Mammary Gland Development in Early Pubertal Female Macaques

Abstract

Ductal morphogenesis of the mammary gland occurs largely in the early pubertal period under the influence of ovarian sex hormones. In this study we characterized developing ductal structures in breast tissue of 14 female prepubertal or early pubertal rhesus monkeys (average age, 2.6 ± 0.1 years). Serum estradiol concentrations ranged from < 5.0 to 61.2 pg/ml, while progesterone concentrations were low in all animals (< 0.6 ng/ml). No corpora lutea were seen on ovarian histology. The most immature mammary glands contained multilayered ducts frequently ending in distinctive terminal end bud structures. Transitional ducts had variably stratified columnar epithelium, often with prominent rounded myoepithelial cells and loose periductal stroma. Mature ducts had simple cuboidal luminal epithelium, flattened myoepithelium, and associated lobular development, typical of the premenopausal breast. Expression of estrogen receptor alpha was significantly higher in immature ducts compared to mature ducts, while progesterone receptor expression, epithelial proliferation, and cytokeratin expression did not show a relationship to the maturity of the ducts. These findings identify dynamic morphologic changes occurring in the primate breast during early puberty and suggest that estradiol (rather than progesterone) is the predominant hormonal stimulus for early ductal and lobular development.

Introduction

The defining feature of mammalian mammary gland development is ductal arborization and differentiation into secretory milk-producing units. For the human mammary gland, this process of “mammopoiesis” begins prenatally with establishment of the rudimentary ductal scaffold in the second trimester (Howard and Gusterson, 2000). This prenatal ductal system develops to a variable extent in utero under the influence of maternal hormones, but then regresses following birth and undergoes few major morphologic changes until puberty (Russo and Russo, 1987; Monaghan et al., 1990). The most dramatic period of ductal morphogenesis occurs in the early pubertal transition, which begins in women approximately 2–3 years prior to menarche (Juul et al., 2006). At this time a combination of ovarian hormones and local growth factors induce elongation, branching, and lobuloalveolar development of the primitive ductal scaffold alongside marked stromal expansion (Monaghan et al., 1990; Howard and Gusterson, 2000; Hennighausen and Robinson, 2005; Hens and Wysolmerski, 2005). These changes culminate in a complex epithelial network of ducts and secretory lobules characteristic of the adult premenopausal breast.

Recent evidence suggests that mammary gland development may have an important role in determining future breast cancer risk. In the human and rodent mammary gland, a proportion of immature ducts end in solid bulbous terminal end bud (TEB) structures which differentiate into alveolar buds and lobular units (Russo et al., 1982; Monaghan et al., 1990). Recent studies have shown that populations of breast epithelial progenitor cells, possibly remnant from nonterminally differentiated TEBs, may later become founder cells for breast cancers (Liu et al., 2005; Russo et al., 2006). Approximately 90% of human breast cancers are thought to arise from epithelial cells in the terminal ductal lobular unit, and early patterns of ductal growth and maturation may be critical in establishing numbers of these susceptible cells (Russo et al., 2000; Hennighausen and Robinson, 2005).

This idea is supported by numerous studies of rodents showing that TEB cells are key sites of carcinogenesis and that TEB differentiation in puberty has a lasting chemoprotective effect (Russo et al., 1982; Russo et al., 2000; Lamartiniere 2002). A similar type of protective differentiating effect may account for the 2- to 3-fold decrease in breast cancer risk resulting from full-term pregnancy early in life (Guzman et al., 1999). Many key processes in breast carcinogenesis such as loss of differentiation, local invasion, and metastasis may also be closely related to pathways of normal ductal development at the molecular level (Hennighausen and Robinson, 2005). For example, signaling molecules such as beta-catenin, Wnt, and epidermal growth factors 1 and 2 have been found to have important roles in both breast development and carcinogenesis (Hovey and Trott, 2004). This evidence highlights how increased understanding of mammary gland development may provide important clues regarding breast disease.

Many of the early studies on mammary gland development used rhesus macaques (Schultz, 1937; Folley et al., 1939; Speert, 1948), due mainly to their close phylogenetic relationship to humans (Magness et al., 2005) and their highly similar reproductive anatomy and physiology. For example, female macaques have a 28-day menstrual cycle, spontaneous menarche and menopause (Gilardi et al., 1997; Kavanaugh et al., 2005), and mammary glands that are similar to women in terms of microanatomy (Folley et al., 1939; Speert 1948; Tsubura et al., 1991; Cline et al., 1997), sex steroid receptor expression (Cline et al., 1997, 1998), responses to exogenous hormones (Cline et al., 1998), and the development of hyper-plastic and neoplastic lesions in older animals (Wood et al., 2006).

An important potential application of the macaque model is the study of early life stage determinants of disease that cannot be directly assessed in human populations or appropriately modeled in rodents (Sternlicht et al., 2006). For the mammary gland, such studies require a more complete understanding of the detailed morphologic changes and specific temporal patterns of development that are not well-documented in the primate breast. The main purpose of the current study was to characterize key features of ductal morphogenesis in the early pubertal breast of female macaques.

Materials and Methods

Animal Subjects

In this study we evaluated mammary glands from 14 prepubertal or early pubertal rhesus macaques (Macaca mulatta) ranging in age from 2.38–3.15 years. All animals were colony-born at the Wake Forest University School of Medicine (WFUSM). The animals had previously been fed soy protein-based diets either with or without soy isoflavones (9.4 mg/kg body weight) in a Latin-square crossover experiment, as described previously (Anthony et al., 1996). The dose of soy isoflavones was similar to that found in commercially available monkey chow diets (Stroud et al., 2006). All procedures involving these animals were conducted in compliance with State and Federal laws, standards of the U.S. Department of Health and Human Services, and guidelines established by the Wake Forest University Animal Care and Use Committee. The facilities and laboratory animal program of Wake Forest University are fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care.

Serum Hormones

Serum estradiol (E2), progesterone (P4), and dehydroepiandrosterone sulfate (DHEAS) concentrations were measured in blood samples collected by femoral venipuncture at 2 time points, ~8 and 4 weeks prior to necropsy. Concentrations of E2, P4, and DHEAS were quantitated by radioimmunoassay using commercially available kits and protocols from Diagnostics Products (Los Angeles, CA) for E2 and DHEAS and Diagnostic Systems Laboratories (Webster, TX) for P4. Assays for E2 and DHEAS were run at the Yerkes Biomarkers Core Laboratory, Yerkes National Primate Research Center (Atlanta, GA) while P4 assays were run at the Wake Forest University Comparative Medicine Clinical Research Center. The normal assay range is 5.0 – 1000 pg/ml for E2, 0.1–40.0 ng/ml for P4, and 5.0–1,000 μg/dl for DHEAS.

Tissue Collection and Processing

For collection of reproductive tissues, animals were sedated with ketamine and euthanized using sodium pentobarbital (100 mg/kg, intravascular), as recommended by the Panel on Euthanasia of the American Veterinary Medical Association (AVMA Panel on Euthanasia, 2001). Euthanasia was performed as part of an unrelated experimental protocol. Mammary tissues, ovaries, and uteri were removed, fixed at 4°C in fresh 4% paraformaldehyde solution made in 0.1 molar sodium phosphate buffer at pH 7.4 (Fisher Scientific, Atlanta, GA), and transferred to 70% ethanol 24 hours later. Each ovary was examined grossly for evidence of corpora lutea.

Fixed tissues were cut at the level of the nipple (mammary gland), mid-longitudinally (ovaries), or transversely at the point of maximal thickness (uteri), dehydrated in graded ethanol, cleared in xylene and embedded in paraffin wax, sectioned at 4 μm thickness, and stained with hematoxylin and eosin (H&E) using standard histologic procedures. Paraffin-embedded ovaries were sectioned by microtome, and three sections spaced at intervals of 25–30 μm were stained with H&E. Each stained ovary section was then evaluated in entirety at objective magnifications of 2x and 10x in an effort to identify evidence of current or prior luteal activity (corpora lutea, corpora hemorrhagica, or corpora albicantia). For whole mount analysis, the right breast plate was fixed in 4% paraformaldehyde and 70% ethanol as described above for histologic specimens. The craniolateral quadrant of mammary fat pad (~10–15 cm²) was dissected from the overlying skin and placed in acetone for two nights or until all fat was removed.

Fat removal was evaluated by gently pressing the tissue between paper towels, allowing 1–2 minutes for the acetone to evaporate from the towels, and then examining the towels visually for any lipid residue. Mammary gland tissues were then rehydrated in 70% ethanol, placed in deionized water for 30 minutes, stained in 0.28% Toluidine blue (0.25 gm diluted in 10 ml of 100% ethanol + 80 ml of water and aged for at least 2 days), soaked in 100% methanol for 90 minutes, cleared in 100% methyl salicylate, and photographed. This technique was modified from Russo et al. (1994). Histologic and whole mount evaluations were performed by a board-certified veterinary pathologist (J.M.C.). Images of columnar cell lesions in ductal mammary gland epithelium of postmenopausal female macaques were taken from archival slides at WFUSM. These images were used to compare morphologic features of columnar cell lesions in older macaques and immature ductal structures in early pubertal macaques.

Immunohistochemistry

Immunostaining procedures were performed on fixed paraffin-embedded mammary gland tissues using commercially available primary monoclonal antibodies for estrogen receptor alpha (ERα, NCL-ER-6F11, Novocastra, Newcastle-upon-Tyne, UK), progesterone receptor (PGR) (NCL-PGR, Novocastra, Newcastle-upon-Tyne, UK), Ki67 (Ki67/MIB1, Dako, Carpinteria, CA), proliferating cell nuclear antigen (NCL-L-PCNA, Novocastra, Newcastle-upon-Tyne, UK), and cytokeratins 14 (clone LL002), 18 (clone DC10), and 19 (clone A53-B) (Lab Vision, Fremont, CA). Antibodies were diluted in 1X Automation Buffer (Biomeda, Foster City, CA) containing 0.5% casein (Sigma, St. Louis, MO); all antibodies were diluted 1:100 except for Ki67 (1:50).

Staining methods included antigen-retrieval, biotinylated rabbit anti-mouse F _c antibody as a linking reagent, alkaline phosphatase-conjugated streptavidin as the label, and Vector Red as the chromogen (Vector Laboratories, Burlingame, CA). Antigen retrieval was performed by heating slides to ~125°C for 30 minutes in an HS900 steamer (Black and Decker, Miramar, FL) in citrate buffer at pH 6.0 (modified from Pertschuk and Axiotis, 2000). Negative control slides were included for each immunostain; for these slides, the protocol was the same as for study slides except that nonimmune serum (from the same species as primary antibody) was used in place of the primary antibody.

Mammary gland tissues from estrogen-treated post-menopausal macaques in a separate study were also run in parallel and qualitatively assessed as external positive control slides. Cell staining was quantified by a computer-assisted counting technique, using a grid filter to select cells for counting (Lindholm et al., 1992) and our modified procedure of cell selection, described previously (Cline et al., 1997). For quantitation of mammary gland immunostaining, ducts were categorized as either immature (multiple luminal cell layers, columnar morphology) or mature (single luminal cell layer, cuboidal morphology), and 100 luminal epithelial cells were counted for each ductal type. Numbers of positively stained cells were expressed as a percentage of the total number examined (Cline et al., 1997).

Morphometry

Mammary gland epithelial area was quantified using both whole mount and histology images. Whole mounts were digitally photographed at a subgross level, and epithelial area was determined by manually tracing major ducts and lobuloalveolar units. Two representative 1 cm² regions were measured using Image Pro-Plus Version 5.1 (Media Cybernetics, Silver Spring, MD), and the average value was recorded for each animal. For histomorphometry, H&E-stained slides were digitized and 3 microscopic fields were randomly selected and examined at a magnification of 20 × and manually traced using NIH Image v1.60 (available at http://rsb.info.nih.gov/nih-image/) (Cline et al., 1998). Mammary gland epithelial area was expressed as a percentage of the total area examined. Endometrial thickness and vaginal keratin thickness were quantitated using similar techniques; these measures were taken in triplicate at sites of maximal perpendicular depth on H&E-stained sections, and the average value was used for each individual.

Statistical Analysis

To evaluate the relationship between mammary gland development and other measures, animals were divided into tertiles (low, middle, or high) based on percent epithelial area on whole mount, and tertiles were compared using a general linear model (SAS Institute Inc., 1999). Correlation between average serum E2 concentration and morphometric measures was evaluated using simple regression analysis. For immunohistochemistry data, a mixed general linear model with repeated measures was used to determine group means and test for significant differences between immature and mature ducts (measured on the same slide). Body weight was used as a covariate for all analyses. All variables were evaluated for their distribution and equality of variances between groups. Of these, Ki67 and PCNA immunostaining values were non-normally distributed and thus evaluated by Wilcoxon Rank Sums nonparametric analysis.

Quantitative endpoints (body weight, serum hormones, epithelial area, and immunohistochemical counts) were screened for effects of dietary protein (soy vs casein lactalbumin) at the time of tissue collection using a mixed general linear model or Wilcoxon analysis; no significant effects of diet were noted (p > 0.1 for all). Ductal epithelium was missing from histologic sections from 1/14 animals, reducing sample size for corresponding measures. Data were analyzed using the SAS statistical package (version 8; SAS Institute, Cary, NC). A two-tailed significance threshold of 0.05 was chosen for all comparisons. Data are reported as mean ± standard error of the mean (SEM).

Results

Pubertal Stage

The mean age of the 14 animals was 2.61 ± 0.08 years, similar to that reported previously for the onset of puberty in rhesus macaques (Speert, 1948). Average body weight was 4.15 ± 0.17 kg.

Serum hormone concentrations ranged from < 5.0 to 61.2 pg/ml for E2, 5.2 to 72.5 μg/dl for DHEAS, and < 0.01 to 0.51 ng/ml for P4. For the 13 confirmed early pubertal animals, mean serum hormone concentrations were 13.7 ± 2.3 pg/ml for E2, 0.07 ± 0.03 ng/ml for P4, and 36.4 ± 5.0 μg/dl for DHEAS. Serum E2 concentrations were < 5.0 pg/ml at both time points in 3/14 animals, while serum P4 concentrations were below detectable limits (< 0.1 ng/ml) at both time points in 12/14 animals (Figure 1). In adult cycling pre-menopausal macaques, E2, P4, and DHEAS concentrations generally range from 5–150 pg/ml, 0.5–6.0 ng/ml, and 5–30 μg/dl, respectively (Stute et al., 2004; Wood et al., 2004). Average serum E2 concentrations (4 and 8 week samples) correlated positively with age (r = 0.64, p = 0.01), body weight (r = 0.69, p = 0.007), uterine weight (r = 0.66, p = 0.01), and endometrial thickness (r = 0.58, p = 0.03), but not with vaginal keratinization (r = 0.24, p = 0.41). Average DHEAS concentrations did not correlate significantly with any of these measures (r < 0.5, p > 0.05 for all) (Figure 1).

No corpora lutea, corpora hemorrhagica, or corpora albicantia were observed grossly or on ovarian histology in any of the animals. All animals had multiple secondary or tertiary ovarian follicles on each ovary (~5–15 per section), producing a distinctive polycystic appearance (Figure 2A). Follicles ranged in size up to ~1.5 mm and often showed evidence of atresia (granulosa cell disaggregation and nuclear fragmentation).

Uteri from 13/14 animals were characterized by stromal edema and straight simple endometrial glands (Figure 2B) with scattered epithelial mitoses. The remaining animal had a rudimentary uterus with no stromal edema or epithelial mitoses, suggestive of prepubertal status. Average values for uterine weight and endometrial thickness were 1.08 ± 0.11 gm and 1.45 ± 0.17 mm, respectively. No evidence of spiral arterioles, glandular arborization, or stromal decidual change (luteal-associated changes) was noted on endometrial histology, while hemorrhage was evident along the endometrial surface in 2 animals. Eversion of the glandular cervix (a progestogenic change) was absent in all animals (Figure 2C), while 3/14 animals had squamous metaplasia of the glandular cervix. Vaginal keratinization averaged 116 ± 21 μm in thickness and was noted qualitatively to be moderate to marked in 10/14 cases.

Morphology of the Early Pubertal Breast

Substantial variation in mammary gland development was found among the 14 animals. On whole mount evaluation, lobuloalveolar development was minimal in 2 animals, mild-to-moderate in 8 animals, and extensive in 4 animals (Figures 3A–3C). Overall epithelial area was 57.6 ± 3.9% of total area on whole mount and 3.5 ± 1.0% on histology. Tertile analysis of animals based on whole mount epithelial area showed a significant positive association between epithelial area and serum E2 (p = 0.04, 4 wk sample) but no other serum or morphologic measures (Table 1). The least developed mammary glands consisted of a simple network of 8–12 large ducts extending ~0.5–1.5 cm from each nipple (Figure 3A). Conspicuous knob-like structures resembling TEBs were often present at the ends of these immature ducts (Figure 3A). With further development, ducts became more elongate and took on a “transitional” appearance characterized by extensive terminal and side branching and scattered small lobules (Figure 3B). In more mature mammary glands, marked lobular development was present along secondary and tertiary ducts, similar to that seen in a sexually-mature premenopausal state (Figure 3C).

On histology, mammary gland sections typically contained a mixture of both immature and mature ductal structures. The least developed mammary glands were composed primarily of primary and secondary ducts with limited branching (Figure 4A). These immature ducts consisted of 3–6 layers of poorly organized and densely packed luminal epithelial cells with scant cytoplasm, small nuclei with dense chromatin, minimal polarity, and no discernible nucleoli; a small fissure for a lumen; and a rim of plump myoepithelial cells sometimes showing distinctive cytoplasmic vacuolation (Figure 4B). The stroma surrounding immature ducts was frequently loose and myxomatous (Figure 4B–4D), possibly as a sign of remodeling associated with ductal migration.

Early stages of ductal elongation and branching were associated with increased luminal canalization and epithelial cell organization. Luminal epithelial cells frequently had greater amounts of pale eosinophilic cytoplasm, larger ovoid nuclei with increased polarity, finely granular chromatin, and distinct nucleoli. Densely cellular TEBs were often present at the end of these more active ducts. Histologic features of TEBs included stratification of luminal cells into 4–8 cell layers and prominent rounded myoepithelial cells (Figure 4C). The stromal clearing around TEBs was in some cases infiltrated by low numbers of eosinophils, lymphocytes, and plasma cells. Where present, mitoses were typically found at the neck region of the TEBs rather than the leading edge.

With increased ductal maturation, luminal epithelial cells assumed a more elongate morphology with variable nuclear polarity (Figure 4D). These transitional-type ducts had increased luminal size and progressive loss of luminal epithelial stratification. Luminal cells showed greater amounts of apical cytoplasm as nuclei became more basally located, ovoid to fusiform shape, and perpendicularly oriented to the basement membrane (Figure 4E–4G). In transitional ducts, myoepithelial cells became more flattened and the surrounding stroma increased in density (Figure 4E–4F). Ducts associated with more extensive lobuloalveolar development had luminal epithelial cells that were typically aligned in a single cuboidal layer with rounded basal nuclei and bordered by an inconspicuous myoepithelium (Figure 4H).

Sex Steroid Receptor Expression

Mammary gland development during puberty is mediated in large part through estrogen receptor alpha (ER) and progesterone receptor (PGR) signaling in the breast epithelium. Both ER and PGR proteins were detected within nuclei of immature ducts (with multilayered epithelium) and mature ducts (with simple cuboidal epithelium) (Figures 5A–5D). Among immature ducts, receptor expression in luminal epithelial nuclei averaged 30.9 ± 3.1% for ER and 22.8 ± 3.7% for PGR. The numbers of nuclei expressing ER was significantly higher in immature ducts compared to mature ducts (p = 0.04) measured on the same slide (Figure 6).

No difference was observed in the expression of PGR between immature and mature ducts (p = 0.80). Intraindividual comparisons showed that immature ducts had higher ER expression in 12/13 animals and higher PGR expression in 8/13 animals. A subset of immature ducts had strikingly high levels of ER staining present within both nuclear and cytoplasmic compartments. Of the 2 animals with minimal to no lobular development on whole mount evaluation, one had no glandular epithelium on histologic sections while the other (depicted in Figure 3A) had very low ductal ER expression (3% positive nuclei), suggesting that the increase in ER expression may coincide with early pubertal developmental changes.

Proliferation Marker Expression

We next evaluated expression of 2 widely used proliferation markers, PCNA and Ki67, in epithelial cells of immature and mature ducts. Among immature ducts, expression ranged from 0 to 39% (mean, 16.9 ± 4.0%) of luminal epithelial nuclei for PCNA and 0 to 10% (mean, 0.9 ± 0.5%) of nuclei for Ki67. Expression of PCNA was sporadic and random in distribution across a given slide (particularly within immature ducts), suggesting that the early pubertal mammary gland may undergo irregular bursts of proliferative activity (Figures 5E–5H). However, no significant differences were found in PCNA (p = 0.23) or Ki67 (p = 0.93) between immature ducts and mature ducts measured on the same slide (Figure 6).

Cytokeratin Expression

Mammary gland sections were also immunostained for cytokeratins (CKs) 14, 18, and 19, which are common markers of epithelial differentiation in the breast (Mikaelian et al., 2006). Strong CK 14 expression was present within mammary gland myoepithelium (lobules and ducts) (Figure 7A), epidermis, sebaceous glands, hairshaft epithelium, and apocrine sweat gland myoepithelium, while no CK 14 staining was present in mammary luminal epithelium (lobules or ducts), apocrine glands, or hair bulbs. A complementary pattern was seen for CK 18 and 19 staining, in which strong diffuse cytoplasmic staining was present throughout mammary luminal epithelium but not in mammary gland myoepithelium, epidermis, hair follicles, or sebaceous glands (Figure 7B–7C). No staining differences were noted qualitatively between mature and immature ductal structures or between ductal and lobular epithelial cells for any of the cytokeratins.

Morphology of Immature Ducts and Columnar Cell Lesions

Columnar cell change is a common mammary gland lesion in older postmenopausal women (Schnitt and Vincent-Salomon, 2003) and macaques (Cline and Wood, 2005–2006). These lesions are characterized morphologically by a pattern of nuclear elongation, alignment perpendicular to the basement membrane, and cell stratification, features similar in appearance to certain changes described for immature and transitional early pubertal ducts (Figure 8).

Discussion

The goal of this study was to document key features of ductal morphogenesis in the primate mammary gland during the early pubertal transition. Fourteen macaques 2.3–3.2 years old were surveyed, and all animals were classified as preovulatory based on ovarian and uterine histology and serum P4 concentrations. Despite this lack of luteal activity, we found substantial interindividual variation in ductal development and lobular differentiation in the mammary gland. Immature developing ducts showed a variety of morphologic features distinct from the more mature mammary gland, including epithelial stratification, columnar luminal epithelial cells with palisading nuclei, prominent rounded myoepithelial cells, and periductal stromal clearing. Transitional ducts also had greater overall expression of estrogen receptor alpha compared to more mature ducts, while no differences in expression of progesterone receptor, proliferation markers, or cytokeratins were noted.

Previous evidence from our laboratory and others has indicated that progestogens are key mitogens for epithelial cell growth in the breast, particularly in lobuloalveolar cells. For example, in the postmenopausal breast the addition of a progestin to an estrogen increases overall epithelial area, breast density, lobular cell proliferation, and breast cancer risk (Cline et al., 1998; Greendale et al., 2003; Chlebowski et al., 2003). In the developing mammary gland, studies of rodents have also shown that inhibiting progesterone receptor expression restricts ductal side-branching and alveologenesis (Ismail et al., 2003). The presence of lobuloalveolar development (which was marked in some cases) in the absence of luteal activity was thus an unexpected finding of the present study. Serum P4 and E2 concentrations and ovarian/uterine histology suggest that all but one of the macaques were in an “estrogen-only” state resulting from multiple anovulatory ovarian follicles. This unique period of early puberty is not well-studied in women but may have important implications for later developmental patterns in the breast and elsewhere.

Earlier studies in cycling premenopausal macaques have found ~10–15% ER-positive and ~5–10% PGR-positive ductal epithelial nuclei (Stute et al., 2004), values which are consistently lower than those reported for immature ducts in this study. We also noted distinct cytoplasmic ER expression among some immature ductal epithelial cells with very high nuclear staining, possibly representing increased ER translation in this subset of cells. These findings suggest that immature ducts of the early pubertal breast may go through periods of exaggerated responsiveness to estrogen, which could help drive the dynamic changes of early puberty. Future studies using prospective serial monitoring, gene expression profiling, and other methods are needed to help confirm this idea.

In contrast to ER immunostaining, expression of the proliferation marker Ki67 was notably low (< 2% of nuclei in all but 1 animal) within ductal epithelium, compared to ~2–8% of ductal epithelial nuclei reported previously in cycling premenopausal macaques (Stute et al., 2004). This low expression contrasted with nuclear Ki67 staining present throughout the basal epidermal and follicular epithelial cells overlying the mammary gland, which served as an internal staining control. We should also point out that mitotic epithelial nuclei, where present, were clearly positive for Ki67 staining, and that similar patterns of immunostaining (high ER, low Ki67) within immature ducts have been observed in mammary gland biopsies taken from a separate study of early pubertal female cynomolgus macaques (data not shown).

The low Ki67 expression is curious given the obvious ductal expansion and remodeling, the proliferative appearance of transitional ducts and TEB regions, and the relatively high expression of PCNA (which persists longer throughout the cell cycle). While reasons are unknown, we speculate that the low expression of Ki67 may relate to particular cell cycle dynamics of Ki67 expression, which can vary across specific stages of the cell cycle (Urruticoechea et al., 2005), and also to the high ER expression, which rarely overlaps with Ki67 in normal breast tissue (Clarke et al., 1997). Patterns of cytokeratin expression did not differ significantly between immature and mature ducts in this study or with patterns noted in the postmenopausal macaque breast (data not shown), suggesting that the epithelial CK phenotype (at least for 14, 18, and 19) is established prior to puberty and maintained throughout subsequent development.

An interesting incidental finding of this study was that certain stages of ductal development resembled a specific type of postmenopausal ductal lesion called columnar cell change or hyperplasia. In addition to morphologic appearance, previous studies also report high positive immunostaining for ER (Tremblay et al., 2005) and notably low expression for the Ki67 antigen (Noel et al., 2006) in columnar cell lesions, a pattern similar to that observed in immature ducts. It is unclear whether these morphologic and immunohistochemical similarities between a normal developmental growth pattern and a spontaneous age-related lesion are biologically related, but we should emphasize that for toxicologic studies using younger macaques (< 4 years of age), the developmental changes described herein should not be confused with hyperplastic ones.

Findings of this study document key stages of ductal morphogenesis in the primate breast occurring in the preovulatory window of development. Our results also indicate that TEB structures may be reliably identified and studied in the macaque breast, representing a unique opportunity to examine hormonal and growth factor pathways involved in the differentiation of this important population of mammary gland cells. Molecular profiling of specific maturation stages of developing ducts may also provide insights into the regulation of epithelial proliferation and differentiation, particularly as it relates to carcinogenesis. Information from this study should be helpful for future studies examining determinants of mammary gland development and how these factors may reduce breast cancer risk later in life.