enzymes (CP450R, 0.8-7.8 nmol cytochrome c min-1 mg-1; Cb5R, 3.5-27.6 nmol min-1 mg-1).(ABSTRACT TRUNCATED AT 250 WORDS)

Antiproliferative activity of the topoisomerase I inhibitors topotecan and camptothecin, on sub- and postconfluent tumor cell cultures.

We have assessed the antiproliferative effects of a 24-hr exposure to the topoisomerase I inhibitors, topotecan and camptothecin, on two colon and one ovarian human tumor cell lines, cultured as subconfluent and as multilayered postconfluent cultures. Chemosensitivity was measured by the sulforhodamine B assay. In general, postconfluent cultures were less sensitive to these agents, yielding GI50S (drug concentrations inhibiting growth by 50%) from 1.2 to more than 6000 times higher than those of subconfluent cultures. Both compounds displayed similar effects on subconfluent cells, inducing complete growth inhibition at concentrations ranging from 0.03 to 0.5 microM. Topotecan, however, was more potent than camptothecin in two out of the three cell lines tested as multilayered postconfluent cultures. Topoisomerase I mRNA expression on postconfluent cultures was 50% lower than on subconfluent cultures in the three cell lines studied. However, we did not detect any reproducible differences in topoisomerase I protein expression and in relaxation activity of supercoiled DNA between the two types of cultures. From accumulation experiments it appeared that the peak concentration of the lactone form of topotecan as well as the area under the concentration-time curve (AUC) were 2-fold higher in the monolayer than in the multilayer cultures. Therefore, the differences in the activity of topoisomerase I inhibitors under our experimental conditions were likely due to a decreased rate of proliferation of postconfluent cells, associated with a reduction in drug uptake.

Chemosensitivity to the indoloquinone EO9 is correlated with DT-diaphorase activity and its gene expression.

EO9, a new bioreductive indoloquinone alkylating agent, requires activation by a two-electron reduction, which can be catalysed by the NAD(P)H:quinone oxidoreductase DT-diaphorase (DTD) (EC 1.6.99.2). Seven human and four murine tumor cell lines from different histological origins were evaluated for their DTD enzyme activity (evaluated using dichlorophenolindophenol and EO9 as substrates), DTD gene expression and chemosensitivity to EO9. In general the cell lines could be divided into two groups: leukemic cells which were relatively resistant to EO9 (IC50 > or = 0.5 microM) and had no measurable DTD activity, and solid tumor cells, which were more sensitive to the drug (IC50 < 0.06 nM) and contained a high DTD activity (> 90 nmol/min/mg). The expression of the DTD gene was measured by semiquantitative PCR in the human cell lines and an excellent correlation between gene expression and enzyme activity was observed (r2 = 0.94). A higher DTD gene expression also correlated with higher chemosensitivity to EO9. Protection of chemosensitivity to EO9 by dicoumarol, a strong and specific inhibitor of DTD activity, was dependent on duration of exposure and concentration of dicoumarol. Inhibition was best observed by short exposure to dicoumarol and EO9 together, demonstrating that bioactivation of EO9 by DTD is essential. In conclusion, DTD activity and expression appear to predict sensitivity to EO9 in a variety of cell lines. Evaluation of activity or expression in patients' tumor samples might predict the response to EO9.

In vitro sequence-dependent synergistic effect of suramin and camptothecin.

Suramin, a hexasulphonated naphthylurea with activity in prostatic cancer, possesses a wide variety of antitumour mechanisms of action, one of which is the inhibition of topoisomerase II. In this in vitro study, suramin was combined with the topoisomerase I inhibitor, camptothecin. Several suramin concentrations (0.2-3000 microM) were combined with camptothecin (0.4 pM-20 microM) in MCF-7 and PC3 human cancer cell line cultures. In addition, we studied the topoisomerase II and I gene expression by northern blot analysis, and the cell cycle distribution by flow cytometry, after exposure to suramin. While there was only an additive effect when suramin and camptothecin were added simultaneously, a remarkable synergism was obtained when camptothecin was added after a 3-day exposure to suramin. Topoisomerase II and I gene expression and the number of cells in S phase were significantly reduced after exposure to suramin. In conclusion, interaction of suramin with camptothecin is schedule-dependent and can be synergistic. These findings might help in identifying optimal combinations of suramin or other topoisomerase II inhibitors, with topoisomerase I inhibitors.

Cytotoxic effects of anticancer agents on subconfluent and multilayered postconfluent cultures.

The cytotoxic effects of conventional (doxorubicin, 5-fluorouracil, cisplatin) and investigational (2',2'-difluorodeoxycytidine, hexadecylphosphocholine, EO9, rhizoxin) anticancer drugs were studied in subconfluent and multilayered postconfluent cultures of human colon and ovarian carcinoma cell lines. Chemosensitivity was assessed 4 days after a 24-h drug exposure with the sulphorhodamine B assay. Except for rhizoxin, all drugs tested yielded an EC50 (drug concentration producing absorbance readings 50% lower than those of non-treated wells) in postconfluent cultures that were higher than an EC50 obtained with subconfluent cultures. Compared with subconfluent cultures, postconfluent cultures showed decreased cellular nucleotide concentrations and ATP/ADP ratios, in addition to an increased percentage of G0/G1 cells. The activity of DT-diaphorase, a reductase involved in the bioactivation of EO9, was similar in sub- and postconfluent cultures. These results indicate similarity of the postconfluent model presented with those obtained with in vivo models and more complex in vitro techniques.