Pancreatic Cancer Combination Therapy Using a BH3 Mimetic and a Synthetic Tetracycline

Pancreatic ductal adenocarcinoma (PDAC) is an extremely aggressive cancer that is predicted to cause almost 40,000 deaths in the United States this year (NCI 2014). PDAC is fairly resistant to most standard therapies and results in a 5-year survival rate of about 4% (1). These dire statistics, combined with the fact that there have been minimal new therapies developed for PDAC over the last decade, highlight the need for new approaches to effectively treat this invariably fatal disease.

The aggressive nature and dismal prognosis of patients with pancreatic cancer results partly from the plethora of molecular changes that occur during PDAC development, one of which is overexpression of the antiapoptotic proteins of the Bcl-2 family (2–5). Cancer cells exploit this overexpression to evade cell death and as a mechanism promoting resistance to diverse chemotherapeutic agents. Apoptosis reflects a balance between pro- and antiapoptotic proteins within cells. The ability of a cancer cell to shift the balance toward survival promotes resistance to toxic factors (6).

Consequently, the antiapoptotic Bcl-2 proteins have emerged as a novel therapeutic target. Although multiple strategies have attempted to target these molecules, BH3 mimetics have shown significant promise. Proapoptotic or antiapoptotic effects in cells arise, ultimately, as a consequence of physical interactions between anti- and proapoptotic Bcl-2 proteins (7, 8). On the basis of modeling predictions of these interactions, small molecules have been developed that mimic the BH3 domain of proapoptotic proteins and bind to the antiapoptotic proteins, thereby impeding their ability to inhibit apoptosis. These drugs represent a novel and exciting new strategy in cancer therapeutic development (6).

Sabutoclax (BI-97C1) is a novel Apogossypol derivative BH3 mimetic developed by Wei and colleagues (9–12). This compound binds to the Bcl-2 antiapoptotic proteins Bcl-2, Mcl-1, Bcl-X_L, and Bfl-1. It was originally identified on the basis of its ability to bind Bcl-X_L with low to submicromolar binding affinity (11). We have previously shown that sabutoclax shows efficacy against prostate and colorectal cancers, two cancers that also overexpress antiapoptotic Bcl-2 proteins (13, 14).

Minocycline is a synthetic tetracycline antibiotic that displays marginal activity against multiple cancers (15–19). However, less than stellar outcomes have diminished enthusiasm for using these drugs in cancer research. The marginal single-agent effects of minocycline against cancer may be due to the fact that it also impedes cell death in the face of toxicity or injury by inhibiting mitochondrial apoptosis and upregulation of Bcl-2 (20–22). In an attempt to develop a unique therapeutic strategy for PDAC, we hypothesized that sabutoclax and minocycline might show therapeutic efficacy against this disease when used in combination because of both the reliance of PDAC on the Bcl-2 proteins for survival as well as the theoretical ability of sabutoclax to counteract the antiapoptotic effects of minocycline, thereby uncovering the true therapeutic potential of this previously overlooked drug.

MIA PaCa-2, PANC-1, BxPC-3, AsPC-1, and HPNE cells were all obtained from the ATCC. LT2 cells were purchased from Millipore. MIA PaCa-2 and PANC-1 were maintained in DMEM plus 10% FBS. BxPC-3 and AsPC-1 cells were maintained in RPMI plus 10% FBS. HPNE and LT2 cells were maintained with media according to the distributor's instructions. Cell lines were expanded and cryopreserved at early passages and new vials were thawed out and used for experiments approximately every 3 months.

Cell lines were derived from the ascites of tumor-bearing KPC mice. At the time of necropsy, ascitic fluid was collected from the mice and centrifuged to pellet tumor cells. The pellet was repeatedly washed in PBS and centrifuged before being resuspended in RPMI supplemented with 4% FBS and placed in culture. These media were used to maintain these cell lines.

For all in vitro studies, sabutoclax (produced by Dr. Maurizio Pellecchia, Sanford-Burnham Medical Research Institute, La Jolla, CA) was diluted in DMSO and minocycline (Sigma) in PBS. For combination treatments, sabutoclax and minocycline were administered to cells simultaneously. zVAD-FMK (20 μmol/L; Promega) was incubated with cells for 1 hour before treatment with sabutoclax and minocycline. Caspase-8–specific inhibitor, z-IETD–FMK (20 μmol/L; BD Pharmingen) was also incubated with cells for 1 hour before treatment with sabutoclax and minocycline.

A total of 5 × 10³ cells were plated in 96-well plates and treated with sabutoclax and/or minocycline for 72 hours. Proliferation was assessed by MTT assay as previously described (23). All data were normalized to the control.

For Trypan blue, 5 × 10⁵ cells were plated in 6-cm dishes, treated as indicated for 48 hours, and then assayed as previously described (24).

A total of 5 × 10⁵ cells were plated in 6-cm dishes and treated as described. After 48 hours, media was collected from each dish and assayed according to the manufacturer's instructions [Cytotoxicity Detection Kit (LDH); Roche].

A total of 1 × 10⁶ cells were plated in 10-cm dishes and cultured in normal media with 0.5% serum for 48 hours. Cells for the zero hour time point were collected and fixed at this point. Remaining plates were kept in either normal media or 750 nmol/L sabutoclax for indicated time points. Once all cells were collected and fixed, they were incubated with propidium iodide and FACS was used for cell-cycle analysis. Cell-cycle studies were done as previously described (25).

MIA PaCa-2 cells were treated with sabutoclax (500 nmol/L), minocycline (50 μmol/L), or a combination of both for 24 hours. Cells were then trypsinized and 1,000 cells were plated into 6-cm plates in triplicate. Cells were allowed to grow and form colonies in normal media for approximately 14 days. Plates were then fixed and stained with Giemsa.

A total of 5 × 10⁵ cells were plated in 6-cm dishes and treated as described previously. After 48 hours, whole-cell lysates were prepared and Western blotting analysis was carried out as previously described (24). Primary antibodies used for these studies are PARP (1:1,000), Stat3 (1:1,000), pStat3 (1:750), Mcl-1 (1:1,000), survivin (1:750), p21 (1:750), p27 (1:1,000), cyclin D1(1:500), caspase-2 (1:1,000), caspase-3 (1:1,000), caspase-6 (1:1,000), caspase-7 (1:1,000), caspase-8 (1:1,000), caspase-10 (1:1,000), caspase-12 (1:1,000), AIF (1:1,000), pRB (1:750; Cell Signaling Technology), EF1-α (1:1,000; Millipore), and actin (1:5,000; Sigma). Densitometric analysis was done using ImageJ software (NIH).

MIA PaCa-2 or PANC-1 cells were transfected with a plasmid expressing a mutated form of Stat3 (pRc.CMV.Stat3Y705F; Addgene). Clones were selected with Neomycin for approximately 2 weeks and then picked and grown up individually. Whole-cell lysates were made and samples were used for Western blotting to characterize clones.

MIA PaCa-2 or PANC-1 cells were transfected with pGL3.CMV.luc (Promega). Transfections used Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol.

Tumors were fixed in formalin, embedded in paraffin, and sectioned for staining. Staining was done as previously described (26) with anti–p-Stat3 (1:100; Abcam), anti-PCNA (1:100; Abcam), and anti–Mcl-1 (1:100; Abcam) per the manufacturer's instructions.

Combination index (CI) values were determined for the combination of sabutoclax and minocycline in MIA PaCa-2 cells. Values were calculated using CompuSyn software (ComboSyn, Inc.) according to the Chou–Talalay method.

A total of 5 × 10⁶ MIA PaCa-2 and 3.5 × 10⁶ PANC-1 cells were used to establish bilateral subcutaneous tumors on the flanks of 8- to 10-week-old male athymic nude mice. Studies were done as previously described (13). Treatment began when tumors reached approximately 100 mm³. Sabutoclax was given at a dose of 1 mg/kg for both studies and was dissolved in a 10:10:80 solution of 100% ethanol:Cremophor:PBS. Minocycline was given at a dose of 20 mg/kg and dissolved in PBS. Both drugs were given via i.p. injection three times per week; n = 5 mice per group

In this model, we injected 1 × 10⁶ MIA PaCa-2-luc cells IP into 8- to 10-week-old male athymic nude mice. We allowed 1 week for tumor cells to attach and grow and then began treatment with PBS, 1 mg/kg sabutoclax, 20 mg/kg minocycline, or both sabutoclax and minocycline. Sabutoclax and minocycline were given via i.p. injection 3× per week. Mice were imaged by bioluminescence imaging (BLI) at 3 weeks after treatment was initiated and then sacrificed at 4 weeks. At time of sacrifice, mice were imaged prenecropsy. After necropsy, organs from a few mice per group were imaged to determine tumor specificity; n = 7 mice per group

A total of 5 × 10⁶ MIA PaCa-2-luc cells were injected i.p. into 8- to 10-week-old male athymic nude mice. We allowed 1 week for tumor cells to attach and grow and then began treatment with PBS, 1 mg/kg sabutoclax, 20 mg/kg minocycline, or both sabutoclax and minocycline. Sabutoclax and minocycline were given via i.p. injection 3× per week. Mice were sacrificed at 4 weeks. The pancreas from each mouse was removed and weighed during necropsy; n = 10 mice per group

KPC (Pdx-1-Cre/K-ras^LSL-G12D/p53^flox/wt) mouse cell line 48 was injected s.c. into both flanks of nontumorigenic KPC mice (Pdx-1-Cre negative/K-ras^LSL-G12D/p53^flox/wt). Tumors were given 1 week to establish before treatment was initiated. Mice were treated with PBS, 1.5 mg/kg sabutoclax, 10 mg/kg minocycline, or both sabutoclax and minocycline. Sabutoclax was dissolved in a 10:10:80 solution of 100% ethanol:Cremophor:PBS. Minocycline was dissolved in PBS. Mice were treated with sabutoclax, minocycline, or both drugs every 2 to 3 days via i.p. injection for a total of 6 injections. Tumors were measured with calipers to obtain tumor volumes; n = 5 mice per group

KPC mice (Pdx-1-Cre/K-ras^LSL-G12D/p53^flox/flox) were started on a sabutoclax and minocycline treatment regimen at 1 month of age. Mice were treated with a combination of sabutoclax (1.5 mg/kg) and minocycline (10 mg/kg) via i.p. injection three times per week. Mice were kept on this treatment until reaching a moribund status. At this point, mice were sacrificed and necropsied. Tumor sections were obtained from these mice and subjected to IHC; n = 12 mice (control group); n = 10 mice (sabutoclax + minocycline group).

The data presented are the mean ± SD of the values from three independent experiments. For in vivo studies, data shown are the mean ± 95% confidence interval. Significance was determined using the Student t test. A P value of <0.05 was considered statistically significant.

Multiple PDAC cell lines and two normal pancreatic cell lines were treated with increasing doses of sabutoclax and assessed for effects on proliferation. The normal cell lines displayed little change in cell growth (Fig. 1A and B), whereas all of the cancer cell lines were growth inhibited. Interestingly, the sensitivity of the cancer cell lines varied, with some being very sensitive to sabutoclax (MIA PaCa-2) and others more resistant (AsPC-1; Fig. 1A). The ability of sabutoclax to induce cell death and apoptosis was also evaluated (Fig. 1C and D). Again, MIA PaCa-2 was most sensitive to sabutoclax. However, despite not showing a potent cell death phenotype, cells like AsPC-1 still showed profound growth inhibition following sabutoclax treatment. Furthermore, sabutoclax displayed greater growth inhibitory effects than the BH3 mimetic ABT-737 (Supplementary Fig. S1).

Sabutoclax induces cell death in MIA PaCa-2 cells, without promoting potent cell death in either AsPC-1 or PANC-1. Despite this, there is still a dramatic reduction in cell growth after drug treatment. To interrogate the mechanism of growth suppression in these cell lines, we performed cell-cycle analysis (Fig. 2A; Supplementary Fig. S2A–S2C). Sabutoclax increased MIA PaCa-2 sub-G₁ phase after 24 and 48 hours (Fig. 2B), without affecting cell cycle, which is consistent with our previous data. In PANC-1 and AsPC-1, however, potent G₁–S phase cell-cycle arrest was evident. These experiments only evaluated cells up to 48 hours posttreatment. A second set of experiments also included 72 and 96 hours time points (Fig. 2C; Supplementary Fig. S2D and S2E) to determine whether longer drug incubations promoted cell-cycle arrest or resulted in a switch to apoptosis. In AsPC-1, the G₁–S phase arrest was sustained throughout all time points evaluated. However, at 72 and 96 hours, PANC-1 showed an increase in the sub-G₁ population of cells, indicating that there is an initial cell-cycle arrest in these cells that later switches to apoptosis (Fig. 2D). These results emphasize the complexity of responses observed in PDAC to a single agent, such as sabutoclax.

Evaluation of cell-cycle protein markers (Fig. 2E) confirmed decreased cyclin D1 and increased p27 expression in AsPC-1 and PANC-1 cells. In addition, there was a dramatic decrease in levels of phosphorylated Rb (Ser780). This corresponded with the observed G₁–S phase arrest. Interestingly, although levels of p21 in AsPC-1 cells increased, p21 levels decreased in PANC-1 cells. This difference may account for the switch from growth arrest to apoptosis in these cells (27,28).

It is now almost axiomatic that to successfully combat cancer multiple targeting strategies may be necessary. Cancer cells develop resistance to initially effective treatments and acquire avoidance mechanisms preventing toxicity. Accordingly, combinatorial approaches attacking multiple pathways in a cancer cell can increase drug efficacy, reduce toxicity, and increase the time to resistance. On the basis of this concept, we sought to identify an agent that would promote synergy when combined with sabutoclax. Minocycline, a commonly used antibiotic, can negatively affect cancer growth and survival (15–19). Despite this, it paradoxically also protects cells in the face of an insult through inhibition of mitochondrial apoptosis and upregulation of Bcl-2 (20–22). For these reasons, we hypothesized that these drugs might work well in combination, given their cancer-selective toxicities as single agents and the potential of sabutoclax to counteract the prosurvival effects of minocycline.

Minocycline is fairly nontoxic to several PDAC cells (MIA PaCa-2 and PANC-1), but is inhibitory in others (AsPC-1 and BxPC-3; Supplementary Fig. S3). In resistant cells, the combination of sabutoclax and minocycline is very toxic. It significantly reduces cell proliferation and overall cell number, and induces cell death to a much greater extent than either agent alone in MIA PaCa-2 (Fig. 3A–D) and PANC-1 cells (Supplementary Fig. S4). CI values also demonstrate synergy between these two compounds (Supplementary Fig. S5). It also promotes increased apoptosis and reduced colony formation in PDAC cells. Importantly, this synergistic effect is not evident in the normal cell line, HPNE (Fig. 3E–G).

We next sought to define the mechanism underlying this synergy. The combination of subtoxic doses of sabutoclax and minocycline induced potent cell death that was partially abrogated by pretreatment with zVAD, indicating that this toxic effect is, in part, caspase mediated. This rescue phenotype is evident in the morphology of treated cells (Fig. 4A; Supplementary Fig. S4C) and through reductions in LDH activity, indicating lower levels of cell death (Fig. 4B; Supplementary Fig. S4B). Furthermore, PARP cleavage was reduced in cells pretreated with zVAD (Fig. 4C; Supplementary Fig. S4A). When specific caspases were evaluated, the combination treatment showed caspase-3 cleavage and loss of full-length caspase-8, which was not found in zVAD-treated samples.

Stat3 activation is clinically relevant for PDAC as constitutive activation of Stat3 has been reported in 30% to 100% of human tumor specimens, and is crucial for PDAC initiation, progression, and maintenance (29). Sabutoclax treatment resulted in a potent loss of Stat3 phosphorylation (Tyr 705), as well as a loss of downstream Stat3 targets, such as survivin (Fig. 4D). Minocycline alone induced similar changes (Fig. 4D). Subtoxic doses of each drug lowered Stat3 activation, but the combination of both drugs almost completely eliminated pStat3 expression (Fig. 4E). To determine whether this would affect the cytotoxicity of the combination, we created MIA PaCa-2 clones stably expressing an activated Stat3 mutant, Stat3 Y705F, for example, C 13 (Fig. 4F). As compared with the parental cell line, C 13 showed enhanced resistance to combination treatment (Fig. 4G). A similar elevated resistance was observed in other Stat3 Y705F–overexpressing MIA PaCa-2 clones (data not shown). Because pretreatment with zVAD partially rescued cells from sabutoclax- and minocycline-induced cell death, we evaluated expression of a variety of caspases to determine their potential involvement not only in death induced by the combination, but which ones might be affected in C 13 cells (Fig. 4H). In parental MIA PaCa-2 cells, caspases-3, -7, and -8 were cleaved. Loss of full-length caspases-2 and -6 was evident, whereas AIF, caspase-12, and caspase-10 (caspase-10 data not shown) were not significantly altered. In C 13 cells, however, we did not see any significant changes in any caspase examined, supporting the observation that loss of Stat3 activation is necessary for caspase-dependent cell death by sabutoclax and minocycline. We also created stable cell lines overexpressing Stat3 Y705F in PANC-1 cells. Similar to what we observed in MIA PaCa-2 cells, constitutive Stat3 activation in these clones protected them from cytotoxicity induced by the combination of sabutoclax and minocycline (Supplementary Fig. S6B and S6C).

Caspase-8 activation is an integral part of the extrinsic pathway of apoptosis. However, it can also be activated independent of this pathway by other caspases (30). To determine whether caspase-8 involvement was due to extrinsic pathway activation or downstream intrinsic pathway activation, we used Z-IETD–FMK to specifically inhibit caspase-8 before drug treatment. Unlike the pan-caspase inhibitor, zVAD, pretreatment with Z-IETD–FMK did not protect cells from sabutoclax- and minocycline-induced cell death (Fig. 4I; Supplementary Fig. S6A). This indicates that caspase-8 activation is dispensable for sabutoclax- and minocycline-induced cell death and is most likely a secondary effect of other caspases. Therefore, our data support the hypothesis that sabutoclax and minocycline induce a cytotoxic phenotype through activation of the intrinsic pathway of apoptosis.

Luciferase-expressing MIA PaCa-2 cells, displaying a similar phenotype as parental cells (Fig. 5A), were established as subcutaneous xenografts in athymic nude mice and were treated with sabutoclax (1 mg/kg), minocycline (20 mg/kg), or both agents. Minocycline treatment alone showed minimal effects, whereas sabutoclax alone showed greater reductions in tumor weight and growth (Fig. 5B and C). Despite this, animals treated with a combination of sabutoclax and minocycline showed a synergistic reduction in tumor growth, with a significantly smaller tumor growth rate, even as compared with sabutoclax alone (Fig. 5C). This was confirmed with BLI (Fig. 5D). A similar experiment was conducted using PANC-1-luc cells, resulting in the same trend, with the combination group showing significant inhibition of tumor growth (Supplementary Fig. S7).

IHC of tumor sections (Fig. 5E) demonstrated higher intensity PCNA staining in control-, sabutoclax-, and minocycline-treated groups as compared with the combination-treated group. Furthermore, combination-treated tumors showed significantly less phosphorylated Stat3 expression, consistent with in vitro observations.

Although subcutaneous xenograft studies are useful in evaluating in vitro observations in an in vivo setting, flank tumors do not accurately mimic natural disease states. To study the effect of our drugs in a more natural setting, we used a quasi-orthotopic model of PDAC (31). We found that MIA PaCa-2-luc cells, when injected i.p., specifically homed to the pancreas, with the majority of tumors or tumor nodules found in this organ. Other locations in which we found tumors (liver, peritoneal lining) are common places for this cancer to metastasize and more closely mimic the clinical picture of this disease.

In an initial set of experiments, we injected 1 × 10⁶ cells i.p. and allowed 1 week for tumor establishment before treating mice with sabutoclax and minocycline, alone or in combination. Consistent with our previous data, minocycline as a single agent did not have any effect and mice presented similarly to control animals. Sabutoclax showed a potent single-agent effect, with fewer animals showing evidence of disease. The combination group showed even more promise (Fig. 6C), with only 1 of 7 mice showing tumors by BLI (Fig. 6A and B). Pancreas weight did not show significant differences, though the trend supported our other observations (data not shown).

A second quasi-orthotopic study was conducted using similar drug-dosing parameters. Tumors, though, were initiated with the injection of 5 × 10⁶ cells as opposed to 1 × 10⁶. In our first study, the use of fewer cells allowed testing whether sabutoclax and minocycline could prevent tumor formation. However, the use of fewer cells also means that the overall tumor burden was less when therapy was initiated, which prevented determining the true magnitude of the effect sabutoclax and minocycline would have in the context of a greater tumor burden, which might more closely imitate the clinical setting.

As anticipated, injecting a greater number of cells resulted in a larger overall tumor burden in the animals. Mice from all groups developed tumors or tumor nodules in the pancreas. However, the weight of the pancreas from mice treated with both sabutoclax and minocycline was significantly less than those from animals in all three other groups (Fig. 6D).

The KPC transgenic PDAC mouse model is used frequently in PDAC research (32–36). This animal spontaneously develops precursor pancreatic lesions, PanINs, which progress and eventually develop into invasive disease. Tumors form as a result of activated K-ras and functional loss of p53 in the pancreas, genetic changes also seen in a high percentage of human tumors. Histologic analysis shows extensive local invasion as well as metastasis in a subset of animals (37).

Pancreatic tumors from these mice overexpress Mcl-1 as shown by IHC (Supplementary Fig. S8A). This is consistent with the disease observed in humans (2–5) and provides a rationale for using sabutoclax in these animals. In addition, cell lines derived from these animals show combination effects with sabutoclax and minocycline (Supplementary Fig. S8B). These cells were injected s.c. into the flanks of control KPC mice (Pdx-1-Cre-negative/K-ras^LSL-G12D/p53^flox/wt) and allowed to grow for approximately 1 week. Mice were treated with sabutoclax, minocycline, or both drugs every 2 to 3 days via i.p. injection for a total of 6 injections. Sabutoclax or minocycline alone did not significantly affect tumor growth. The combination of drugs, however, significantly inhibited tumor growth as compared with controls (Fig. 6E). IHC of tumor sections demonstrated decreased staining for PCNA and pStat3 in combination-treated tumors (Fig. 6F and G).

In addition, a survival study was done using KPC mice (Pdx-1-Cre/K-ras^LSL-G12D/p53^flox/flox) to determine the effects of sabutoclax and minocycline. Mice were treated with sabutoclax (1 mg/kg), minocycline (10 mg/kg), or a combination of both three times a week via i.p. injection starting at around 1 month of age and continuing until animals reached a moribund state. Mice receiving combination treatment showed a significant survival advantage as compared with control mice (Fig. 6H). Tumors from these animals were isolated, sectioned, and stained for pStat3Y705. In correlation with our other studies, tumors from animals treated with sabutoclax and minocycline showed significantly less staining as compared to control tumors (Supplementary Fig. S9). Importantly, no toxicity was seen from either drug in any of the animals studies conducted.

Pancreatic cancer is one of the most lethal cancers remaining largely untreatable. Moreover, advances in therapy have been minimal over the last 15 years, due in part to the aggressive nature of PDAC and the difficulty in developing selective and effective therapeutics. We presently describe an efficacious novel drug combination for PDAC that uses a new BH3 mimetic and uncovers the hidden therapeutic potential of minocycline as an anticancer agent.

Considering overexpression of the antiapoptotic Bcl-2 proteins in PDAC, we initially evaluated the efficacy of sabutoclax, a BH3 mimetic that targets these antiapoptotic proteins inhibiting their function. Sabutoclax was effective as a single agent. It induced cancer-inhibitory effects in multiple genetically diverse PDAC cell lines. Potent apoptosis-inducing activity was evident in some PDAC cell lines and in those that only showed minor increases in death, we found instead a potent cell-cycle arrest. Tumor heterogeneity is a problem manifested to some degree in all cancers, but is especially relevant in PDAC, contributing to the global resistance seen in this cancer to conventional chemotherapeutics. A beneficial aspect of sabutoclax is its anticancer activity in multiple PDAC cell lines, irrespective of genetic background.

Despite the potential of sabutoclax as a single-agent, there is a pressing need for combination therapy in the clinical setting. Cancer is an adaptive disease and effectively combatting it requires a multifaceted approach. Because of this, we sought to find a second drug that would potentially synergize with sabutoclax to further promote its clinical applicability. We focused on the antibiotic, minocycline, which has a small literature base supporting a novel role for this drug in the field of cancer (15–19). Despite these reports, translational cancer research using tetracyclines never expanded as only minor efficacy was evident with these drugs. In addition to its classical role as an antibiotic, minocycline has also been studied extensively as a neuroprotective agent. Zhu and colleagues (22) showed that this neuroprotection develops, in part, through the inhibition of caspase-9 and caspase-3 activation after exposure to death-inducing stimuli due to the inhibition of cytochrome c release from mitochondria. Minocycline has also been shown to lead to increases in the expression of antiapoptotic Bcl-2 proteins like Bcl-2 (20, 21). We hypothesized that these properties might actually mask the true potential of minocycline as a cancer therapeutic and contribute to the disappointing results seen in the past. However, we hypothesized that these properties might make it an ideal combination partner for sabutoclax. Sabutoclax works by directly acting at the level of the mitochondria specifically inducing apoptosis driven by cytochrome c release. In combination with sabutoclax, the block on apoptosis in cells treated with minocycline may be released, thereby leading to synergistic anticancer properties.

Our study has shown that sabutoclax plus minocycline is exceptionally effective against pancreatic cancer (Fig. 6I). Subtoxic individual doses of each drug produced a dramatic synergistic cytotoxic effect. Furthermore, we found that this combination resulted in a loss of Stat3 activation, which was not seen in response to sabutoclax treatment in nontumorigenic LT2 cells (data not shown). This may help to explain the absence of sensitivity seen in these cells. Reintroduction of a constitutively active Stat3 mutant into cancer cells rendered them resistant to the combination of sabutoclax and minocycline. This combination shows promise in vivo using multiple mouse models, both immune-compromised and immune-competent transgenic models of PDAC, showing significant anticancer activity without any signs of gross toxicities.

In summary, we describe an innovative combinatorial therapeutic approach with remarkable activity against pancreatic cancer cells in vitro and in vivo in three animal PDAC models. Considering the paucity of effective therapies for PDAC, it is clear that new approaches are mandatory to impact clinically on this invariably fatal cancer. The ability to target the Bcl-2 family for inactivation, using sabutoclax, and combining this with a simple synthetic tetracycline antibiotic, such as minocycline, opens up new areas of research with the potential to lead to an effective therapy for pancreatic cancer.

Maurizio Pellecchia has ownership interest (including patents) in AncoreX Therapeutics. No potential conflicts of interest were disclosed by the other authors.

Conception and design: B.A. Quinn, B. Azab, P.B. Fisher

Development of methodology: R. Dash, B. Azab, P. Bhoopathi, J. Wei

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B.A. Quinn, S. Sarkar

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B.A. Quinn, S. Sarkar, B. Azab, P.B. Fisher

Writing, review, and/or revision of the manuscript: B.A. Quinn, S.K. Das, L. Emdad, D. Sarkar, P.B. Fisher

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Sarkar, S.K. Das, M. Pellecchia

Study supervision: L. Emdad, D. Sarkar, P.B. Fisher

The present study was supported in part by National Cancer Institute grants R01 CA127641 (P.B. Fisher) and R01 CA168517 (M. Pellecchia and P.B. Fisher) and a VCU Massey Cancer Center (MCC) development award. D. Sarkar is a Harrison Scholar in the MCC and a Blick Scholar in the VCU School of Medicine. P.B. Fisher holds the Thelma Newmeyer Corman Chair in Cancer Research in the MCC.

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.