Background
Hypoxia in solid tumors and the affected bone marrow of hematologic malignancies is a prevalent feature of cancer. Cells in the hypoxic tumor microenvironment are more resistant to radiotherapy and to most antiproliferative cancer drugs, and also acquire a more malignant and metastatic phenotype [
1]. One therapeutic approach being developed for the treatment of cancer is hypoxia-activated cytostatic or cytotoxic prodrugs [
2].
TH-302 is a hypoxia-activated prodrug of bromo-isophosphoramide (Br-IPM) that is reduced at its 2-nitroimidazole group and selectively activated under the severe hypoxic conditions commonly found in tumors, but not typically observed in normal tissues [
3]. Br-IPM is a potent DNA alkylating agent, and kills tumor cells by creating DNA crosslinks [
4]. Preclinical data demonstrate that TH-302 exhibits anti-tumor activity both as a monotherapy as well as in combination with other cancer therapies [
5-
7]. Clinically, TH-302 has been investigated in several early stage trials [
8-
11] and is currently being evaluated in Phase III trials in soft-tissue sarcoma in combination with doxorubicin and pancreatic cancer in combination with gemcitabine (NCT01440088 and NCT01746979, respectively).
There are two major cell-cycle checkpoint systems for detecting and responding to DNA damage: the G
1/S and intra-S checkpoints system to prevent the replication of damaged DNA, and the G
2/M checkpoint to prevent segregation of damaged chromosomes. The majority of tumors are deficient in the G
1/S DNA damage checkpoint due to tumor suppressor p53 mutations. Pharmacological inhibition of the remaining intact G
2/M checkpoint, e.g. through Chk1 inhibition, should lead to enhanced tumor cell death, as compared with p53 proficient normal tissue [
12]. It has been shown that inhibition of Chk1 signaling using small molecule inhibitors, dominant negative constructs, interference RNA (RNAi), or ribozymes leads to abrogation the G
2/M checkpoint, impaired DNA repair, sensitization of p53-deficient cells to apoptosis, and an increase in tumor cell death [
13-
15]. Of particular note, Chk1 inhibitors have also been designed as prodrugs for selective activation in the hypoxic regions of tumors [
15,
16].
Chk1 also regulates homology-directed repair (HDR), as DNA damage-induced HDR is dependent on Chk1-mediated Rad51 phosphorylation. Chk1 inhibition leads to impaired Rad51 foci formation, a key step in HDR [
17,
18]. Abrogation of Chk1 function leads to persistent unrepaired DNA double-strand breaks (DSBs). Chk1 inhibition results in premature mitotic entry in response to DNA damaging agents thus resulting in increased phosphorylated histone H3, a marker of mitosis [
19]. In addition, Chk1 pathway plays an important role in protecting cells from caspase-3-mediated apoptosis [
20,
21]. Reports have shown that cells with reduced levels of Chk1 were found to be more prone to apoptosis [
14,
21,
22]. More recently, it has been reported that Chk1 may have prognostic and predictive significance in breast cancer [
23].
Chk1 inhibition can potentiate the cytotoxicity of radiation and genotoxic therapies [
24-
29]. Chk1 inhibitors have been widely studied and a select number of compounds have reached early clinical trials. Notable among these are the ATP-competitive inhibitors LY2603618, PF477736, AZD7762, SCH90077617, and LY260636818 [
5], the latter three of which have progressed to Phase II clinical trials. Here we describe the combination therapeutic efficacy profile of Chk1 inhibitors with the hypoxia-activated prodrug TH-302 in
in vitro and
in vivo preclinical models.
Methods
Reagents and cell lines
TH-302 was from Syngene, AZD7762 and LY2603618 were from Selleck Chemicals, and PF477736 was from Tocris Bioscience. RIPA was from Sigma. Protease inhibitor cocktails were from Thermo Scientific. ChemiGlow substrate was from Proteinsimple. ECL reagent, Rad51 mAb, actin mAb, goat anti-rabbit HRP, goat anti-mouse HRP, and cell cycle reagents were from EMD Millipore. γH2AX mAb was from Epitomics. Antibodies against phospho-Histone H3, phospho-Cdc2 Y15 antibody, total Chk1, phospho-Chk1 (S296) were from Cell Signaling. Total Cdc2 p34 antibody was from Santa Cruz Biotechnology. FITC-conjugated goat anti-mouse secondary antibody and AlamarBlue cell viability reagent were from Life Technologies. Comet assay kit was from Trevigen. Caspase Glo 3/7 assay system was from Promega. Isogenic p53 proficient and deficient cell line pairs were from Horizon Discovery. All other cell lines were from ATCC.
In vitro proliferation
Exponentially growing cells were seeded 24 h prior to addition of test compounds. After compound addition, the plates were incubated for 2 h under either normoxia (21% O
2) or hypoxia (N
2) supplied with 5% CO
2 at 37°C. After wash, cells were cultured for an additional 70 h in fresh medium containing Chk1 inhibitors under normoxia (21% O
2), and the viable cells were quantified using either AlamarBlue or ATP assay and normalized using either vehicle for single treatment or Chk1 inhibitor alone for combination treatment. IC
50 was calculated using Prism software. The level of synergism by the Chou-Talalay method was expressed as the Combination Index (CI) calculated using CalcuSyn software [
30].
Cell cycle analysis
Cells were treated with 0.1 μM of either PF477736 or AZD7762 and TH-302 for 2 h under either normoxia (21% O2) or hypoxia (N2). Following wash, cells were cultured for additional 22 h in the presence of Chk1 inhibitor under normoxia. Cells were fixed in 75% ethanol and cell cycle distribution was determined using Cell Cycle reagent and Guava flow cytometry (EMD Millipore).
Single cell gel electrophoresis comet assay
After seeding cells for 24 h, TH-302 and 0.1 μM of AZD7762 were added and incubated for 24 h under either normoxia (21% O2) or hypoxia (0.1% O2). For cross-linking assessment experiments, cells were treated with 20 μM of bleomycin for 1 h starting at the end of the TH-302 treatment period. Comet assay was performed with Trevigen’s single-cell electrophoresis system. The data was analyzed using Comet Assay IV software from Perceptive Instruments.
Detection of γH2AX
HT29 colon cancer cells were treated with vehicle or TH-302 for 2 h under either normoxia (21% O2) or hypoxia (N2) conditions with or without 0.1 μM of AZD7762, and then continuously incubated for additional 4 h in the presence of AZD7762 for the combination group and AZD7762 monotherapy group. Cells were permeabilized with 1% Triton X-100 and incubated with γH2AX monoclonal antibody for 2 h and goat anti-mouse-FITC for 1 h. Cells were imaged using a Nikon TS-100 fluorescent microscope.
Caspase activity
HT-29 cells were exposed to TH-302 and 0.1 μM of AZD7762 for 2 h under either normoxia (21% O2) or hypoxia (N2). After wash, cells were continuously cultured for additional 46 h in the presence of 0.1 μM of AZD7762. Luminescence-based caspase activity assay was performed based on the manufacturer’s (Promega) instructions.
Western blot
HT29 cells were exposed to TH-302, AZD7762, or combined TH-302 and AZD7762 for 2 h under either normoxia (21% O2) or hypoxia (N2). After removal of TH-302, cells were continuously incubated with AZD7762 for additional 46 h. Cell extracts were prepared and protein concentrations were determined. Proteins were detected after SDS-PAGE and Western blotting with ChemGlow detection system (ProteinSimple) using antibodies recognizing autophospho-Chk1 (S296), total Chk1, phospho-Histone H3, phosphorylated Cdc2 Y15, total Cdc2, Rad51, and actin.
In vivo antitumor activity
Female nude mice (4-6 weeks; Nu-Foxn 1nu NU/NU, Charles River Laboratories) were tagged with microchips (Locus Technology) for identification. All animal studies were approved by the Institutional Animal Care and Use Committee at Threshold Pharmaceuticals.
HT29 cells were mixed with 50% Matrigel and 0.2 ml/mouse were subcutaneously implanted to the flank area of the animals. When the tumor size reached 150 mm3, mice were randomized into experimental groups.
TH-302 was dissolved in saline (0.9% NaCl) at 5 mg/ml and filtered prior to animal dosing. AZD7762 was formulated in 1% DMSO, 11.3% cyclodextrin in water for injection. Maximum tolerated dose (MTD) for TH-302 in combination with AZD7762 was determined in a small number of non-tumor bearing nu/nu mice. The MTD was defined as the highest possible dose resulting in no animal deaths, less than 20% weight loss for any one animal in an experimental group, no significant changes in general clinical signs, and no abnormal gross anatomical findings after necropsy. The doses of compounds used in all studies were no higher than MTD.
Two dosing regimens of TH-302 were employed in the study. With the TH-302 intermittent dosing regimen, TH-302 was dosed intraperitoneally (ip) at 100 mg/kg, twice/wk x 2 wks. AZD7762 was dosed i.p. at either 20 or 12.5 mg/kg, four times/wk x 2 wks. For the combination therapy study, two different dosing sequences were investigated: (1) TH-302 was given first, and 4 h and 24 h later AZD7762 was administered (the ‘TAA’ sequence); and (2) AZD7762 was given first, and 4 h and 24 h later followed by TH-302 and AZD7762, respectively (the ‘ATA’ sequence). With the TH-302 daily dosing regimen, TH-302 was dosed ip at 50 mg/kg, with a regimen of QDx5/wk x 2wks. AZD7762 12.5 mg/kg was given under same regimen as TH-302. When the combination was scheduled, TH-302 was administered 4 h prior to AZD7762 (the ‘TA’ sequence).
Tumor growth and body weight were measured twice a week. Tumor volume was calculated as (length x width2)/2. Drug efficacy was assessed as Tumor Growth Inhibition (TGI) and Tumor Growth Delay (TGD). TGI was defined as (1-ΔT/ΔC) x 100, where ΔT/ΔC presented the ratio of the change in mean tumor volume of the treated group and of the control group. TGD was calculated as the extra days for the treated tumor to reach 1000 mm3 as compared to control group (TGD1000). Animals were culled when individual tumor size was over 2000 mm3 or individual tumor size was over 1000 mm3 if mean tumor volume exceeded 1000 mm3 in the group. Conditional survival was defined as the time that an animal reached the endpoint of a tumor size of 1000 mm3. Kaplan-Meier plots were constructed based on the percentage animals surviving in each group as a function of time. Median time (MT) is the time at which half the animals in the group had a tumor size less than 1000 mm3. The antitumor activity was evaluated as follows: T/C % = MT of treated group/MT of control group × 100. Results were also expressed as the percentage of increased life span (ILS, T/C of treated group–100). Statistical significance between the groups was evaluated by the log-rank test.
Data are expressed as the mean ± SEM. One-way analysis of variance with Dunnett post-comparison test (GraphPad PRISM 4) or two-tail student’s t test were used for analysis. A P level < 0.05 was considered statistically significant.
Histology and immunohistochemistry
300-600 mm3 HT29 xenograft tumors were used in the pharmacodynamics studies. Six animals per group were treated with vehicle, AZD7762 25 mg/kg, ip, on Day 1 and Day 2, TH-302 150 mg/kg, ip on Day 1 or the combination of AZD7762 and TH-302. In the combination treatment, two dosing sequences were used: TAA and ATA (as above). Tumors were harvested on Day 3 (which was 24 h after the second AZD7762 treatment), and fixed in 10% neutral buffered formalin and embedded in paraffin. 5 μm thick paraffin sections were cut and adhered to poly-L-lysine-coated glass slides.
After deparaffinization and rehydration of the slides, antigen was retrieved. Endogenous peroxidase was quenched by Peroxidaze 1 and non-specific binding was blocked by Background Sniper (both Biocare Medical). Slides were incubated with rabbit monoclonal anti γH2AX (Epitomics, 1:3000), or rabbit polyclonal anti-phospho Chk1 (phospho S345, Chk1-S345, Abcam, 1:25) or Caspase 3 (Cell Signaling Technology, 1:300) for 1 h at RT followed by secondary HRP-conjugated anti rabbit IgG (Epitomics).
Image analysis
γH2AX, phospho-Chk1-S345 or Caspase 3 positive cells were counted at 400x magnification. Ten fields per section were used. The percentage of positive cells was calculated as number of γH2AX or Chk1-S345 positive cells/number of total cells in the field x 100%. P value <0.05 was considered significant. One-way analysis of variance with Dunnett’s test (GraphPad PRISM 4) was used to compare the significance of the multiple groups. A student’s t-test was used to find the significance between two groups.
Discussion
Both the
in vitro and
in vivo preclinical results described here demonstrate the potentiation of TH-302 efficacy by the addition of Chk1 inhibitors in the context of p53 deficiency. Our findings suggest that TH-302 cytotoxicity may be enhanced through at least two Chk1-dependent mechanisms. The first mechanism is abrogation of DNA damage-dependent cell cycle arrest, which is supported by the finding that Chk1 inhibition sensitizes the p53
-/- cells but not the p53
+/+ cells to TH-302. Substantial literature supports that Chk1 inhibitors selectively sensitize tumor cells with p53 deficiency to DNA damaging agents [
24-
28]. More than 50% of human tumors have mutations in p53 [
31]. Cells deficient in p53 cannot undergo p53-dependent apoptosis and are resistant to drugs that induce a p53-dependent apoptosis. Such resistance has been reported in many p53 null or mutated human cancer cell lines as well as in clinical samples [
32]. Indeed inhibition of Chk1 in tumor cells, either by siRNA knockdown of Chk1 protein expression [
19], or by small molecules inhibiting its kinase activity, demonstrated its role in the potentiation of the cytotoxic activity of DNA damaging agents [
18,
19] This model is also supported by the findings that Chk1 inhibition preferentially sensitizes p53 deficient human cancer cells, but not p53 functional cells to gemcitabine, radiation and 5-fluorouracil [
29,
33,
34].
The second mechanism of enhanced TH-302 cytotoxicity by Chk1 inhibitor is related to HDR inhibition. Based on our previously published study, HDR plays the key role in the response and repair of TH-302-induced DNA cross-links [
4]. Enhanced TH-302 activity was also observed in Rad51 knockout cell lines and in triple-negative breast cancer cell lines exhibiting an HDR-deficient (BRCA-like) phenotype [
35,
36]. It has been proposed that Chk1 is required for HDR [
17], which normally occurs in the S and G
2 phase [
37]. p53-mutated cells lack a G
1 checkpoint, and thus they may be more dependent on HDR [
26]. Thus, it would be anticipated that Chk1 inhibition would predominantly affect HDR in p53-mutated cells [
17]. The requirement for HDR inhibition in TH-302 sensitization by Chk1 inhibitors is shown by a lack of TH-302 sensitization by Chk1 inhibition in HDR-deficient cells. However NHEJ-proficient and -deficient cells exhibited a similar Chk1 inhibitor-involved sensitization to TH-302. Furthermore, our findings demonstrate that Chk1 inhibitors can down-regulate TH-302-induced overexpression of Rad51, and subsequently restore cell sensitivity to TH-302. The current findings suggest that Chk1 inhibition may offer considerable benefit to TH-302 in Rad51-overexpressing tumors. Pancreatic ductal adenocarcinoma is a cancer type where overexpression of Rad51 has been described [
38]. Chk1 is involved in HDR by directly phosphorylating Rad51 and in the recruiting of Rad51 to sites of DNA damage [
17,
39,
40]. It has also been reported that a Chk1 inhibitor reduces Rad51-mediated HDR [
17]. To determine whether Rad51 is involved in the enhancement of TH-302 by AZD7762, we examined the Rad51 expression in cells exposed to TH-302, AZD7762, or combined TH-302 and AZD7762. Rad51 levels increased in cells treated with TH-302. Rad51 protein levels were not affected by AZD7762. However, AZD7762 abolished Rad51 upregulation mediated by TH-302. Although both inhibition of cell cycle arrest and HDR are associated with TH-302 sensitization by Chk1 inhibitors, the relative importance of these effects remains to be determined.
Upon DNA damage, Chks are activated and promote cell cycle arrest at G2 phase. G2 arrest correlates with an increase of Cdc2 inhibitory phosphorylation at its T14 and Y15 sites. Flow cytometry data confirmed that Chk1 inhibitors abrogate TH-302-induced G2 arrest in HeLa cells and S arrest in HT 29 cells. Consistent with flow cytometry data, induction of pY15 Cdc2 was observed following TH-302 treatment in vitro and this signal was abolished and phosphorylation of histone H3 was enhanced in the co-treatment group of TH-302 with Chk1 inhibitors.
Co-treatment of TH-302 and AZD7762 caused a dramatic increase in DNA breaks as measured by γH2AX staining and directly assessed with the single cell electrophoresis comet assay. Co-treatment of TH-302 and AZD7762 under normoxia produced a concentration-dependent and a greater tail moment compared to under hypoxia. This observation is consistent with a high level of DNA cross-linking of the broken DNA fragments under hypoxia, leading to slower migration of larger molecular weight fragments.
Chk1 plays a key role in protecting cells from apoptosis in response to many types of DNA damage [
21]. Down-regulation of Chk1 has been shown to selectively induce apoptosis in cancer cells [
41]. Our data showed that co-treatment of TH-302 and AZD7762 induced apoptosis although neither TH-302 nor AZD7762 alone induced a high level of apoptosis.
Preclinical studies have shown that AZD7762 potentiates DNA and replication-targeted therapies, including cisplatin, gemcitabine, irinotecan, and paclitaxel [
24,
29,
42-
44]. Similarly, the antitumor efficacy and levels of the corresponding biomarkers, γH2AX and caspase-3, from the combination of TH-302 and AZD7762 was significantly increased in the p53 mutant HT29 human tumor xenograft model. γH2AX was induced by AZD7762 or TH-302 alone and to a greater level when AZD7762 and TH-302 were combined. The induction of γH2AX by AZD7762 alone was consistent with the findings of Mitchell
et al. [
24], and which may be the result of replication stress [
18,
45]. Of note, the higher dose AZD7762 (20 mg/kg) group showed a sequence-dependent superior efficacy when combined with TH-302, which was consistent with the biomarker findings showing enhanced downstream PD effects with the same sequence. AZD7762 given first, followed by TH-302, and then followed by another dose of AZD7762 might be an optimal dosing sequence for future preclinical and clinical studies. Taking together the
in vivo efficacy profile and pharmacodynamic results, the HT29 xenograft data reported here support the hypothesis for a selective enhancement of TH-302 antitumor activity by the co-administration of a Chk1 inhibitor. As TH-302 predominantly and selectively targets hypoxic cells [
4,
6] even though all the tumor cells are chemosensitized by AZD7762, a triplet combination of AZD7762, TH-302, and a chemotherapeutic targeting the normoxic compartment may lead to an even superior efficacy profile [
5]. A similar therapeutic strategy can also be applied with hypoxia-activated Chk1 inhibitors [
15,
16] in combination with conventional cancer drugs targeting normoxic cells.
The results presented here support the following model for enhanced TH-302 activity by Chk1 inhibition. Hypoxia-activated TH-302 fragments in a one-electron reductase-dependent and hypoxia-selective manner, and releases the bis-alkylating effector Br-IPM, causing DNA cross-linking. Upon the recognition of the DNA damage by the DNA damage response (DDR), the variant histone H2AX is phosphorylated; yielding γH2AX, and Chk1 is activated. Chk1 phosphorylates Cdc2 at Y15, blocking phosphorylation of histone H3 and arresting the cell cycle at G2/M and S phase as well as allowing cells to undergo HDR-mediated DNA repair. When Chk1 is inhibited by a small molecule inhibitor, for example AZD7762, Chk1 inhibition triggers phosphorylation of histone H3, reduces Cdc2 phosphorylation, and leads to abrogation of cell cycle arrest. This allows arrested cells to progress directly into mitosis in cells co-treated with TH-302. Furthermore, Chk1 inhibition downregulated TH-302-induced upregulation of DNA repair protein Rad51, a key component of the specific DNA repair pathway. TH-302 treatment in the context of Chk1 inhibition activates caspases and induces apoptosis.
Competing interests
The authors declare that they have no competing interests. All authors are employees of Threshold Pharmaceuticals.
Authors’ contributions
FM, DB, JDS, QL, DA and YW performed research and conducted data analysis. CPH and MDM provided final approval of the manuscript. FM and CPH are accountable for all aspects of the work. All authors read and approved the final manuscript.