Selenized milk casein in the diet of BALB/c nude mice reduces growth of intramammary MCF-7 tumors

Estrogen receptor-positive MCF-7 breast cancer cells were cultured in Eagle’s Minimum Essential Medium (ATCC Catalog No. 30–2003) supplemented with 0.01 mg/mL human insulin and 10% fetal bovine serum. Cells were incubated at 37°C and 5% CO₂ in air atmosphere. Passage was conducted when cells reached confluency every two to three days. Cells were collected for injection at 80% confluency by centrifuging at 125 xg for 5 minutes. Cells were suspended in 50% Matrigel™ Basement Membrane Matrix (BD Biosciences, Mississauga, ON), 50% media at a final concentration of 3 × 10⁶ cells/100 uL.

To generate low- and high-Se caseins, Holstein dairy cows were fed 0 or 4.5ppm (dry basis) selenium as Se-yeast (Sel-Plex, Alltech Inc., Kentucky, USA) on top of a basal diet of 0.15 ppm Se from Na₂SeO₃. Diets were fed for 3 weeks, after which milk was collected and stored at 4°C until pasteurized and skimmed at 70°C with a flow rate of 1.5L/min. The skim milk was cooled to 45°C and an acid casein precipitate was formed by adding lactic acid (88% food grade) to a pH of 4.6. The casein precipitate was washed twice with cold deionized water, collected on cheese cloth, and drained overnight. The casein was separated into trays and freeze dried at −20°C for 3 days. The product was then ground and stored at 4°C.

Final Se concentrations in the low- and high-Se caseins, measured by fluorometry (AOAC, 2005) were 0.87 and 9.3 ppm, respectively. These low- and high-Se caseins were then sent to Research Diets Incorporated to be mixed with a standard rodent to produce four diets with varying Se levels. AIN-76 diets (Research Diets Inc., New Brunswick, NJ) containing 0.16, 0.51, 0.85, and 1.15 ppm Se (dry basis) were produced by mixing low- and high-Se caseins, to a final concentration of 20% casein in each diet (Table 1). After mixing, diets were gamma-irradiated and stored at 4°C. Proximate nutrient composition of diets was determined at a commercial laboratory according to AOAC (2005) methods.

All mice were housed and cared for according to guidelines established by the Animal Care Committeeat the University of Guelph and the Canadian Council on Animal Care. Seventy-two female, athymic, BALB/c nude mice, 6–8 weeks old, were purchased from Charles River Laboratories (Senneville, QC). Mice were housed in autoclaved, ventilated cages and provided with autoclaved water. Water and food were offered ad libitum. Food consumption was not measured, however body weights were measured biweekly throughout the trial. They were exposed to a 12-hour light/12-hour dark cycle and fed standard mouse chow (Research Diets, New Brunswick, NJ) once daily at 1800 h. After 1 week of acclimatization, mice were xenotransplanted under isoflurane anaeasthesia with 3 x 10⁶ MCF-7 cells in the mammary fat pad and 90-day release 17β-estradiol pellets (0.5 mg/pellet; Innovative Research of America, Sarasota, FL) subcutaneously between scapulae.

Beginning one week post-surgery, tumor volumes were assessed daily using caliper measurements. Tumor volumes (V) were calculated as V = l × w²/2, where l is length and w is width of the tumor. Once tumor volume reached approximately 60 mm³ mice were randomized to one of the four treatment diets. Only mice with tumors that reached a palpable volume within 3 weeks after implantation were included in the trial. The animal trial involved 72 mice; 18 mice per diet. 7, 8, 4, and 5 mice from treatments 1, 2, 3, and 4, respectively, were euthanized during this trial. These mice were prematurely euthanized for exhibiting criteria for euthanasia; either clinical signs or because tumor size exceeded 10% of body weight. Each cage of 5 mice was furnished with 2 plastic feeders that were replaced at 1200 h daily with 10 g powdered diet.

After 10 weeks on experimental diets, mice were sacrificed by CO₂ asphyxiation, and tumors and livers were excised, weighed and frozen in liquid nitrogen prior to storage at −80°C. A section of tumor was also stored in 10% formalin solution for at least 24 hours, after which it was embedded in paraffin wax.

Western blots were performed on tumor tissue from each subject to measure levels of cyclin D1, Bcl-2, and Bax. Tumor tissues were homogenized in 2 μL RIPA lysis buffer/mg tissue, containing 10 μL/mL protease inhibitor. Protein concentrations in tissue homogenates were measured according to Bradford (1976) with bovine serum albumin as the standard. Membranes were blocked for one hour with 5% (wt/vol) skim milk at 4°C and incubated with rabbit anti-cyclin D1 (1:1000 dilution, Cell Signaling, Danvers, MA), rabbit anti-Bcl-2 (1:1000 dilution, Cell Signaling, Danvers, MA), rabbit anti-Bax (1:1000 dilution, Cell Signaling, Danvers, MA) or rabbit anti-β-tubulin (1:1000 dilution, Cell Signaling, Danvers, MA) on a rocking platform. After washing with TBST (TBS, 1% (vol/vol) Tween 20), membranes were incubated with peroxidase-conjugated anti-rabbit secondary antibody (1:2000 dilution, Cell Signaling, Danvers, MA) for 1 hour at room temperature on a rocking platform. Immunoreactive proteins were visualized by chemiluminescence (ECL Western Blotting Detection Reagents) using horseradish peroxidase-linked secondary antibody (anti-rabbit immunoglobulin, 1:2000 dilution). β-tubulin was used to normalize protein loads between blots.

Tumor samples were freeze-dried and digested in nitric acid in a 90°C sand bath. The digestate was then brought up to volume (10mL) and analyzed using Inductively Coupled Plasma Mass Spectroscopy (AOAC, 2005). Selenium values are reported on a dry matter basis.

To compare tumor growth characteristics across treatments, trajectories of individual tumor volume (V_t) versus time (t) were fitted with the Gompertz equation,

where V_max is the asymptotic upper bound that the tumor volume approaches as time approaches infinity, and b and c describe the growth. The maximum specific growth rate (c) occurs at the inflection point, t_inflection = ln(b)/c. Best estimates for parameters V_max, b and c were obtained by minimizing residual sums of squares between predicted and observed tumor volumes using Excel Solver. The parameter estimates were then used to generate predicted volumes at weekly intervals after assignment to diet. Goodness of fit to each curve was assessed by root mean square prediction error (rMSPE), as a percentage of the mean observed tumor volume, calculated as:

where pred_i is the i-th prediction, obs_i is the i-th observation, and n is the number of observations.

Specific growth rate (a) at any point in time can be calculated using the following:

The specific growth rate at 70 d of dietary treatment was calculated as:

The general linear models procedure of SAS was used to detect differences in observations between treatments by one-way analysis of variance. To eliminate discrepancies in initial growth rates, mice put on treatment three weeks or longer post-inoculation were excluded from the dataset. Tumors were also excluded based on a confidence grade, where a grade of 1 was assigned to regular-shaped tumors, and a grade of 0 to those with an irregular shape for which volume could not be accurately measured. Tumors with a grade of 0 were then excluded from analysis. Tumors were also excluded if their final volume was below 30 mm³, indicating inadequate estrogen implantation. Number of tumors excluded from 0.16, 0.51, 0.85, and 1.15 ppm treatments were 3, 3, 1, and 1, respectively. Orthogonal linear and quadratic contrasts of dietary Se level were determined with coefficients calculated in PROC IML of SAS to match the measured Se level. Because of a large variation around mean tumor growth characteristics within treatments, the proportion of tumors with growth characteristics above or below specified threshold values were calculated for each treatment (see Figure 1). Characteristics considered were final observed tumor volume (V_final), c, V_max, t_inflection, and a_final. The maximum specific growth rate (c) may also be described as the specific growth rate at the inflection point (a_inflection). Linear effects of dietary Se level on the proportions were detected using an exact Cochran-Armitage trend test in PROC FREQ of SAS. P-values less than 0.05 were reported as significant.

Growth of mammary tumor volumes during the 10 weeks of dietary treatment exhibited exponential growth to a plateau (Figure 2). There was a linear decrease in mean tumor volume at 70 days with increasing Se intake (P = 0.040; Figure 3a) and a tendency for final tumor mass to decrease (P = 0.090; Figure 3b). Final tumor volume decreased 35% between 0.16 and 1.15 ppm Se. Within each treatment, some tumors reached their maximum volume by 70 days, while others were still in a quasi-exponential growth phase at 70 days. In addition to variation in the time at which plateau was reached, there was a large variability in the plateau volume itself. Chignola et al. [13] speculated that this growth variability is an intrinsic property of each individual tumor. Even multicellular tumor spheroids grown under controlled culture conditions in vitro exhibit large growth variability [13]. In order to account for these variables, growth curves were fit with the Gompertz equation which generates estimates of the plateau volume (V_max), and volumes (Equation 1) and specific growth rates (Equation 3) at any time point. Root MSPE averaged 16% of mean tumor volume with no difference between treatments. One mouse on 0.51 ppm Se was excluded from Gompertz analysis due to a physiologically implausible best-fit b-value > 100,000. There was no effect of Se-casein on mean V_max or other Gompertzian growth parameters of the mammary tumors (Table 2). Using the fitted Gompertz parameters, tumor volumes were predicted at weekly intervals and averaged by treatment to generate tumor growth curves for each Se inclusion level (Figure 4). Mean volumes at days 56, 63 and 70 on diet were significantly different between treatments, with decreasing volumes as Se-casein intake increased.

Effects of Se-casein on tumor growth were also evaluated by counting the number of large and fast-growing tumors on each treatment. The proportion of tumors with a final volume above 500 mm³ (Figure 1a) and the proportion of tumors with a maximum growth rate above 0.03 d^-1 (Figure 1b) were both linearly decreased by Se inclusion level (P _exact< 0.02). Proportions of tumors with a large V_max, t_inflection, or specific growth rate at 70 days were not affected by treatment (Figure 1c - e).

Average Se content of tumors increased linearly with Se intake from Se-casein (P< 0.001; Figure 5). The proportion of apoptotic cells in the 4 largest tumors on each treatment, which was assayed by DNA nick-end labeling (Figure 6a - d), also increased 2.4-fold with increasing Se levels (P = 0.007; Figure 6e).

The expression of Bcl-2 protects cells against apoptosis, while Bax opposes the action of Bcl-2 and aids in the induction of apoptosis. The ratio of Bax to Bcl-2 is often considered a rheostat to control the level of apoptosis occurring in tissue. Western blot analysis of tumor tissue showed no significant treatment effects on the expression of Bcl-2 or Bax proteins, or the ratio of Bax:Bcl-2 (Figure 7a - c).

The level of cyclin D1 expression was chosen as a marker of cell proliferation, as it is commonly over-expressed in breast cancer cells [14, 15]. Western blot analysis of tumor tissue from each mouse subject did not reveal a significant difference in cyclin D1 expression between treatments (Figure 7d).