Retinoblastoma Tumor Suppressor Gene Expression Determines the Response to Sequential Flavopiridol and Doxorubicin Treatment in Small-Cell Lung Carcinoma Free

Small-cell lung cancer (SCLC) accounts for ∼13% of the lung cancers (1, 2). Despite the high response rate of SCLC to many anticancer agents, during the last 30 years there has been only a modest change in survival, particularly for patients with extensive-stage disease. Although the initial complete response rate is ∼80% in extensive-stage disease, the commonly used regimens have achieved a median survival of less than 12 months (2, 3). Comparative trials have shown that multidrug regimens, alternating agents, or dose-dense high-dose regimens are no better than cisplatin plus etoposide, a standard two-drug regimen (2, 4–6). An understanding of the molecular pathogenesis of SCLC may lead to new treatment strategies.

The retinoblastoma (pRb) tumor suppressor gene is inactivated in more than 90% of SCLC (7, 8) and, consequently, is the most common molecular abnormality in SCLC. pRb has a critical role in the proliferative response of the cells to mitogenic signals (9–11). Phosphorylation of the Rb protein is regulated in a cell cycle–dependent manner. Hypophosphorylated Rb protein assembles transcriptional repressor complexes on promoters of E2F-regulated genes and inhibits cell cycle progression. Cells respond to mitogenic signals by sequential activity of cyclin–cyclin-dependent kinase complexes and phosphorylation of pRb. This chain of events disrupts the Rb interaction with E2F1-3 family members and releases the cells into cell cycle progression. Thus, pRb is considered to be the master regulator of the G₁-S phase transition (9–11).

Doxorubicin, a widely used chemotherapeutic for SCLC, has cytotoxic activity in all phases of the cell cycle, more predominantly during the S phase. The aromatic chromophore portion of doxorubicin interacts with two base pairs of the DNA by intercalation (12, 13). This inhibits the progression of the enzyme topoisomerase II that unwinds DNA for transcription. Doxorubicin-stabilized topoisomerase II complex prevents resealing of the DNA double helix. Irreparable DNA damage induces p53 activation and consequently apoptosis (14). Doxorubicin cytotoxicity mechanisms also involve inhibition of RNA polymerase and cytochrome c oxidase, and chelation of iron and generation of reactive oxygen species (15).

Flavopiridol is a novel flavone that inhibits multiple cyclin-dependent kinases (CDK) that govern cell cycle progression at the G₁-S and G₂-M boundaries (16–19). Flavopiridol is an antiproliferative and an apoptotic agent that also has antiangiogenic properties (5, 20–23).

We previously reported in a sarcoma model that flavopiridol blocked pRb-expressing cells (i.e., normal replicating cells) at the G₁-S boundary, whereas the cells lacking pRb or with nonfunctional pRb kept cycling (24). Therefore, cells with pRb were protected from the effects of doxorubicin after flavopiridol pretreatment, whereas cells lacking pRb were not. The flavopiridol then doxorubicin treatment combination had a synergistic cytotoxic effect on pRB lacking human osteosarcoma cell line (SaOs-2) whereas their pRb-expressing counterparts had a G₁ arrest with flavopiridol exposure and had a decreased cytotoxic response to following doxorubicin exposure (24).

The purpose of the present study was to test this observation in SCLC, with particular reference to induction of synergistic cell kill in a de novo pRb-negative cell line, H69, and a pRb-expressing cell line, H865, and a pRb small interfering RNA (siRNA) knockdown H865 clone. We show that de novo SCLC cells are relatively sensitive to both flavopiridol and doxorubicin, and flavopiridol and doxorubicin combinations exert both in vitro and in vivo synergistic effects. Ablating the Rb expression in H865 cell line with stable pRb siRNA transfection did not alter tumor growth kinetics neither in vitro nor in vivo and the relative drug sensitivity of the cell lines; however, the H865 pRb knockdown cell line gained significant sensitivity to the sequential flavopiridol-doxorubicin combination.

Cell lines and drugs. A SCLC cell line, NCI H69, that lacked pRb was purchased from the American Type Culture Collection and was cultured in RPMI 1640 (Life Technologies) supplemented with 2 mmol/L l-glutamine, 1.5 g/L sodium bicarbonate, 4.5 g/L glucose, 10 mmol/L HEPES, 1.0 mmol/L sodium pyruvate, and 10% heat inactivated fetal bovine serum (FBS). The NCI H865 cell line, expressing functional pRb, was generously provided by Adi Gazdar (Southwestern Medical Center at Dallas, Dallas, TX) and was cultured in DMEM/F-12 containing 10% FBS.

Flavopiridol was obtained from the Cancer Therapy Evaluation Program as a pure powder, and doxorubicin hydrochloride was obtained from Sigma-Aldrich Co. Both drugs were dissolved in DMSO and stock solutions were stored at −20°C.

Retinoblastoma gene knockdown in the H865 cell line. The H865 cell line was transduced with Rb-specific siRNA coding pFIV U6/H1 vector (System Biosciences) to generate stable pRb-negative H865 clones. Two different siRNA sequences known to down-regulate Rb mRNA >90% were chemically synthesized (25). The sequences for the siRNA template were as follows: (a) 5′-aaagGATACCAGATCATGTCAGA-3′, 3′-CTATGGTCTAGTACAGTCTaaaa-5′ and (b) 5′-aaagCCCAGCAGTTCGATATCTA-3′, 3′-GGGTCGTCAAGCTATAGATaaaa-5′. Oligonucleotides were annealed, phosphorylated, and then inserted into pFIV-H1/U6 siRNA expression vector that contained both ampicillin and puromycin resistance genes (System Biosciences). Vector siRNA template insertions were confirmed with sequencing. 293TN cells were transfected with pPACK packaging plasmid mix and lentiviral siRNA constructs using Lipofectamine (Invitrogen Corporation). The pFIV U6/H1 construct expressing siRNA against luciferase mRNA was used as a control. All vectors were pseudotyped with vesicular stomatitis virus G protein envelope.

The H865 cell line was transduced with high-titer vector stocks (>10⁷ tu/mL) in the presence of polybrene (8 μg/mL), and then single cell–derived clones were generated with puromycin (4 μg/mL) selection. Seventeen clones were picked with cloning rings and their genomic DNA was isolated. Vector insertion was evident in all of the clones with woodchuck hepatitis virus posttranscriptional regulation element (WPRE)–specific PCR. The primers for PCR were as follows: forward primer 5′-TCATGCTATTGCTTCCCGTATGGCTTTC-3′, and reverse primer 5′-GGATTGAGGGCCGAAGGGACGTAG-3′.

In vitro cytotoxicity assay. For single-drug cytotoxicity, cells were plated in 96-well plates as follows: 1 × 10⁴ per well for H865 parental cell line and its derivatives and 5 × 10⁴ per well for H69. Serial dilutions of each drug or equivalent volumes of DMSO were then added in quadruplicate to the plates. After 24 h of treatment, drugs were removed and cells were incubated in a drug-free medium for 3 more days. On day 5, tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H tetrazolium, inner salt; MTS] and the electron coupling reagent phenazine methosulfate were added to the wells according to the manufacturer's protocol (Promega Co.). After 4 to 6 h of incubation, a microplate reader (Molecular Devices Corp.) was used to read the absorbance of the wells at 492 nm. Inhibitory concentration values at 50% (IC₅₀) were calculated with Hill's equation by using the Graph Pad Prism software v3.0.

The effect of sequential exposure to flavopiridol and doxorubicin was assessed by exposing cells to graded concentrations of drugs at fixed ratios. The schedules tested were as follows: (a) flavopiridol for 24 h and then doxorubicin for 24 h; (b) doxorubicin for 24 h and flavopiridol for 24 h; and (c) flavopiridol for 24 h, 2 d of washout period and then doxorubicin for 24 h. Fixed ratios were flavopiridol IC₁₀ and doxorubicin IC₅₀, or flavopiridol IC₂₅ and doxorubicin IC₅₀. Serial dilutions of drug concentrations were done as follows: flavopiridol (1/8, 1/4, 1/2, 1, 2, 4, 8 ×IC₁₀) and doxorubicin (1/8, 1/4, 1/2, 1, 2, 4, 8 ×IC₅₀). The fractional survival for each combination was calculated by dividing the mean absorbance value of drug-treated wells to the mean absorbance value of control wells. The median effect lines were analyzed by the method of Chou and Talalay, using CalcuSyn software (Biosoft; ref. 26). Combination index (CI) values were calculated under the assumption that drugs act at different targets (i.e., mutually nonconclusive drug interaction). With this methodology, synergy was indicated by CI < 1, additive by CI = 1, and antagonism by CI > 1. For combination experiments, each condition had six to eight replicate wells and experiments were repeated at least thrice (correlation coefficients for median-effect lines >0.9).

Western blot analysis. The clones were then tested for pRb expression with Western blotting. Cell lysate proteins were separated by 8% SDS-PAGE and then transferred to Westran S polyvinylidene difluoride membranes (Whatman). After blocking with 5% nonfat dried milk solution, the membranes were incubated first with primary antibodies (i.e., mouse monoclonal IgG1 to pRb, Santa Cruz Biotechnology) or purified mouse anti-human underphosphorylated pRb or polyclonal α-tubulin antibody (Sigma Chemical Co.), and then with secondary antibody, goat anti-mouse IgG (Santa Cruz Biotechnology), conjugated to horseradish peroxidase. The blots were visualized by enhanced chemiluminescence detection kit (Amersham Biosciences). We compared the Rb protein levels from different clones by using chemiluminescence detection (ChemiDoc XRS; Bio-Rad Laboratories) and Quantity One 1-D Analysis software.

Cell cycle analysis. H69, H865 parental, and H865 Rb siRNA clone 12 cells were dispersed into 12-well plates (2 × 10⁵/well, triplicate) 24 h before exposure to flavopiridol and doxorubicin. Single drug exposures were for 24 h and then cells were transferred into fresh growth medium. For combination studies, flavopiridol and doxorubicin doses were scheduled as IC₂₅ of flavopiridol day 1 and IC₅₀ doxorubicin on day 2 (flavopiridol then doxorubicin) or in reverse order (doxorubicin then flavopiridol). Following drug treatment, cells were harvested at either 48 or 72 h time points and then stained with propidium iodide/RNase staining buffer (BD Pharmingen) according to the manufacturer's protocol. Samples were acquired with a FACSort flow cytometer (Becton Dickinson) and then the DNA distribution for cell cycle analysis was determined with ModFit v3.0 software (Becton Dickinson).

Annexin V staining for apoptosis. Apoptotic populations of the H865 parental and pRb knockdown clone 12 cell lines were compared after drug exposure in three independent experiments. Flavopiridol and doxorubicin single drug incubations were limited to 24 h, and for combinations each drug exposure was also 24 h. At the end of 48 h, all samples were harvested and washed with PBS. For each condition, 10⁵ cells were stained with 5 μL Annexin V-FITC (BD Pharmingen) and 5 μL 7-amino-actinomycin D (BD Pharmingen, Franklin Lanes NJ). Cells were then analyzed with a FACSort flow cytometer (Becton Dickinson). The percentage of early apoptotic cells was determined as percentage of Annexin V-FITC–positive but 7-amino-actinomycin D–negative cell populations using the CellQuest software.

In vivo studies. Athymic nude male mice were purchased from The Jackson Laboratory at 5 to 6 wk age. When the animals were 7 to 8 wk old, they were injected with human NCI-H69 (pRb negative), parental H865 (pRb positive), or pRb siRNA–transduced H865 cells (clone 12). A suspension of 10⁷ cells in 0.1 mL of RPMI 1640 was injected s.c. into the right flanks of mice. In 2 wk when the tumor diameters reached ∼0.5 cm, the mice were randomly assigned into six groups: (a) group 1, control (no drugs, 1.0% DMSO/sodium chloride solution); (b) group 2, flavopiridol (5 mg/kg/d, every other day for 3 d); (c) group 3, doxorubicin (4 mg/kg/d, every other day for 3 d); (d) group 4, flavopiridol (5 mg/kg/d, on days 1, 3, and 5) followed by doxorubicin (4 mg/kg/d, on days 2, 4 and 6); (e) group 5, doxorubicin (4 mg/kg/d, on days 1, 3, and 5) followed by flavopiridol (5 mg/kg/d, on days 2, 4, and 6); and (f) group 6, flavopiridol (5 mg/kg/d, on days 1, 5, and 9) followed by delayed doxorubicin (4 mg/kg/d, on days 4, 8, and 12) exposure. Tumor size and body weight were assessed every other day. Tumor volume was calculated as tumor volume (mm³) = length × (width)²/2. Toxicity was monitored by weight loss. Mice were sacrificed when the diameter of tumor exceeded 1.5 cm.

We used the Cox proportional hazards models to model the survival of the animals (27). All models were fitted using the R statistical system version 2.6.0 (28). For comparing the effects of drug treatment modalities on tumor growth, we calculated the time to reach 1.5 cm³ tumor volume for each animal, and the mean values for each group were compared by one-way ANOVA test and Bonferroni's multiple comparison test. However, substantial toxicity of certain drug treatment options and toxicity-related animal deaths required analyzing the data further for understanding the survival differences among groups. We considered two outcome measures that were measured over time in each animal: tumor volume and total body weight change as a percentage of the initial animal weight. The tumor volume measurements were log transformed to enable the fitting of linear relationships of this variable with time. To account for repeated measurements on individual animals and for the variability of the outcome/time relationships from animal to animal, we used random-effects models (29). Slopes and 95% confidence intervals for linear relationships of each of the two variables and for each of the six treatments with respect to time were computed from the fitted models.

Expression of pRb in SCLC cell lines. All three PCR reactions with WPRE, U6/H1, and puromycin resistance gene–specific primers (data are not shown for the last two primer sets) were positive for all vector-transduced H865 clones and negative for H69 and parental H865 cell lines. Figure 1A depicts the PCR reaction for WPRE, including control, H865 luciferase (Luc) siRNA clone, and H865 Rb siRNA carrying clones 16 and 12.

Out of 17 H865 Rb siRNA clones, only two (clones 12 and 17) had undetectable levels of Rb protein, whereas the other 15 clones had pRb levels of 5% to 100% compared with the pRb level of the original H865 cell line. Western blot analysis revealed that parental H865, the control Luc siRNA-transduced H865 clone, and the Rb siRNA-transduced clone 16 expressed pRb whereas H69 cells and H865 Rb siRNA clone 12 had undetectable levels of Rb protein (Fig. 1B).

Flavopiridol and doxorubicin sensitivity of the cell lines. Flavopiridol has been reported to induce growth inhibition and cytotoxicity in SCLC cell lines (30). The loss of cell cycle regulation in SCLC cells, largely due to loss of Rb function, results in a defective G₁-S restriction point in these cells and makes them unresponsive to normal cell cycle regulation. Inhibiting the CDK pathway with flavopiridol interferes with cell cycle progression in these cells and induces either a cytostatic or an apoptotic effect. SCLC cell lines have been observed to be highly sensitive to flavopiridol-induced apoptosis. In our study, the de novo Rb-negative cell line H69 was the most sensitive cell line to both flavopiridol (IC₅₀ 167.5 ± 24.5 nmol/L) and doxorubicin (IC₅₀ 121.4 ± 41.1 nmol/L; Fig. 1C).

The IC₅₀ values for doxorubicin for the parental H865 cell line and its derivative clones with FIV integration were similar (970 ± 133.3, 1,067 ± 90.3, and 932 ± 225.7 nmol/L for H865 parental, H865 Rb siRNA clone 12, and H865 Rb siRNA clone 16, respectively). However, FIV-integrated H865 clones became more resistant to flavopiridol than the parental H865 cell line (Fig. 1C; P < 0.03). The H865 cell line transduced with control Luc siRNA FIV was also more resistant to flavopiridol (IC₅₀ 750 ± 187 nmol/L) than the parental H865 cell line, whereas doxorubicin sensitivity was not significantly different from the other H865 cell lines (IC₅₀ 1,297 ± 345 nmol/L, P = 0.626).

Sequential flavopiridol and doxorubicin combination has a synergistic cytotoxic effect on Rb-negative SCLC cell lines. Flavopiridol combinations with a variety of chemotherapeutic agents are observed to have either synergistic or antagonistic effects depending on their administration schedule (31). Flavopiridol administration preceding doxorubicin was observed to result in synergistic cell kill in an osteosarcoma cell line (24). The effect of sequential administration on the cytotoxic interaction between flavopiridol and doxorubicin was tested with three different schedules: 24 hours of flavopiridol and then 24 hours of doxorubicin exposure, doxorubicin then flavopiridol exposure in reverse order, and 24 hours of flavopiridol followed by 48-hour washout period and then 24 hours of doxorubicin exposure. For combination studies, fixed drug IC_x ratios were as follows: flavopiridol IC₁₀ and doxorubicin IC₅₀, or flavopiridol IC₂₅ and doxorubicin IC₅₀. Figure 2 presents the results for flavopiridol IC₂₅ and doxorubicin IC₅₀ fixed ratios but the other drug combination ratio (flavopiridol IC₁₀ and doxorubicin IC₅₀) had similar results (data not shown). Combination indexes presented in Fig. 2 are the means of three independent experiments and each experimental plate had six to eight replicates for each serial dilution point. In the H69 cell line, all flavopiridol and doxorubicin combinations were synergistic; however, in the H865 Rb siRNA clone 12, the greatest reduction in IC₅₀ values were observed when 24 h of flavopiridol incubation was followed by 24 h of doxorubicin exposure. Meanwhile, for all Rb protein–expressing cells [i.e., H865 parental, H865 with Luc siRNA (data not shown), and H865 Rb siRNA clone 16], all flavopiridol and doxorubicin combinations were antagonistic (Fig. 2). Although the flavopiridol sensitivity of the H865-derived cell lines was decreased, the Rb knockdown H865 derivatives became more sensitive to the flavopiridol and doxorubicin combination.

The flavopiridol and doxorubicin combination induces different cell cycle changes in Rb-negative and Rb-positive SCLC cell lines. To explore the cell cycle effects of flavopiridol and doxorubicin and their combination, H69 (Supplementary Fig. S1A-E), H865 Rb siRNA clone 12 (Supplementary Fig. S1F-J), and H865 parental cells (Supplementary Fig. S1K-O) were treated with flavopiridol IC₂₅ alone, doxorubicin IC₅₀ alone, flavopiridol IC₂₅ then doxorubicin IC₅₀, or doxorubicin IC₅₀ then flavopiridol IC₂₅. Cell line relevant IC₂₅ and IC₅₀ values that were used in these experiments were similar to flavopiridol and doxorubicin doses summarized in Fig. 2. Cells were stained with propidium iodide at the 6-, 12-, 24-, 48-, and 72-hour time points; the 48-hour time point results are shown in Fig. 3. In all cell lines, doxorubicin alone induced a substantial increase in the percent of cells in G₂-M (Fig. 3A-C; P < 0.01). In the H865 parental cell line, flavopiridol induced G₁ accumulation (G₁ fraction for control and flavopiridol treated cells were 44 ± 2% and 56.6 ± 0.8%, respectively) starting from as early as 24 hours (P < 0.05). None of the flavopiridol and doxorubicin combination sequences were able to change doxorubicin-triggered G₂-M accumulation in the H865 parental cell line. In contrast, flavopiridol did not alter the cell cycle distribution of the Rb-negative H69 cells. Cell cycle modeling of the ModFit program failed to detect the flavopiridol then doxorubicin combination–induced G₁-S block in Rb-negative cell lines. The combination triggered both apoptotic sub-G₁ and shifted S-phase peaks that substantially overlapped with the G₁ peak (Supplementary Fig. S1D). There was a prominent shift of the channels corresponding to N and 2N compared with the other H69 cell cycle plots. Nevertheless, flavopiridol then doxorubicin induced prominent S-phase accumulation in H69 cells (Fig. 3A). In the H865 Rb siRNA clone 12 Rb, doxorubicin alone and doxorubicin followed by flavopiridol exposure resulted in a G₂-M block (Fig. 3C); however, interestingly, these cells responded to flavopiridol then doxorubicin treatment with a G₁-S accumulation, similar to H69 cells.

Flavopiridol then doxorubucin exposure induces the highest amount of apoptosis in the Rb knockdown H865 cell line. Flavopiridol then doxorubicin treatment–related cell cycle changes, in conjunction with summarized in vitro synergism data, made us inquire as to whether this flavopiridol dysregulated G₁-S transition was increasing the doxorubicin exposure of the cells and subsequently increasing DNA damage–induced apoptosis.

Apoptotic populations of the H865 parental and its Rb knockdown derivative cell line were compared after flavopiridol and/or doxorubicin exposure (Fig. 4). H865 Rb siRNA clone 12 and H865 parental cells were treated with flavopiridol IC₂₅ alone, doxorubicin IC₅₀ alone, flavopiridol IC₂₅ then doxorubicin IC₅₀, or doxorubicin IC₅₀ then flavopiridol IC₂₅. The cells were exposed to the same drug doses that were summarized in Fig. 2. Flavopiridol exposure induced statically significant apoptosis in only the H865 Rb siRNA clone 12 (upon flavopiridol exposure the apoptotic cell population increased from 9.1 ± 2.5% to 18.8 ± 1.7%), but the most marked change was with flavopiridol then doxorubicin sequential treatment (apoptotic population increased to 37 ± 7.7%; P < 0.02; Fig. 4). Treatment did not alter the apoptotic cell populations for H865 parental cells; however, interestingly, we observed less apoptosis (5.4 ± 1.2%) with the doxorubicin then flavopiridol sequence in the parental cell line (Fig. 4).

We performed similar apoptosis experiments with the H69 cell line; however, they were very sensitive to the Ca²⁺-containing Annexin V binding buffer, and even nontreated cells lost 75% of their viability within 15 minutes of incubation time.

Therapeutic efficiency of sequential flavopiridol and doxorubicin in nude mice bearing human SCLC xenografts.In vivo growth kinetics of the cell lines H69, H865 parental, and H865 Rb siRNA clone 12 were different from their in vitro growth patterns. H69 cells grow dominantly as spheres in culture and their generation time varies between 4 and 5 days. This cell line is well known for its extracellular matrix dependence for survival (32). H69 cells grew faster in vivo than the other two cell lines after injection of 10⁷ cells, although both of the H865 cell lines have an average generation time of 2 days in vitro. The animals bearing the H69 tumor had a median survival of 38 days, whereas the animals with either H865 parental or H865 Rb siRNA clone 12 had median survivals of 42 and 50 days, respectively. In both H69 and H865 Rb siRNA clone 12 xenograft models, flavopiridol then doxorubicin treatment modality provided the longest median survivals of 50 and 68.5 days, respectively (Fig. 5A and C; P < 0.001 and P < 0.003, respectively). We calculated the change in tumor size in log volume per day with a random-effects model, as well as mean time to reach 1.5 cm³ tumor volume (Fig. 5D). For the H69 xenograft, treatment with flavopiridol alone treatment modality had the steepest log tumor volume change per day (mean 0.22 per day with a 95% confidence interval range of 0.186-0.257 per day), whereas flavopiridol followed by delayed doxorubicin had similar log volume change as the control (0.20 and 0.197 per day, respectively). Doxorubicin alone treatment decreased the growth rate to 0.17 log tumor volume change per day (95% confidence interval range of 0.145-0.19 per day). Flavopiridol/doxorubicin combinations had similar log volume changes of 0.15 (P < 0.05) that were significantly lower than the other groups. However, doxorubicin then flavopiridol treatment (−0.30 log percent weight change per day, range −0.4 to −0.20 per d) was more toxic (P < 0.01) than both doxorubicin alone (−0.175 per day with a range of −0.29 to −0.06 per day) and flavopiridol then doxorubicin sequential combination (−0.24 per day with a range of −0.34 to −0.19 per day).

For in vivo studies, a nontoxic flavopiridol dose that would induce mainly a G₁ block was used. We did not observe any overt total body weight loss with flavopiridol treatment in any of the xenograft models. Flavopiridol treatment alone did not alter the tumor growth rate in the H865 xenografts and slightly increased the tumor growth rate in the H69 model (P < 0.03; Fig. 5A-D). Doxorubicin alone (−0.42 per day, with a range of −0.60 to −0.25 per day percent weight change in log) and doxorubicin then flavopiridol treatment (−0.55 per day, with a range of −0.70 to −0.39 per day) modalities were the most toxic treatment modalities in the H865 Rb0 model (P < 0.001). Flavopiridol then doxorubicin treatment had less toxicity than doxorubicin then flavopiridol treatment in both H69 and H865 Rb knockdown models. Although the flavopiridol then delayed doxorubicin sequencing had less toxicity than the other sequential treatments, this sequence did not show significant tumor growth control in any of the tumor models (Fig. 5D). In the H865 parental xenograft model, tumor growth and survival rates of the drug treatment groups were similar to that of the no treatment group (P > 0.05; Fig. 5B and D).

Therefore, we observed a significant survival advantage in Rb-negative SCLC xenograft models with the sequential flavopiridol then doxorubicin treatment program, as a result of its better tumor growth delay as well as less toxicity.

SCLC cells are not only devoid of G₁-S cell cycle restriction point due to loss of functional retinoblastoma protein expression, but may also have a defective G₁-S DNA damage check point, and defective G₂-M and M check points, due to loss of p53 function, mutation, or aberrant methylation of the different checkpoint regulator genes (7, 8). This high degree of variability of cell cycle components and their regulators in SCLC makes the predictions about response to cell cycle–directed agents very complex. Almost 60% to 80% of the SCLC patients express abnormal p53 without any chemotherapy response correlation or prognostic significance (33) and in vitro, wild-type p53 restoration does not modify the sensitivity of a lung cancer cell line to a given chemotherapeutic agent (34). Flavopiridol is known to induce apoptosis in certain cell types independent of Rb and p53 pathway alterations (35, 36). As a single agent, flavopiridol has unsatisfactory or limited value in cancer treatment but it has been observed to be most effective when combined with other chemotherapeutic agents (21, 37–41). Flavopiridol not only disrupts cell cycle progression by interfering with the CDK cascade but also reduces the levels of different antiapoptotic proteins (42–47). SCLC cell lines have been noticed to have enhanced sensitivity to flavopiridol exposure relative to non-SCLC cell lines (30, 36). Flavopiridol could also induce apoptosis in primary cells, but that dose is usually much higher than the dose required for SCLC apoptosis (30). SCLC cells have defects in both G₁-S and G₂-M checkpoint regulation, and S-phase synchronization increases flavopiridol-induced apoptosis in SCLC cell lines (30). In the presence of functional pRb, inhibition of the CDK cascade, i.e., cyclin D1/cdk4,6 or cyclinE/cdk2, by CDK inhibitors dominantly arrest the cells at G₁ and makes them less responsive to S-phase–acting agents (31). In cells that display a high-grade G₁ arrest with flavopiridol, like our H865 parental model, the response to drug treatment was observed to be more cytostatic rather than apoptotic (30). Doxorubicin is known to have significant activity in SCLC, alone or in combination (48–50). With this knowledge, we aimed to combine flavopiridol and doxorubicin so that the effects of either agent are not inhibited but rather augmented.

We previously observed in pRb-negative osteosarcoma cells (SaOs-2) that flavopiridol pretreatment made them more sensitive to doxorubicin, whereas a subline that was transfected with pRb arrested at G₁-S with flavopiridol treatment and had decreased cytotoxicity to doxorubicin (24). These observations encouraged us to test the efficiency of different flavopiridol and doxorubicin combinations in SCLC, a tumor that has a high frequency of pRB loss. The proposed mechanism was that in the presence of an intact G₁-S checkpoint, cells with functional pRb would be blocked in G₁-S by flavopiridol and subsequent exposure to doxorubicin would have a lessened effect as the drug exerts its maximum toxicity in late S-G₂. Furthermore, as normal renewal tissue cells also have an intact G₁-S checkpoint, toxicity to these tissues might also be lessened.

Cell cycle data showed substantial S-phase accumulation and apoptotic sub-G₁ population with this schedule, but because of innate sensitivity of the H69 cell line to the Annexin binding buffer, we were not able to quantitate apoptosis with Annexin V staining. It is known that in vitro survival of H69 cells increases in the presence of extracellular matrix proteins, and they become less sensitive to chemotherapeutics (32). In the nude mouse-human H69 xenograft model, flavopiridol alone treatment and the schedule of flavopiridol then delayed doxorubicin were the worst in terms of tumor volume control, but were also the least toxic. The doxorubicin then flavopiridol schedule exerted some in vitro synergism, but the toxicity of this schedule was far worse than all the of treatment options. Log tumor growth change per day transformation of the tumor volumes revealed similar tumor growth rates both for flavopiridol then doxorubicin and doxorubicin then flavopiridol combinations, but treatment-related toxicity caused doxorubicin then flavopiridol treatment group to have a significantly different survival than the flavopiridol then doxorubicin–treated animals. H69 exerts different growth kinetics and increased drug resistance in vivo (32), and that may be a plausible explanation for the results obtained for H69 cells in vitro compared with in vivo. The in vivo extracellular matrix support might have allowed H69 cells to become less sensitive to flavopiridol and doxorubicin in xenografts and may have somewhat attenuated the response to flavopiridol and doxorubicin combinations from what we expected from the in vitro synergism data.

For understanding the role of Rb deficiency in response to flavopiridol and doxorubicin combination, we also used a Rb-positive H865 SCLC model for testing our hypothesis. The caveat is, of course, that each tumor has its own distinct phenotype and consequently sensitivity to doxorubicin varied considerably from one cell type to another. Further, tumor growth rates varied, making comparison of the results of in vivo treatment more difficult. Therefore, we followed these studies by knocking down pRb in the H865 cell line to compare this sequence in a cell line that presumably would only differ by the level of pRb expression. The synergistic effect noted with the flavopiridol sequence was coupled with lessened toxicity in these SCLC lines. The H865 cell line was intrinsically more resistant to both flavopiridol and doxorubicin, and knocking down Rb with a specific siRNA vector did not change doxorubicin sensitivity. H865 Rb siRNA clones had a variable amount of Rb expression. The in vitro growth kinetics of the H865 Rb siRNA clone was similar to that of H865 parental cell line, but the siRNA Rb clone grew somewhat slower in vivo. None of the treatment modalities exerted any in vivo tumor growth control for the H865 parental cell line. Doxorubicin alone and doxorubicin then flavopiridol combinations were substantially more toxic than flavopiridol followed by doxorubicin treatment. Seven weeks after one cycle of flavopiridol then doxorubicin treatment, 50% of the animals were still alive. Although this model is very artificial and might not carry the complexities of de novo Rb-deficient SCLC, it confirms and extends the observation previously reported for an osteosarcoma cell line and provides a rationale to test sequential low-dose flavopiridol followed by doxorubicin in SCLC and other malignancies that lack functional pRB and are sensitive to doxorubicin.

Low-dose flavopiridol did not exert toxicity but moderately reduced the toxicity of doxorubicin in all of our xenograft models. We did not aim for high-grade cell cycle arrest and apoptosis with flavopiridol-alone treatment (30). With a rather lower dose of flavopiridol, we wanted to induce a relatively longer G₁-S block in normal tissue cells for protection during doxorubicin treatment.

In our experimental model, we used both a de novo Rb-deficient SCLC cell line, H69, and a Rb knockdown clone, H865 Rb siRNA clone 12, for evaluating the Rb-dependent effects of flavopiridol and doxorubicin combinations. Rb status in H865 cells did not affect cell growth in vitro and had minimally prolonged tumor growth in the H865 Rb siRNA clone 12 model. The cellular response to drug treatments was highly dependent on the Rb status of the cells both in vitro and in vivo. Flavopiridol then doxorubicin sequential treatment had strong in vitro synergism (CI < 0.7) only in Rb-deficient cells, H69, and H865 siRNA clone 12, but was antagonistic in Rb-positive H865 cells. Doxorubicin then flavopiridol treatment had no effect on pRb-negative cells but decreased apoptosis in Rb-positive H865 parental cells in culture. Flavopiridol treatment alone accelerated tumor growth in de novo Rb-negative H69 cell line but not in Rb-positive H865 and its Rb knockdown derivative, H865 Rb siRNA clone 12, cells. Additionally, flavopiridol delayed doxorubicin treatment accelerated tumor growth both in Rb-negative and Rb-positive cells. These observations might suggest that tumor growth acceleration occurred in an Rb-independent manner in these models. The flavopiridol then doxorubicin treatment led to increased survival only in Rb-negative xenograft models, H69 and H865 Rb siRNA clone12, by both decreasing toxicity and increasing tumor cell kills.

Collectively, these results provide evidence that de novo Rb status and timing of flavopiridol and doxorubicin combination treatment greatly affects the response. This might have important clinical implications because it suggests that tumor growth could actually be enhanced by the treatment in specific clinical scenarios. The relevance of these observations to the response to this flavopiridol followed by doxorubicin sequence in SCLC patients requires further investigation with clinical trials.

No potential conflicts of interest were disclosed.