Background
Colorectal carcinoma (CRC), like most solid malignancies, comprises heterogeneous tumors with predominant hypoxic components. The adaptive tissue responses to hypoxic stress involve increased resistance to apoptosis (programmed cell death) as well as altered DNA damage repair and mutation rates, and thereby genomic instability [
1‐
3], ultimately leading to compromised efficacy of DNA-damaging therapies (chemotherapy and radiation). Moreover, mutations in genes such as
KRAS,
BRAF, and
PIK3CA commonly result in constitutive activation of cellular signaling mediated by mitogen-activated protein kinases (MAPK) and phosphatidylinositol 3-kinase–protein kinase B (PI3K–AKT) [
4,
5]. These pathways converge at the mechanistic target of rapamycin (mTOR), which regulates cell growth and survival [
6] and makes the mTOR complex an attractive target for CRC therapy. Consequently, a number of mTOR inhibitors have entered clinical trials.
There is however evidence of crosstalk between the mTOR-conducted signaling and other signaling pathways which will allow tumor cells to escape mTOR-inhibitory therapy [
7,
8]. Targeting of multiple pathways has therefore been considered. Recent findings showed that the combination of the mTOR inhibitor AZD8055 with ABT-263, an inducer of apoptosis, promoted cell death in CRC cell lines with
KRAS or
BRAF mutation [
9], a timely result given CRC entities harboring these mutations are refractory to current targeted therapies. ABT-263 and its structurally related compound ABT-737 are potent inhibitors of the anti-apoptotic proteins Bcl-2, Bcl-xL, and Bcl-w, but not of Mcl-1, and induce apoptosis in cancer cells [
10,
11]. Overexpression of Mcl-1 is associated with resistance to ABT-737, and inhibition of Mcl-1 has proven to sensitize cancer cells to ABT-737 [
12‐
14]. Interestingly, hypoxia has been shown to promote ABT-737-mediated apoptotic cell death in small-cell lung carcinoma, CRC, and hematologic cell lines via down-regulation of Mcl-1 [
15‐
17].
Since no information is available regarding the concurrent inhibition of anti-apoptotic proteins and mTOR-mediated pro-survival signaling under CRC tumor hypoxia, we investigated response to treatment with ABT-737 and AZD8055, in this report referred to as combo-Rx, in a panel of hypoxic CRC cell lines harboring various typical mutations.
Methods
Cell lines, culture conditions, and reagents
Fourteen human CRC cell lines (kindly provided by Prof. Kjersti Flatmark, Oslo University Hospital, Oslo, Norway or purchased from the American Type Culture Collection, Manassas, VA, USA) were first determined for mutations in
KRAS,
BRAF, and
PIK3CA by Ion Torrent PGM™ sequencing, and mutation profiles were in agreement to already published data [
18‐
20]. All cell lines except Caco-2 were kept in RPMI 1640 medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10 % fetal bovine serum (Gibco by Life Technologies, Grand Island, NY, USA) and 2 mM L-glutamine (GE Healthcare, PAA Laboratories, Pashing, Austria). The Caco-2 cells were kept in DMEM medium (Sigma-Aldrich) containing 15 % serum. The cell lines were routinely tested and found free of mycoplasma infection. For all assays, cells were seeded and left to adhere overnight to reach exponential growth at start of experiments. Cells were incubated under normoxic (21 % O
2) or hypoxic (0.2 % O
2) conditions, the latter obtained using the hypoxic chamber Invivo2 300 (Ruskinn Technologies, Leeds, UK). The mTOR inhibitor AZD8055, the PI3K/mTOR inhibitor BEZ235, the Bcl-2 family protein inhibitor ABT-737, and the pan-caspase inhibitor Z-VAD (all by Selleckchem.com, SMS-gruppen, Rungsted, Denmark) were dissolved in dimethyl sulfoxide (Sigma-Aldrich). Control cells received the vehicle.
Cell viability assay
Depending on the cell line, 12,000-20,000 cells were seeded per well in 96-well Costar plates (Corning Incorporated, Corning, NY, USA). Cells were given ABT-737 or AZD8055, separately or combined, in increasing concentrations (0.10-10 μM; combo-Rx designates 10 μM of both compounds), the combination of ABT-737 and BEZ235 (10 μM of both compounds), or vehicle. When expedient, the cells were pre-treated for 45 min with Z-VAD (20 or 50 μM). Cell viability was determined after 24 or 72 h by adding CellTiter 96®AQueous One Solution Reagent according to the manufacturer’s instructions (the MTS assay; Promega, Madison, WI, USA). Absorbance was measured using Varioscan (Thermo Electron, Waltham, MA, USA). Values were corrected for background absorbance, and values for treated cells are reported as percentage cell viability to corresponding control cell values. Presented results are from between three and seven independent experiments, each plated at least in triplicate.
Western blot analysis
Cells were seeded in Nuncleon T25 flasks (Thermo Fisher Scientific, Roskilde, Denmark) and were treated as indicated, and protein lysates from both floating and adherent cells were harvested as previously described [
21]. Equal amounts of protein (20 μg) were separated by NuPAGEBis-Tris (Novex by Life Technologies, Carlsbad, CA, USA), transferred by electrophoresis to Immobilon® membrane (Millipore Corporation, Billerica, MA, USA), and probed with antibodies against hypoxia-inducible factor type 1α (HIF-1α; BD Transduction Laboratories, Franklin Lakes, NJ, USA) and carbonic anhydrase IX (CAIX; kindly provided by Prof. Silvia Pastorekova, Slovak Academy of Sciences, Bratislava, Slovak Republic), and against Mcl-1, Bcl-2, Bcl-xL, caspase-3, mitogen-activated protein kinase3/1 (ERK1/2), p-ERK1/2(Thr202/Tyr204), AKT, p-AKT(Ser473), ribosomal protein S6 kinase beta-2 (S6), and p-S6(Ser235/236) (Cell Signaling Technology, La Jolla, CA, USA). Anti-α-tubulin (Calbiochem/EMD Chemicals Inc., San Diego, CA, USA) and Amido Black (Sigma-Aldrich) total protein staining were used as loading controls. Secondary antibodies were from Dako Denmark AS (Glostrup, Denmark). Peroxidase activity was visualized using SuperSignal West Dura Extended Substrate (Thermo Scientific, Rockford, IL, USA). Sufficient amount of lysate from each sample was prepared to run three gels. The parallel blotting membranes were considered identical, and different proteins were visualized on different membranes for practicality. All Western blot experiments were performed as three biological replicates.
RNA interference
Mcl-1 expression was inhibited using short hairpin (sh)RNA (clone ID NM_021960.3-953s1c1; Sigma-Aldrich), and control cells were generated using non-target sequence (product number shc002v; Sigma-Aldrich). The manufacturer’s instructions were followed apart from extending the lentiviral incubation period to 48 h.
Microscopy
Cells were seeded in Nuncleon T25 flasks and treated as indicated for up to 72 h. When expedient, the cells were pre-treated for 45 min with Z-VAD. Phase-contrast images were processed at the start of experiment and further after 24, 48, and 72 h by Olympus IX81 (Olympus Europa Holding GmbH, Hamburg, Germany).
Kinase activity profiling
The Tyrosine Kinase PamChip® Array technology (PamGene International B.V., ‘s-Hertogenbosch, The Netherlands) enables profiling of composite tissue kinase activities [
4]. The array contains peptides that are kinase substrates and consisting of 13 or 14 amino acids with tyrosine residues for phosphorylation. Protein lysates used for Western blot analysis were also incubated on the arrays for kinase activity profiling. Substrate phosphorylation intensities were measured using the Evolve software (PamGene International B.V.). Applying BioNavigator software (PamGene International B.V.), endpoint signal intensities generated from bound fluorescent anti-phosphotyrosine antibody were converted to numerical values. The primary array data are available in the ArrayExpress data repository (
http://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-3870/) by accession number E-MTAB-3870. Background signals were subtracted, and negative signal intensities were managed by subtracting the 1 % quantile of all data and setting the remaining signal intensities less than 1 to the value of 1. Data were log
2-transformed before mean signal intensity of three replicates that were analyzed for each experimental condition was calculated for each peptide substrate. The resulting data from each type of treatment was compared to the relevant control (vehicle-treated cells) for assessment of increase or decrease in substrate phosphorylation level. Substrates associated with PI3K-AKT and/or MAPK pathways were retrieved from PathCards (
http://pathcards.genecards.org/), applying the super-pathway definitions ‘PI3K-AKT signaling pathway’ and ‘MAPK signaling pathway’.
Statistical analysis
Differences between groups were analyzed using two-tailed Student’s t-test. p-values less than 0.05 were considered statistically significant. In the assessment of combination effects on cell viability, we chose to evade calculations based on median-effect equation for multiple drug interactions because the ABT-737 single-agent effects did not appear with typical dose-response curves.
Discussion
A number of recent findings [
4,
9,
16,
17] led us to investigate the effects of ABT-737, an inhibitor of anti-apoptotic Bcl-2 family proteins, in combination with the mTOR inhibitor AZD8055 in a panel of 14 hypoxic CRC cell lines. Combo-Rx (i.e.
, the combination treatment) suppressed viability of 13 of the cell lines, albeit ABT-737 did not significantly potentiate the inhibitory effect of single-agent AZD8055 in six of the models. On further mechanistic investigations, the hypoxic
KRAS/PIK3CA-mutant HCT-116 and HCT-15 cell lines (both with low endogenous expression of the anti-apoptotic Mcl-1 protein and showing augmented inhibition of viability following the addition of ABT-737 to AZD8055) responded to combo-Rx by induction of apoptosis (as assessed by various experimental approached in the HCT-116 cells) and with the simultaneous strong Mcl-1 up-regulation and activation of MAPK/PI3K-conducted signaling. A ubiquitous activation of hypoxic kinase signaling by combo-Rx was also confirmed in the HCT-116 cells. In contrast, in hypoxic
KRAS-mutant LoVo,
BRAF/PIK3CA-mutant RKO, and wild-type Colo320DM cell lines (all with high endogenous Mcl-1 expression and being resistant to the additional effect of ABT-737 to AZD8055), combo-Rx did not elicit apoptotic or pro-survival responses. Collectively, this data revealed complex responses to the concurrent inhibition of anti-apoptotic proteins and mTOR-mediated signaling in hypoxic CRC cell lines, where pro-survival responses were elicited in parallel with the intended anti-proliferative effects in
KRAS/PIK3CA-mutant entities in particular, a finding that should be of note if considering the combinatory targeting of multiple pathways in CRC treatment.
As recently shown [
9], when combined with ABT-263 (structurally related to ABT-737), AZD8055 via the specific suppression of Mcl-1 sensitized CRC cell lines with
KRAS or
BRAF mutation to undergo apoptosis, a timely result as such CRC entities are refractory to current targeted therapies. In vitro studies have shown promising treatment effects of both types of agents, but concerns have been raised with regard to lacking therapeutic efficacy of mTOR inhibitors in solid tumors [
11,
25]. Both AZD8055 and the anti-apoptotic inhibitors ABT-263 and ABT-737 have reached early-phase clinical trials. However, only a few reports exist on the use of ABT-263 or ABT-737 in hypoxic tumor models [
15,
16], and as pointed out by Harrison and co-workers [
16], whether Mcl-1 is up- or down-regulated may be cell type- and oxygen concentration-dependent. To our knowledge, little information is currently available on effects of AZD8055 or other mTOR inhibitors under hypoxic conditions, with the exception of BEZ235 which has been shown to sensitize hypoxic breast and prostate cancer cells to radiation [
26,
27].
Invariably, CRC comprises heterogeneous tumors with predominant hypoxic components [
4], which is important to take into consideration with established as well as novel therapies. The present study showed that combo-Rx significantly suppressed viability of hypoxic CRC cells, but in six of 14 cell lines there was no additional inhibitory effect of ABT-737 to that of single-agent AZD8055. A similar finding was obtained with the combination of ABT-737 and BEZ235 (potentiation was not seen in five of 12 cell lines) and when AZD8055 in a lower concentration of 0.10 μM was given together with ABT-737 (four of seven hypoxic cell lines were resistant to an additional effect of the anti-apoptotic inhibitor). In the five cell lines where mechanistic investigations were undertaken (the
KRAS-mutant HCT-116, HCT-15, and LoVo models and the
BRAF/PIK3CA-mutant RKO and wild-type Colo320DM cells), combo-Rx under hypoxic conditions caused dual phenotypic responses in terms of concurrent apoptotic and pro-survival effects in the HCT-116 and HCT-15 cell lines, as demonstrated through the specific examination of the anti-apoptotic Mcl-1 protein, microscopy of cell cultures, and both targeted and comprehensive analysis of kinase signaling. Importantly, both of these cell lines also harbor
PIK3CA mutation. These findings suggest that CRC cancers with co-occurring
KRAS and
PIK3CA mutations, which is not a frequent entity [
5], may be particularly susceptible to parallel apoptotic and pro-survival effects with this combination treatment.
The MAPK and PI3K-AKT signaling pathways merge at the mTOR complex, which promotes cell survival through phosphorylation of S6 and the resulting increase in Mcl-1 protein translation [
28]. Intriguingly, in hypoxic HCT-116 and HCT-15 cells, combo-Rx strongly increased expression of Mcl-1 and p-S6, which under normoxia and in agreement with previously reported data [
9] showed the opposite response following the addition of AZD8055 to ABT-737. The Tyrosine Kinase PamChip® Array approach enabled the investigation of more general kinase activity responses. Using this technology, ABT-737 was shown to repress kinase activities known to be important for CRC survival in hypoxic HCT-116 cells. Under normoxic conditions, ABT-737 treatment has been shown to sensitize CRC and rhabdomyosarcoma cell lines for AZD8055-directed apoptosis [
9,
29]. It is therefore notable that in our experimental setups, a number of array substrates associated with MAPK/PI3K-conducted signaling (RASA1, RAF1, PIK3R1, and PDPK1) were phosphorylated by lysates from hypoxic HCT-116 cells given combo-Rx. Moreover, array substrates reflecting proteins that are fundamental in the angiogenic response to hypoxia, particularly PDGFRB, were also highly phosphorylated. In angiogenesis, PDGFR is required for the formation of a functional pericyte coverage of regenerating endothelium within the tumor stroma [
30]. Importantly, the
KRAS-mutant HCC2998 cell line, which was sensitive to combo-Rx for viability but devoid of
PIK3CA mutation, and the resistant wild-type Colo320DM model, showed repressed or unchanged global kinase activities to the experimental perturbations. Again, the findings indicate a particular susceptibility of
KRAS/PIK3CA-mutant CRC entities to unfavorable responses to the combined inhibition of anti-apoptotic proteins and mTOR signaling. Of final note, in the context of interpreting the ex vivo kinase substrate data, some important considerations should be kept in mind [
4]. One is that hypoxia elicits a multitude of adaptive signaling responses, which as such are challenging to portray, and another is that phosphorylation of each individual substrate on the kinase target array reflects the net result of an extensive network of multiple kinase activities.