Background
Small molecule inhibitors against checkpoint kinases constitute a promising class of targeted cancer therapeutics and many are currently under preclinical or even clinical evaluation. The role of checkpoint kinases is to respond to stress, typically damaged DNA or aberrant chromosomal structure, and stop the cell division process long enough for the damage to be repaired. These “checkpoints” prevent cells from dividing and perpetuating mutations or chromosomal anomalies that would otherwise lead to cellular lethality. The rationale for inhibiting checkpoint kinases is to accumulate irreparable and fatal genetic lesions by compromising the DNA damage response (DDR) and forcing premature or untimely cell division. Notable examples include the mitotic checkpoint kinases Aurora A and B, checkpoint kinase 1 (CHK1), CHK2, ATR, and WEE1.
Several CHK1 inhibitors have been employed in early stage clinical trials [
1,
2]. Notably, MK-8776 (also referred to as SCH-900776), a CHK1-selective inhibitor, is under evaluation in phase I studies in combination with gemcitabine or cytarabine [
3]. Only one inhibitor of WEE1 has been explored clinically. MK-1775, a potent and selective inhibitor of WEE1, achieved favorable phase I pharmacokinetic and pharmacodynamic endpoints in combination with carboplatin, cisplatin, and gemcitabine, and is under further investigation as a chemosensitizer in a phase II trial [
4].
CHK1 is an essential serine/threonine kinase involved in S- and G2/M-phase checkpoints [
5‐
9], replication initiation and fork stability [
10‐
12], homologous recombination repair [
13,
14], and entry into mitosis in normal cycling cells [
15]. Importantly, CHK1 is necessary for unperturbed DNA replication and cell cycle coordination even in the absence of any exogenous insult [
16]. The cytotoxic nature of CHK1 knockdown or inhibition, either alone or in combination with DNA-damaging therapeutics, has been described extensively (for review, see [
2]).
WEE1 is an essential tyrosine kinase that is also involved in S and G2/M checkpoints. WEE1 directly phosphorylates and inhibits CDK1 and CDK2 at the conserved tyrosine 15 residue, affecting entry into mitosis as well as coordination of DNA replication events. WEE1 is therefore critical for properly timing cell division in unperturbed cells, and loss of WEE1 results in chromosomal aneuploidy and accumulated DNA damage [
17]. Additionally, WEE1 is critical to S- and G2/M-phase checkpoint responses following DNA damage as well as in unperturbed cells [
18‐
20]. Interfering with WEE1 has been shown to repress cancer cell proliferation and sensitize theme to the anti-tumor growth effects of DNA-damaging chemotherapeutics or radiation therapy [
21‐
28].
Considering the overlapping roles of WEE1 and CHK1 in mitotic entry, DNA replication, and the DDR, we sought to determine whether inhibition of these two kinases was redundant or complimentary. We demonstrate here that combination of a CHK1 inhibitor, MK-8776, and a WEE1 inhibitor, MK-1775, results in synergistic inhibition of cell proliferation in several human tumor cell lines. Minimal concentrations of the drugs required to block cell proliferation lead to a greater than additive increase of γH2AX, a marker of DNA double strand breaks (DSB). This occurs primarily in S-phase cells, suggesting that the unique combination of CHK1 and WEE1 inhibitors disrupts DNA replication and its associated checkpoint. Pharmacodynamic (PD) analysis in xenograft tumors supports this notion, showing an increase in both the percentage of cells containing DNA damage as well as the duration of the DDR. Consistent with the PD data, we demonstrate that the combination of CHK1 and WEE1 inhibitors leads to greater-than-additive tumor growth inhibition in two human tumor xenograft models. Collectively, these data demonstrate the synergistic anti-tumor effects of pharmacological WEE1 and CHK1 inhibition and highlight the potential of this unique combination in treating human cancer independently of chemotherapeutic drugs.
Conclusions
Using small molecule inhibitors currently under early clinical development, we have shown that simultaneous inhibition of the WEE1 and CHK1 kinases results in synergistic potentiation of each drug for a variety of cell types in proliferation assays. Knockout of WEE1 results in embryonic lethality before day 3.5 [
17], and knockdown of WEE1 is known to inhibit proliferation of several cancer cell lines in vitro [
27,
28]. Similarly, anti-proliferative effects of CHK1 inhibition via siRNA or pharmacologic inhibition have been described [
30]. The increased potency of MK-1775 and MK-8776 when combined supports the notion that WEE1 and CHK1 have non-overlapping activity. Potentially predictive biomarkers for each class of drug have been described for their chemosensitizing effects, including p53 status for both WEE1 and CHK1 [
2,
31], WEE1 expression levels for WEE1 [
28], and cyclin B levels for CHK1 [
32]. Interestingly, synergy between MK-1775 and MK-8776 did not correlate with the p53 status of the cell line, though overall sensitivity to the drugs might favor p53 mutant lines. Furthermore, three of the seven lines described in Figure
1 are wild type for p53 (A2780, LoVo, and A427). Further examination of other putative markers such as expression of WEE1, CHK1, or cyclin B1, will be important future questions to address in understanding the cellular context of WEE1 and CHK1 inhibitor activity.
Mechanistic studies suggest that WEE1 and CHK1 inhibitors combine synergistically due to, at least in part, alterations of the cell cycle and compounded DNA damage (Figures
2,
3, and
4). Though both MK-1775 and MK-8776 are chemosensitizers that potentiate the anti-proliferative effects of DNA damaging chemotherapeutics, it is also known that knockdown or inhibition of either WEE1 or CHK1 alone leads to DNA damage. Therefore, it is likely that MK-1775 and MK-8776 work together in an analogous fashion as they do in combination with genotoxic agents to prevent proper checkpoint response and damage control. Importantly, DNA damage incurred by WEE1 and CHK1 inhibition occurs primarily in S phase and requires CDK activity, consistent with findings that disruption of either WEE1 or CHK1 individually leads to S-phase arrest, slowed DNA replication, and induced DNA damage. Increased accumulation and duration of DNA damage by MK-1775 and MK-8776 was observed in vivo, and accordingly the combination led to inhibition of tumor growth in xenograft models. WEE1 and CHK1 inhibition was unable to prevent tumor regrowth, however, suggesting either that not all cells are affected or that following drug treatment cells are able to sufficiently repair damaged DNA. Along these lines, we were unable to find robust evidence of apoptosis both in vitro and in vivo (data not shown).
The WEE1 inhibitor MK-1775 is known to reduce phosphorylation on tyrosine 15 of CDK1/2, resulting in increased CDK1/2 activity [
26]. Inhibition of CHK1 increases the activity of the protein phosphatases CDC25A/B/C, thereby reducing phosphorylation of tyrosine 15 and indirectly increasing CDK1/2 activity. We hypothesized, therefore, that combined inhibition of WEE1 and CHK1 could result in an additive inhibition of phospho-CDK1/2
Y15. However, we were unable to observe a substantial decrease in phospho-CDK1/2
Y15 beyond the effect of MK-1775 alone, suggesting that CHK1 inhibition by MK-8776 compliments inhibition of WEE1 through mechanism(s) and target(s) distinct from CDK1/2.
The synergistic antiproliferative effect of combined WEE1 and CHK1 inhibition was also noted by Davies et al. [
33] and Carrassa L et al. [
34]. Each of these studies identified the WEE1 gene as an siRNA target that could sensitize to either a CHK1 inhibitor (Davies et al.) or a CHK1 siRNA (Carrassa et al.) in solid tumor cell lines. Davies et al. reported synergy between WEE1 and CHK1 inhibitors in four cell lines, three of which are reported p53 wild type [
33]. Similarly, Carrassa et al. reported synergy in seven cell lines regardless of p53 status [
34]. This manuscript extends earlier findings into 37 cancer cell lines using compounds that are currently under early stage clinical development. Our findings align with those reported demonstrating that the mechanism underlying synergy between WEE1 and CHK1 inhibition is ubiquitous as well as with the finding that p53 status does not affect this synergy.
Davies et al. reported an absence of premature mitosis in the HEL92.1.7 cell line, though this experiment was conducted with an excess of WEE1 and CHK1 inhibitors required for inhibition of cell proliferation (compare Figures
2C and
5 in [
33]). Carrassa et al. conducted mechanistic studies in one cell line, OVCAR-5, and concluded that premature mitosis accompanied the simultaneous inhibition of WEE1 and CHK1 inhibition [
34]. It was unclear in that study whether concentrations of inhibitors used to study biochemical correlates coincided with the concentrations required to inhibit proliferation. By examining the effects of MK-1775 and MK-8776 at the lowest concentrations needed to achieve antiproliferative activity, individualized for multiple cell lines, we are able to demonstrate that DNA damage rather than premature mitosis seems to be the primary cause of synergistic cytoxicity (Figure
4), though we do find that select cell lines, i.e. HT-29, may undergo premature mitosis as well. Importantly, these findings were corroborated in vivo where LoVo xenograft tumor samples demonstrated synergistic increases in the DNA damage markers γH2AX and pCHK1
S345 but not in the mitosis marker pHH3 (Figure
7 and data not shown). Collectively these data argue that nonoverlapping functions of the WEE1 and CHK1 kinases during S- phase are responsible for the widespread and strong synergy observed following their inhibition.
Our studies describe synergy achieved by simultaneous inhibition of the WEE1 and CHK1 kinases and, together with the work of Davies et al. [
33] and Carrassa et al. [
34], provide pharmacologic evidence that the two kinases have unique and nonoverlapping activities. Combined treatment with MK-1775 and MK-8776 demonstrates synergistic DNA damage and anti-tumor efficacy at tolerated doses, suggesting possible clinical use of the drugs in combination. The robust and ubiquitous nature of the synergy may suggest potential toxicity in normal tissue and therefore identification of mechanisms underlying sensitivity will be important in understanding the potential clinical application of this combination.
Methods
Cell culture and compounds
All cell lines were obtained from American Type Culture Collection, except A2780 cells which were obtained from Sigma, and cultured under vendors’ recommended conditions. The HCT116 and RKO isogenic cell lines were obtained from Horizon Discovery, LTD. The chemical name of MK-1775 is (2-allyl-1-[6-(1-hydroxy-1-methylethyl) pyridin-2-yl]-6-{[4-(4-methylpiperazin-1-yl) phenyl]amino}-1,2-dihydro-3H-pyrazolo[3,4-d] pyrimidin-3-one), and its chemical structure is described elsewhere [
26]. The chemical name of MK-8776 is
(R)-(−)- 6-Bromo-3-(1-methyl-1H-pyrazol-4-yl)-5-piperidin-3-yl-pyrazolo[1,5-a]pyrimidin-7-ylamine and has been described previously as SCH-900776 [
35]. SCH-727965 has also been previously described in the literature [
29].
Cell viability assay
For each experiment, cells were seeded in duplicate 96-well white walled plates at 4,000 cells per well. After overnight incubation, cells were treated with combinations of DMSO as vehicle control, MK-1775, and MK-8776 for 72 hours. Cell viability was determined by measuring ATP with Vialight Plus (Lonza) according to manufacturer’s instructions. Drug potency was calculated as the ratio of relative light units (RLUs) in compound treated wells over DMSO-treated control wells and expressed as % DMSO control. Compound EC50s were calculated in GraphPad Prism using a 4 parameter variable slope sigmoidal dose response curve fit.
Flow cytometry
Cells were treated with indicated concentrations of MK-1775, MK-8776, both, or an equivalent volume of vehicle for a fixed time period. At time of harvest, cells were counted and then fixed in ice cold 70% ethanol overnight before staining with anti-phospho-histone H2AX (S139; γH2AX) antibody conjugated to FITC (from Millipore kit 17–344), anti-phospho-histone H3 (S28; pHH3) antibody conjugated to Alexa Fluor® 647 (BD Biosciences 558217), and propidium iodide (BD Biosciences 550825). Samples were read on the LSR II (BD Biosciences 347545) flow cytometer, and data were analyzed using FlowJo software version 7.5.5.
Animals and xenograft studies
CD-1 Nu/Nu female mice aged 5–6 weeks were obtained from Charles River Laboratories and housed in our animal care facility at standard laboratory conditions and fed 2018S autoclaveable diet and water ad libitum. The protocol was approved by Merck’s Institutional Animal Care and Use Committee. Mice were inoculated with 5 x 106 LoVo cells in 100 μL (1:1 Matrigel:PBS) subcutaneously (S.C.) into the right flank. When tumor volume reached 200 mm3 (+/−50) mice were pair-matched so each group had a similar mean and standard deviation. Tumor volume and body weights were recorded bi-weekly. Mice received 4 treatment cycles of twice daily dosing (BID) for 2 days receiving either vehicle, MK-1775 (50 mpk), and/or MK-8776 (50 mpk) For pharmacodynamic assays, mice were dosed with 60 mpk of each compound.
In vivo pharmacodynamic assays
Xenograft tumors were fixed in 10% formalin, paraffin-embedded and sectioned at 5 μm. Tumor sections were immunostained with rabbit monoclonal anti-phospho-CHK1 (S345) antibody (1:300 dilution; Cell Signaling); rabbit polyclonal anti-gamma-histone H2AX antibody (1:2000 dilution; Bethyl); rabbit polyclonal anti-phospho-CDC2 Y15 antibody (1:960; R&D Systems) and rabbit monoclonal anti-Ki67 antibody (1:200; Epitomics). Labeled antigens were visualized using Omni Map anti-rabbit HRP and peroxidase substrate (Ventana Medical Systems). Slides were digitized using an Aperio ScanScope XT Image System and immunostained cells were quantified using Aperio Imagescope software. The percentage of cells showing immunostaining in each tumor was calculated relative to the number of total cells with necrotic regions excluded.
Bliss synergy calculations
The Bliss independence (BI) model is used to define the effect of two drugs assumed to act through independent mechanisms [
36]. BI is described by the equation
E
i
= (
E
A
+
E
B
) - (
E
A
x
E
B
), where
E
i
is the predicted effect (percentage of inhibition) by the combination of drugs
A and
B if they were to act additively and independently, and
E
A
and
E
B
are the observed effects (percentage of inhibition) of each drug alone, respectively. When observed inhibition exceeds predicted inhibition, the two compounds are considered to act synergistically.
Competing interests
AG, XQ, NM, YL, JL, IF, YB, AB, and SS are employees of Merck Research Laboratories. The author(s) declare that they have no competing interests.
Authors’ contributions
AG and MM both participated in proliferation assays, carried out flow cytometry for γH2AX, and performed Western blot analyses of cell line studies. YL and JL performed the combination screen of MK-1775 and MK-8776 on solid tumor cell lines. BR, YB, and IF performed synergy analyses of the screen. MH performed xenograft efficacy studies and xenograft sample collection. XQ and NM conducted and quantitated immunohistochemistry assays on xenograft tumor samples. AB contributed to study design and concept. CT contributed to study design, concept, and manuscript preparation. SS participated in proliferation assays, study design and coordination, figure construction, and drafting of the manuscript. All authors read and approved the final manuscript.