Novel ABCG2 Antagonists Reverse Topotecan-Mediated Chemotherapeutic Resistance in Ovarian Carcinoma Xenografts Free

Epithelial ovarian cancer continues to be the most lethal gynecologic malignancy, and chemotherapeutic resistance remains a major barrier to treatment. Unfortunately, most ovarian cancers are not diagnosed until advanced disease (grade 3 and 4), and these patients only have a 5-year survival rate of approximately 20% (1–4). Despite the fact that response rates to cytoreductive surgery and first-line chemotherapeutic treatments (platinum compounds and taxanes) approach 75% (2), ovarian carcinoma has a recurrence rate that exceeds 75% (1, 5, 6). In addition, recurrent tumors are characteristically more resistant to chemotherapy, with response rates only at approximately 20% (1, 2, 7). Furthermore, the limited efficacy of chemotherapy in recurrent disease has been attributed to the development of multiple-drug resistance (MDR; refs. 1, 3–5, 7, 8). Because of continued treatment difficulties encountered in drug-resistant ovarian adenocarcinoma and resulting mortality, it is important to identify compounds that reverse chemotherapeutic resistance.

MDR may be mediated by a variety of mechanisms (1, 4, 7), although the most significant and prevalent mechanism in ovarian cancer is cellular extrusion of chemotherapy by the ATP-binding cassette (ABC) transporter proteins (1, 3, 4, 7). Although the family of ABC transporters is relatively large (48 ABC encoding genes; ref. 9), only three transporters comprise the major mechanism of MDR in cancer: ABCB1, ABCC1, and ABCG2 (10–12). In ovarian cancer, some members of the ABCC subfamily have been implicated in platinum agent and taxane resistance, whereas ABCG2 has been demonstrated to confer topotecan resistance (1, 8, 13–15). Consistent with these cellular observations, one clinical trial demonstrated ABCG2 expression in all patients with primary or recurrent ovarian carcinoma (16). In addition, a Gynecologic Oncology Group phase III trial investigating ABCB1, ABCC2, and ABCG2 expression on clinical outcomes in ovarian cancer revealed that women expressing the specific C421A ABCG2 polymorphism, approximately 20% of studied patients, had a shorter progression-free survival time following chemotherapeutic treatment (17). Therefore, ABCG2 plays an important clinical role in conferring MDR to ovarian cancer.

There are a variety of ABCG2 inhibitors that have been investigated in vitro and in vivo with many types of cancers exhibiting MDR. However, there are only limited ABCG2 antagonists that have been examined in solid tumors despite its clinical importance as an MDR mechanism (18–24). Thus far, only fumitremorgin C and Ko143 (a fumitremorgin C analogue) have been investigated in preclinical animal studies utilizing topotecan for ovarian carcinoma (18–20). Fumitremorgin C is neurotoxic, which hinders its clinical use (19). Although low toxicity was observed with Ko143 when administered as oral doses in mice, the question of its potential toxicity in humans due to its relation to fumitremorgin C is a possible hurdle in advancement to the clinic (18). Another compound with some activity against ABCG2, WK-X-34, was tested in ABCB1-overexpressing, daunorubicin-resistant ovarian cancer, but it appeared to reverse MDR mainly through its inhibition of ABCB1 (20). Furthermore, GF120918 (elacridar), an ABCB1 inhibitor, was tested in mice using ovarian cancer cells resistant to doxorubicin and was later discovered to have some activity against ABCG2 (25). A single phase II clinical trial investigating inhibition of ABCG2 by lapatinib in recurrent ovarian cancer has been performed (16). However, there was no observed clinical benefit, and the trial was canceled due to substantial adverse hematologic events. As a result, there is a critical need to identify inhibitors with low toxicity and high potency for ABCG2.

Although ABCG2 has a broad diversity of substrates, the most well known with relation to MDR in ovarian carcinoma are topotecan, mitoxantrone, platinum agents, paclitaxel, and doxorubicin (1, 17). Topotecan is a second-line therapy used adjunctively with primary cytotoxic agents in ovarian carcinoma (5). Topotecan is one of the most studied agents for treatment of relapsed ovarian carcinoma, and according to present evidence, there is no current difference in safety or effectiveness when compared with other second-line nonplatinum chemotherapeutic regimens (26, 27). Therefore, topotecan is a useful drug, and its cellular resistance is valuable to examine as a model for ABCG2-mediated MDR in ovarian cancer.

Because of the importance of discovering new ABCG2 inhibitors, our study focused on investigating three novel ABCG2 inhibitors with similar chemical scaffolds for the treatment of MDR in ovarian cancer. Each compound was effective in reversing MDR in cell culture and in a murine ovarian carcinoma model with relatively low toxicity. Neither topotecan nor ABCG2 inhibitor alone had a major effect on tumor response to treatment. However, when topotecan was combined with low doses of an ABCG2 inhibitor, we observed substantial tumor size reduction and necrosis. The median toxic doses of each ABCG2 inhibitor in vitro were all greater than 1 μmol/L. In addition, our compounds appear to have low toxicity in vivo with a higher potency (each with sub-10 nmol/L median inhibitory concentrations) than other ABCG2 inhibitors previously described in ovarian cancer (18–24). To determine the optimal administration route, we administered each ABCG2 antagonist in combination with topotecan by intratumoral, retro-orbital, and intraperitoneal injection. When accounting for tumor necrosis and fibrosis along with tumor size reduction, we observed a significant decrease in tumor viability compared with the controls (P < 0.01). These studies are an important basis for transitioning this new class of ABCG2 inhibitors to the clinic as a potential remedy for MDR in ovarian cancer.

Igrov1 human ovarian carcinoma cells and an Igrov1-derived cell line overexpressing ABCG2, Igrov1/T8, were generously provided by Dr. Douglas Ross at the University of Maryland (Bethesda, MD) and were received in 2011. Cells were cultured in RPMI1640 medium (Gibco, Invitrogen Corporation) supplemented with 10% FBS (Gibco), 1% penicillin/streptomycin (Gibco), and 5 ng/L ciprofloxacin (Bayer Pharmaceuticals) and maintained at 37°C in an incubator with an atmosphere of 95% air and 5% CO₂. The Igrov1/T8 ABCG2-overexpressing phenotype was maintained by exposure to 950 nmol/L topotecan hydrochloride hydrate (Sigma) for 1 hour per week on the day when the cells were not expanded. We synthesized and identified three ABCG2-selective efflux inhibitors, CID44640177, CID1434724, and CID46245505, by high-throughput flow cytometry as reported previously (28, 29). For the purposes of this article, we will refer to the ABCG2 inhibitors CID44640177, CID1434724, and CID46245505 as 177, 724, and 505, respectively.

Drug resistance was quantified by comparing the percent confluency of Igrov1/T8 cells and Igrov1 parental cells maintained in 950 nmol/L topotecan and the effective concentration of topotecan that resulted in 50% cell death (EC₅₀) of Igrov1 parental and Igrov1/T8 cells. To determine that Igrov1/T8 cells were resistant to topotecan, Igrov1 parental and T8 cells were plated at 50% confluency on day 0 in a concentration of 950 nmol/L topotecan. Percent confluency was estimated under light microscopy every day for 14 days. Once the Igrov1/T8 cells grew to 100% confluency (day 7), they were split down to 50% confluency on that day to prevent overgrowth and cell death. A conversion factor of 2 was used for Igrov1/T8 confluency on days 7 to 14 for an estimate of continued cell growth. To determine the EC₅₀ of resistant and parental cells, Igrov1 parental and ABCG2-overexpressing Igrov1/T8 cells were separately placed into 6-well plates, such that the total number of cells in each well was approximately 1.5 × 10⁵ in a total volume of 2 mL (7.5 × 10⁴ cells/mL). Topotecan was added to the Igrov1 parental and Igrov1/T8 wells at increasing concentrations from 0 to 3 μmol/L, and cell viability was determined on days 3 and 7. Similar experiments were used to evaluate the cytotoxicity and efficacy of each ABCG2-selective inhibitor by examining the median lethal dose (TD₅₀) or the IC₅₀ of each compound (28).

This assay was performed according to the Soft Agar Assay described by Provost and Wallert laboratories (Minnesota State University, Moorhead, MN) with minor modifications. Igrov1 parental and Igrov1/T8 cells were plated at 1,250 cells per well in 24-well plates. The cells were treated with topotecan, 177, 724, or 505 alone and in combinations of ABCG2 inhibitor with varying concentrations of topotecan. Cells were then incubated under standard cell culture conditions at 37°C and 5% CO₂ for 48 hours. Next, the cells were washed in PBS, trypsinized, centrifuged, and replated on top of agar mixed in a half RPMI with 20% FBS and half 1% agarose solution according to the protocol. After 14 days, colonies consisting of 20 or more cells were counted following staining with 0.005% crystal violet. The combination index (CI) values for each topotecan/inhibitor combination were calculated according to the formula below:

Where, I₁ and I₂ represent the concentrations of two different compounds that individually achieve X% target inhibition and CI₁ and CI₂ represent the concentration of those same compounds that achieve X% target inhibition in combination.

Total RNA was amplified and reverse transcribed to cDNA using an Ambion WT Expression Kit (Ambion), which was then fragmented and labeled using the Affymetrix GeneChip WT Terminal Labeling Kit (Affymetrix; ref. 30). The resulting cDNA was then hybridized to a Human Gene Array 1.1 ST Array (Affymetrix) and subsequently washed and stained. Each step was carried out according to the manufacturer's protocol. The microarray gene chips were scanned using an AffyMetrix GeneChip Scanner 3000, and the normalized data were analyzed using GeneSpring software (Agilent). The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (31) and are accessible through GEO Series accession number GSE85582.

Total RNA was extracted from ovarian carcinoma mice xenografts, mouse liver, and Igrov1/T8 and Igrov1 parental cells in culture using the Qiagen RNeasy Mini Kit (Qiagen Inc.). Purified, DNase-1–treated total RNA was then reverse transcribed using the RETROscript RT Kit (Ambion). ABCG2 was quantified using the Light Cycler SYBR Green I Master Real-Time PCR Mix (Roche Applied Science) in a 30 μL reaction. The primer sequences used for gene amplification were ABCG2 sense – 5′-TGGCTGTCATGGCTTCAGTA-3′ and antisense – 5′-GCCACGTGATTCTTCCACAA-3′; ref. 31). cDNA from mouse skeletal muscle was used as a negative control whereas cDNA from cultured Igrov1/T8 cells was used as a positive control. GAPDH was used as an endogenous control. qRT-PCR reactions were carried out on the LightCycler 480 Real-Time PCR System (Roche Applied Science) in technical triplicates using previously reported cycling parameters (32). The second derivative maximum method was used to determine crossing points of each amplification curve (33). qRT-PCR cDNA products were also examined by 2% agarose gel electrophoresis using a Bio-Rad Sub-Cell GT Cell (Bio-Rad) according to the manufacturer's instructions.

We previously demonstrated that the JC-1 fluorochrome is a substrate of ABC transporters (34). Igrov1 parental and Igrov1/T8 cells were stained and prepared for flow cytometry using the JC-1 Mitochondrial Membrane Potential Detection Kit (Cell Technology) according to the manufacturer's instructions. Compound 177, 724, or 505 was added to one vial with stained Igrov1 parental and Igrov1/T8 cells immediately prior to flow cytometry. Cells were then run on a Becton Dickinson FACScan automated flow cytometer including negative and positive controls.

CB-17 SCID mice were obtained from Charles River Laboratories and were allowed to acclimate to the animal facility for 1 week prior to use in the study. The mice were kept on a 12-hour light/dark cycle and had free access to food and water. Mice were handled in accordance with the Guide for Care and Use of Laboratory Animals under the approval of the Institutional Animal Care and Use Committee. Approximately 7.5 × 10⁶–1 × 10⁷ Igrov1 parental cells or Igrov1/T8 cells suspended in 200 μL Matrigel (BD Biosciences) were injected into each hind-limb region of the mice. Both hind limbs were injected in some mice, whereas others had only one hind limb injected with cells. Tumors formed between 2 and 4 weeks following injection. At approximately 4 weeks when the tumors reached our established detectable size range, 50–250 mm³, the tumor xenografts were injected with topotecan at 0 and 24 hours by three different administration routes: intratumoral, retro-orbital, and intraperitoneal. 177, 724, and 505 were also injected at 48 and 72 hours by one of the three administration routes to verify MDR and nontoxicity of the tumor xenografts. Subsequently, topotecan combined with the reversal agent was administered every 24 hours for 4 to 5 days or until tumors regressed below our limit of detection (∼15 mm³). The dosage of drug for each injection was determined on the basis of analysis of in vitro EC₅₀, TD₅₀, and IC₅₀ results. Tumor measurements were acquired using a Scienceware Digi-Max slide caliper (Sigma-Aldrich), and tumor volume was calculated by the equation: L × (W²/2), where L is the length in the rostral–caudal direction and W is the width measured transversely (19, 28). Statistical significance was evaluated using a two-tailed t test and reported as a P value. Mice were then euthanized, and any remaining tumor was cut into histologic sections, stained with H&E, and verified by light microscopy. Fibrosis and/or necrosis of remaining tumor sections was quantifiably measured by two investigators, initially blinded to each other's results, using a Lovin Field Finder Micro Slide Grid (Electron Microscopy Sciences). An average of both investigators' measurements was calculated for each drug therapy group and incorporated into the size reductions of the tumors.

The cell line Igrov1 was previously made drug resistant to Asta Z and Adriamycin through sequential passage in increasing doses of those drugs and shown to be resistant to topotecan by Maliepaard and colleagues (35, 36). To quantify the degree to which Igrov1/T8 was resistant to topotecan, we performed two sets of experiments. First, we compared the growth of Igrov1/T8 versus the drug-sensitive parental line, Igrov1, in a fixed concentration of topotecan over a 2-week period (Fig. 1A). In a concentration of 950 nmol/L topotecan, cells were plated at 50% confluency, and over 14 days, Igrov1/T8 cells grew while the parental line showed 100% cell death by day 5. In the second set of experiments, we determined the concentration at which 50% of cells would die (EC₅₀) after 7 days of culture using different concentrations of topotecan (Fig. 1B). This analysis showed an EC₅₀ of 1.4 μmol/L for the drug-resistant Igrov1/T8 in comparison with 7.4 nmol/L for the Igrov1 parental cell line, a 189-fold difference.

To determine the mechanism of drug resistance, several experiments were performed. First, we examined the expression of all known ABC proteins in both the Igrov1 parental and Igrov1/T8 drug-resistant cell lines using microarray analysis (Fig. 2A). Comparison of the array data indicated that ABCG2 was the most differentially expressed, having an 87-fold increase in Igrov1/T8 expression in comparison with Igrov1. This differential expression was confirmed by RT-PCR (Fig. 2B).

We then performed a functional efflux assay using JC-1, a fluorochrome that is extruded through ABCG2 (34). The relative fluorescence corresponding to the presence (right shift) or absence (left shift) of JC-1 was measured using flow cytometry. Accordingly, ABCG2 Igrov1 parental cells had a higher amount of fluorescence than ABCG2-overexpressing Igrov1/T8 cells, indicating retention of JC-1 in Igrov1 parental cells and extrusion of JC-1 in Igrov1/T8 cells. Following the addition of each ABCG2 antagonist, JC-1 retention increased inside the Igrov1/T8 cells. Fluorescence of the Igrov1 parental cells was essentially unaffected by the presence of any of the ABCG2 inhibitors. These results indicate that ABCG2 efflux activity is reversed by the presence of 177, 724, and 505 (Fig. 2C and D).

We have previously identified three potential compounds (177, 724, and 505) that inhibit ABCG2 activity (28, 29). To determine whether these compounds restored chemosensitivity to the drug-resistant cells, we measured the cell viability of Igrov1/T8 cells in the presence of increasing concentrations of these chemosensitizers and a fixed concentration of 100 nmol/L topotecan (Fig. 3). This concentration of topotecan, if used alone, results in 100% cell death of the parental cell line, but has no effect on the survival of the chemoresistant Igrov1/T8 cell line. 177, 724, and 505 showed an ability to restore chemosensitivity as measured by the concentration at which there was 50% cell death (IC₅₀). The doses of each ABCG2 inhibitor were escalated to a point where near 100% cell death was observed after 7 days. Compound 177 was the most potent ABCG2 antagonist resulting in an IC₅₀ of 2.1 nmol/L, although compounds 724 and 505 had similar potency indicated by their IC₅₀s of 4.7 and 9.8 nmol/L, respectively.

In addition, we performed a colony formation assay with chemoresistant Igrov1/T8 cells pretreated with topotecan in combination with an ABCG2 inhibitor for 48 hours (Fig. 4). Topotecan in combination with compound 177, 724, or 505 resulted in fewer colonies at lower concentrations of topotecan than topotecan alone (4–91 colonies vs. 109 colonies at 100 nmol/L topotecan). Therefore, ABCG2 inhibitors lower the dose of topotecan required to inhibit the growth of topotecan-resistant Igrov1/T8 cells. CI is typically used to demonstrate that two drugs work synergistically. In our case, the compounds are blocking the ability of a cell to exude topotecan. Therefore, our CI values indicate the relative effectiveness of these compounds to reverse chemoresistance. CI values were calculated for each combination of topotecan and ABCG2 inhibitor, and all were less than 1, indicating their ability to restore chemosensitivity by inhibiting ABCG2 (Fig. 4). Nonresistant Igrov1 cells grew few to no colonies in the presence of greater than or equal to 10 nmol/L topotecan alone and in combination with 177, 724, or 505, whereas exposure to 177, 724, or 505 alone resulted in up to 203 viable colonies, which decreased to lows of 120, 11, and 112 colonies when exposed to the highest concentrations of 177, 724, and 505 administered: 1, 10, and 1 μmol/L, respectively (data not shown).

To insure that our results were not due to toxicity produced by increasing doses of the chemosensitizer, we also examined the cell viability over a range of chemosensitizer concentrations and determined that the concentrations at which there was 50% cell death (e.g., median toxic dose; TD₅₀) was greater than the concentrations examined. In all cases, the TD₅₀ was greater than 1 μmol/L and at least 100-fold the IC₅₀, indicating that the chemosensitizers were nontoxic (data not shown).

As a first step to demonstrate that these compounds would have efficacy in vivo, chemosensitizers and topotecan were administered conjunctively by intratumoral injection over a 5-day period in a murine ovarian cancer xenograft model (19). By this time, an average reduction in tumor size in excess of 60% was measured in all cases (Fig. 5). When 177, 724, and 505 were injected alone every day for 3 days at therapeutic concentrations (100, 500, and 100 nmol/L, respectively), they did not have any effect on the tumor size (data not shown). In addition, the mice did not exhibit any significant weight changes or other observable toxic side effects over the course of therapy with the ABCG2 inhibitors, topotecan, or their combination.

As chemosensitizers administered in combination with topotecan were effective at reducing tumor size, we then sought to determine the most effective route of administration. We administered chemosensitizer combined with topotecan by intratumoral, retro-orbital, and intraperitoneal injection every day for 5 days. Intratumoral and intraperitoneal routes of administration showed the most effective tumor size reductions by day 5, ranging from 62% to 66% and 16% to 69%, respectively (Table 1), with the most tumor reduction achieved by the administration of 724 intratumorally and 505 intraperitoneally.

To verify that residual masses were tumor and/or fibrosis, we examined each tumor histologically. When viewed by light microscopy, all residual masses contained substantial necrosis and fibrosis in addition to some tumor. The most substantial amount was seen in the intratumoral injection group, although the largest amount of change in tumor size reduction was seen in the retro-orbital group (2%–14% to 27%–51%; Table 1). We also normalized mass reduction for fibrosis and/or necrosis and found that the amount of remaining tumor was significantly less than what was measured grossly, especially by the retro-orbital and intraperitoneal administration routes. By this analysis, all routes of administration resulted in significant tumor reduction (all P < 0.05), with compound 724 resulting in the most reduction by the intratumoral and intraperitoneal administration, at 80% and 77%, respectively.

MDR remains a large barrier to improved patient outcomes in ovarian cancer. ABCG2 contributes to MDR by active efflux of chemotherapeutic substrates, including topotecan, and is preferentially overexpressed in topotecan-resistant ovarian carcinoma more than 1,000-fold compared with other ABC transporters (1, 13). The importance of ABCG2 in ovarian cancer is further emphasized by the fact that certain polymorphisms lead to a shorter progression-free survival time, although the functional effects of these polymorphisms remain unknown (17). This indicates that ABCG2 expression may play an important role in clinical outcomes and survival, making it an important target for pharmacotherapy in MDR ovarian cancer.

In the current study, we examined the role of three novel piperazine-substituted pyrazolo[1,5]pyrimidine substructure ABCG2 antagonists that we previously identified (28, 29) and their role in restoring chemosensitivity to MDR ovarian cancer. Compounds 177, 724, and 505 were found to effectively inhibit ABCG2-mediated efflux of topotecan in ovarian cancer cells with sub-10 nmol/L IC₅₀s, suggesting that each compound is a potent ABCG2 antagonist drug candidate. The three other ABCG2 inhibitors, fumitremorgin C, Ko143, and GF120918, that have been investigated in topotecan-resistant ovarian cancer in vivo, have IC₅₀s near the micromolar range (18, 19). Following the administration of our ABCG2 antagonists with topotecan by varying routes, topotecan-resistant ovarian carcinoma xenografts demonstrated significant tumor reduction (P < 0.05). In our initial observations, only the size of the mass was recorded. However, substantial tumor necrosis and fibrosis was observed on histology and taken into account when quantifying overall tumor reduction, leading to much larger overall reductions than could be grossly observed.

The relative IC₅₀s of each drug tested here were similar but showed greater differences in their in vivo potencies. All were in the sub-10 nmol/L range and effectively inhibited ABCG2-mediated efflux of topotecan, but their potencies in vivo did not directly correlate with those observed in vitro. We used a sigmoidal curve fit by a nonlinear regression technique to calculate the IC₅₀ values, which closely reflects the behavior of biological dose–response systems (37, 38). Although small inherent error exists in this model, it may be improved by increasing the number of data points. IC₅₀ values may not have a one-to-one correspondence in vivo due to potential nonlinear pharmacokinetics, extensive first pass metabolism, or pharmacodynamics that alter bioavailability (39). Multiple studies have reported differences between in vitro and in vivo efficacy of antitumor drugs in ovarian cancer and ABCG2-mediated resistance (40–42). Predicting the correlation between in vitro and in vivo activity remains challenging due to these potential confounding factors, and the variation in potency in vivo that was observed in this study is likely related to pharmacokinetics, pharmacodynamics, or other issues beyond their IC₅₀s (43). Although a full biological analysis and comparison of the isogenic cell lines would provide additional information into other potential molecular differences, our study was focused on the effects of our most promising compounds.

Compounds 177, 724, and 505 did not appear to promote overt toxicity. The TD₅₀ values from our previous work (28, 29) suggest that these compounds were suitable for use in vivo because of their apparent low toxicity at the doses used for administration. No matter how high we escalated the dose of compound 177, 724, and 505 alone, we did not see any tissue necrosis or death. In addition, we did not observe any tremors or convulsions in the mice that would have been suggestive of neurotoxicity as was previously reported with fumitremorgin C (18). We did not observe any significant weight loss and histologic evaluation of the liver, heart, kidneys, and bone marrow did not show any signs of necrosis, suggesting the absence of any significant systemic side effects. Thus, these compounds appear to carry a low toxicity risk, which is important when considering their addition to a relatively toxic chemotherapeutic regimen.

Targeting ABCG2 represents a less-studied mechanism for restoration of chemosensitivity in MDR ovarian cancer. Currently, only about 20% of recurrent tumors respond to second-line therapy (1, 2, 7). Despite the fact that ABCG2 has been well described as the key mechanism for resistance of ovarian cancer to topotecan (1, 13, 44), relatively few attempts to develop drugs at this important target have been made. Moreover, the compounds that have been created and tested in ovarian cancer have not progressed, typically due to toxicity (16, 18–20, 45). Both fumitremorgin C and its analogue, Ko143, were discovered to inhibit ABCG2 and tested in ovarian carcinoma murine models. However, fumitremorgin C is neurotoxic in vivo and therefore not suitable for use in the clinic (19). Although no toxicity has been described with Ko143, it is structurally similar to fumitremorgin C, and concern for in vivo toxicity may hinder its clinical use (18). The only compounds inhibiting ABCG2 efflux that have been tested in clinical trials for ovarian cancer are GF120918 (elacridar) and lapatinib (16, 45). Elacridar was found to increase topotecan bioavailability; however, it was not shown to affect tumor burden in its phase I trial. Lapatinib, which is primarily a tyrosine kinase inhibitor with some effect on inhibiting ABCG2-mediated efflux, has been well studied with its potential to reverse MDR (16, 32, 46). Lapatinib is currently the only compound affecting ABCG2 activity that has been examined in a phase II clinical trial for recurrent ovarian cancer (16). However, the trial was canceled due to substantial adverse hematologic events and no observed benefit. Furthermore, antitumor effects of ABCG2 antagonists, elacridar and WK-X-34, explored in ovarian cancer were both found to act primarily through inhibition of ABCB1 (P-gp; refs. 20, 22, 25). Although both ABCB1 and ABCC1 are important in mediating MDR, they do not play as large a role in topotecan-resistant ovarian cancer as does ABCG2. Even though the compounds in this study have some activity toward ABCB1 at high concentrations, the concentrations used in the current study were far below these levels, and the compounds would thus be expected to be specific for ABCG2. Although ABCB1 has a higher level of expression in the drug-resistant Igrov1/T8 cell line as compared with the Igrov1 cell line, we previously reported that compounds 177, 724, and 505 are more selective for ABCG2 over ABCB1 using a JC-1 fluorescent dye substrate marker (28, 29). Therefore, the increased level of expression of ABCB1 is not expected to account for the chemoreversal observed. As ABCB1 helps to protect a variety of tissues, such as the bone marrow stem cells, from the toxic adverse effects of chemotherapeutics, ABCG2-selective inhibitors may be preferred in some clinical applications.

Notably, our studies compare different routes of administration and their effect on reduction in tumor growth. Although route of administration may have a substantial effect on tumor growth inhibition, this level of comparison is not frequently reported (47). Of 17 in vivo studies investigating potential ABCG2 inhibitors for various types of cancer, only three compared two different routes of administration, namely oral to either intravenous or intraperitoneal (18, 25, 48). Furthermore, no studies investigating ABCG2 in vivo have compared the antitumor effects of three different routes of administration. In our formal comparison of intratumoral, intraperitoneal, and retro-orbital routes of administration, the greatest antitumor effect was achieved by intratumoral and intraperitoneal. Only one study has compared oral with intraperitoneal using Ko143 as the ABCG2 antagonist and did not report differences in the antitumor effect but only described that Ko143 was nontoxic by both routes and enhanced topotecan bioavailability by oral (18). This group also described that oral bioavailability of topotecan was enhanced with oral GF120918 administration. These results suggest that the oral route may lead to predominant inhibition of intestinal ABCG2, with less of the drug available for targeting ABCG2 overexpression in the tumor. Therefore, intraperitoneal may be the preferred route to oral for reversing chemoresistance. It is well known that material delivered via an intraperitoneal injection is absorbed much slower than by an intravenous injection (47), and in peritoneal malignancies, such as ovarian carcinoma, this allows the tumor to be exposed to the drug therapy for a longer period of time with less systemic effects. Consequently, first-pass metabolism might not play as large of a role with intraperitoneal injections, whereas the first-pass metabolism seen with the oral route might still affect drug efficacy. On the other hand, our study used subcutaneous hind-limb xenografts for enhanced accessibility and ease of tumor measurements during the treatment course. Therefore, altered pharmacokinetics, pharmacodynamics, or first-pass metabolism might have contributed to the differences that we observed. Alternatively, lower doses of inhibitor reaching the tumor xenograft following systemic absorption may have been sufficient to result in tumor death. In general, a more dramatic effect is seen with intraperitoneal administration in the treatment of ovarian cancer when compared with intravenous (49–51). Thus, intraperitoneal remains the administrative route of choice in ovarian cancer and is a reasonable route of administration to use for the compounds described here.

Each compound presented here appears suitable for GMP-based preclinical studies and first-in-human studies due to their high in vivo potency, low IC₅₀s, and low toxicity. All compounds were able to effectively reverse ABCG2-mediated topotecan resistance with compound 724, resulting in the largest tumor reduction by both intratumoral and intraperitoneal routes of administration. However, intraperitoneal administration would be the most useful as intratumoral administration is not clinically practical. Future studies should focus on the development of compound 724 for use in human clinical trials with recurrent ovarian carcinoma failing second-line treatment, which encompasses approximately 80% of women with recurrent disease (1, 2, 7). Because of its unique structure and action against ABCG2, the addition of compound 724 to a chemotherapeutic regimen for topotecan-resistant ovarian cancer may allow for increased rates of clinical remission and progression-free survival in this lethal disease.

No potential conflicts of interest were disclosed.

Conception and design: R.S. Larson

Development of methodology: D.M. Lovato, R.S. Larson, J.W. Ricci

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.W. Ricci, D.M. Lovato

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.W. Ricci, D.M. Lovato, R.S. Larson

Writing, review, and/or revision of the manuscript: J.W. Ricci, D.M. Lovato, V. Severns, L.A. Sklar, R.S. Larson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.M. Lovato, V. Severns, R.S. Larson

The University of Kansas Specialized Chemistry Center synthesized compounds CID44640177, CID1434724, and CID46245505 for our previous project, and these were used in the current study.

This work was supported by the NIH Clinical and Translational Science Award grant 5UL1TR000041-05 (to R.S. Larson).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.