Targeting of Nuclear Factor κB Pathways by Dehydroxymethylepoxyquinomicin, a Novel Inhibitor of Breast Carcinomas: Antitumor and Antiangiogenic Potential In vivo

Over the past few years, extensive studies conducted into breast cancer have led to the recognition that overexpression of HER2, estrogen receptor (ER), and mutation of genes including p53, BRCA1, and BRCA2 play important roles in the development and progression of breast cancer via induction of multiple angiogenic, proapoptotic regulators including nuclear factor κB (NF-κB). NF-κB was initially discovered as a heterodimeric protein consisting of p65 and p50. The members of Rel and NF-κB family include p50 (p105), p65 (RelA), p52 (p100), RelB, and c-Rel. Members p105 and p100 are processed into p50 and p52, respectively. Mammary NF-κB mainly consists of RelA/p50. NF-κB is the transcription factor that promotes the transcription of inflammatory cytokines, cell adhesion molecules, and inhibitor of apoptosis-associated proteins (1–4). Ligand-stimulated tumor necrosis factor-α (TNF-α) receptor 1 activates NF-κB-inducing kinase through TNF-α receptor-associated death domain protein, receptor interacting protein, and TNF-α receptor–associated factor-2 (5). NF-κB-inducing kinase then activates IκB kinase (IKK), which is composed of catalytic subunits (IKKα and IKKβ) and a regulatory subunit (IKKγ/NEMO). IKK induces the phosphorylation and degradation of IκB-α, and consequently, the liberated NF-κB molecules enter the nucleus. For the TNF-α-induced activation of NF-κB, IKKβ/γ is essential but IKKα is not. In contrast, it has recently been suggested that IKKα is essential for an alternative NF-κB-inducing kinase-dependent pathway (6–8). The lymphotoxin-β (LT-β) receptor signaling pathway utilizes both IKKβ/γ and IKKα for the activation of NF-κB. The RelB/p100 complex, which is usually inactive in the cytoplasm, is activated by IKKα after the stimulation with LT-β (9). Phosphorylated p100 is processed to p52, and this RelB/p52 complex is then translocated to the nucleus (10). Thus, translocation of activated NF-κB is mainly regulated by two distinct pathways, one triggered by TNF-α receptor 1 and the other by the LT-β receptor pathways.

NF-κB is often constitutively activated in breast carcinomas (11, 12), bladder carcinomas (13), prostate carcinomas (14), and melanomas (15). A previous study showed that constitutive activation of NF-κB contributed to the progression of breast cancer to hormone-independent growth (12), the inhibition of NF-κB may enhance the therapeutic efficacy for malignant phenotypes such as those with less sensitivity to apoptosis or those that show hypersecretion of angiogenetic cytokines.

Therefore, low–molecular weight inhibitors of NF-κB should be useful as anticancer agents for breast carcinomas. In the course of our search for inhibitors of NF-κB function, we have designed new NF-κB inhibitors of low molecular weight by reference to the structure of the antibiotic epoxyquinomicin C (16). One of the designed compounds, dehydroxymethylepoxyquinomicin (DHMEQ; Fig. 1), inhibited TNF-α-induced activation of NF-κB, and was effective in suppressing rheumatoid arthritis in an in vivo model (16). Subsequently, DHMEQ was found to inhibit nuclear translocation of NF-κB in TNF-α-treated human T cell leukemia Jurkat and monkey kidney COS-1 cells (17). DHMEQ has been already shown to provide a potential for apoptosis induction in prostate cancer (18).

In the present study, we looked into the effect of DHMEQ on the tumorigenic activity of human breast carcinomas with respect to the cellular phenotypes related to malignancy and on tumor growth in vivo.

Materials. DHMEQ was synthesized in our laboratory as described before (19). Recombinant human TNF-α, recombinant human LT-α ₁/β₂, and Adriamycin were purchased from Sigma Chemical Company (St. Louis, MO). Rabbit polyclonal anti-IκB-α, mouse monoclonal anti-p65, mouse monoclonal anti-p50, and rabbit polyclonal anti-RelB antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).

Cell Culture. MDA-MB-231, MCF-7, and MCF-7/Adr cells were grown in DMEM supplemented with 10% fetal bovine serum, 200 μg/mL kanamycin, 100 units/mL penicillin G, 600 μg/mL l-glutamine, and 2.25 g/L NaHCO₃. For trypan blue exclusion analysis, cells were seeded into 24-well plates at 5 × 10⁵ cells/well and treated with various concentrations of DHMEQ for 24 hours. Then the number of cells stained with trypan blue dye was counted.

Electrophoretic Mobility Shift Assay. Nuclear extracts were prepared according to the method of Andrews and Faller (20). In brief, MCF-7 cells treated or not with DHMEQ were incubated for 2 hours at 37°C and then further incubated for 30 minutes in the absence or presence of TNF-α (20 ng/mL). The cells were harvested, washed with PBS, suspended in 400 μL of buffer A [10 mmol HEPES (pH 7.8), 10 mmol KCl, 2 mmol MgCl₂, 0.1 mmol EDTA, 1 mmol DTT, 0.1 mmol phenylmethylsulfonyl fluoride], and incubated on ice for 15 minutes. Nuclei were pelleted by centrifugation for 5 minutes at 14,000 rpm, resuspended in 40 μL of buffer C [50 mmol HEPES (pH 7.8), 50 mmol KCl, 300 mmol NaCl, 0.1 mmol EDTA, 1 mmol DTT, 0.1 mmol phenylmethylsulfonyl fluoride, 25% glycerol (v/v)], incubated on ice for 20 minutes, and centrifuged for 5 minutes at 14,000 rpm at 4°C. The supernatant was used as a nuclear extract. Binding reaction mixture containing 5 μg protein of nuclear extract, 2 μg poly dI-dC and ³²P-labeled probe were incubated for 20 minutes at room temperature. DNA-protein complexes were separated from free DNA on a 4% native polyacrylamide gel in 0.25 mmol Tris-borate EDTA buffer. The DNA probe used for NF-κB binding was the double-stranded oligonucleotide containing the κB site from the mouse κ light chain enhancer (5′-ATGTGAGGGGACTTTCCCAGGC-3′).

Western Blotting Analysis. Cells were lysed with lysis buffer [20 mmol Tris (pH 8.0), 150 mmol NaCl, 2 mmol EDTA, 100 mmol NaF, 400 μmol/L Na₃VO₄, 1% NP40, 1 μg/mL leupeptin, 1 μg/mL antipain, 1 mmol phenylmethylsulfonyl fluoride]. Each extract (100 μg protein) was fractionated on a polyacrylamide/SDS gel, and then transferred to a polyvinylidene difluoride membrane (Amersham). The membrane was incubated overnight at 4°C for blocking in TBS buffer [20 mmol Tris-HCl (pH 7.6), 137 mmol NaCl] containing 5% skim milk. After having been washed thrice with 0.1% Tween 20 in TBS, the membrane was incubated for 1 hour at room temperature with the anti-IκB or anti-RelB antibody in TBS buffer. After three more washes with the TBS-Tween buffer, the membrane was incubated for 1 hour at room temperature with anti-rabbit immunoglobulin sheep antibody linked to horseradish peroxidase (Amersham). Immunoreactive proteins were visualized by use of the ECL detection system (Amersham).

Detection of Cytokines. MDA-MB-231, MCF-7, and MCF-7/Adr cells (1 × 10⁵ per well) were seeded into 24-well plates. After 24 hours, the culture medium was removed, and cells were washed twice with PBS (−), after which 1 mL of serum-free medium was added to each well, along with various concentrations of DHMEQ. For nonstimulated MDA-MB-231 and MCF-7 cells, the cells were treated with DHMEQ for 6 hours. In the case of stimulation with TNF-α, the cells were pretreated with DHMEQ for 2 hours, and then TNF-α (20 ng/mL) was added and the incubation continued for 4 hours. For MCF-7/Adr cells, the duration of stimulation of TNF-α combined with DHMEQ was 3 hours. Collected culture media were centrifuged at 15,000 rpm for 2 minutes, and the supernatants were used as samples for an (ELISA). Concentrations of the interleukin (IL)-6 and IL-8 in the collected samples were determined by using commercially available ELISA kits (Techne, Minneapolis, USA) according to the manufacturer's instructions.

In vivo Antitumor Activity. All procedures involving animals and their care in this study were approved by the Animal Care Committee of Tokyo Metropolitan Komagome Hospital in accordance with institutional and Japanese government guidelines for animal experiments. Male BALB/c severe combined immunodeficiency (SCID) mice were obtained from Sankyo Laboratory Service, Co. (Tokyo, Japan). MDA-MB-231 (1 × 10⁶) or MCF-7 cells (5× 10⁶) were implanted s.c. in the flank of each nude mouse. Once the animals had developed palpable MDA-MB-231 tumors, they were randomly assigned into four groups. Then DHMEQ suspended in 0.5% chloromethyl cellulose was given i.p. for 8 weeks to three of the groups (12 mg/kg once a week, 4 mg/kg thrice a week, and 12 mg/kg thrice a week), and the vehicle only medium was injected into the fourth group (control). In each mouse inoculated with MCF-7 cells, an estrogen pellet was implanted s.c. 7 days before the cell inoculation. Once the animals had developed palpable MCF-7 tumors, they were randomly assigned into two groups. Then DHMEQ was given for 8 weeks to one group (4 mg/kg thrice a week), and the vehicle only medium was injected into the other group (control). Each experimental group consisted of five mice. Animals were carefully monitored, and tumor size as well as body weight was measured weekly. Tumor volume was calculated according to the formula a² × b × 0.5, where “a” and “b” are the smallest and largest diameters, respectively.

Immunostaining. Resected tumors from SCID mice were immediately frozen in liquid nitrogen and stored at −80°C. Frozen samples in OCT compound were sectioned with a cryostat at a 7-μm thickness and mounted on silane-coated glass slides. Tissue sections were incubated at room temperature for 30 minutes with anti-CD31 antibody (1:200 dilution; ref. 21) for the detection of tumor vasculature or with M30 (1:200 dilution) for the detection of apoptosis or mouse monoclonal antibody M30 (Roche, Basel, Switzerland) against caspase-cleaved neoepitope of cytokeratin 18 was used to recognize apoptotic cancer cells (22). Slides were processed by using a commercial Elite ABC kit (Vectastain, Vector Laboratories, Burlingame, CA). Diaminobenzidine was used as the final chromogen, and Meyer's hematoxylin was used as a counterstain.

Statistical Evaluation. The potential of DHMEQ for inhibition of in vivo tumor growth was analyzed by using Scheffe F test. Statistical tests were done with a StatView software package (Abacus Concepts, Inc.), and findings were considered significant when the P value was < 0.05.

Inhibition of Constitutively Activated NF-κB by DHMEQ in MDA-MB-231 Cells. NF-κB was constitutively activated in MDA-MB-231 in the EMSA assay, as shown in Fig. 2A. DHMEQ completely inhibited the activation at 10 μg/mL for about 8 hours. DHMEQ was previously shown to inhibit nuclear translocation of p65 in COS-1 cells (17). Therefore, as expected, DHMEQ did not change the phosphorylation or the amount of IκB in MDA-MB-231 cells (Fig. 2B).

Inhibition of TNF-α-Induced NF-κB Activation by DHMEQ in MCF-7 Cells. NF-κB was not activated in MCF-7 cells without stimulation. TNF-α was shown to activate NF-κB in MCF-7 in the EMSA assay, and DHMEQ inhibited the activation in a dose-dependent manner at 3 and 10 μg/mL (Fig. 3A). DHMEQ did not inhibit TNF-α-induced phosphorylation and degradation of IκB-α as shown in Fig. 3B. However, the synthesis of IκB-α at 60 minutes was strongly inhibited by DHMEQ treatment, indicating that liberated NF-κB could not induce transcription of the IκB-α gene due to failed nuclear translocation.

Inhibition of LT-α₁β₂-Induced NF-κB Activation by DHMEQ in MCF-7 Cells. We also determined whether DHMEQ inhibited the nuclear translocation of RelB/p52, which is a major form of NF-κB in the LT-β receptor pathway. LT-α₁β₂ (20 ng/mL) induced the activation of NF-κB in MCF-7 cells, as shown in Fig. 4A. This activation was apparently inhibited by DHMEQ at 10 μg/mL. DHMEQ inhibited the activation of NF-κB in the stimulated MCF-7 cells without influencing the processing of p100 (Fig. 4B); i.e., the production of p52 by the stimulation with LT-α₁β₂ was not influenced by DHMEQ. However, synthesis of p100 was inhibited at 8 hours after the addition. Fig. 4C clearly shows that DHMEQ inhibited the activation of NF-κB in stimulated MCF-7 cells, possibly by inhibiting the nuclear translocation of RelB up to 4 hours. Thus, DHMEQ inhibited the activation of NF-κB by inhibiting the nuclear translocation of both RelA/p50 and RelB/p52.

Inhibition of IL-6 and IL-8 Secretion by DHMEQ. MDA-MB-231 cells were found to secrete IL-6 constitutively, and the addition of TNF-α further accelerated the secretion (Fig. 5A and B). DHMEQ inhibited the constitutive IL-6 secretion of MDA-MB-231 in a dose-dependent manner (Fig. 5A). Even when cells were stimulated with TNF-α (20 ng/mL), 10 μg/mL of DHMEQ completely inhibited the IL-6 secretion (Fig. 5B). MCF-7 cells failed to produce detectable IL-6 protein even after stimulation (data not shown). MDA-MB-231 cells constitutively expressed a high level of IL-8 (Fig. 5C), whereas ER-positive MCF-7 cells, expressed a low level of IL-8 (Fig. 5E). DHMEQ inhibited the constitutive expression of IL-8 in MDA-MB-231 cells (Fig. 5C). TNF-α treatment enhanced the levels in both cell lines. DHMEQ inhibited these elevated IL-8 secretions in both kinds of cells (Fig. 5D and E).

Antitumor Effects of DHMEQ in Mice. We then tested the antitumor effects of DHMEQ on tumors arising from MDA-MB-231 and MCF-7 cells inoculated into SCID mice. I.p. administration of DHMEQ at 12 mg/kg, thrice a week, resulted in a significant decrease in the MDA-MB-231 tumor volume found for the controls (Fig. 6A). In the MCF-7 xenograft model, the mice receiving DHMEQ at 4 mg/kg, thrice a week, also showed a significant decrease in tumor growth when compared with the controls (Fig. 6B). DHMEQ treatment at the dosage used was well tolerated, leading to no gross or histopathological changes in liver, kidney, stomach, intestine, or lung attributable to toxicity, and to no body weight loss in the treated animals when compared with that of the controls (data not shown).

Induction of Apoptosis and Antiangiogenesis in Xenografted Tumors Caused by DHMEQ. Tumor vessels were identified by staining frozen sections of tumor tissue with antibody against CD31, a specific marker of angiogenesis. There was a large reduction in the number of CD31-positive vessels in the tumors of DHMEQ-treated mice as compared with that of tumors in the controls (Fig. 6C,, top). Apoptosis was then evaluated by staining with antibody M30, which reacts with caspase-cleaved neoepitope of cytokeratin 18. The expression of apoptosis in the DHMEQ treatment group was higher than that in the control group (Fig. 6C , bottom).

Inhibition of TNF-α-Induced NF-κB Activation by DHMEQ in MCF-7/Adr Cells. MCF-7/Adr cells highly expressing MDR1 were resistant to Adriamycin compared with the parental cells (Fig. 7A), but there was no constitutive activation of NF-κB in MCF-7/Adr cells (Fig. 7B). DHMEQ almost completely inhibited TNF-α-induced activation of NF-κB in these cells. The expression level of IL-8 from MCF-7/Adr cells was not different from that of parental cells (Fig. 7C). DHMEQ almost completely inhibited TNF-α-induced secretion of IL-8 from MCF-7/Adr cells.

DHMEQ inhibited NF-κB activation and secretion of inflammatory cytokines such as IL-6 and IL-8 in breast carcinoma cells. Continuous elevation of IL-6 levels was reported as a factor indicative of a poor prognosis in heavily pretreated patients with recurrent breast cancer (23). In MCF-7 cells, overexpression of human manganese superoxide dismutase, which may inhibit NF-κB, suppresses tumor growth both in vitro and in vivo by down-regulating genes responsible for tumor malignant phenotype including IL-6 (24). Elevated expression of IL-8 was reported to mediate invasiveness, angiogenesis, and metastasis of breast cancer cells in association with ER status and to influence prognosis (25, 26). ER-positive cells often expressed low levels of IL-8, whereas ER-negative ones often expressed high levels of IL-8 (27, 28). In our experiment, ER-negative MDA-MB-231 cells constitutively expressed a high level of IL-8 (Fig. 5C) compared with ER-positive MCF-7 cells (Fig. 5E). TNF-α treatment enhanced the levels in both cell lines. DHMEQ inhibited TNF-α-stimulated IL-8 secretion in both kinds of cells (Fig. 5D-E). Thus it is possible that the suppression of MDA-MB-231 and MCF-7 tumors in vivo by DHMEQ may be partly due to the inhibition of cytokine secretions.

NF-κB has also been shown to up-regulate the expression of several proangiogenic genes, directly or indirectly, including urokinase-type plasminogen activator, matrix metalloproteinase 9, and vascular endothelial growth factor. Therefore, the suppression of NF-κB could also be an antiangiogenic treatment for cancer. Copper deficiency, which inhibits NF-κB, contributes to a global inhibition of NF-κB-mediated transcription of proangiogenic factors such as vascular endothelial growth factor, fibroblast growth factor-2, IL-1, IL-6, and IL-8 in breast cancer (29). DHMEQ is likely to decrease these proangiogenic factors.

LT-β was discovered as a glycoprotein receptor that binds to LT-α/TNF-β at the cell-surface membrane (30). When LT forms a homotrimer, LT-α₃, it binds to TNF-α receptors 1 and 2; but when LT forms a LT-α₁β₂ heterotrimer, it binds to the LT-β receptor. Using MCF-7 cells, we found that DHMEQ inhibits both TNF-α receptor 1 and LT-β-mediated activations of NF-κB (Fig. 4A-C).

Chemoresistance toward Adriamycin, which is a key drug for the treatment of breast cancer, is a serious problem. The expression of P-glycoprotein encoded by the multidrug resistance (MDR1) gene is often elevated by clinical use of Adriamycin. MCF-7/Adr cells are the typical resistant cells expressing large amounts of MDR1. We found that DHMEQ was equally effective for the inhibition of NF-κB and cytokine secretion in MCF-7/Adr cells. Therefore, DHMEQ should not be a substrate of MDR1.

Thus, DHMEQ inhibited activation of NF-κB and cytokine secretion in breast carcinoma cells including ER-negative and ER-positive ones. It was also effective in MDR1-expressing breast carcinoma cells. DHMEQ had an excellent therapeutic potential in both antiangiogenesis and apoptosis induction in xenograft tumors of breast carcinomas. Previously, it was shown to inhibit animal models of prostate carcinoma (18), rheumatoid arthritis (16), and renal inflammation (31) without any toxicity. DHMEQ may be a candidate of chemotherapeutic agent for breast carcinomas targeting NF-κB.