The human non-small cell lung carcinoma lines, A549 & A427 were obtained from American Type Culture Collection (Manassas, VA, USA) and maintained in basic medium as described [13]. In addition, mammary tumor cell line, 4T1-Luc, was obtained from American Type Culture Collection (Manassas, VA) and maintained as described [14].After growing to 100% confluence, cells were trypsinized (0.5 mL trypsin in 5 mL PBS) and incubated for 5 -10 min. The cells were then harvested by low speed centrifugation, washed in 1 mL media and seeded at 3 X 10⁴cells per 200 µL of medium per well in a 96-well plastic plate (Nunc Inc. Rochester, NY). A portion of cells (1 X 10⁵ cells/well) was also grown overnight on Lab Tek 8-chamber slides (Nunc Inc. Rochester, NY) for caspase-activity and live/dead assays.

The effect of paclitaxel on tumor cell growth was evaluated using Cell Titer 96®Aqueous One Solution Cell Proliferation Assay (Promega). This assay permits the detection of metabolic activity of proliferating cells without the required solubilization steps in the MTT assay. Briefly, 24 h drug-treated cells were cultured in a 37˚C humidified incubator for 2 h with 20 µL of CellTiter96®Aqueous One Solution Reagent containing tetrazolium compound [3-(4, 5-dimethyl-thiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt; MTS] and an electron coupling reagent, phenazine ethosulfate (PES). The absorbance per well was measured at 490 nm using a micro-plate reader (Perkin Elmer Wallac Victor™ 1420 Multilabel Counter).

Cell viability and integrity after drug treatment was assessed by two assays. With the live-dead viability kit (Invitrogen), drug-treated cells were incubated with ethidium homodimer (EthD-1) for the last 45 min of the drug exposure period. After this treatment, the media was removed and slides were prepared for microscopy by mounting with prolong gold antifade® reagent (Invitrogen, Molecular Probes). For clonogenic survival assay, drug-treated cells were cultured in a 37˚C humidified incubator for 10 days to allow colony formation. Drug exposure times varied from 2 to 14 days for time course study. Cells were then fixed with methanol, stained with solution containing 20% methanol and 0.5% crystal violet, and clusters made up of 100-cells were counted as surviving colonies. The percentage of surviving colonies relative to untreated cells was determined by the proportion of colonies produced from cells after drug treatment. Plating efficiency was 80% ± 10%.

To quantify the levels of basal and induced caspase-3 protein and corresponding activity in different tumor cell lines, we used a fluorometric enzymatic assay, EnzChek® Caspase-3 (Molecular Probes, Eugene, OR). Cell extracts from paclitaxel-treated (25 µM) and untreated A549, A427 and 4T1-Luc were treated with final substrate concentration of 200 μM and incubated at room temperature for 1 h. Fluorescence was measured with a micro plate reader using excitation at 360± 40nm, and emission at 460 ± 40 nm. Appropriate controls were performed to determine background fluorescence of the substrate. A standard AMC-substrate curve was used to quantify the levels of enzyme-cleaved substrate from the measured fluorescence values.

Statistical analysis was performed using the Student’s t-testand nonlinear regression analysis in GraphPad Prism Software® (version 4.0 for windows, San Diego, CA).

Studies of the human non-small cell lung carcinoma cell lines, A549-T12 and A549-T24, show that while these paclitaxel resistant cells are cultured and maintained in nanomolar levels of drug, the parental A549-cell line from which they are derived is, in contrast, sensitive to nanomolar drug doses [15-20]. To test the ability of nanomolar- concentrations of drug to inhibit cell proliferation, we investigated the viability of the A549 and A427cell lines after exposure to 1- 1000 nM of paclitaxel using a modified MTT assay [19,21].We observed that 24 h treatment of A549 and A427-cells with concentrations ≤ 200 nM did not impede cell viability. Cells showed a minor response at concentrations ≥ 500 nM, with A427 being slightly more susceptible than A549-cells in the 500 t0 1000 nM range (Fig. 1A). Overall, proliferation of either cell line was not affected by nanomolar concentrations of paclitaxel. However, increasing paclitaxel doses to the micromolar scale remarkably decreased proliferation of A549 and A427 cells. The half maximal effective concentration (EC-50) for these cells was determined to be 22.5 μM and 12.5 μM, respectively (Fig. 1B and C). In addition, we measured the response of the 4T1-luc cell-line to equivalent micromolar drug concentrations. In contrast to A549 and A427-cells, 4T1-luc cells showed a strong dose response with EC-50 of 8.5 μM (Fig. 1D). At concentrations >40 μM, all three cell lines succumbed to death without distinct behavioral differences. However, disparate susceptibility of the three cell lines was pronounced at a dose range of 10 – 25 μM. While the anti-proliferative effect of paclitaxel was enhanced in 4T1-Luc cells, A427 and A549 cells were relatively less affected by the drug (Fig. 1E).

Fig. (1). Breast and lung cancer tumor cell lines exhibit differential susceptibility to micromolar doses of paclitaxel at equal drug exposure time. Cell viability was determined by a modified MTT assay. After 24-h drug exposure with indicated concentrations, cells were incubated with 20 µl of MTS reagent for 2-h. Absorbance values were measured at 490 nm. Experiments were done in triplicate for each cell type, with appropriate controls. The half maximal effective concentration (EC50) of paclitaxel for each cell type was determined by nonlinear regression analysis using GraphPad Prism Software® (version 4.0 for windows, San Diego, CA). Data represent normalized mean values. (A) Nanomolar concentrations of paclitaxel have little effect on tumor cell viability especially at lower concentrations. However a dose response is observed in both tumor cell types, with A427 cells showing slightly increased sensitivity than A549-cells at higher nanomolar drug concentrations (500 nm and 1000 nm).Open circles represent negative control, crosses and diamonds represent positive control, black rectangles represent samples. (B) EC50 for A549 cells was determined to be 22.5 µM. (C) EC50 for A427 was determined to be 12.5 µM and (D) EC50 for 4T1-Luc was determined to be 8.5 µM. At high drug doses > 40 µM, no significant difference in tumor cell line response to paclitaxel is observed.However, clear differences in cell type sensitivity to drug dosage is observed at concentrations < 20 µM. Of the 3 tumor cell lines,A549 and A427 show low sensitivity to the drug at low concentrations, with 4T1-Luc cells appearing to be the most sensitive in this dose range. (E) Anti-proliferative effect of paclitaxel is enhanced in 4T1-Luc cells, but not in A427 and A549 cells at high drug doses. Proliferation of tumor cells after 24-hour drug treatment was analyzed by MTS assay (Molecular Probes), as described in methods section. Experiments were done twice in triplicates. N=6. At concentrations ≥ 10 µM, a drastic reduction in the number of proliferating cells was observed in 4T1-luc, in comparison with A549 and A427 cells which showed significant cell proliferation. (F) 4T1-luc tumor cells show increased aggregated and clustered colony formation. Clonogenic survival assay was performed to evaluate the behavioral response of tumor cells to an extended duration of drug exposure. A time course study was performed to determine colony formation at a fixed drug concentration (12.5 μM) over varying exposure times (2- 14 days). Each data point was derived from the results of at least three independent experiments. N= 9. Data show mean ± SEM of three experiments, each in triplicate. Relative to untreated controls, a time dependent-decline in the number of colonies formed was observed in 4T1-luc cells, but not in A549 or A427 cells (not shown). In these latter cell lines, cells were more dispersed and very few aggregated colonies (cell clusters consisting of more than 100-cells) were observed. (G) 4T1-luc cells form colony aggregates in a dosedependent fashion, relative to A549-cells which show reduced cell colonies at higher drug concentrations. Overall, however, more proliferating A549 cells than proliferating 4T1-luc cells were present at higher concentrations.

The relatively higher number of persistent proliferative cells in A549 and A427-cells suggests that both tumor cell lines potentially have an intrinsic resistance to paclitaxel. To further assess the behavioral response differences between the lung and breast tumor cell lines, we performed clonogenic survival assays, where tumor cells were treated with varying concentrations of the drug for an extended period (2-14 days). A time dependent-decline in the number of 4T1-luc cells, relative to the untreated controls, was observed (Fig. 1F). Although all three cell lines showed survival dose response, 4T1-luc tumor cells formed more aggregated and clustered colonies. In contrast, A549 and A427 cells appeared more dispersed, with less aggregated colonies (i.e. one or two clusters made up of more than 100-cells) (Fig. 1G).This suggests that 4T1-luc cells perhaps respond more to prolonged drug therapy by forming aggregates while A549 cells appear to resist paclitaxel treatment by spreading over the plate surface.The differences in chemotherapeutic response among the 3 cell lines clearly point to the underlying differences in their resistance mechanisms to paclitaxel [22-24].

Although it is fairly well known that most tumor cells undergo paclitaxel-induced apoptotic death, it is equally well established that paclitaxel induces other non-apoptotic programmed cell death mechanisms in different tumor cell types [20,21,25]. To ascertain and distinguish between apoptotic and non-apoptotic death, we used a Live/Dead® Viability assay to evaluate the quintessential cellular features of death such as plasma membrane blebbing, DNA fragmentation, and cell shrinking. We particularly examined drug-treated cells for the presence of fragmented nuclei and compromised plasma membranes as indicators of death. We observed a dose-dependent fragmentation of nuclear DNA as well as damage to the cell membrane. This phenotype was highly pronounced in 4T1-luc tumor cells with extensive fragmentation at 50 μM and 25 μM of drug, visualized as red punctate dots in dead cells stained by EthD-1 (Fig. 2, A-F). Very few live cells were observed at these concentrations and cells had a mix of necrotic and apoptotic characteristics. Nonetheless, the extent of membrane disintegration and nuclear damage in 4T1-luc cells was more moderate at 12.5 μM and 6.25 μM of paclitaxel (Fig. C & D). Like 4T1-luc cells, A549 cells showed dose-dependent DNA fragmentation, but with less damaged cell membranes (Fig. 3, A-F). Moreover, dead cells had fewer and less dispersed punctate nuclear fragments (Fig. 3A, insert). Interestingly, we observed a few mitotic cells among the apoptotic A549 cell population, which were immobilized at the mitotic phase in the presence of paclitaxel. (Fig. 3A, D & F). The absence of this distinct population of cells in 4T1-luc cells clearly highlights different death response mechanisms in the two tumor cell lines. Previous studies reported that paclitaxel suppressed microtubule dynamics and disrupted the functioning of the mitotic spindle [26-29], suggesting a slow cell cycle progression at the metaphase/anaphase checkpoint that eventually induces cell death. The features in A549-cells fit perfectly with this pattern, albeit, we also observed a mix of apoptotic morphologies in this cell line. Our results demonstrate that 4T1-luc breast tumor cells respond to paclitaxel by initiating extensive nuclear fragmentation and inducing both necrotic and apoptotic forms of death. On the other hand, paclitaxel-induced death in A549 cells occurs partly through interference with cell cycle progression, thus, pointing to different phases of drug-induced apoptosis in the two tumor cell lines.

Fig. (2). Paclitaxel induces nuclear fragmentation and cell death in 4T1-Luc tumor cell lines. The extent of DNA damage observed in tumor cells was dose dependent. Arrows point to fragmented nuclei. Inserts highlight nuclear fragments and damaged DNA within cells. Scale bar indicates 90 µm. Cell morbidity was determined by Live-Dead Viability Assay. Drug treated cells were incubated with ethidium homodimer (EthD-1), which stains fragmented nuclei of dead cells. EthD-1 dye can only penetrate cells with compromised membranes and emits red fluorescence (ex/em - 495 nm/ 635 nm) on binding to nucleic acids. Exclusion of the dye by intact membrane of live cells distinguishes dead cells from live ones. (A – F) - Apoptosis in 4T1-Luc cells. Cells treated with high paclitaxel doses, 50 µM (A) and 25 µM (B), show extensive nuclear fragmentation and damaged membranes. (C) At 12.5 µM and (D) 6.25 µM, moderate nuclear fragmentation is observed. (E) Reflects background fluorescence from negative control cells. (F) Apoptotic induction in positive control (ethanol-treated) cells. Wavy black line traces boundary of damaged cell membrane.

Fig. (3). A549 tumor cells exhibit both apoptotic and mitotic morphology in response to paclitaxel treatment. Arrows point to fragmented nuclei. Inserts and boxed regions highlight fragmented nuclear DNA and mitotic cells. These distinct populations of mitotic cells were not observed in 4T1-luc tumors. Scale bar indicates 90 µm. Cell death was determined by Live-Dead Viability Assay as described. Exclusion of EthD-1 dye by intact membrane of live cells distinguishes dead cells from live ones. Images are shown in red and grey channels. (A – F) -Cell death in A549 cells. Fragmented nuclei are visible in cells treated with high paclitaxel doses, 50 µM (A) and 25 µM (B), as well as in cells treated with moderate doses, 12.5 µM (C) and 6.25 µM (D). (E) No dead populations are observed in untreated cells; mitotic cells are visible in the grey channel. In contrast to the morphology of 4t1-luc tumor cells, mitotic proliferating A549 cells may contribute to A549 tumor resistance at low doses of paclitaxel. (F) Apoptotic induction in positive control (ethanol-treated) cells.

Based on reports that somatic mutations of caspase-3 gene in human cancers cause selective inactivation of cysteinyl proteinases to compromise apoptotic death [28-31], we hypothesized that differences in the intrinsic and drug-induced activity of effector caspases contribute to the acute paclitaxel-induced death response in 4T1-luc versus the benign response in A549 and A427 tumors. We predicted that 4T1-luc cells would have relatively higher basal and drug-induced caspase-3 activity than A549 or A427 tumor cells. We found that while all 3-lines showed some levels of basal enzymatic activity, A427-cells exhibited the least activity with 40 µM of cleaved substrate, followed by A549-cells with 54 µM of cleaved substrate, and then 4 T1-Luc cells, with 68 µM of substrate (Fig. 4A). A decrease in activity in the presence of a caspase-3 inhibitor corroborated the specificity of the reaction. On measuring caspase-3 activity in extracts from drug-treated cells, we found increased activity in 4T1-luc cells but a decrease in A549 and A427-cells (Fig. 4B). This suggests that while 4T1-luc cells activate caspase-3 on exposure to drug, A549 and A427 cells somehow repress caspase-3 activation. In contrast to the drug-treated 4T1-luc cells that showed a two-fold increase in caspase-3 activity in comparison to basal activity levels, A549 and A427 showed approximately a 3-fold and 4-fold decrease in enzymatic activity, respectively, relative to basal levels (Fig. 4C). This result provides evidence that increased activation of caspase-3 confers sensitivity of 4T1-1uc cells to the drug, whereas repressed caspase-3 activation potentially accounts for the resistance of A549 and A427 cells to paclitaxel.

Fig. (4). Tumor cell lines exhibit differential basal and drug-induced caspase-3 activity. The intrinsic and drug-induced caspase-3 activity of 3-cell lines were determined by a fluorimetric enzymatic assay that measures cleavage of a caspase-3 specific substrate, Z-DEVDAMC,where Z represents a benzyloxycarbonyl group, and AMC, 7-amino-4-methylcoumarin. The substrate fluoresces weakly in the UV range (Ex/Em~330/390 nm) but yields strong blue-fluorescent product (Ex/Em ~ 342/441 nm) upon proteolytic cleavage and allows continuous monitoring of the activity of caspase-3 and its related proteases in cell extracts. A standard AMC-substrate curve was used to quantify the levels of enzyme-cleaved substrate from the measured fluorescence values. To confirm that the observed fluorescence signal was due to the activity of caspase-3-like proteases, control treatments were incubated with 10 M of a reversible caspase-3 inhibitor, Ac-DEVD, prior to treating cells with the substrate. Experiments were performed twice in triplicate. N= 6. Data show mean ± SEM of 2 experiments done in triplicate. (A) While all 3-lines showed some level of basal enzymatic activity, A427-cells showed the least activity with 40 µM of cleaved substrate, followed by A549-cells, with 54 µM of cleaved substrate. 4 T1-Luc cells showed the highest levels of activity with 68 µM of cleaved sub-strate. A decrease in activity in the presence of a caspase-3 inhibitor confirms the specificity of the reaction. The underlying difference in caspase-3 activity may account for the observed resistance patterns after drug exposure. (B) Paclitaxel induces differential caspase-3 activity in tumor cell lines. Whereas 4T1-Luc cells showed an increase in levels of cleaved product (120 µM, 60% of added substrate) after exposure to drug, A549 and A427-cells showed a decrease in the levels of cleaved product, with 20 µM and 11 µM, respectively, of cleaved substrate. This suggests that while 4T1-luc cells activate caspase-3 on exposure to drug, A549 and A427 cells somehow repress caspase-3 activation. (C) In comparison to basal enzymatic activity levels, paclitaxel induces a two-fold increase in caspase-3 activity in 4T1-luc cells as measured by increased levels of cleaved substrate. In contrast, A549 and A427 show approximately a 3-fold (2.7) and 4-fold (3.64) decrease in enzy-matic activity, respectively. This increased activation of caspase-3 underlies the sensitivity of 4T-1uc cells to drug, whereas, repressed cas-pase-3 activation potentially accounts for the resistance of A549 and A427 cells to paclitaxel