Oral Administration of a Gemini Vitamin D Analog, a Synthetic Triterpenoid and the Combination Prevents Mammary Tumorigenesis Driven by ErbB2 Overexpression

HER2 (also known as ErbB2) is a member of the ErbB family of transmembrane receptor tyrosine kinases, including EGF receptor (EGFR, also known as ErbB1 or HER1), ErbB3 (HER3) and ErbB4 (HER4; ref. 1). ErbB2, which presents in an activated conformation without known ligands, functions as a signal amplifier of other ErbB family receptors by forming heterodimers and regulates cell proliferation, survival, and apoptosis (2). In breast cancer, the amplification of ErbB2 is one of the major genomic abnormalities and has been reported in approximately 20% of invasive breast cancer cases (3). The ErbB2-targeted therapies, such as trastuzumab and lapatinib, have substantially improved clinical outcomes of breast cancer patients with ErbB2-overexpressing tumors (1). The overexpression of ErbB2 has also been reported in premalignant lesions of the breast (4). Moreover, recent studies implicated ErbB2 in the early development of breast cancer (5, 6), suggesting ErbB2 is a molecular target for prevention as well as treatment of ErbB2-positive breast cancer.

1α,25(OH)₂D₃, a hormonally active form of vitamin D, modulates proliferation, survival, and differentiation of various types of cells, in addition to its role as a regulator of calcium and phosphorus homeostasis (7, 8). 1α,25(OH)₂D₃ has been shown to inhibit the growth of various types of breast cancer cells, including the ErbB2-positive subtype (9). Moreover, knockout of the vitamin D receptor (VDR) facilitated the development of mammary tumors in MMTV-ErbB2/neu mice, indicating the inhibitory role of 1α,25(OH)₂D₃ on ErbB2-overexpressing mammary tumorigenesis (10). Previously, we reported that novel Gemini vitamin D analogs, which were synthesized to increase anticancer efficacy and to decrease hypercalcemic toxicity of 1α,25(OH)₂D₃, repressed the tumor growth of both estrogen receptor (ER)–positive and -negative subtypes of breast cancer in vivo (11). In addition, the Gemini vitamin D analog BXL0124 inhibited the growth of ErbB2-overexpressing mammary tumors in MMTV-ErbB2/neu transgenic mice, which spontaneously develop mammary tumors driven by ErbB2 overexpression (12).

Oleanolic acid, is a common natural triterpenoid that shows weak anticancer and anti-inflammatory activities in vivo (13). Numerous synthetic oleanolic acid derivatives were produced to improve the activity of the parent substance. One of the most potent synthetic derivatives, 2-cyano-3,12-dioxooleana-1,9(11)-dien-28-oic acid (CDDO) was further modified to produce further derivatives, including CDDO-Me (methyl ester), CDDO-Im (imidazolide), CDDO-MA (methyl amide), and CDDO-EA (ethyl amide; ref. 14). CDDO inhibited proliferation and induced apoptosis of ErbB2-overexpressing breast cancer cells (15, 16). Kinase activity and phosphorylation of ErbB2 were repressed by CDDO in the ErbB2-overexpressing breast cancer cells (15). In animal studies, CDDO-Me markedly delayed mammary tumor development in ER-negative ErbB2-overexpressing and BRCA1-deficent mouse models (17, 18). Although CDDO-Im also showed potent anticancer activity against ER-negative breast cancer cells (19), its activity on ErbB2-overexpressing breast cancer has not been investigated.

Homo- or heterodimerization of ErbB2 with other ErbB family receptors allows ErbB2 to activate wide ranges of downstream signaling pathways, including Janus-activated kinase 2 (JAK2), STAT3, mitogen-activated protein kinase (MAPK), and PI3K/Akt (3). The complexity of the ErbB2 signaling pathway emphasizes the importance of using multiple therapies that target different components of the ErbB2 signaling pathway (3). In studies with MMTV-ErbB2/neu mice, the Gemini vitamin D analog BXL0124 was shown to target the ErbB2/Akt/Erk signaling pathway (12), whereas derivatives of CDDO were reported to inhibit activation of STAT3 (17). Moreover, CDDO was shown to directly interact with ErbB2 (18), which might contribute to the regulation of ErbB2 by CDDO. Previous studies with leukemia cells showed that combined treatments of 1α,25(OH)₂D₃ or its analog with CDDO-Im had synergistic effects on the induction of monocytic differentiation (20, 21). In the present study, we investigated the inhibitory effects of the Gemini vitamin D analog BXL0124, the synthetic triterpenoid derivative CDDO-Im and the combination on both prevention and treatment of mammary tumorigenesis of MMTV-ErbB2/neu transgenic mice.

Gemini vitamin D analog BXL0124 [1α,25-dihydroxy-20R-21(3-hydroxy-3-deuteromethyl-4,4,4-trideuterobutyl)-23-yne-26,27-hexafluro-cholecalciferol, (22)] was provided by BioXell, Inc. and dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich). CDDO-Im (23, 24) was provided by Dr. Michael B. Sporn (Dartmouth Medical School, Hanover, NH) and dissolved in DMSO. Dilutions of BXL0124 and CDDO-Im in sesame oil (Sigma-Aldrich) were used for oral administration in vivo.

Female MMTV-ErbB2/neu transgenic mice (6–7 weeks old) were purchased from The Jackson Laboratory. MMTV-ErbB2/neu mice were given 0.3 μg/kg body weight of BXL0124 for 56 weeks, which is the dose and duration of treatment used previously to show antitumor activity without toxicity (12). Short-term or single treatment of CDDO-Im as high as 100 μg/kg body weight has been shown to be well tolerated in mice (20, 25). For long-term treatment in the present study, MMTV-ErbB2/neu mice were administrated 3 μmol/kg body weight of CDDO-Im (≈1.6 μg/kg body weight). The dose selection was based on previously reported studies (20, 25). In the prevention study protocol, 15-week-old mice received by oral administration the vehicle control [DMSO: sesame oil = 3: 97 (v:v) in 0.1 mL volume, n = 19], BXL0124 (0.3 μg/kg body weight, n = 19), CDDO-Im (3 μmol/kg body weight, n = 20), or the combination of BXL0124 (0.3 μg/kg body weight) and CDDO-Im (3 μmol/kg body weight; n = 19) three times a week. The body weight and tumor size of each animal were measured weekly. The mice with large (tumor sizes ranging from 0.85–1.18 cm³) tumors were sacrificed earlier. All remaining mice were sacrificed when they were 56 weeks old.

In the treatment study protocol, mice were kept without treatment until they developed mammary tumors between 23 and 30 weeks of age. Following the development of mammary tumors, the sizes of tumors were measured, and mice were paired and separated into control or treatment groups with similar average tumor sizes with a final animal number of 11 or 13, respectively. Then, the mice were treated orally with the vehicle control or combination of BXL0124 (0.3 μg/kg body weight) and CDDO-Im (3 μmol/kg body weight) six times a week for 3 weeks.

The tumor size of each animal was also measured three times a week. Tumors were harvested and weighed at necropsy. The tumors were fixed in 10% formalin and transferred to 70% ethanol for immunofluorescent staining or flash-frozen and stored in −80°C for Western blot analysis or RNA analysis. All animal studies were carried out in accordance with the protocol approved by the Institutional Animal Care and Use Committee at Rutgers, the State University of New Jersey (Piscataway, NJ; Protocol Number: 04–001).

The detailed procedure of the determination of calcium concentration in serum samples (POINTE Scientific, INC.) was described previously (11).

The tumors were embedded in paraffin (Electron Microscopy Sciences) and then sectioned at 4 μm. The tumor samples were incubated in PBS containing 10% donkey serum to block nonspecific binding and then incubated overnight at 4°C with a combination of primary antibodies against pErbB2 (1:100; Cell Signaling Technology) and ErbB2 (1:40; Cell Signaling Technology), diluted in SignalStain antibody diluent (Cell Signaling Technology). Subsequently, the tumor samples were incubated with fluorophore-conjugated secondary antibodies (Alexa Fluor 488 or 546; 1:200; Life Technologies) and TO-PRO-3 iodide nuclear antibody (1 μmol/L; Life Technologies) at room temperature for 60 and 15 minutes, respectively. The images were taken using confocal microscope (Eclipse Ti; Nikon Instrument Inc.) with laser at 488 nm (pErbB2), 546 nm (ErbB2), and 633 nm (TO-PRO-3).

The detailed procedure was described previously (26). The primary antibodies recognizing ErbB1, pErbB1, pErbB2, pErbB3, Erk1/2, pErk1/2, Akt, pAkt, JAK2, pJAK2, pSrc (Tyr416), pSrc (Tyr527), STAT3, pSTAT3, c-Myc, and RXRα were from Cell Signaling Technology; ErbB3, p21, CycD1, Bcl2, and PPARγ were from Santa Cruz Biotechnology; ErbB2 and VDR were from Thermo Scientific Pierce; proliferating cell nuclear antigen (PCNA) was from EMD Millipore, β-actin was from Sigma-Aldrich. Secondary antibodies were from Santa Cruz Biotechnology. Densitometric quantification of Western blot analysis data was conducted using ImageJ program (NIH, Bethesda, MD), and each protein level was normalized to the protein level of corresponding β-actin.

The procedure was described previously (27); the labeled primers for all genes in the experiments were obtained from Life Technologies.

Log-rank test was used to test the difference in tumor-free distributions, and survival distributions between treatment groups. Chi-square test was used to evaluate the differences in proportions of mice sacrificed at the end of experiment (week 56) instead of during the experiment. The statistical analyses of tumor weights, body weights, and serum calcium levels were conducted using one-way ANOVA with Dunnett's adjustment. Data with two samples were analyzed with the Student t test. All the statistical calculations were conducted using SAS (Version 9.2) and GraphPad Prism 4.0 (GraphPad Software Inc.). The data are presented as the mean ± SE. P values less than 0.05 were considered significant.

The effects of BXL0124, CDDO-Im, and the combination on the development of ErbB2-overexpressing mammary tumors were investigated by using the MMTV-ErbB2/neu transgenic mouse model. MMTV-ErbB2/neu mice were treated with vehicle control, BXL0124, CDDO-Im, or the combination, beginning at 15 weeks of age. The development of mammary tumors was delayed by the treatments with BXL0124, CDDO-Im, and the combination (Fig. 1A). The differences in tumor-free distributions were significant (P = 0.0294) between the combination and the control groups, borderline significant (P = 0.0731) between the CDDO-Im and the control group, and not significant (P = 0.1542) between the BXL0124 and the control group (Fig. 1A). The percentage of tumor-free mice at the end of the experiment at week 56 was significantly higher with the combination group (52.6; P = 0.0436), but not with BXL0124 (36.8; P = 0.2832) and CDDO-Im (45.0; P = 0.1128) groups, compared with the control group (21.1; Fig. 1A). The survival distribution of mice sacrificed before the end of the experiment due to large tumors was significantly different between the CDDO-Im group and the control group (P = 0.0437) and the combination and the control groups (P = 0.0449), whereas the difference of survival distributions between the BXL0124 and the control group (P = 0.0789) was borderline significant (Fig. 1B). The estimated median time until mice were sacrificed because of large tumors (>0.85 cm³) was 48 weeks for the control group, whereas the median times for all other groups were over 56 week (Fig. 1B). Although BXL0124 (0.63 ± 0.14 g), CDDO-Im (0.44 ± 0.11 g), and the combination (0.43 ± 0.12 g) showed reduced average tumor burden as compared with controls (0.89 ± 0.14 g) at necropsy, only the reductions of tumor burden by CDDO-Im and the combination were statistically significant (P < 0.05; Fig. 1C). Average body weight and serum calcium level were not affected by BXL0124, CDDO-Im, and the combination, indicating all the drug treatments were not toxic to animals at the given concentration for the given duration (Fig. 1D and E).

Because activation of ErbB2 is one of the key factors that drive tumorigenesis of ErbB2-overexpressing breast cancer, the effects of BXL0124, CDDO-Im, and the combination on the activation of ErbB2 (determined by the phosphorylated form of ErbB2) were investigated in MMTV-ErbB2/neu mammary tumors. As shown in the pictures at ×40 magnification, the activation of ErbB2 (pErbB2, shown as green) was prominent at the leading edge of MMTV-ErbB2/neu mammary tumor where the tumor interacted with surrounding stromal tissues (Fig. 2A). The edge areas of tumors were magnified (×400), and the pictures of pErbB2 (green), total ErbB2 (red), nuclei (blue), and merged images of the three were shown (Fig. 2B). BXL0124, CDDO-Im, and the combination markedly decreased protein levels of activated ErbB2 (pErbB2), whereas the protein level of total ErbB2 was not affected by the treatments (Fig. 2B). The merged images of pErbB2, ErbB2, and nuclei for the control and combination treatment groups were enlarged for better presentation (Fig. 2C).

ErbB2 heterodimerizes with ErbB1 or ErbB3 and transactivation has been shown to play a critical role in tumorigenesis of ErbB2-overexpressing breast cancer (28). Therefore, we investigated the effects of BXL0124, CDDO-Im, and the combination on the activation of ErbB2 as well as ErbB1 and ErbB3 in the mammary tumors of MMTV-ErbB2/neu mice. Six tumors were randomly selected after excluding large (>0.9 cm³) or small (<0.2 cm³) tumors, and levels of given proteins were analyzed as individual samples (Fig. 3, left). The same amount of the six individual samples (3 μg/sample) were combined for each group and analyzed as pooled samples (Fig. 3, right). The protein level of activated ErbB2 was markedly decreased by BXL0124, CDDO-Im, and the combination, whereas the protein level of total ErbB2 was not changed by any of the treatments (Fig. 3). BXL0124, CDDO-Im, and the combination repressed the protein level of activated ErbB1 (pErbB1), and the protein level of total ErbB1 was also reduced by CDDO-Im and the combination (Fig. 3). The protein level of activated ErbB3 (pErbB3) was also repressed by all three treatments, whereas the protein level of total ErbB3 remained unchanged by any of the treatments (Fig. 3).

The effects of BXL0124, CDDO-Im, and the combination on the mRNA level of ErbB family receptors in MMTV-ErbB2/neu mammary tumors were investigated (Fig. 4). The fold change of average mRNA level for each group relative to control group was shown (Fig. 4). BXL0124, CDDO-Im, and the combination did not have significant effects on the mRNA level of ErbB2 (Fig. 4A). The mRNA levels of ErbB1 and ErbB3 were not significantly changed by BXL0124 or CDDO-Im, whereas the combination significantly decreased the mRNA levels of ErbB1 (P < 0.01) and ErbB3 (P < 0.05; Fig. 4A). The mRNA levels of EGF and Heparin Binding-Epidermal Growth Factor (HB-EGF), key ligands of ErbB1, as well as neuregulin 1 (Nrg1) and neuregulin 2 (Nrg2), key ligands of ErbB3, were measured to investigate the effect of BXL0124, CDDO-Im, and the combination on ErbB ligands. BXL0124 significantly decreased the mRNA level of EGF in ErbB2-overexpressing mammary tumors (P < 0.01; Fig. 4B). The mRNA level of HB-EGF was significantly decreased by CDDO-Im (P < 0.05) and the combination (P < 0.05; Fig. 4B). The mRNA level of Nrg1 was low [cycle number of quantitative PCR (qPCR) > 30] in all ErbB2-overexpressing mammary tumors (data not shown), and the mRNA level of Nrg2 did not show significant changes by any of the treatments (Fig. 4B).

We investigated the effects of BXL0124, CDDO-Im, and the combination on the downstream signaling pathways of ErbB receptors in MMTV-ErbB2/neu mammary tumors (Fig. 5). The protein levels of activated extracellular signal–regulated kinase (Erk)1/2 (pErk1/2) and Akt (pAkt), two key downstream molecules of ErbB2 signaling, were markedly inhibited by BXL0124, CDDO-Im, and the combination, whereas the protein level of total Erk1/2 and Akt were not affected (Fig. 5A). BXL0124, CDDO-Im, and the combination also markedly decreased the activation of JAK2 (determined by phosphorylated-JAK2, pJAK2) and Src [determined by phosphorylated-Src at Y416 position, pSrc (Y416)], intracellular kinases that regulate downstream targets of ErbB signal (Fig. 5A). The protein level of total JAK2 and the inactive form of Src [pSrc (Y527)] did not show changes in any of the treatment groups (Fig. 5A). Activation of STAT3 was inhibited by CDDO-Im and the combination, whereas BXL0124 moderately increased the activation of STAT3 (determined by phosphorylated-STAT3, pSTAT3; Fig. 5A).

The effects of BXL0124, CDDO-Im, and the combination on protein levels of downstream targets of the ErbB2 signaling pathway were further investigated. c-Myc, a transcriptional regulator of multiple genes involving cell proliferation, was markedly decreased by BXL0124, CDDO-Im, and the combination (Fig. 5B). BXL0124, CDDO-Im, and the combination increased p21, whereas the three treatments decreased cyclin D1 (Fig. 5B). PCNA was repressed by BXL0124 and the combination (Fig. 5B). Bcl2, a key regulator of cell survival, was also markedly decreased by BXL0124, CDDO-Im, and the combination (Fig. 5B). The effects of BXL0124, CDDO-Im, and the combination on their receptors, VDR for BXL0124 and PPARγ for CDDO-Im were investigated. There were no effects on VDR, but PPARγ was increased by the combination (Fig. 5B). VDR forms heterodimers with retinoid X receptor α (RXRα) to regulate the expression of VDR-targeted genes (29). PPARγ also forms heterodimers with RXR, and the RXR agonists, such as bexarotene, potentiate the biologic function of PPAR ligands, including CDDO or CDDO-Im (20, 30). Therefore, we determined the protein level of RXRα, which was not affected by BXL0124, CDDO-Im, or the combination (Fig. 5B). The proposed ErbB signaling pathways regulated by BXL0124, CDDO-Im, and the combination is shown in Fig. 5C.

In the prevention study, MMTV-ErbB2/neu mice were treated before the development of mammary tumors, and long-term treatment with the combination of BXL0124 and CDDO-Im showed the most potent chemopreventive activity. In the treatment study, MMTV-ErbB2/neu mice were kept without treatment until they developed mammary tumors. Following the development of mammary tumors, the tumor size was measured, and mice were assigned to control (n = 11) or treatment group (n = 13) with a similar average tumor size (0.07 cm³ for control and 0.08 cm³ for treatment) in each group. Mice were then treated with control or the combination of BXL0124 (0.3 μg/kg BW) and CDDO-Im (3 μmole/kg BW) for 3 weeks. The 3-week treatment with the combination did not affect the growth rate of MMTV-ErbB2 mammary tumors (Fig. 6A). There were no distinctive differences between the control and treatment groups for the average tumor weight (Fig. 6B), body weight (Fig. 6C), and serum calcium level (Fig. 6D). Five tumor samples from each group were combined and analyzed to study the effects of short-term treatment with the combination on the ErbB2 signaling pathway and its downstream targets. Protein levels of total and activated ErbB2, Erk1/2, and Akt were not affected by the combination (Fig. 6E). The protein levels of downstream target molecules, c-Myc, p21, CycD1, and Bcl2 were also not changed by the combination treatment (Fig. 6F).

The present study showed that the Gemini vitamin D analog BXL0124, the synthetic triterpenoid CDDO-Im and the combination delayed the development of mammary tumors in MMTV-ErbB2/neu mice (Fig. 1). Although both BXL0124 and CDDO-Im markedly inhibited the activation of ErbB2, BXL0124 and CDDO-Im showed different activities on the suppression of downstream signaling pathways of ErbB2 (Figs. 3 and 5). The combination was more effective than the individual compounds for inhibiting the activation of ErbB2 (Fig. 3). Moreover, the combination had additive effects on the suppression of the multiple downstream signaling pathways of ErbB2 (Fig. 5), emphasizing the advantage of the combination on the inhibition of the ErbB2 signaling pathway.

The study by Lee and colleagues showed that BXL0124 significantly decreased the mammary tumor burden in MMTV-ErbB2/neu mice (12). Previously, MMTV-ErbB2/neu mice were administered BXL0124 by intraperitoneal injection (12), whereas MMTV-ErbB/neu mice were given BXL0124 by oral gavage in the present study. BXL0124 decreased mammary tumor burden by 30% (Fig. 1), although the effect was lower than previously reported. The different routes of BXL0124 administration could influence the bioavailability of BXL0124 in animals, which might be the reason for the lower efficacy of BXL0124 shown in the present study. The recent study by Liby and colleagues showed that CDDO-Me, equivalently potent as CDDO-Im with better stability, prevented mammary tumorigenesis in MMTV-ErbB2/neu mice, and its inhibitory effect was synergistic in combination with the rexinoid LG100268 (LG268; ref. 17). Studies with breast cancer cell lines showed that CDDO-Me and LG268 had different molecular targets and biologic functions (17). In the present study, the combination of BXL0124 and CDDO-Im significantly delayed development of MMTV-ErbB2/neu mammary tumors, although its inhibitory effect was not significantly better than the effect of single treatment with CDDO-Im (Fig. 1). Interestingly, BXL0124 suppressed the p-ERK1/2/c-Myc/PCNA proliferation pathway, whereas CDDO-Im inhibited the pSrc/pSTAT3/Bcl2 survival pathway (Fig. 5C). In addition, the doses of BXL0124 and CDDO-Im used in the present study were much lower than their maximum-tolerated doses, showing no signs of toxicity or weight loss during the in vivo experiment. Synthetic triterpenoids, including CDDO-Im, act on different molecular targets depending on the concentration (14). Therefore, further investigation with higher doses of CDDO-Im than the dose used in the present study might be necessary to determine synergistic effects in combination with other drugs. CDDO-Me also significantly arrested the growth of established mammary tumors in MMTV-ErbB2/neu mice in the previous report (17), which was not achieved with CDDO-Im in the present study. Liby and colleagues used higher dose of CDDO-Me in the treatment protocol (100 mg/kg diet ≈ 25 mg/kg body weight) than the dose used in the prevention protocol (60 mg/kg diet ≈ 15 mg/kg body weight; ref. 17). In the present study, the dose of CDDO-Im used in the treatment and prevention protocols was same (3 μmol/kg body weight ≈ 1.6 μg/kg body weight). In addition, MMTV-ErbB2/neu mice were fed with diet containing CDDO-Me in the study by Liby and colleagues (17), whereas MMTV-ErbB2/neu mice were given CDDO-Im by oral gavage in the present study. Therefore, it might be difficult to directly compare the anticancer activities of CDDO-Me and CDDO-Im based on the results of these two studies. Further evaluation with same dose of CDDO-Me and CDDO-Im by the same route of administration might be necessary to make conclusive comparison.

Pharmacodynamic study by Yates and colleagues showed that CDDO-Im could be readily absorbed through the digestive system of mice after oral administration and was pharmacologically active in inducing Nrf2-regulated genes in many organs, including liver, lung, kidney, and intestine (25). In the present study, we also showed that CDDO-Im could be orally delivered and was biologically active to prevent ErbB2 overexpression-driven mammary tumorigenesis (Fig. 1). Although CDDO-Im is one of the most potent synthetic triterpenoids in various bioassays in vitro and in vivo, it was found to be unstable in human plasma (20, 31, 32). Moreover, the solid form of CDDO-Im was decomposed within 1 year when stored at 4°C (33). 1-[2-cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]-4-ethynylimiadazole, one of the recently developed N-acylimidazole analogs of CDDO, was shown to be nearly as potent as CDDO-Im, whereas its solid form was more stable than that of CDDO-Im (33). Further preclinical evaluation of the stable N-acylimidazole analog of CDDO is presently ongoing for potential clinical development as a chemopreventive or chemotherapeutic agent (33).

The transactivation of ErbB2 via heterodimerization with ErbB1 is one of the key mechanisms to potentiate the activation of its downstream signaling such as MAPK and PI3K/Akt (2). A study using the normal mammary epithelial cell line, MCF10A, showed that overexpression of wild-type ErbB2 was insufficient to induce invasive transformation of MCF10A cells in three-dimensional culture, which required additional stimulation by EGF (6). In addition, transactivation of ErbB2 and ErbB1 by EGF caused the resistance of ErbB2-overexpressing breast cancer against ErbB2-targeting antibodies (34). In the present study, the protein levels of activated ErbB2 but not total ErbB2 were markedly decreased by BXL0124, CDDO-Im, and the combination (Fig. 3), indicating that inhibition of ErbB2 activation rather than inhibition of ErbB2 expression was a key anticancer mechanism of the compounds. We also found a strong correlation between ErbB2 activation and ErbB1 activation (Fig. 3), suggesting the transactivation of ErbB2 and ErbB1 in MMTV-ErbB2/neu mammary tumors. In addition, we found that the treatments significantly inhibited mRNA levels of ErbB1 ligands, EGF, and HB-EGF (Fig. 4B), indicating ligand-induced transactivation of ErbB2 and ErbB1 as a critical target of the treatment.

c-Myc is a key transcriptional regulator of cellular proliferation, differentiation, and apoptosis, and its amplification has been frequently found in human breast cancer (35). In ErbB2-overexpressing breast cancer cells, c-Myc was shown as one of key effectors of ErbB2 signaling to sustain ErbB2-driven cell proliferation (36). In human breast cancer, the strong correlation between gene amplification of ErbB2 and c-Myc was reported, and patients with amplification of both ErbB2 and c-Myc had substantially worse clinical outcomes than patients with single-gene amplification (37, 38). A recent trastuzumab trial, HER2-positive National Surgical Adjuvant Breast and Bowel Project (NSABP) B31, showed that c-Myc alterations were associated with different clinical benefits with trastuzumab in patients with early-stage ErbB2-overexpressing breast cancer (39). In the present study, c-Myc was highly expressed in MMTV-ErbB2/neu mammary tumors, and the treatments with BXL0124, CDDO-Im, and the combination markedly decreased the protein level of c-Myc (Fig. 5B). In addition, the regulation of c-Myc by the treatments was associated with altered protein levels of downstream targets of c-Myc, p21, and PCNA, to regulate cell proliferation (Fig. 5B and C), emphasizing the importance of c-Myc as key downstream effector of ErbB2 signaling.

The amplification of ErbB2 has been reported in both premalignant ductal carcinoma in situ and invasive ductal carcinoma, suggesting the involvement of ErbB2 signaling in both early development and late progression of mammary tumorigenesis (3). Moreover, preclinical studies reported that chemotherapeutic drugs targeting ErbB receptors, such as gefitinib (ErbB1) and lapatinib (both ErbB1 and ErbB2), also had preventive effects on the development of ErbB2-overexpressing mammary tumors (40, 41). Therefore, we investigated the effects of the combination of BXL0124 and CDDO-Im on ErbB2-overexpressing mammary tumorigenesis in prevention and treatment studies. Our results showed that the combination treatment was effective in blocking or delaying the development of ErbB2-overexpressing mammary tumors, whereas it was not active on established mammary tumors. One explanation is that the activation of ErbB2 at an early stage of tumorigenesis may be inhibited by the treatment, leading to reduction of mammary tumor formation. Once the tumors are established, however, ErbB2 is overexpressed and highly activated such that the combination treatment would not reverse the activation of ErbB2 signaling. It is also possible that the duration of treatment was too short to elicit any effect. The short-term treatment did not induce apoptosis that can contribute to the reduction of established tumors. Higher dose of CDDO-Im should be tested for the activity of the combination on the established MMTV-ErbB2/neu mammary tumors, as high concentrations of CDDO-Im were required to induce apoptosis in breast cancer cells (19, 42). Further long-term treatment or dose-escalation studies may be necessary to confirm the inactivity of the treatment on established tumors.

ErbB2 has been a critical target for patients with breast cancer with ErbB2-overexpressing tumors, and ErbB2-targeting therapies such as trastuzumab have markedly improved clinical outcomes of patients with ErbB2-overexpressing breast cancer. However, there is still a need to search for chemopreventive agents for the ErbB2-positive subtype of breast cancer. Moreover, with the complexity of the interconnected signaling pathway of ErbB2 and other ErbB family receptors, drugs or combination of drugs targeting different components of the ErbB2 signaling pathway might be more effective. In the present study, we show that both BXL0124 and CDDO-Im targeted multiple components of the ErbB2 signaling pathway and delayed the development of ErbB2-overexpressing mammary tumors. The combination of BXL0124 and CDDO-Im showed additive effects on the inhibition of the ErbB2 signaling pathway, suggesting BXL0124, CDDO-Im and the combination as potentially useful chemopreventive agents in ErbB2-overexpressing breast cancers.

K.T. Liby has a commercial research grant from Reata Pharmaceuticals and ownership interest in triterpenoid patents. M.B. Sporn has a commercial research grant from Reata Pharmaceuticals and ownership interest in some patents. N. Suh has other commercial research support including research funding from Reata Pharmaceuticals and ownership interest in a patent. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J.Y. So, J.E. Wahler, A.K. Smolarek, N. Suh

Development of methodology: J.E. Wahler, A.K. Smolarek, N. Suh

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.Y. So, J.E. Wahler, T. Yoon, A.K. Smolarek, N. Suh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.Y. So, J.E. Wahler, T. Yoon, A.K. Smolarek, Y. Lin, W.J. Shih, N. Suh

Writing, review, and/or revision of the manuscript: J.Y. So, A.K. Smolarek, Y. Lin, K.T. Liby, M.B. Sporn, N. Suh

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H. Maehr, M. Uskokovic, M.B. Sporn

Study supervision: N. Suh

The authors thank the Laboratory of Animal Service at the Department of Chemical Biology for animal care.

This work was supported in part by the NIH [R01 CA127645; to N. Suh], and the Trustees Research Fellowship Program at Rutgers, The State University of New Jersey [to N. Suh].

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.