Pharmacologic inhibition of Pim kinases alters prostate cancer cell growth and resensitizes chemoresistant cells to taxanes

Pim kinases are a family of serine/threonine kinases that regulate cell survival (1). This family of kinases is composed of three different isoforms (Pim 1, Pim 2, and Pim 3) that are highly homologous at the amino acid level, yet differ partially in their tissue distribution (2, 3). Although the cellular localization of Pim kinases is mainly cytoplasmic, some evidence suggests that they can also be found in the nucleus and more recently on the cell surface (4, 5). The first family member identified, pim-1, was initially discovered as a preferential integration site for the Moloney murine leukemia virus and has subsequently been associated with cell cycle progression, apoptosis, and tumorigenesis (6, 7).

Unlike other serine/threonine kinases, the Pim family members are tightly regulated at the level of transcription and translation (8). Their expression is mediated by the Janus-activated kinase/signal transducers and activators of transcription signaling pathway, which is activated by various cytokines and hormones (9). Pim kinases are known to suppress apoptosis by directly phosphorylating and inhibiting the proapoptotic Bcl-2 family member, Bad (10–12). Additionally, Pim-1 kinase can regulate cell cycle progression at the G₁-S and G₂-M transition phases by its ability to phosphorylate Cdc25A (13), p21^Cip1/WAF1 (14), and p27^Kip1 (15) as well as Cdc25C (16), respectively.

The Pim kinase family members are considered oncogenes and have been implicated in tumorigenesis either alone or functioning synergistically with c-Myc (17–19). Initially, elevated levels of Pim kinases were discovered in leukemia and lymphoma tumors (3, 7, 20); however, more recently, they were found to be overexpressed in solid tumors, including pancreatic cancer and prostate cancer (11, 21). Several independent studies have shown that Pim-1 is upregulated in human prostate cancer clinical samples and in various animal models (18, 21, 22). Additionally, interleukin-6, a cytokine known to be elevated in sera from patients with advanced prostate cancer, has been shown to induce Pim-1 kinase through activation of the Janus-activated kinase/signal transducers and activators of transcription signaling pathway in prostate cancer systems (23, 24). Recent studies have also correlated Pim-1 kinase with chemoresistance in prostate cancer cells, which is a common occurrence in more aggressive, hormone-refractory prostate cancers (25, 26).

It is believed that the oncogenic and prosurvival potential of the Pim kinases contributes to tumorigenesis and promotes drug resistance in prostate cancer. Therefore, recent efforts have been placed on the identification of Pim kinase inhibitors as potential therapeutic targets of cancer (1, 27–31). The purpose of this study was to characterize a novel small molecule inhibitor of Pim kinases, SGI-1776, in prostate cancer model systems to determine whether targeting Pim kinase activity alone or in combination with a taxane-based chemotherapeutic regimen could have a significant impact on the treatment of prostate cancer. We show that SGI-1776 is a fairly selective inhibitor of the Pim kinases and can cause a reduction in prostate cancer cell viability and alter overall cytotoxic responses to taxanes. Specifically, we reveal that SGI-1776 can disrupt cell cycle progression, induce apoptosis, promote taxane sensitivity, and overcome acquired chemoresistance by inhibiting multidrug resistance (MDR)1 activity in specific prostate cancer models. These findings suggest that Pim kinase inhibition may prove to be a beneficial strategy for treating prostate cancer and in particular those that are nonresponsive to taxane-based chemotherapy.

Kinase inhibition was measured by radiometric assays using the KinaseProfiler service at Millipore.⁷

The LNCaP, PC3, DU145, and 22Rv1 prostate cancer cell lines were acquired from American Type Culture Collection and cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS). The tumorigenic RWPE2-W99 cell line (American Type Culture Collection) was maintained in Keratinocyte Serum-Free Medium (Life Technologies Invitrogen) supplemented with bovine pituitary extract and human recombinant epidermal growth factor (EGF). The LAPC-4 and C4-2B cell lines were acquired through an MTA from University of California at Los Angeles and ViroMed Laboratories, respectively (32). The LAPC-4 cell line was maintained in Iscove's modified medium and 5% FBS, whereas the C4-2B cell line was cultured in RPMI 1640 and 10% FBS.

The pBK/CMV-Pim-1 plasmid (33) used in these studies was a generous gift from Dr. Nancy Magnuson (Washington State University, Pullman, WA). The stable 22Rv1 cell line overexpressing Pim-1 cDNA (22Rv1-Pim1) was generated by Lipofectamine 2000 transfection of 22Rv1 cells with the pBK/CMV-Pim-1 plasmid followed by selection in 400 μg/mL G418.

The paclitaxel-resistant 22Rv1 cell line (22Rv1-T) was generated in vitro by serially passaging the parental 22Rv1 line in the presence of increasing concentrations of paclitaxel (0–400 ng/mL). After a period of ∼4 mo, the paclitaxel-insensitive derivative of 22Rv1 emerged and is currently being maintained in the presence of 5 ng/mL of paclitaxel.

Approximately 1,000 cells per well were plated in ScreenMates 384-well tissue culture–treated microplates (Thermo Fisher Scientific) in 1% serum (except for RWPE2-W99 cells, which were plated in complete Keratinocyte serum-free media conditions). The following day, increasing concentrations of SGI-1776 and/or paclitaxel were added to the respective wells. After a specified period of time, cell viability was assessed by adding 25 μL of ATPlite reagent (Perkin-Elmer) to each well. Luminescence was determined by reading the 384-well plates on the VICTOR Light Luminescence Counter machine (Perkin-Elmer). The results are based on an average of five wells per experimental sample.

C4-2B and 22Rv1 cells were cultured in 1% serum overnight and then treated with SGI-1776 for 24 h. Cells were harvested and fixed in cold methanol, followed by the addition of propidium iodide (50 μg/mL) and RNase for 30 min to determine the percentage of cells in each phase of the cell cycle using a flow cytometer.

RNA was extracted from human prostate cancer cell lines using the Qiagen RNeasy kit (QIAGEN). Individual Taqman primer and probe sets were purchased from Applied Biosystems at a 20× concentration (see Table 1). One hundred nanograms of RNA per reaction per well were combined with a mastermix of Amplitaq Gold PCR Master Mix, the target primer set, ROX, Mulitscribe RNase, and RNase inhibitor (Applied Biosystems). Reactions were run at 50°C for 15 min, 95°C for 10 min, and then cycled 40 times at 95°C for 15 s, and 60°C for 1 min. The data were then analyzed as the ΔΔCt method on the SDS RQ manager software (Applied Biosystems). All data were normalized to 18S and displayed as relative expression.

Cells were lysed using radioimmunoprecipitation assay buffer [1 mmol/L Na3VO4, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L NaF, and protease inhibitor cocktail (Sigma)]. Proteins (70 μg) were separated and transferred on to a polyvinylidene difluoride membrane, which were then probed with the following primary antibodies: anti–phospho p21 (Thr145) at 1:200 (Santa Cruz Biotechnology, Inc.), anti-p21 at 1:200 (Santa Cruz Biotechnology, Inc.), anti-phospho BAD (Ser112) at 1:2,000 (Cell Signaling), anti-BAD at 1:1,000 (Cell Signaling), Mdr-1 (D-11) at 1:200 (Santa Cruz Biotechnology, Inc.), anti–poly ADP ribose polymerase at 1:1,000 (Cell Signaling), anti–Pim-1 at 1:200 (Santa Cruz Biotechnology, Inc.), and anti–β-actin at 1:10,000 (Sigma).

Cells were cultured in 1% serum overnight and then incubated with SGI-1776 for 24 h. Apoptosis was determined by measuring caspase-3 activity using the Caspase-3 Fluorometric Protease Assay (Invitrogen). Data were normalized to untreated samples and displayed as relative caspase-3 activity.

Assay was done according to manufacturer's instructions using a calcein acetoxymethyl ester fluorescence–based assay (Vybrant Multidrug Resistance Assay kit, Molecular Probes). Approximately 2.5 × 10⁵ 22Rv1-T cells were used per reaction and increasing concentrations of SGI-1776 and/or paclitaxel were tested in triplicate. Data are presented as a percentage of calcein retention in treated versus untreated MDR (22Rv1-T) cells. This assay was also done on the parental (22Rv1) cells as a comparison (data not shown).

Primarily, an unpaired student's t test was done to determine statistical significance in the presented studies. However, when determining whether the cytotoxic effects of two agents were additive or synergistic, a two-way ANOVA analysis was applied. For all analyses, P value of <0.05 was considered statistically significant and displayed as P < 0.05 (*), P ≤ 0.001 (**), and P ≤ 0.0001 (***).

The small-molecule inhibitor targeting Pim kinases known as SGI-1776 was provided by SuperGen, Inc. SGI-1776, a member of the imidazo[1,2-b]pyridazine chemical class, was selected for its ability to exploit an extended ATP-binding and solvent pocket, thereby making it a selective Pim family kinase inhibitor (Fig. 1A).

SGI-1776 was identified by in silico screening of 1.5 million compounds from focused and/or diverse libraries using Pim-1 kinase crystal structure in complex with AMP-PNP as a template. This discovery process identified a promising lead scaffold, imidazo[1,2-b]pyridazine, which was then used to optimize the potency and selectivity of compounds against the Pim-1 target through specific modifications at R₁ and R₂ groups predicted by our proprietary CLIMB-based approach. A novel synthetic route was developed starting from 1,4-dibromo-trans-2-butene. We prepared the 3-bromo-6-chloro-imiadazo[1,2-b]pyridazine scaffold in three steps (Fig. 1A). Suzuki coupling with various substituted aryl boronic acids, followed by introduction of aliphatic amines, led to the synthesis of >30 novel chemical entities as Pim-1 kinase inhibitors. Initial SAR studies revealed that the introduction of a hydrophobic moiety at R₁ and small ring substituent groups with 1 or 2-CH₂ spacers at the 6-amine position were vital for Pim-1 kinase inhibition.

The lead compound SGI-1776 showed potent activity against Pim-1 kinase and exhibited promising selectivity against a panel of kinases (34). Figure 1B illustrates a dose-dependent decrease in Pim-1, Pim-2, and Pim-3 activity with increasing doses of SGI-1776. The IC₅₀ of SGI-1776 for the three isoforms of the Pim kinase family, Pim-1, Pim-2, and Pim-3, were in the low nanomolar range (7 ± 1.8 nmol/L, 363 ± 27.6 nmol/L, and 69 ± 9.2 nmol/L, respectively; Fig. 1B). The specificity of SGI-1776 was further confirmed by screening SGI-1776 against a panel of >300 kinases (KINOMEscan, Ambit Biosciences). In addition to the three Pim kinases, SGI-1776 exhibited inhibitory activity against two other kinases: Flt-3 (IC₅₀, 44 nmol/L) and Haspin (IC₅₀, 34 nmol/L).

Initially, we chose to evaluate SGI-1776 for activity in prostate cancer by determining its response in a collection of prostate cancer cell lines. Before assessing the cytotoxic effects of this inhibitor, we first confirmed Pim-1, Pim-2, and Pim-3 expression by quantitative real-time reverse transcription-PCR (Fig. 2A). Differential expression of the Pim family members was seen across the various prostate cancer model systems as also reported by other groups (25, 27). After verifying the presence of the target, we used the ATP viability assay and determined that SGI-1776 reduces cell viability in the examined prostate cancer cell lines in a dose-dependent manner (Fig. 2B). The androgen-independent cell lines, 22Rv1, PC3, and C4-2B, showed the highest sensitivity to SGI-1776, with approximate IC₅₀s of 2 ± 0.05 μmol/L, 3 ± 0.07 μmol/L, and 3 ± 0.02 μmol/L, respectively. The LNCaP cell line showed the next greatest response (5 ± 0.03 μmol/L), followed by the DU145 (8 ± 0.03 μmol/L) and RWPE2-W99 (9 ± 0.09 μmol/L) cell lines, whereas LAPC-4 (>10 ± 0.14 μmol/L) cells exhibited a minimal response (Fig. 2).

We next wanted to investigate whether SGI-1776 was capable of altering specific substrates of Pim kinase (i.e., p21 and Bad) in prostate cancer cells, which may ultimately contribute to prostate cancer growth inhibition. We focused our subsequent studies on the 22Rv1 and C4-2B cells, which show the maximal response in presence of SGI-1776. Additionally, both lines are hormone refractory and express the androgen receptor (AR), making them clinically relevant model systems for studying prostate cancer.

Pim-1 kinase supports cell cycle progression by phosphorylating the cell cycle inhibitor, p21^Cip1/WAF1, at the Thr145 residue (14). Increasing concentrations of SGI-1776 reduced phospho-p21^Thr145 protein levels, whereas total p21 remained constant (Fig. 3A). Maximal inhibition of phospho-p21^Thr145 was observed at 5 and 7.5 μmol/L in 22Rv1 and C4-2B cells, respectively. In addition, treatment with SGI-1776 for 24 hours caused a G₁ arrest in a dose-dependent manner, inhibiting the natural progression to S phase (Fig. 3B). For C4-2B, the percentage of cells in the G₁ phase went from 72 ± 3.28% in the untreated group, to 81 ± 1.24% with the addition of 2.5 μmol/L SGI-1776, and 95 ± 0.16% with 7.5 μmol/L SGI-1776 (P < 0.05). The 22Rv1 line had approximately 60 ± 0.03% of untreated cells in the G₁ phase and then following treatment with 2.5 and 7.5 μmol/L SGI-1776, the percentage of cells in G₁ increased to 70 ± 2.00% and 81 ± 1.17%, respectively (P < 0.05).

Besides the ability of Pim-1 to help regulate the cell cycle through the G₁-S transition, this serine/threonine kinase has also been shown to have antiapoptotic potential through its phosphorylation of Bad at serine residue 112 (10). Western blot analysis performed on C4-2B and 22Rv1 cells treated with increasing doses of SGI-1776 revealed a clear dose-dependent decrease in phosphorylated Bad^Ser112 protein levels, whereas total Bad levels remained unchanged (Fig. 3C). A reduction in phosphorylation of Bad ultimately leads to the release of cytochrome c from the mitochondria into the cytosol, which triggers the activation of caspase-9 and caspase-3, resulting in cell death (10, 35). To confirm this mechanism of action, cells were exposed to increasing doses of SGI-1776 for 24 hours and apoptosis was determined by measuring the induction of caspase-3 activity compared with the untreated samples. As evident in Fig. 3D, inhibition of Pim kinase activity with increasing concentrations of SGI-1776 caused a dose-dependent increase in caspase-3 activity levels in both C4-2B and 22Rv1 cells, signifying the induction of apoptosis. There was a concomitant increase in levels of cleaved poly(ADP-ribose) polymerase-1, a substrate of caspase-3, further indicating apoptosis induction with SGI-1776 treatment (Supplementary Fig. S1).

To further validate that SGI-1776 can specifically target Pim kinases, which trigger the observed cytotoxic effects seen in prostate cancer cell lines, we stably overexpressed recombinant Pim-1 in the 22Rv1 cell line and compared the overall SGI-1776 treatment responses to that of the parental line. Initial characterization of the 22Rv1-Pim1 cell line confirmed Pim-1 overexpression at the protein level (Supplementary Fig. S2) as well as showed increased cell growth compared with the 22Rv1 line (Fig. 4A).

To address the specificity of SGI-1776 as a Pim kinase inhibitor, we examined the effects of SGI-1776 treatment on p-p21^Thr145 inhibition, which is a direct substrate of Pim-1. Treatment of the 22Rv1-Pim1 cells with increasing doses of SGI-1776 revealed a dramatic inhibition in p-p21^Thr145 abundance at 1 μmol/L SGI-1776 compared with a minimal, if any, effect in the 22Rv1 line at this dose (Fig. 4B). It is worth noting that a significant decrease in p-p21^Thr145 in the parental 22Rv1 cells was first observed between 2.5 and 5 μmol/L of SGI-1776 (Fig. 3A) under similar experimental conditions. It is also important to mention that Pim-1 overexpression considerably increased levels of p-p21^Thr145 in the 22Rv1-Pim1 cell line as is expected with an enhanced Pim kinase activity.

A side by side comparison of cell viability after 72 hours of treatment with increasing concentrations of SGI-1776 showed an approximate 3-fold reduction in IC₅₀ in the 22Rv1-Pim1 line compared with the 22Rv1 line, suggesting that overexpression of the Pim-1 target increases SGI-1776 sensitivity and resulting cytotoxicity (Fig. 4C). Similarly, treatment of 22Rv1-Pim1 cells with SGI-1776 also led to a greater fold induction of apoptosis compared with 22Rv1 cells (1.6-fold at 2.5 μmol/L SGI-1776; P < 0.0001) as measured by a caspase-3 fluorometric protease assay (Fig. 4D).

Taken together, these data show that SGI-1776 inhibits Pim-1 kinase in androgen-independent 22Rv1 prostate cancer cells, which results in G₁ cell cycle arrest followed by an apoptotic induction.

Previously, it has been shown that Pim-1 acts as a survival factor in the presence of chemotherapeutic agents (25). In addition to the evidence presented above, which illustrates the inhibitory effects of SGI-1776 as a single agent, we were interested in determining whether SGI-1776 is capable of sensitizing prostate cancer cells to chemotherapy treatment. We treated 22Rv1 cells with increasing doses of paclitaxel in the presence or absence of SGI-1776 for 48 hours followed by an assessment of cell viability. We established sensitivity to paclitaxel and SGI-1776 as single agents as evidenced by a dose-dependent reduction in viability (Fig. 5A). We observed an additional cytotoxic effect when SGI-1776 was combined with paclitaxel treatment at specific doses. For example, 2.5 μmol/L SGI-1776 and 2.5 nmol/L paclitaxel resulted in 63% and 62.6% viability, respectively, as single agents. Yet, when the same concentrations of each compound were added together, the cell viability decreased to 42.2%, signifying a marginal benefit with combination therapy (P < 0.0001; Fig. 5A). Similar results were observed in presence of docetaxel and SGI-1776 (data not shown). Although Pim kinase inhibition with SGI-1776 did result in a minor enhancement of taxane-sensitization, a two-way ANOVA analysis revealed that this therapeutic combination was neither additive nor synergistic in prostate cancer cells.

With the 22Rv1 cell line being the most sensitive to SGI-1776 treatment, we generated a taxane-resistant form of this line (22Rv1-T) to examine the effects of Pim kinase inhibition on acquired resistance to paclitaxel. As seen in Fig. 5B, the 22Rv1-T cell line is clearly resistant even at high doses (100 nmol/L) of paclitaxel with little decline in cell viability observed. This result is compared with the 22Rv1 parental cell line, which is sensitive to paclitaxel as evidenced by a reduction in cell viability in a dose-dependent manner. We confirmed that the 22Rv1-T line was also resistant to increasing doses of docetaxel, which is a current standard treatment for advanced hormone-refractory prostate cancer (36); however, a slight decline in viability was observed at the higher doses (i.e., starting around 50 nmol/L) of the compound (Fig. 5B). The approximate IC₅₀s for the 22Rv1 cell line treated with paclitaxel and docetaxel were 4.5 ± 0.035 nmol/L and 2.5 ± 0.034 nmol/L, respectively. On the other hand, the 22Rv1-T cell line had IC₅₀s of >100 nmol/L when treated with either taxane.

After confirmation of resistance, we further characterized the paclitaxel-resistant line to elucidate the mechanism of chemoresistance. Several reports have alluded to MDR mechanisms as the cause for resistance to chemotherapy agents in prostate cancer (37, 38). Therefore, we investigated whether MDR in the 22Rv1-T line was mediated by ATP-binding cassette (ABC) transporters such as MDR1, breast cancer resistance protein (BCRP), or multidrug resistance associated protein 1 (MRP1). Both real-time reverse transcription-PCR and Western blot analysis showed overexpression of MDR1 mRNA and protein levels in the 22Rv1-T cell line compared with the parental 22Rv1 line (Supplementary Fig. S3A). BCRP and MRP1 were not overexpressed in the taxane-resistant prostate cancer cell line (data not shown).

Following the characterization of the 22Rv1-T cell line, we investigated the effects of Pim kinase inhibition on cell viability in this prostate cancer model of acquired taxane resistance. 22Rv1 and 22Rv1-T cells were treated with increasing concentrations of SGI-1776 for 72 hours. Assessment of viability in these two lines revealed that both are sensitive to SGI-1776, with the 22Rv1 parental line displaying a better overall response to treatment with the inhibitor than 22Rv1-T. The approximate IC₅₀s for the 22Rv1 and 22Rv1-T cell lines treated with SGI-1776 were 2.5 and 5 μmol/L, respectively (Fig. 5C).

We further verified that SGI-1776 was indeed modulating the downstream targets of Pim kinases in the 22Rv1-T cell line. Phosphorylation levels of the antiapoptotic factor Bad were determined using identical conditions as depicted with the parental 22Rv1 cell line (Fig. 3C). Analogous to 22Rv1, there was a dose-dependent decrease in phosphorylated Bad^Ser112 protein levels with SGI-1776 treatment with maximal inhibition occurring between 5 and 7.5 μmol/L, whereas total Bad levels remained unchanged (Fig. 5D).

Recently, it has been suggested that targeting Pim-1 activity is capable of resensitizing drug-resistant tumor cells to chemotherapeutic agents (38). Accordingly, we predicted that SGI-1776–mediated Pim kinase inhibition would be able to overcome acquired taxane resistance in the 22Rv1-T model. A comparative time course analysis (0, 24, 48, and 72 hours after drug treatment) in 22Rv1 and 22Rv1-T cells in the presence and absence of SGI-1776 and paclitaxel as single agents or in combination was done. As illustrated in Fig. 6B, the 22Rv1-T cell line showed resistance to 50 nmol/L of paclitaxel over time compared with the 22Rv1 line, which exhibited a clear sensitivity to an equivalent dose of paclitaxel (Fig. 6A). Most striking is the evidence that presence of SGI-1776 resensitizes the 22Rv1-T cells to 50 nmol/L paclitaxel equivalent to similar levels observed in the 22Rv1 paclitaxel-sensitive line. Specifically, the 22Rv1-T cell line does not respond to 50 nmol/L of paclitaxel alone and shows a minimal response to 2.5 μmol/L SGI-1776; however, when these two agents are combined, the cell viability decreases by as much as 40% in a time course experiment (P < 0.0001; Fig. 6B). In addition, SGI-1776 was able to resensitize the taxane-resistant cells to docetaxel at similar concentrations (data not shown).

Next, we showed that SGI-1776 was capable of inducing apoptosis in the acquired taxane-resistant model system. A DNA-specific dye, Hoechst 33342, was used to visualize the nuclear morphology in the 22Rv1-T cells after 24 hours of drug treatment. Supplementary Fig. S4 depicts condensed chromatin, a characteristic of apoptotic nuclei, in 22Rv1-T cells in the presence of 2.5 μmol/L SGI-1776 alone, which is dramatically enhanced when 2.5 μmol/L SGI-1776 is combined with 50 nmol/L of paclitaxel. In addition, a more extensive quantitative analysis of apoptosis was done by examining caspase-3 activity levels in 22Rv1-T cells treated with two different doses of SGI-1776 in the presence of increasing concentrations of paclitaxel. Again, we confirm that the 22Rv1-T cells are resistant to increasing doses of paclitaxel as evidenced by a lack of change in caspase-3 levels across the various treatments. More importantly, a gradual increase in caspase-3 activity is detected as SGI-1776 is combined with increasing doses of paclitaxel (Fig. 6C). For example, addition of 2.5 μmol/L SGI-1776 (142 ± 3%) increases caspase-3 activity by 1.42-fold compared with the untreated cells (100 ± 2%). However, when 2.5 μmol/L SGI-1776 is combined with 50 nmol/L paclitaxel, the caspase-3 activity level is increased by 2.7-fold (269 ± 8%; P < 0.0001).

After demonstrating the ability of SGI-1776 to resensitize multidrug-resistant prostate cancer cells to taxanes, it was important to gain insight on the mechanism by which this Pim kinase inhibitor is able to reverse the resistance process. SGI-1776 treatments did not alter MDR1 mRNA expression levels in 22Rv1-T cells (Supplementary Fig. S3B); however, it did inhibit MDR1 activity, which is the likely mechanism for reversing taxane resistance. Specifically, we used an assay that is based on the fluorogenic dye calcein acetoxymethyl ester as a substrate for the efflux activity of MDR1 and measures the intensely fluorescent calcein that accumulates in the cytoplasm. In 22Rv1-T cells, the overexpression of MDR1 prevents a majority of calcein acetoxymethyl ester to cross the cell membrane and be hydrolyzed, thus reducing the amount of fluorescent calcein retained inside the cell compared with the non-MDR1 parental 22Rv1 line. Yet, when 22Rv1-T cells were treated for 15 minutes with varying concentrations of SGI-1776, a dose-dependent increase in fluorescent calcein retention was observed (Fig. 6D). On the contrary, paclitaxel-treated cells were no different from the untreated cells with respect to calcein accumulation. Combination of paclitaxel with SGI-1776 did not further enhance calcein accumulation. Cyclosporin A, a known competitive inhibitor of MDR1-drug binding, was used as a positive control and showed a similar percentage of calcein retention to that obtained with 10 μmol/L of SGI-1776. These data indicate that SGI-1776, in addition to serving as a Pim kinase inhibitor, can overcome acquired taxane resistance by inhibiting MDR1 activity and triggering apoptosis.

Pim kinases are intimately involved in tumor progression and several inhibitors targeting these kinases have recently been identified (27, 28, 30, 31, 39). A majority of the Pim kinase inhibitors described to date are specific for the Pim-1 isoform and have mainly been analyzed in leukemia systems and to a lesser degree in prostate cancer (40). A large proportion of the prostate cancer literature concerning Pim kinases concentrates mainly on Pim-1 expression, although a few studies do report involvement of the other Pim kinase family members (22, 31, 41). Despite the fact that the current Pim-1 kinase inhibitors exhibit potential anticancer effects, targeting all three members of the Pim family may be necessary to realize an improved therapeutic response. For this reason, we have identified a selective small molecule inhibitor, an imidazo[1,2-b]pyridazine derivative, against the entire Pim kinase family and present its specific cytotoxic capabilities within several in vitro prostate cancer systems in this study.

SGI-1776 was determined to have low nanomolar affinity for all three Pim family members. It also showed cross-reactivity with two other kinases (FLT3 and Haspin) present in the 300-kinase panel screened. It cannot be ruled out that the observed cytotoxic effects may, in part, be due to FLT3 inhibition; however, our data indicate that SGI-1776 is quite specific for Pim kinases. Furthermore, it is also known that Pim-1 acts downstream of FLT3 (42). With the use of this potent inhibitor, we showed that inhibition of Pim kinase activity in prostate cancer cells compromised cell viability by inducing cell cycle arrest and apoptosis. SGI-1776 treatment caused molecular alterations consistent with Pim kinase inhibition such as targeting Pim kinase substrates at very specific phosphorylation sites, as seen with Bad at Ser¹¹² and p21 at Thr¹⁴⁵. A few other kinases are believed to phosphorylate these exact residues (i.e., RSK1 and protein kinase A), but none of them emerged as SGI-1776 targets in the broad-scale kinase screening assay. Additionally, SGI-1776 was capable of severely inhibiting the enhanced prosurvival signals induced by recombinant Pim-1 overexpression, further supporting the specificity of SGI-1776 for Pim kinases. Another important point that emerged from our analysis is that SGI-1776 selectivity is determined by the activity of the Pim kinases rather than the actual expression levels, which is unlike other Pim-1 kinase inhibitors that suggest a correlation between Pim expression levels and efficacy of the compound (4, 27).

Interestingly, a correlation was noted in that androgen-independent prostate cancer cell lines seem to be more sensitive to the cytotoxic effects of SGI-1776 compared with the androgen-dependent cell lines. It has been shown that Pim-1 kinase can indirectly modulate AR transcriptional activity in prostate cancer (43). Given that modifications in AR activity have been associated with the transition to an androgen-independent phenotype in prostate cancer progression, it may be valuable to use SGI-1776 to address the role of AR in influencing the antiapoptotic effects of Pim kinases in AD versus androgen-independent systems in future studies.

A current standard treatment for patients with advanced or hormone-refractory prostate cancer is taxane-based chemotherapy (44). However, the exposure of prostate cancer cells to such cytotoxic agents activates a number of survival pathways that can limit the efficacy of such treatments (45). Recently, it has been suggested that one of the potential prosurvival signals being activated during this stress response is the Pim-1 kinase family member (26). Not only does the treatment of prostate cancer cells with docetaxel induce Pim-1 expression, it has also been shown that overexpression of Pim-1 protects prostate cancer cells from the apoptotic effects of these cytotoxic drugs (25, 26, 38). This brings up the question of whether modulation of Pim kinase activity by a small-molecule inhibitor could sensitize prostate cancer cells to taxane-based therapies. As evident by our studies in the 22Rv1 prostate cancer cell line, we suggest that targeting Pim kinases through SGI-1776 treatment in combination with taxanes can generate an additional effect with respect to cell viability. Consistent with our data, Xie et al. (25, 38) showed a similar result when the Pim-1 isoform was inhibited through siRNA techniques and combined with chemotherapy drugs. Our ability to achieve greater clinical benefit, in the model systems studied, with a combination therapy that uses lower overall doses of a pharmacologic agent against Pim kinases and paclitaxel than required as single agents suggests additional avenues for treatment of this disease.

Although chemotherapy as a first line treatment may initially be successful at inducing apoptosis, eventually the cancer cells become resistant to its effects. Understanding the mechanism for chemoresistance in prostate cancer cells and targeting candidate markers to sensitize these cells to a chemotherapy regimen is important for effectively treating prostate cancer patients. To investigate this issue, we used the acquired taxane-resistant 22Rv1-T cell line to ascertain the role of Pim kinases in governing the chemotherapeutic response of prostate cancer cells. Indeed, inhibiting Pim kinase activity with SGI-1776 was able to overcome paclitaxel resistance by resensitizing the 22Rv1-T prostate cancer cells to taxane treatment. A previous study also showed that knockdown of Pim-1 by siRNA was able to overcome resistance to chemotherapeutic agents in drug-resistant prostate cancer cell lines (38). The authors suggest that the potential for Pim-1 to confer resistance is in part due to its ability to phosphorylate the half ABC transporter, BCRP/ABCG2. Incidences of BCRP overexpression have also been noted in other multidrug-resistant cell lines (46). However, BCRP expression was not found to be induced in our 22Rv1-T cell line, rather another member of the ABC transporter family, MDR1, was shown to be overexpressed. Therefore, the ability to overcome drug resistance mediated by MDR1 through pharmacologic inhibition of Pim kinases is a novel finding. Several prostate cancer studies have shown the involvement of MDR1 in the development of drug resistance, including one study that found overexpression of MDR1 in two established paclitaxel-resistant prostate cancer cell lines (47). Drug resistance is a complex process that involves several multidrug-resistant proteins and varies between systems. Through further investigation into the process leading to reversal of chemoresistance, we discovered that SGI-1776 is an inhibitor of MDR1 efflux activity. The proapoptotic effects of this Pim kinase inhibitor in combination with its ability to block MDR1 activity provide strong mechanistic evidence for the ability of SGI-1776 to resensitize chemoresistant cells to taxanes. Further exploration into the manner in which SGI-1776 interferes with MDR1 activity is still needed as well as determining whether this phenomenon is applicable to other drug-resistant phenotypes (i.e., MRP1 and BCRP) in prostate cancer.

The study presented here is a comprehensive examination of a Pim family small-molecule inhibitor displaying anticancer activity in prostate cancer cell lines that are sensitive and resistant to chemotherapy. Given the complexity of prostate cancer development, targeting a single kinase may not be sufficient to generate an effective clinical response, suggesting combination therapy inhibiting multiple signaling cascades may be a more beneficial strategy. In this report, we provide evidence that Pim kinase inhibition using SGI-1776 in combination with chemotherapy treatment can promote taxane sensitivity in prostate cancers and may overcome acquired taxane resistance by inhibiting MDR1 activity. Preliminary studies in leukemia have shown in vivo efficacy with SGI-1776 treatment. Studies are ongoing in prostate cancer xenografts to show antitumor activity and modulation of chemoresistance by SGI-1776 in vivo. Currently, this compound is being investigated in human clinical trials for hormone- and docetaxel-refractory prostate cancer as well as relapsed or refractory non-Hodgkin lymphoma.

D. Bearss, G. Berk, S. Kanekal, S. Redkar, P. Taverna: employees, SuperGen, Inc. No other potential conflicts of interest were disclosed.

We thank Sabrina Lei for her assistance with data review and analysis, Rajashree Joshi for her advice on drug formulations, Doug Evans for the generation of the taxane-resistant cell line, and Jim Mirocha for the ANOVA statistical analysis.