Abstract
Background: Triple-negative breast cancer (TNBC) is an aggressive form of breast cancer currently lacking targeted therapies. Our previous work demonstrated a therapeutic synergism with gemcitabine (GEM) and the CHK1 inhibitor (AZD7762) combination treatment in a TNBC cell line. We hypothesized that the response to this combination therapy would differ among heterogeneous TNBC patients and that addition of a SMAC mimetic (TL32711) could improve efficacy. Materials and Methods: Therapeutic responses to GEM, GEM/AZD7762, and GEM/AZD7762/TL32711 combinations were investigated by XTT assays and western blotting of cell cycle and apoptosis-related proteins in ten TNBC cell lines. Results: TNBC cell lines harboring low levels of endogenous CHK1, cIAP1 and cIAP2 were responsive to GEM alone, whereas cell lines demonstrating a minimal increase in phospho-S345 CHK1 after treatment were responsive to GEM/AZD7762 or GEM/AZD7762/TL32711 combination. Conclusion: The response of TNBC cells to particular therapies varies and will require development of predictive biomarkers.
Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer that accounts for approximately 15% of all breast cancer cases with poor survival rates (1). The overall 5-year survival rate remains poor primarily due to the lack of targeted therapies for this form of breast cancer (2-4). Additionally, multiple subtypes of TNBC have been described indicating that different types of therapies may be required for subtypes of TNBC (5).
Most TNBC tumors harbor mutations or deletions of p53 (7), an important mediator of the cellular response to DNA damage. Under conditions where p53 function has been lost, the cell-cycle checkpoint regulator checkpoint kinase 1 (CHK1) becomes an especially critical mediator of cellular response to DNA damage and has intensively been studied as a therapeutic target in p53-deficient cancer cells (8, 9). In p53-deficient cells, where DNA damage has occurred, CHK1 blocks cell-cycle progression at the G2-M phase, providing cells with the opportunity to repair damaged DNA, maintaining genomic integrity and preventing apoptosis (10). Our previous work identified CHK1 and ribonucleotide reductase 1 and 2 (RRM1/2, targeted by gemcitabine (GEM)) as genes critical for TNBC growth and further demonstrated that GEM and CHK1 inhibitor, UCN-01 or AZD7762, act synergistically in inhibiting cell growth of some TNBC cells (11).
We, however, hypothesized that the response of the GEM/CHK1 inhibitor combination is different among TNBC patients and this combination might not be sufficient to induce apoptosis in some TNBC cells with elevated levels of anti-apoptosis proteins such as inhibitor of apoptosis proteins (IAPs). IAPs play key roles in preventing apoptosis by inhibiting caspases and by activating death receptor-mediated NF-κB pathways (12). The major IAPs in mammalian cells are cellular IAP1 and 2 (cIAP1 and cIAP2) and X-linked IAP (XIAP). cIAP1 and cIAP2 are mainly involved in the activation of TNF receptor-mediated NF-κB pathways that can result in inhibition of apoptosis, whereas XIAP is an inhibitor of the proteolytic activity of caspases (12,13). The second mitochondria-derived activator of caspases (SMAC) mimetics including TL32711 (birinapant) are small molecules that trigger proteasomal degradation of IAPs such as cIAP1 and cIAP2 (14). We hypothesized that a SMAC mimetic might promote GEM/CHK1 inhibitor-mediated apoptosis by blocking IAPs in susceptible TNBC cells.
In this study, we examined the effects of GEM, AZD7762 and TL32711, alone or in combination, on cell proliferation and cell death in ten p53-mutated TNBC cell lines. We found that low levels of endogenous CHK1, cIAP1, and cIAP2 correlated with a good response to GEM-only treatment, whereas a minimal increase in the level of phosphorylated CHK1 at S345 following GEM/AZD7762 or GEM/ AZD7762/TL32711 combination treatment correlated with good responsiveness of TNBC cells.
Materials and Methods
Reagents and cell culture. GEM and AZD7762 were purchased from Sigma-Aldrich (St. Louis, MO, USA) and Selleckchem (Houston, TX, USA), respectively. TL32711 was provided by TetraLogic Pharmaceuticals (Malvern, PA, USA). Human breast cancer cell lines HCC1806, HCC38, MB468, BT549, Hs678T, MB231, HCC70, HCC1937, HCC1395, and BT20 were purchased from the American Type Culture Collection (Manassas, VA, USA). All cell lines were grown in RPMI medium 1640 supplemented with 10% (v/v) fetal bovine serum and 100 mM penicillin-streptomycin at 37°C, which were purchased from Invitrogen (Carlsbad, CA, USA). TNF-α neutralizing antibody and caspase inhibitor z-VAD-FMK were purchased from Cell Signaling (Danvers, MA, USA) and Promega (Madison, WI, USA), respectively.
Cell proliferation assay (XTT assay). Cells were seeded in 96-well plates at 2×103 cells /well in triplicate and treated with the drugs 24 h after seeding. XTT assay was performed 3 days after drug treatment. Cellular viability was assessed by incubating cultures with XTT mixed with PMS (N-methyl dibenzopyrazine methyl sulfate) (Sigma-Aldrich), measured by a Tecan plate reader (Männedorf, Switzerland), and calculated relative to experimental negative controls. Data points represent an average of three samples per treatment and experiments were repeated at least three times.
Western blot analysis. Total protein was extracted from cells using RIPA buffer (Sigma-Aldrich) supplemented with protease inhibitor and phosphatase inhibitor cocktails (Thermo Scientific, Rockford, IL, USA), quantified using BCA Protein Assay Kit (Thermo Scientific), separated on NuPage 4-12% gel (Invitrogen), and transferred to an Immobilon-P transfer membrane (Millipore, Billerica, MA, USA). Protein bands were visualized using SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) or Pierce ECL Western Blotting Substrate (Thermo Scientific). All primary antibodies were purchased from Cell Signaling. GAPDH and the secondary antibodies were purchased from Millipore (Billerica, MA, USA) and GE Healthcare (Marlborough, MA, USA), respectively.
Results
Gemcitabine-sensitive TNBC cells express low levels of CHK1, cIAP1, and cIAP2 proteins. To determine the potential heterogeneous responses of TNBC cells to treatment with GEM alone, we treated ten TNBC cell lines (Table I) with the relatively low concentration of 10 nM GEM (data not shown) to provide an improved opportunity to assess the efficacy of combination therapies. After treatment for 3 days with GEM, cell growth of HCC1806 and HCC38 cell lines was dramatically reduced by ~75-80% whereas the other eight cell lines did not demonstrate a major response, although BT549 had a modest (~50%) response (Figure 1A). To determine whether cells were dying from treatment, we measured the apoptosis markers cleaved PARP1 and caspase-3 after treatment for 1 day (Figure 1B). Only the HCC1806 and HCC38 cell lines underwent apoptosis that was correlated with the reduced cell growth.
Given the lack of response of most cell lines to low-dose GEM, we hypothesized that some of the cells might prevent the activation of apoptosis by G2/M-checkpoint-related proteins including CHK1 and/or by anti-apoptosis-related proteins, including cIAP1, cIAP2, XIAP, survivin, and MCL-1 (15). Interestingly, CHK1, cIAP1, and cIAP2 expressions are very low in the GEM-sensitive cell lines HCC1806 and HCC38 compared to the more resistant cell lines (Figure 1C). These findings suggest that low expression of CHK1, allowing cell-cycle progression despite DNA damage by GEM, and low expression of cIAP1 and cIAP2, leading to reduced activation of anti-apoptosis pathways, might serve as potential biomarkers associated with GEM-induced apoptosis in TNBC.
AZD7762 sensitizes cells to GEM in a subset of TNBC cell lines inducing lower CHK1 phosphorylation. Based on our results above and our previous work (11), we further delineated the effect of combining the CHK1 inhibitor AZD7762 with GEM to induce apoptosis in the 8 TNBC cell lines that were not very sensitive to GEM. 100 nM AZD7762, a concentration not affecting cell growth as a single treatment, was chosen for combinational treatments (data not shown). The response of cells to GEM/AZD7762 combination was examined using the XTT cell proliferation assay (Figure 2A) and apoptosis markers (Figure 2B). The GEM/AZD7762 combination resulted in more than a 60% decrease in cell growth for BT549, MB468, Hs578T, MB231 and HCC70 but little decrease for HCC1937, HCC1395 and BT20 (Figure 2A). Apoptosis was clearly induced only in the BT549 and MB468 cell lines with the combination treatment as evidenced by cleaved caspase-3 and PARP1 (Figure 2B). As expected, both the HCC1806 and HCC38 cell lines (which are highly sensitive to GEM alone) exhibited a greater apoptotic response than did the BT549 and MB468 cell lines with the combination treatment.
We expected that the GEM/AZD7762 combination treatment would induce more DNA damage and ATR activity resulting in increased S345 CHK1 phosphorylation (8, 10, 11). Interestingly, the S345 CHK1 phosphorylation was relatively low in the sensitive cell lines HCC1806, HCC38, BT549 and MB468 (Figure 2B). Thus, high S345 CHK1 phosphorylation may indicate a poor response to the GEM/AZD7762 combination in TNBC and could potentially be a pharmacodynamics marker of response.
GEM/AZD7762/TL32711 combination treatment enhances apoptosis in a subset of TNBC cell lines. The GEM/AZD7762 combination treatment was not effective in inducing apoptosis for some TNBC cell lines (Hs578T, MB231, HCC70, HCC1937, HCC1395, and BT20). Therefore, we hypothesized that the efficacy of the GEM/AZD7762 combination therapy could be improved in cell lines expressing high cIAP1 levels through the inhibition of IAPs mediated by the addition of the SMAC mimetic TL32711. Based upon dose titrations, 500 nM TL32711 was chosen (data not shown). The response of the GEM/AZD7762/TL32711 triple combination treatment was compared to that of GEM/AZD7762 dual combination and TL32711 single treatments using the XTT assay in the 8 TNBC cell lines (Figure 3A). As expected, BT549 and MB468 cell lines demonstrated no significant effect in response to the addition of TL32711. Among the cell lines resistant to the GEM/AZD7762 combination treatment, cell growth using the triple combination was dramatically decreased in Hs578T, MB231, and HCC70 cell lines, but remained relatively high in HCC1937, HCC1395, and BT20.
We selected MB231 and HCC70 as sensitive cell lines to the triple combination and HCC1937 and HCC1395 as non-sensitive cell lines for further studies. As expected, these non-sensitive cell lines did not undergo apoptosis as measured by caspase-3 and PARP-1 cleavage (Figure 3B). Interestingly, treatment of MB231 cells with TL32711 alone resulted in greatly increased apoptosis as previously reported (16). We next investigated TL32711 activity in sensitive and non-sensitive cell lines by assessing the TL32711 mediated degradation of its targets, cIAP1 and cIAP2. As expected, cIAP1 protein levels were dramatically decreased and cIAP2 protein levels were somewhat decreased by TL32711 in all cell lines (Figure 3B). This suggests that TL32711 is capable of inhibiting the activity of IAP proteins in both the sensitive and non-sensitive cell lines.
It has been demonstrated that the anti-tumor effects of SMAC mimetics are dependent upon the TNF receptor pathway that is stimulated by TNFα leading to caspase-3 activation (13, 17). To investigate whether GEM/AZD7762/ TL32711 treatment-induced apoptosis is TNFα and/or caspase-3 dependent, we inhibited caspase-3 using z-VAD-FMK and neutralized TNFα using a TNFα antibody (Figure 4A). Induction of apoptosis was completely blocked by inhibiting caspase-3 in both MB231 cells treated only with TL32711 and HCC70 cells treated with the triple combination as in HCC38 cells treated with the GEM/AZD7762 combination, indicating that TL32711-mediated apoptosis is caspase-3 dependent. In addition, apoptosis in MB231 and HCC70 cells was also reduced by neutralizing TNFα, which was different from the GEM/AZD7762 combination-induced apoptosis in HCC38 cells, suggesting that TL32711-mediated caspase-3 activation is TNF pathway dependent.
We showed above that a low level of S345 CHK1 phosphorylation in cells treated with the GEM/AZD7762 combination resulted in increased apoptosis (Figure 2B). CHK1 phosphorylation levels were, therefore, measured before and after the triple combination treatment in both sensitive and non-sensitive cell lines (Figure 4B). Interestingly, the phospho-CHK1 level was decreased in the sensitive cell lines MB-231 and HCC70 with the addition of TL32711, but did not change in the non-sensitive cell lines HCC1937 and HCC1395 with TL32711 (compare GEM/AZD7762 to GEM/AZD7762/TL32711). In addition, neutralization of TNFα restored the phospho-CHK1 levels to the levels seen with the GEM/AZD7762 combination in the sensitive cell lines. These results suggest that TL32711 inhibits CHK1 phosphorylation induced by GEM and AZD7762 by an unknown mechanism that is TNF pathway-dependent and this may further contribute to the efficacy of the triple combination treatment.
Relative anti-proliferative and apoptosis activity of gemcitabine in triple-negative breast cancer cell lines. (A) Proliferation of cells treated with 10 nM gemcitabine for 3 days was measured using the XTT assay. Cell viability of the treated cells was compared with non-treated cells of each cell line. Values are the average of three determinations±SD. (B) Detection of apoptosis marker proteins in the cells treated with 10 nM gemcitabine for 24 h. C: Control, G: gemcitabine, CL: cleaved length, FL: full length. (C) Endogenous expression of anti-apoptosis proteins. Ponceau staining was used as loading control.
Discussion
The fact that most TNBC tumors harbor p53 mutations provides avenues to identify synthetic lethal approaches based upon the loss of p53 function. Targeting p53-deficient cells through inhibition of CHK1, that inhibits the G2-M checkpoint, provides a complimentary means to enhance cell-cycle progression before cells have the opportunity to repair DNA damage (8-10), leading to mitotic catastrophe and death of tumor cells (18), as we and others have previously shown (11, 19, 20).
Relative anti-proliferative and apoptosis activity of gemcitabine and AZD7762 as a single or combination treatment in triple-negative breast cancer cell lines. (A) Proliferation of cells treated with 10 nM gemcitabine, 100 nM AZD7762, or gemcitabine/AZD7762 combination for 3 days was measured using the XTT assay. Cell viability of treated cells was compared to non-treated cells of each cell line. Values are the average of 3 determinations±SD. (B) Detection of apoptosis markers and cell cycling proteins in the cells treated with 10 nM gemcitabine and 100 nM AZD7762 for 24 h. C: Control, G: gemcitobine, GA: gemcitabine + AZD7762, CL: cleaved length, FL: full length. Ponceau staining was used as loading control.
Since TNBC represents several sub-groups (5, 21), we extended our GEM/CHK1 inhibitor studies to determine how effective the GEM/CHK1 inhibitor treatment performs using ten TNBC cell lines. Not surprisingly, the sensitivity of TNBC cells to the GEM/AZD7762 combination was variable among the ten TNBC cell lines. These results suggest that the combinational treatment with a genotoxic agent and a CHK1 inhibitor may not be efficacious for all TNBC patients, and that additional strategies need to be implemented.
Since GEM has often been used as a single treatment for breast cancer patients (22), we first investigated the sensitivity of TNBC cell lines to GEM-only treatment. The response to GEM only treatment was quite variable among the cell lines. Interestingly, HCC1806 and HCC38 cell lines expressing lower endogenous levels of CHK1 than the other cell lines were more sensitive to GEM (Figure 1). CHK1 is a cell-cycle checkpoint kinase controlling the pre-M phase arrest in response to DNA damage and is required to facilitate DNA repair. Thus, CHK1 is considered to be a critical effector of the DNA damage checkpoint (8-10). Previous reports have demonstrated that CHK1 levels in various cancer cells is high and that overexpression of CHK1 is associated with therapeutic resistance in non-small cell lung cancer cells (23, 24). High CHK1 expression may contribute to DNA repair by blocking progression to the M-phase in the cell lines less-sensitive to GEM (25, 26).
The response to the GEM/AZD7762 combination treatment was also variable among the eight cell lines that did not respond to GEM alone, with good apoptotic responses in BT549 and MB468 cell lines and modest responses in Hs578T and MB231 cell lines (Figure 2). We found that, in contrast to single treatment with GEM, the endogenous expression level of CHK1 did not correlate with the response of cells to the GEM/AZD7762 combination treatment. CHK1 is phosphorylated on Ser317 and Ser345 through activation of ATR in response to DNA damage (8-10). DNA damage leading to phosphorylation of CHK1 at these sites stimulates CHK1 function. Surprisingly, a large increase in S345 CHK1 phosphorylation following treatment was correlated with poor responsiveness to the GEM/AZD7762 combination treatment (Figure 2B).
Relative anti-proliferative and apoptosis activity of gemcitabine/AZD7762 combination, TL32711 single, and gemcitabine/AZD7762/TL32711 combination treatments in triple-negative breast cancer cell lines. (A) Proliferation of the cells single-treated with 500 nM TL32711 (T) or co-treated with 10 nM gemcitabine and 100 nM AZD7762 (GA), or with 10 nM gemcitabine, 100 nM AZD7762, and 500 nM TL32711 (GAT) for 3 days was measured using the XTT assay. Cell viability of the treated cells was compared with the non-treated cells of each cell line. Values are the average of 3 determinations±SD. (B) Responses of IAPs and apoptosis marker proteins in MB231, HCC70, HCC1937, and HCC1395 cells following a single 500 nM TL32711 (T) or combination 10 nM gemcitabine/100 nM AZD7762 (GA), or 10 nM gemcitabine/100 nM AZD7762/500 nM TL32711 (GAT) treatment, respectively, for 24 h. C: Control, CL: cleaved length, FL: full length. GAPDH was used as loading control.
Previous studies have demonstrated that targeting IAP proteins (cIAP1 and cIAP2) with SMAC mimetics affords synergistic cytotoxicity with common chemotherapeutic agents including GEM (27-29). cIAP1 and cIAP2 are mediator proteins in TNF-receptor signaling and involved in TNFα-mediated NF-κB activation, resulting in blocking apoptosis (12). DNA damaging induced by a DNA damaging agent may activate TNF-receptor signaling by up-regulating TNFα to prevent cell death (30). It is well known that SMAC mimetics induce degradation of cIAPs and stimulate apoptosis (14). We explored whether the apoptotic response could be augmented using the SMAC mimetic TL32711 in TNBC cells that did not show a strong response to GEM/AZD7762 therapy. Although TL32711 uniformly suppressed cIAP1 and cIAP2 in these TNBC cells, some cell lines including HCC1937, HCC1395, and BT20 remained resistant to the GEM/AZD7762/TL32711 combination (Figure 3).
Response of apoptosis markers and cell-cycle proteins in triple-negative breast cancer cell lines following TL32711 only or gemcitabine/AZD7762/TL32711 combination treatment with the caspase-3 inhibitor or anti-TNFα antibody. (A) Detection of apoptosis marker proteins in MB231 and HCC70 cells treated with 500 nM TL32711 (T) or 10 nM gemcitabine/100 nM AZD7762/500 nM TL32711 combination (GAT) with and without 10 μM caspase-3 inhibitor z-VAD-FMK (zVAD) or 100 ng/ml anti-TNFα antibody (TNFab), respectively, for 24 h. HCC38 cells are a control. (B) Detection of CHK1 phosphorylation in MB231, HCC70, HCC1937, and HCC1395 cells treated 10 nM gemcitabine/100 nM AZD7762/500 nM TL32711 combination (GAT) with and without 10 μM caspase-3 inhibitor z-VAD-FMK (zVAD) or 100 ng/ml anti-TNFα antibody (TNFab). C: Control, CL: cleaved length, FL: full length. GAPDH was used as loading control.
There exist potential scenarios that might account for the lack of improved responses with TL32711 in HCC1937, HCC1395, and BT20 cell lines. If cells rely on non-canonical NF-kB signaling for survival, they may be less prone to stimulate apoptosis in the absence of cIAP1 and cIAP2 as has been demonstrated for multiple myeloma cells (31). Second, IAPs also block the formation of the ripoptosome complex to inhibit cell death pathways (32). It is possible that dysregulation of ripoptosome members or mutations of protein-protein interaction regions may still suppress the formation of the ripoptosome complex even with degradation of cIAP1 and 2, resulting in inhibition of apoptosis (33). Additional studies are needed to understand these processes.
Interestingly, cell lines that were sensitive to the GEM/AZD7762/TL32711 combination, demonstrated a reduction in S345 CHK1 phosphorylation in a TNFα-dependent manner compared to the increase in S345 CHK1 phosphorylation observed in response to only the GEM/AZD7762 combination (Figure 4B). In contrast, S345 CHK1 phosphorylation was unchanged in response to the addition of TL32711 to GEM/AZD7762 in the non-sensitive cell lines. These results are consistent with our demonstration that the cells that showed a good response to the GEM/AZD7762 combination treatment exhibited a lower S345 CHK1 phosphorylation level after treatment (Figure 2B).
In this study, we demonstrated that the responses to CHK1-based combination therapies were heterogeneous in p53-deficient TNBC cells underscoring the need for predictive biomarker identification. TNBC cells with a low expression level of CHK1 and cIAP1 and cIAP2 demonstrated good sensitivity to treatment with GEM alone. Additionally, cells demonstrating a minimal or no increase in the S345 CHK1 phosphorylation in response to GEM/AZD7762 or GEM/AZD7762/TL32711 combination treatment demonstrated a good apoptotic response. These results may have potential implications for the management of TNBC patients attending a clinical trial of CHK1 inhibitor-based therapies. Sub-classification of TNBC into phenotypic subsets based upon endogenous biomarker levels or initial biomarker responses to specific therapies may, thus, help predict chemotherapy efficacy (35).
Acknowledgements
The Authors are grateful to members of the Green lab for useful discussions and to Stan Lipkowitz and Elise Kohn for reviewing the manuscript. This work was supported by the intramural research program of the Center for Cancer Research, National Cancer Institute, NIH, Bethesda, MD.
Footnotes
This article is freely accessible online.
Conflicts of Interest
The Authors declare that no competing interest exists.
- Received April 4, 2016.
- Revision received April 14, 2016.
- Accepted April 28, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
