Virgin Sprague-Dawley female rats (180–200 g) bred in our laboratory were used. The animals were maintained in a light- (lights on from 06.00 to 20.00 h) and temperature-controlled room (22–24°C); rat chow (Cargill, Córdoba, Argentina) and tap water or 6-N-propyl-2-thiouracil (PTU) solution were available ad libitum.

Animal maintenance and handling were performed according to the NIH guide for the Care and Use of Laboratory Animals (NIH publication no. 86-23, revised 1985 and 1991) and the UK requirements for ethics of animal experimentation (Animal Scientific Procedures, Act 1986). All the experimental procedures were approved by the Animal and Ethics Committee of the School of Medicine of the National University of Cuyo, Mendoza, Argentina.

In order to study the effect of hypothyroidism or hyperthyroidism on mammary carcinogenesis, the rats were treated per os with a single dose of DMBA (15 mg/rat) at 55 days of age and were then divided into two experimental groups and their respective controls: hypothyroid rats (HypoT, n=26) and untreated rats (EUT, n=12); hyperthyroid rats (HyperT, n=23) and vehicle-treated rats (VEH, n=5). The day of DMBA administration was considered as day 0 of the experiment.

Hypothyroidism was induced by the administration of PTU (Sigma-Aldrich, Buenos Aires, Argentina) at a concentration of 0.1 g/l in the drinking water, starting on day 3. The respective control group (EUT) was maintained under standard conditions during the study period.

Hyperthyroidism was induced by daily subcutaneous administration of L-thyroxine (T₄; Sigma-Aldrich). T₄ was dissolved in saline solution (0.42 N of NaOH) and injected s.c. daily at a dose of 0.25 mg/ml/kg starting on day 3. Control age- and weight-matched female rats received VEH s.c.

Since no statistical differences were observed between the EUT and VEH groups in any of the studied variables, we grouped and analyzed them as a single control group.

Body weight was monitored weekly. Food intake was assessed on week 10 of the experiment and expressed as the weight of rat chow consumed per rat and per day. The animals were assessed every 72 h for the appearance of the first palpable tumor. The latency, incidence and progression of tumors were determined in all groups. The rats were sacrificed by decapitation at 10:00 h on the day of diestrus when the tumors reached a volume >1,000 mm³ or at the end of the experiment on day 200 when they did not develop mammary tumors. All the animals were weighed and their vaginal smears observed before sacrifice. Trunk blood samples were collected and allowed to clot at room temperature. Serum was separated and stored at −20°C until assayed for hormone determinations. Immediately after decapitation, intra-abdominal fat was removed, weighed and expressed as a percentage of total body weight. A piece of normal mammary gland and the tumors were removed for histopathological and immunohistochemical analysis.

DMBA (Sigma-Aldrich) was dissolved in sunflower oil at a concentration of 5 mg/ml and administered at a dose of 15 mg/rat on day 0 by an intragastric probe, 3 h after food and water deprivation to ensure a complete absorption of the drug. Food and water were replaced 2 h after DMBA administration. DMBA has been extensively used at that dose to study the possible mechanisms responsible for the beginning of the transformation of normal breast tissue into a tumor (11).

Thyroid stimulating hormone (TSH), PRL and GH concentrations were measured by double antibody radioimmunoassay using materials generously provided by A.F. Parlow and the National Hormone and Pituitary Program (NHPP, Harbor-UCLA Medical Center, Torrance, CA, USA). The hormones were radio-iodinated using the chloramine T method and purified by passage through Sephadex G75. The results were expressed in terms of the rat TSH RP-3, rat PRL RP-3 or rat GH RP-2 standard preparations. Assay sensitivity was 0.5 μg/l serum and the inter- and intra-assay coefficients of variation were <10% for all hormones.

E₂, Pg, T₃ and tetraiodothyronine (T₄) concentrations in sera were measured by radioimmunoassay using commercial kits for total hormones (TKE 21 and TKPG1 double antibody radioimmunoassay; Siemens Healthcare Diagnostics, Inc., Los Angeles, CA, USA and RK-6CT1 and RK-1CT1 double antibody radioimmunoassay; Institute of Isotopes Ltd., Budapest, Hungary; respectively). Inter- and intra-assay coefficients of variation were <10%.

Serum leptin concentrations were determined by a specific radioimmunoassay developed in the Instituto Multidisciplinario de Biología Celular (IMBICE) and validated for rat and mouse leptin against a commercial kit (catalog no. RL-83K; Linco Research, Inc., St. Charles, MO, USA) as previously published (12). Briefly, synthetic murine leptin (PrePro Tech, Inc., Rocky Hill, NJ, USA) was used for both labeled peptide and standard and for the development of anti-leptin serum. Leptin was radio-iodinated by the chloramine T method and purified by elution using a Sephacryl S-300 (Sigma-Aldrich) column. The anti-leptin serum was developed by rabbit immunization with murine leptin coupled to BSA. The intra- and inter-assay coefficients of variation were 5–8 and 10–13%, respectively.

The latency of mammary tumors was considered as the time between DMBA administration and the appearance of the first palpable tumor. Incidence was calculated as the percentage of rats that had tumors within the study period in respect to the total number of rats per group. We used a caliper to measure the major (DM) and minor (dm) diameters every 72 h and calculated the tumor volume (TV = dm² × DM/2). Tumor progression was assessed estimating tumor growth rate [GR = TV/(day of sacrifice − day of appearance of first tumor)].

Small pieces of the tumor and inguinal mammary gland (contralateral to the tumor) from each rat were processed for histopathologic studies by fixing in buffered formalin, dehydrating in ethanol and embedding in paraffin wax. Sections (3–5 μm) were cut with a microtome and stained with hematoxylin and eosin (H&E) to define the histopathological changes in the mammary glands and to classify tumors according to published criteria (13). Images were captured with an Eclipse E200 microscope fitted with a digital still camera Micrometrics SE Premium (both from Nikon Corp., Japan) under magnification of ×100, ×400 and ×600.

The quantification of the percentages of stroma, mostly composed by adipocytes, and epithelial tissue in the mammary gland was performed by measuring the area occupied by the adipose tissue or epithelium in 8–10 fields of each preparation from all rats using the ImageJ 1.42q software available at the NIH site (http://rsb.info.nih.gov/ij). Each area was expressed as a percentage of the whole field as previously published (14).

The microscopic analysis was carried out by two independent observers. The apoptotic and mitotic indices were calculated by counting the total number of apoptotic bodies and mitotic figures, respectively, in the histological sections stained with H&E in 10 ×400 magnification-fields from each animal. The mitotic/apoptotic ratio (M/A ratio) was calculated by dividing the mitotic index by the apoptotic index.

Serial sections (3–5 μm) were mounted onto 3-aminopropyltriethoxysilane (Sigma-Aldrich)-coated slides for subsequent IHC analysis. The primary antibody used in this study was a monoclonal mouse anti-proliferating cell nuclear antigen (PCNA; Dako Cytomation, Glostrup, Denmark) at 1:500. An antigen retrieval protocol using heat was used to unmask the antigens (30 min in citrate buffer, 0.01 M, pH 6.0). Tissue sections were incubated with the primary antibody overnight at 4°C in humidity chambers. A commercial kit to detect mouse and rabbit antibodies was used (Dako EnVision system, horseradish peroxidase, diaminobenzidine; Dako, Carpinteria, CA, USA). Slides were lightly counterstained with hematoxylin to reveal the nuclei, examined and photographed. The percentage of positive nuclei was obtained based on an average of 700 cells counted per sample, at a ×400 magnification. The immunostaining of the tumor cells was semi-quantitatively scored as: 0, no staining; 1, nuclear staining of <10% of tumor cells; 2, staining between 11 and 33% of tumor cells; 3, staining between 34 and 65% of tumor cells; 4, staining of >66% of tumor cells. These scores were obtained by two independent observers blinded regarding the clinical evaluation, and a few conflicting scores were resolved by consensus.

Apoptosis was measured by a modification of the TUNEL assay using the ApopTag Plus Apoptosis In Situ Detection kit (Oncor, Gaitherburg, MD, USA), as reported previously (15). The TUNEL apoptotic index was calculated as the percentage of positive nuclei, based on an average of 700 cells counted per case, at ×400 magnification.

Values are expressed as means ± SEM of 17–26 animals/group. All statistical analyses were performed using GraphPad Prism 5.01 software (GraphPad Software, Inc., San Diego, CA, USA). Differences in the distribution of variables between the three studied groups (HypoT, HyperT and controls) were assessed using one-way analysis of variance (ANOVA I) or the Kruskal Wallis test depending on the normality of the variables as evaluated by the Kolmogorov-Smirnov test. Two-way analysis of variance (ANOVA II) was used for analysis of the effects of the PTU or T₄ treatments on body weight over time. Post hoc comparisons between means were conducted by Bonferroni’s test or Dunn’s multiple comparison test. Student’s t-test was used when only two groups were compared. When variances were not homogeneous, logarithmic transformation of data was applied. Incidence and percentages of mammary adipose area were analyzed by Chi-square. Survival curves were compared using the log-rank Mantel-Cox test. The relation between selected variables was performed by Pearson’s correlation coefficient. Differences were considered significant when the probability was 5% or less.

The chronic thyroid condition of the animals was evaluated by measuring serum T₄, T₃ and TSH concentrations. The rats treated with PTU had low circulating T₄ (20.96±1.10 ng/ml, p<0.001) and T₃ (0.42±0.06 ng/ml, p<0.05) concentrations and also markedly increased levels of TSH (4.98±0.43 ng/ml, p<0.01) compared to the EUT rats (33.29±2.62, 0.61±0.06 and 0.84±0.06, respectively); all of them sensitive indicators for hypothyroidism. In contrast, L-thyroxine administration increased T₄ (55.88±6.93 ng/ml, p<0.05) and T₃ (0.91±0.10 ng/ml, p<0.05) and decreased circulating TSH (0.52±0.02 ng/ml, p<0.01) confirming the state of hyperthyroidism.

Chronic treatment with PTU significantly retarded the growth of the HypoT animals as reflected in the decreased weight and reduced levels of circulating GH (Table II) in comparison with the HyperT rats and the untreated controls. HypoT rats had an increase in body weight similar to HyperT and EUT rats during the first 4 weeks of treatment with PTU. Afterwards, they stopped gaining weight until the end of the study showing a significant difference in body weight when compared with the other two groups (p<0.0001, data not shown). HyperT rats gained weight steadily similar to the EUT rats. Daily food intake was significantly lower in HypoT rats compared to HyperT and EUT rats (12.6±0.2 vs. 23.2±1.2 and 19.1±0.7 g/rat, respectively; p<0.0001).

The mass of abdominal fat expressed as a percentage of the body weight has been previously used as a measure of body composition in rats (16). No statistically significant change was observed in this percentage due to PTU (11.3±1%) or T₄ (9.1±0.6%) treatments compared with controls (11.5±0.7%), although the HyperT rats had a tendency to show lower values.

The effect of THs on the development of the mammary gland was evaluated by histological observation and measurement of the areas occupied by parenchyma or stroma. Table I shows that the mammary glands of EUT rats had a normal appearance with few ducts surrounded by a small amount of fibrous connective tissue and abundant adipose tissue. The mammary glands of HyperT and HypoT rats were also normal but in a few cases were associated with mammary benign pathologies such as ductal stasis. Moreover, hyperplasia, intraductal papilloma and adenosis were observed in the HyperT rats (Table I).

Fig. 1A–C show representative microphotographs of H&E-stained mammary tissue from HypoT, EUT and HyperT rats, respectively. Fig. 1D shows the relative percentages of the adipose area. In EUT and HyperT rats, ~80% of the mammary gland was composed of stroma, which consisted of fat and connective tissue. The rest of the gland was represented by the parenchyma including lobules type 1, lobules type 2 and ductal structures (Fig. 1B and C). In HypoT rats, a scarce lobe-alveolar development of the mammary gland (~10%) and an increased percentage of fat in the stroma (~90%, p<0.001 compared to HyperT and EUT rats) were observed (Fig. 1D), suggesting that hypothyroidism, but not hyperthyroidism, induced changes in the parenchyma-stroma ratio.

Table I shows the histopathological characteristics of the tumors in the three groups. On one hand, tumors of EUT and HyperT rats were all of low grade considering the characteristics of the nuclei, the presence of ducts and the number of mitotic figures. They had a predominant in situ component with a small invasive component. Necrosis was scarce and they had moderate inflammatory infiltration. Thirty-three percent of EUT and 47% of HyperT rats developed secondary tumors. On the other hand, malignant tumors from HypoT rats were histologically more aggressive with a predominant invasive component and a small in situ component. Necrosis was moderate and the host response was lower than in the tumors from HyperT and EUT rats.

The latency of onset of tumors was longer (p<0.05) (Fig. 2A) and the incidence was lower (p<0.0001) (Fig. 2B) in the HypoT rats when compared to the EUT and HyperT rats. Hypothyroidism retarded tumor growth (p<0.05) (Fig. 2C) and increased tumor-free survival (p<0.001) (Fig. 2D). No relationship between tumor growth rate and the histological type of the neoplasia was apparent in any of the groups.

To assess the effect of THs on tumor progression, we performed a microscopic analysis of mitosis and apoptosis. The mitotic index was similar in the three groups (Fig. 3A) while the apoptotic index tended to increase in tumors of the HypoT rats (p=0.07) (Fig. 3B). In order to investigate if the balance between mitosis and apoptosis influences tumor development we calculated the mitotic/apoptotic ratio. This parameter was significantly lower in HypoT rats when compared to EUT and HyperT animals (p<0.05) (Fig. 3C), and was positively correlated with the tumor growth rate (r=0.715; p=0.002) (Fig. 3D) suggesting that the difference in the M/A ratio had biological significance.

We further studied apoptosis by TUNEL assay, a more sensitive technique, and we found an increased number of TUNEL-positive cells in tumors of the HypoT rats compared to the numbers in the HyperT and EUT animals (p<0.05) (Fig. 4A–D). In order to validate the results obtained by counting mitotic figures, we additionally evaluated mitosis by IHC of PCNA in the tumors. No statistically significant differences between the treatments were observed (Fig. 4E–H).

The influence of PTU and T₄ treatments on circulating hormone concentrations is shown in Table II. Both treatments decreased circulating levels of E₂ compared to EUT rats (p<0.0001).

Serum PRL and Pg concentrations were similar between the three studied groups. No significant differences were observed in circulating values of GH, PRL, E₂ and Pg between tumor-bearing and tumor-free rats in the HypoT, HyperT and EUT groups.

Leptin concentrations were significantly decreased in HypoT rats whether they were calculated per milliliter of serum or per gram of abdominal fat mass (p<0.05, Table II). This suggested that adipocyte secretion activity was altered in HypoT rats since the body fat mass was similar in all groups. Interestingly, serum leptin levels were higher in HypoT rats with tumors than in those without tumors (1.3±0.4 vs. 0.6±0.1 ng/ml, respectively; p<0.05). No significant differences were observed in EUT or HyperT rats, regarding tumor presence.

GH is required for mammary development since it stimulates ductal growth, proliferation and secretion (26). The effect of GH on mammary development is in part mediated by insulin-like growth factor 1 (IGF-1), and the action of E₂ and Pg are dependent upon IGF-1 (27). GH stimulates IGF-1 secretion in the liver and in the mammary stroma, therefore a paracrine role of GH modulating the effects of other hormones on proliferation and differentiation of mammary epithelium must not be ruled out (28). The present results demonstrated that hypothyroidism diminishes the secretion of GH in the rat, supporting previous studies (29).

Clinical evidence suggests that high circulating levels of GH/IGF-1 are associated with an increased risk of subsequently developing breast cancer (30). In studies in vitro, human GH was found to promote mammary carcinoma cell proliferation in an autocrine/paracrine manner (31) and resulted in a phenotypic conversion of human mammary carcinoma cells into a more invasive phenotype (32). In our in vivo study, we observed retardation in body growth associated with a significant decrease in serum GH levels in HypoT rats, which also had a lower incidence of mammary cancer.

Breast development occurs through a process of ductal elongation, branching and sprouting of ductules or alveoli, a process that requires extensive cell proliferation and penetration of the ductal epithelium into the stroma (33). Ductal elongation is directed by E₂, GH, IGF-1 and epidermal growth factor (EGF), whereas ductal branching and alveolar budding are influenced by additional factors such as Pg, PRL and THs. The stroma of the mammary gland is a complex structure composed of an extracellular matrix and a variety of cell types including endothelial cells, inflammatory cells, fibroblasts, fibroblast-like cells and adipocytes (34). This compartment can also be referred to as the mammary fat pad (34) since it mainly contains adipose tissue (35).

The response by the ductal epithelium to various hormones and growth factors is modulated by epithelial-stromal interactions, which seem to be bidirectional (36). It is expected that the marked variations in the epithelial-stroma ratio influence the bidirectional connection leading to specific modifications in gene expression that may account for the different susceptibility or risk to develop breast cancer (33).

In the present study, we observed that hypothyroidism retarded the growth of the ductal system and induced an increase in mammary gland fat deposit. In accordance, Vonderhaar and Greco (37,38) noticed that the gland of primiparous hypothyroid adult mice retained a primitive ductal appearance, i.e. that the epithelial component consisted of a sparse ductal system with few branches filling about one-fourth of the fat pad. Coincident with our results, they also described that while glands from hyperthyroid and euthyroid virgin animals preserved a small degree of ductal branching with primitive alveoli, the glands from hypothyroid animals showed less ductal branching and were devoid of alveoli (37). Thus, the decrease in mammary tumor incidence in our HypoT rats may be due to the reduced development of the mammary gland and the decrease in the parenchyma-stroma ratio.

THs have important effects on energy balance, since they influence both energy intake and expenditure. Severe hypothyroidism in the rat causes a sharp drop in food intake and metabolism, and this consequently suppresses the growth of DMBA-induced mammary cancer (11). Accordingly, in the present study, we observed a decreased food intake and body weight in HypoT rats in comparison with the controls and the HyperT animals. Calorie restriction was found to decrease proliferation and angiogenesis and increases apoptosis in mammary tumors (39).

Many changes related to energy homeostasis, body weight and food intake are associated with alterations in the function of adipose tissue. Leptin is the obese gene product secreted exclusively by adipocytes. Accumulating evidence shows that leptin is an important pro-angiogenic, pro-inflammatory and mitogenic factor, whose actions are reinforced through crosstalk with cytokines/growth factors (40). Increasing leptin levels activate the thyroid, GH, and gonadal axes (41). Thus, it is conceivable that leptin may have a role in different states of thyroidal disease. This relationship has been the subject of several studies, but no uniform picture has yet emerged from the results. Valcavi et al(42) reported significantly reduced serum leptin concentrations before and during replacement therapy in patients with hypothyroidism. Other authors did not observe alterations in serum leptin in different hypothyroid conditions (43) and Leonhardt et al(44) described increased leptin concentrations in patients with primary hypothyroidism or thyroid carcinoma compared with euthyroid controls.

In hypothyroid Wistar rats, a decrease in metabolizable energy intake and energy expenditure together with a shift in lipid and protein partitioning was found after 7 and 15 days of treatment with PTU (11). Consequently, body lipid percentage significantly increased compared to euthyroid rats. Our results in Sprague-Dawley rats showed that, even though the animals treated with PTU had a lower body weight than the controls, the percentage of body fat was similar in the three studied groups. However, adipose tissue of the HypoT rats appeared to be dysfunctional, attending to the lower amount of leptin expressed per gram of fat and the diminished serum levels of leptin in these animals. This fact can be one of the possible mechanisms related to the increased latency in HypoT rats since leptin has been shown to inhibit apoptosis and to stimulate growth (45) and also angiogenesis in breast carcinogenesis (46). However, in our in vivo study, we did not observe changes in tumor cell proliferation as evaluated by the mitotic index and by a more sensitive method, PCNA immunostaining. In turn, in cancer cells of our HypoT rats apoptosis was augmented in an inverse relation to serum leptin levels. In support of our results, it has been described that leptin inhibits apoptosis and this action is correlated with increased expression of anti-apoptotic protein Bcl-2 (47).

To summarize, dysfunctional body fat secreting lower levels of leptin may be another factor responsible for the delayed development of mammary tumors in HypoT rats. We cannot rule out the possible role of the stroma in the production of adipokines as tumor growth factors. Ongoing studies in our laboratory aim to evaluate the involvement of leptin locally produced by the peritumoral fat pad in the autocrine/paracrine regulation of tumor growth.

In the present study, we demonstrated that chronic hypothyroidism induced by PTU treatment modifies certain hormone patterns. E₂, which plays an essential role in mammary gland development and carcinogenesis, was diminished in HypoT rats. E₂ and Pg are known to promote proliferation and differentiation in the normal breast. Most breast cancers are initially hormone dependent and it is well accepted that E₂ plays a crucial role in their development and progression (48). E₂ exerts carcinogenic effects by simultaneously increasing the number of DNA replication errors by stimulating cell proliferation and gene expression; and through its oxidative metabolism that forms DNA damaging species (49).

There is evidence that Pg and its related signaling pathways are important players in the induction, progression and maintenance of the neoplastic phenotype in the mammary gland (50). However, in our in vivo study we did not find any relationship between circulating Pg and mammary carcinogenesis in rats with different thyroid disorders.

On the other hand, PRL secretion during diestrus was not affected in our experimental thyroid conditions. It has been previously shown that hypothyroidism can increase serum levels of PRL during estrous without changing basal secretion on diestrous day (29). In the present study, based on the similar values of circulating PRL between the treatments, we cannot attribute a relevant role of this hormone on mammary carcinogenesis. Moreover, we did not observe any significant differences in serum PRL in rats with and without tumors; further supporting the minor role of PRL in breast cancer development in HypoT rats.

In conclusion, our results to date show that hypothyroidism alters animal growth, breast morphology, body composition and adipocytoquine secretion and serum E₂ enhancing apoptosis, consequently retarding mammary carcinogenesis in rats. Additional studies by us are currently underway to investigate the involved molecular mechanisms.