Efficacy of fluvastatin and aspirin for prevention of hormonally insensitive breast cancer

Sub-confluent (175,000 cells/well) preneoplastic mammary epithelial cells (MCF10.NeoT and MCF10.AT1) and MCF10.DCIS were plated on coverslips in 6-well dishes and treated with 10 μM fluvastatin with or without the isoprenoid geranyl geranyl pyrophosphate (GGPP) (10 μM), mevalonic acid (100 μM), or squalene (50 μM) for 48 h. Total cellular proteins were extracted by lysing cells in RIPA buffer, described before [16].

Four-week-old SV40C3 TAg hemizygous female mice (strain 013591, FVB-Tg(C3-1-TAg) cJeg/JegJ) were bought from The Jackson Laboratory. Mouse genotype was confirmed by qPCR-based testing for SV40T antigen. In hemizygous mice, the SV40 transgene is activated at about 4 weeks of age, causing p53 mutations and inactivation of the RB pathway. These changes lead to the development of atypia (starting at ~ 8 weeks), DCIS (by ~ 12 weeks), and eventually invasive breast cancer (as early as 16 weeks) [17]. After 1–2 weeks of quarantine and acclimatization, 5–6 week old mice, of approximately equal weight, were randomized into 4 groups. All experimental procedures were performed per guidelines of the institutional animal care and use committee.

In vivo studies were conducted in 2 batches of 5 mice/gp in each batch and 4 groups. Each of the four groups with a total of 10 mice/group were treated with (i) fluvastatin alone (10 mg/kg/day in water, which is equivalent to 48.6 mg (for an average adult weighing 60 Kg) and is well within the 20–80 mg/day range of fluvastatin to humans or (ii) aspirin alone (20 mg/kg/day dissolved in 200 μl of ethanol that was resuspended in 200 ml of water, final concentration of 0.001% ethanol), which is equivalent to 97.3 mg in adult humans, and is within the prescribed dose of 81–325 mg/day in humans or (iii) fluvastatin (10 mg/kg/day) + aspirin (20 mg/kg/day)-also referred as combo group, or (iv) vehicle control (0.001% ethanol). Statin and aspirin treatments were administered to mice in their drinking water, which was changed every other day. Water consumption/cage/day was logged, and the drug concentration adjusted accordingly to achieve the intended doses. Since the goal of this study was to investigate the chemopreventive effects of statin, aspirin, or both on breast cancer development, the drug treatments were started at an age 5–6 weeks of age before the onset of any visible lesions of breast cancer and continued for 16 weeks. Mice were weighed once a week and palpated twice a week for mammary tumors beginning at age 10 weeks until 21 weeks. Tumor/lesion size were measured using Vernier calipers. Any palpable mass of 3 mm or larger was considered a tumor and marked the onset of tumor in treatment groups. Tumor incidence at the time of necropsy was defined as the number of animals with macroscopic mammary lesions upon dissection at 22 weeks of age. While tumor burden was measured by tumor weight and the number of tumors/mice also referred as multiplicity of tumors/animal.

At the end of the 16-week treatment (age ~ 22 weeks), mice were euthanized regardless of tumor burden, and the mammary glands were harvested and weighed. From each mouse, 6 mammary glands (cervical, abdominal, and inguinal sites on the right and left sides of the body) were harvested and processed separately. Both tumor containing and tumor free glands were harvested. Each mammary gland was bisected, with one half fixed in formalin and embedded in paraffin (FFPE) and the second half snap-frozen in liquid nitrogen. FFPE blocks were cut into 5-μm-thick sections. From each block, the first and fifth sections were used for hematoxylin/eosin staining for histological confirmation of palpable tumors, and the second and third sections were used for the TUNEL assay and immunohistochemistry, respectively. RNA was extracted from snap-frozen samples for qPCR analyses.

Apoptosis was measured at the single-cell level in the paraffin-embedded tissue sections or in single-cell suspension using the In Situ Cell Death Detection Kit (Sigma-Aldrich) labeled with TMR red fluorescence following the manufacturer’s instructions.

The tissue sections were first deparaffinized and serially rehydrated, permeabilized at room temperature, and incubated with TUNEL mixture at 37 °C for 1 h in the dark. Nuclei were stained with DAPI and sections were mounted using ProLong Diamond Antifade Mountant (ThermoFisher). TUNEL staining was analyzed using fluorescence microscopy. Eight slides (each containing at least 1 gland) per treatment group were stained. For each gland, multiple high-power (40×) images were taken. Alexa 488-labeled TUNEL-stained cells were overlaid with DAPI-stained nuclei (representing all cells). The average number of cells undergoing apoptosis was determined by counting the number of TUNEL-positive cells in a high-power field and normalizing with the total number of DAPI-positive cells.

Apoptosis was measured in the vehicle and drug treated cells by staining cells with annexin V and PI using a Thermo Fisher kit following the manufacturer’s protocol. This assay is based on externalization of phosphatidyl serine (PS) on the plasma membrane, a hallmark of an apoptotic cell. Annexin V has a strong binding affinity towards PS and annexin V antibody is employed to detect cells that are undergoing apoptosis. Propidium iodide that is impermeant to cells is used in combination with annexin V to distinguish between apoptotic and necrotic cells. Dead/necrotic cells lose their membrane permeability and thus are positive for both annexin V and PI. MCF10.AT1 cells were treated with the various drugs for 2 days, then trypsinized, and the cell pellet was washed with PBS. Next, cells were incubated with Alexa Fluor™ 488 dye–labeled anti Annexin V antibody solution followed by staining with propidium iodide. Apoptotic cell fraction was determined by calculating annexin V -Alexa 488–labeled cells as a proportion of total live cells (unstained cells).

For total RNA extraction, snap-frozen mammary tissues were transferred to 2-ml tubes containing ceramic beads (OMNI International) in the presence of 0.5 ml Trizol (Bio-Rad) and were homogenized using an Omni Bead Ruptor 24 (OMNI International). The samples were lysed by running 2 cycles at the speed of 5.65 m/sec for 45 s, with a 30-s interval between each cycle. The tissue homogenates were spun at 10,000 rpm for 5 min at 4 °C, and the supernatant was processed using a Trizol (Bio-Rad) extraction method to extract RNA. To study mRNA levels, first-strand cDNA was synthesized using an iScript kit (Bio-Rad) as described previously [15]. mRNA expression of HMGCR, AMPK, and reference control -ribosomal protein L19 was measured by SYBR green-based chemistry (Bio-Rad) using the delta Ct (ΔΔCt) method as described before [15, 18, 19]. The expression levels of the genes of interest were measured by normalizing their signal (Ct) relative to loading control L19 and calculating the fold change in treated samples compared to vehicle control as the reference. The primer sequences for detecting mouse transcripts are provided in Supplementary Table 1.

The levels of activated pAMPK were measured in the formalin fixed paraffin-embedded mammary tissue slides from the vehicle and fluvastatin + aspirin combination-treated mice by performing immunohistochemistry assay. Briefly, the slides were deparaffinized, antigen retrieval was performed and the pAMPK antibody was incubated at 1:150 dilution for 2 h followed by incubation with a secondary antibody labeled with DAB. Hematoxylin staining was performed to stain the nuclei. The cytoplasmic expression of pAMPK, which is known to be inversely correlated with histological grade and poor prognosis [20], was measured in the mammary glands by using Aperio imagescope cytoplasmic algorithm. The algorithm was fine-tuned with the help of pathologists to correctly measure the tissue staining. The staining score was evaluated and is presented as a scale from 0 to 3+, where 0 indicates no staining and 3+ strongest staining in the areas of interest that included mammary ducts. Ten slides per treatment group were stained and scored for pAMPK expression that is expressed as % cells stained/μm².

Total cellular proteins were subjected to SDS-PAGE gel electrophoresis as described previously [16]. The proteins were transferred to nitrocellulose membrane, which was then probed with PARP antibody that detects both total PARP (116 KDA) and cleaved PARP (89 KDa). The levels of cleaved PARP protein were quantified using odyssey imaging system and normalized with β-actin to determine the relative expression between vehicle and treated cells.

Chi square and Fisher exact tests were applied to the tumor incidence data. The Wilcoxon rank sum test was used to compare the distributions of blood cholesterol levels between vehicle and fluvastatin treatment groups. The rest of the statistical analyses were performed using Student’s unpaired t test. All tests were two-sided. p values equal to or less than 0.05 were considered statistically significant.

To investigate the effect of fluvastatin treatment alone or in combination with aspirin on the development of mammary tumors, SV40C3 TAg mice were treated with these agents for 16 weeks, and tumor incidence, tumor burden (as measured by weight and multiplicity of tumors/animal), and tumor latency were assessed between treatment groups (Fig. 1). Overall net change in body weight during16 weeks of treatment did not differ significantly between treatment groups (Supplementary Fig. 1). Fluvastatin treatment for 16 weeks (6 week–22 weeks) caused a 50% reduction in tumor incidence to 40%, as compared to 80% in all other treatment groups as revealed by macroscopic tumors present at the time of necropsy (all, p = 0.06) (Fig. 2a). In contrast to our in vitro data that suggested synergy [15] between the aspirin and fluvastatin combination, we observed no reduction in tumor incidence in mice treated with the combination of the 2 agents.

In addition to a lower tumor incidence, fluvastatin treatment resulted in a 50% reduction in the tumor burden as measured by multiplicity of tumors (0.6 tumors/mouse vs 1.2 tumors/mouse in vehicle-treated animals, Fig. 2b), although this difference between vehicle and fluvastatin-treated mice did not reach statistical significance (p = 0.2). Similar to the incidence data, the average tumor burden per mouse was comparable between the aspirin-treated (1.4 tumors/mouse), fluvastatin + aspirin–treated (1.3 tumors/mouse), and vehicle-treated (1.2 tumors/mouse) groups (Fig. 2b).

In mice that formed tumors, tumor burden as measured by weight was significantly reduced in the fluvastatin-treated animals relative to the vehicle-treated group (0.05 g vs 0.217 g, respectively; p = 0.04; Fig. 2c). In contrast, there was a non-significant 28–30% reduction in tumor weight in aspirin- and combination-treated groups relative to vehicle-treated animals (Fig. 2c).

Lastly, tumor latency was noted to be prolonged in fluvastatin- and combination-treated SV40C3 TAg mice, as suggested by tumor growth curves (Fig. 2d). The average times to spontaneous development of palpable mammary tumors were 16.8 weeks for controls, 20 weeks for fluvastatin treatment (p = 0.01 relative to vehicle), 18.75 weeks for aspirin treatment (p = non-significant), and 19.2 weeks for combination treatment (p = 0.02 relative to vehicle (Fig. 2d). Collectively, these results suggest that fluvastatin as a single agent is most effective at prevention, increasing the latency of tumor development and, importantly, reducing the frequency of tumor development. Although a combination of fluvastatin and aspirin also delayed the onset of tumors, the addition of aspirin to fluvastatin did not further abrogate tumor formation as compared to fluvastatin alone and in fact, provided no benefit in preventing tumors when compared to vehicle treatment.

The growth-inhibitory properties of fluvastatin have been reported to be due to increased cell death in cell lines and mouse models [23‐25]. To determine whether increased apoptotic cell death was indeed responsible for reduction in tumor growth in fluvastatin-treated mice, we performed TUNEL staining on a set of mammary glands (n = 8 samples per group) at the end of the 16-week treatment. Compared with the rate of TUNEL-positive apoptotic cells in vehicle control mice (3.62%), we found a significant increase in apoptotic cells in fluvastatin-treated mice (5.36%; p = 0.04) but not in aspirin- or combination-treated mice (3.20% and 2.82%, respectively, p = not significant) (Fig. 4). These results suggest that not only did aspirin alone not increase apoptotic cell death but that in combination with fluvastatin, it abrogated the effects of fluvastatin. These cellular findings are in line with the tumor incidence, tumor burden, and tumor latency data noted across the treatment groups.

The cholesterol biosynthesis pathway leads to the production of cholesterol and non-sterol compounds such as GGPP (Supplementary Fig. 3). The homeostasis of GGPP and cholesterol is regulated differently, where cholesterol levels are sensed very precisely and remain unchanged due to multivalent feedback responses, whereas GGPP levels are more variable, depending on differential demand for intermediates of the cholesterol biosynthesis pathway [26‐28]. Recent studies suggest that statins have cholesterol-independent effects [26‐28]. We were therefore interested in identifying whether growth-inhibitory effects of statins are mediated through GGPP in our tumor model system. We hypothesized that fluvastatin-mediated induction of apoptotic cell death in tumor tissue is driven by depletion of GGPP and turned to a cell line-based model system to test this concept. MCF10.AT1 cells were treated for 48 h with fluvastatin alone or in combination with either GGPP; mevalonate (MVA), an upstream substrate; or squalene, a precursor of cholesterol downstream of GGPP and MVA. As shown in Fig. 5, we found fluvastatin to induce apoptosis (5.40% in fluvastatin vs 1.52% in vehicle control treated cells, p = 0.01) as measured by FACS assay. Replenishing GGPP to fluvastatin-treated cells blocked apoptotic cell death (5.40% in fluvastatin vs 1.63% in fluvastatin + GGPP, p < 0.01) (Fig. 5) and cells look as healthy as vehicle-treated cells (Supplementary Fig. 4). Supplementation with mevalonate and GGPP completely reversed fluvastatin-induced cell death to levels seen in vehicle control (1.63%, 1.79% for GGPP and MVA, respectively, p < 0.05 compared to 5.40% in cells treated with fluvastatin alone). Partial reversal of fluvastatin-induced apoptosis was seen with the addition of squalene (2.83%).

In order to confirm if the apoptotic cell death caused by fluvastatin is characterized by PARP cleavage- a hallmark of apoptosis- we performed western blotting and measured PARP cleavage by fluvastatin treatment in MCF10.NeoT and MCF10.AT1 cells. We found fluvastatin treatment to cause PARP cleavage (10.57 fold in MCF10.NeoT and 7.0 -fold in MCF10.AT1 cells), and GGPP and MVA to completely inhibit the PARP cleavage (Fig. 6a, Supplementary Fig. 5). Consistent with FACS results, western blotting also showed the squalene to partially reverse the apoptosis caused by fluvastatin as indicated a 25.20% and 59.30% reduction in cleaved PARP in MCF10.NeoT and MCF10.AT1 cells, respectively (Fig. 6a, Supplementary Fig. 5).

These findings support the idea that fluvastatin-induced reduction of GGPP is a major cause of reduced apoptosis in statin treated transgenic mice and in turn may be the primary driver of the efficacy of statin chemoprevention.

The interaction and activation of FAS-FASL is an important step that is responsible for initiating apoptosis. FAS-FASL pathway activation is finely regulated and transcriptional upregulation of FASL is one of the key regulatory mechanisms [29]. Statin inhibition of RhoA GTPAse, that is prenylated through GGPP, is known to activate FASL expression [30] (Supplementary Fig. 3). Therefore, we tested if fluvastatin-induced apoptosis is mediated via increase in FASL transcript expression and found a twofold induction (p = 0.06) in FASL expression in mammary tissues of the fluvastatin-treated mice relative to vehicle control tissue (Fig. 6b) suggesting a role of FAS-FASL pathway in fluvastatin-induced apoptosis.