Comparison of stem/progenitor cell number and transcriptomic profile in the mammary tissue of dairy and beef breed heifers

Bovine mammary stem cells (MaSC) are a source of ductal and lobulo-alveolar tissue during the development of the mammary gland and its remodeling in repeating lactation cycles. We hypothesize that the number of MaSC, their molecular properties, and interactions with their niche may be essential in order to determine the mammogenic potential in heifers. To verify this hypothesis, we compared the number of MaSC and the transcriptomic profile in the mammary tissue of 20-month-old, non-pregnant dairy (Holstein-Friesian, HF) and beef (Limousin, LM) heifers. For the identification and quantification of putative stem/progenitor cells in mammary tissue sections, scanning cytometry was used with a combination of MaSC molecular markers: stem cell antigen-1 (Sca-1) and fibronectin type III domain containing 3B (FNDC3B) protein. Cytometric analysis revealed a significantly higher number of Sca-1posFNDC3Bpos cells in HF (2.94 ± 0.35 %) than in LM (1.72 ± 0.20 %) heifers. In HF heifers, a higher expression of intramammary hormones, growth factors, cytokines, chemokines, and transcription regulators was observed. The model of mammary microenvironment favorable for MaSC was associated with the regulation of genes involved in MaSC maintenance, self-renewal, proliferation, migration, differentiation, mammary tissue remodeling, angiogenesis, regulation of adipocyte differentiation, lipid metabolism, and steroid and insulin signaling. In conclusion, the mammogenic potential in postpubertal dairy heifers is facilitated by a higher number of MaSC and up-regulation of mammary auto- and paracrine factors representing the MaSC niche.


Introduction
The bovine mammary gland is a unique organ with regard to its frequently repeating cycles of growth and involution throughout the life of an animal. Although the general processes controlling mammogenesis have been extensively studied, the knowledge on the role of stem cells and their renewal during mammary gland development is still insufficient. Mammary stem cells (MaSC) are defined as cells that can generate the ductal and lobular components of the mammary epithelial tree, complete with all the cell types of the mammary epithelium, as well as having the ability to self-renew (Stingl 2009). Stem cells allow the mammary epithelium to expand intensively during puberty and pregnancy, preparing the gland for milk production and secretion during lactation (Daniel and Smith 1999). To confirm the presence of MaSC, many in vitro and in vivo studies on rodents were conducted with the use of transplantation experiments, electron microscopy, functional techniques, flow cytometry, scanning cytometry, microarrays, and mammosphere cultures. Unfortunately, until now, a universal molecular stem cell marker for the identification of these cells has not been found. The most successful approach used to identify mouse MaSC has been based on a combination of surface markers: CD24 (heat-stable antigen), CD29 (β1 integrin), CD49f (α6 integrin), CD61 (β3 integrin), and Sca-1 (stem cell antigen-1) Stingl et al. 2006;Han et al. 2006). In the mouse, CD24 is a pan-epithelial marker that functions as a crude epithelialstromal discriminator (Sleeman et al. 2006;Stingl 2009). However, in the human mammary gland, CD24 is a luminal cell marker with a similar distribution to the luminal cellspecific glycoprotein MUC1. Thus, the most useful combination of molecular markers for isolating human MaSC comprised the epithelial cell adhesion molecule [EpCAM; also known as epithelial specific antigen (ESA) and CD326], CD49f, and, to a lesser degree, the luminal cell-specific glycoprotein MUC1 (Eirew et al. 2008;Villadsen et al. 2007). The cells expressing the above-mentioned markers were shown to form mammary repopulating units (MRU), which, when transplanted into cleared mammary fat pads of recipient mice, were able to repopulate the fat pad and recreate the structure of the mammary gland Eirew et al. 2008;Villadsen et al. 2007). Although a cleared fat pad technique was also described for ruminant species (Hovey et al. 1999), utilization of the technique has been very limited due to inherent differences between the composition of stroma in rodents and ruminants. Mouse stroma is composed mainly of adipocytes, whereas stromal tissue of the bovine mammary gland is fibrous (Sheffield 1988;Ellis et al. 2012). Furthermore, the global structure of the mammary gland differs significantly between rodents and ruminants. Murine mammary epithelium is a tree-like system of ducts terminated by numerous alveoli, whereas in ruminants, mammary alveoli and converging ducts form terminal duct lobular units (TDLU), which are gathered in a form of lobes.
In the attempt to define the bovine MaSC population, some promising results were obtained from the experiments based on the ability of these cells to retain the bromodeoxyuridine (BrdU) label for an extended period of time (Capuco 2007;Capuco et al. 2009). Stem cells were demonstrated to retain labeled DNA because of their selective segregation of template DNA strands during mitosis. These cells, described as label-retaining epithelial cells (LRECs), were detected immunohistochemically and quantified (Capuco 2007). The studies showed that the size of the bovine LREC population averaged 0.4 %, but could be doubled by xanthosine treatment, due to xanthosine-evoked suppression of the p53 function, which resulted in the promotion of asymmetric stem cell division (Capuco et al. 2009).
In a recently published study, Rauner and Barash (2012) utilized fluorescence-activated cell sorting (FACS) to distinguish and characterize four populations of cells within the bovine mammary gland on the basis of the expression of CD24 and CD49f surface markers. The authors have demonstrated that putative stem cells (puStm) were CD24 med CD49f pos , had basal localization, and preserved the ability to generate organized clones with duct-like cell alignment in culture. The populations of putative progenitors (puPgt: CD24 high CD49f neg ) and basal cells (CD24 neg CD49f pos ) were located downstream of puStm in the hierarchy, with puPgt having bi-potent characteristics, whereas the basal population complemented the puStm population to form the basal comp a r t m e n t . F i n a l l y, l u m i n a l c e l l s ( L u m ) w e r e CD24 med CD49f neg and complemented the puPtg cells in comprising the luminal compartment, and localized in the lower boundary of the luminal lineage.
Our group previously utilized stem cell antigen-1 (Sca-1) in the search for putative stem/progenitor cells in the bovine mammary gland (Motyl et al. 2011). The results demonstrated that bovine mammary epithelial cells expressing Sca-1 comprised about 2 % of the total cell number in the mammary tissue. Sca-1 pos cells were ERα-negative, indicating a possibility that Sca-1 pos cells form a population which localizes in the mammary stem/progenitor niches, and is important for the renewal of the bovine mammary gland during development and tissue regeneration. Transcriptomic analysis of the Sca-1 pos cell population in comparison with Sca-1 neg cells showed that the differentially expressed genes were involved in biological processes, such as signal transduction, development, protein metabolism and protein modifications, cell structure, motility, immunity, and defense (Motyl et al. 2011).
It is very likely that the number and unique morphological and molecular features of stem cells predispose the mammary gland to a certain type and dynamics of growth. Probably, numerous signals from the extracellular matrix (ECM) also affect the pattern of growth of the mammary gland. In fact, local tissue microenvironment composed of progenitor cells, basement membrane, ECM, stromal cells, and soluble factors, such as hormones and growth factors, create a functional signaling niche that directs cellular activity via direct contact or paracrine signaling (Bussard and Smith 2011). Thus, the number of stem cells, as well as the composition of the microenvironment, may determine the rate of development of the mammary gland. It is generally accepted that the intensity of proliferation and the productivity of the mammary gland in beef breeds and dairy breeds differs significantly (Keys et al. 1989;Akers et al. 2006). Up to now, no research has been carried out showing the relationship between the bovine mammary stem/progenitor cell number and interbreed differences in the intensity of mammary tissue outgrowth and development. The present study was undertaken to fulfill this gap by the quantification of stem/progenitor cells in the mammary tissue of non-pregnant 20-month-old Holstein-Friesian (HF) heifers-a typical dairy breed-and Limousin (LM) heifers-a typical beef breed. Moreover, a comparison of transcriptomic profiles of mammary tissue was performed in order to identify genes which could facilitate the formation of favorable tissue environment for self-renewal and differentiation of stem/progenitor cells. It seems essential to determine how the number of stem/progenitor cells and the influence of intramammary auto-and paracrine factors may impact growth, morphology, and productivity of the mammary gland of cattle with different phenotypes. Exploration of these issues may allow the use of the specific features of stem cells to control the development of the mammary gland, leading to a higher productivity of dairy cows, shortening the recovery time of the gland, controlling its defense mechanisms, and maybe even manipulating the milk composition.

Tissue sampling
Mammary tissue was obtained at a slaughterhouse from the udders of individual 20-month-old non-pregnant HF heifers (n=10) and LM heifers (n=10), free of clinical signs of mastitis. Udders were removed and mammary tissue was excised from the well-distinguished parenchymal region near the border with the mammary stroma. The samples were collected and immediately fixed, using different procedures, for scanning cytometry and microarray assay.
Immunofluorescent staining of tissue sections for scanning cytometry For immunofluorescent staining, samples were fixed in 4 % phosphate-buffered formalin and, after 48 h, stored in 70 % ethanol (POCH S.A., Gliwice, Poland) until further processing. Tissues were dehydrated and paraffin-embedded, according to standard histological technique. The paraffin blocks were cut into sections (5 μm), which were mounted on silanized microscope slides. Next, slides were deparaffined in xylene and hydrated in a graded series of ethanol to phosphate-buffered saline (PBS). For antigen retrieval, tissue sections were heated in a microwave (650 W) in 400 ml of 10 mM citrate buffer (ph 6.0), according to Capuco (2007). Next, they were rinsed with PBS.
For scanning cytometry, tissue sections were labeled with fibronectin type III domain containing 3B (FNDC3B) rabbit polyclonal antibody (Santa Cruz Biotechnology), followed by incubation with secondary anti-rabbit Alexa Fluor 647conjugated antibody (Life Technologies, USA), and with mouse anti-Sca-1-FITC-conjugated antibody (BD Pharmingen, USA). Nuclei were counterstained with Hoechst 33342. Scanning cytometry analysis was performed using an Olympus Scan^R screening station (Olympus Polska, Sp. z o. o., Warsaw, Poland), and combined analysis software (SCAN^R Analysis version 1.3.03). The results were statistically evaluated using Microsoft Excel 2003 software (Microsoft Corporation, Redmond, WA, USA) to calculate the mean number of Sca-1 pos FNDC3B pos cells±standard deviation (SD). Statistical significance was calculated by oneway analysis of variance (ANOVA) comparing the number of Sca-1 pos FNDC3B pos cells in mammary tissue samples from LM and HF heifers (GraphPad Prism software version 5.00). A p-value≤0.05 was regarded as significant.

Microarray analysis
For microarrays, mammary tissue samples were snap-frozen in RNAlater (Sigma Aldrich, Poland) and stored at −80°C. The total RNA was isolated from 50 mg of each tissue sample using a Total RNA kit (A&A Biotechnology, Poland), according to the manufacturer's protocol. Isolated RNA samples were dissolved in RNase-free water. The quantity of RNA was measured using NanoDrop (NanoDrop Technologies, USA). Next, RNA samples were treated with DNase-I to eliminate the possibility of DNA contamination, and subsequently purified using the RNeasy MinElute Cleanup Kit (Qiagen, Germany). Finally, RNA was analyzed using a BioAnalyzer (Agilent, USA) to measure the final RNA quality and integrity. Samples with RIN between 7 and 8.5 were used for further analysis. Amplification and labeling of target RNA was done using the Quick Amp Labeling Kit (Agilent, USA) in order to generate complementary RNA (cRNA) for oligo microarrays used in gene expression profiling and other downstream analyses. Prior the labeling procedure, equal amounts of RNA (500 ng) from each mammary tissue were pooled (HF pool from 10 samples and LM pool from 10 samples). We chose to use the pooling of RNA because the number of animals in each compared group was relatively small (n=10) and this method allowed us to decrease the variability between animals within a group (HF and LM). The gene expression of the HF heifers was compared against the gene expression of the LM heifers' mammary tissue. Samples were examined in four repetitions (two dye-swaps to eliminate the effect of label factor). Taking the average of all labeled arrays, the dye effect on any particular gene was cancelled. The hybridization was performed with bovinespecific gene expression Agilent microarrays (4x44K format, no. of probes = 43603) using the Gene Expression Hybridization Kit (Agilent, USA), according to the manufacturer's protocol. The results obtained constitute only a comparison of the "average" expression in one group versus another (HF vs. LM heifers). Although the design used in the analysis was based on four technical repetitions of the two pooled samples, the results obtained were used only for direct comparison, and were not correlated with any individual data, thus, they are acceptable from the methodological point of view (Kendziorski et al. 2003).

Signal detection, quantification, and analysis
Acquisition and analysis of hybridization intensities were performed using a DNA microarray scanner (Agilent, USA). Then, the results were extracted using Agilent's Feature Extraction Software with normalization and robust statistical analyses. Analysis of datasets was performed using GeneSpring software (Agilent, USA). Raw data were preprocessed to remove variability across and within array samples. To minimize non-biological variability across arrays, raw data were first log2 transformed. All data were filtered by flags present in all samples. The unpaired t-test with Benjamin-Hochberg false discovery rate (FDR) <5 % correction was applied (with p-value cut-off ≤0.05). For further analysis, we chose genes with significant changes in expressions of over 1.3-fold change. The area of the analyses covered in this publication has been deposited in the NCBI's Gene Expression Omnibus and is accessible via GEO Series accession number GSE47816.
Gene functions were identified using the Gene Ontology (GO) annotation of genes in the Functional Annotation Chart, available through the online Database for Annotation, Visualization and Integrated Discovery (DAVID). DAVID was used to analyze gene sets significantly differing in expression between the mammary glands of HF and LM heifers. Analysis of signaling pathways was carried out using Single Experiment Analysis (SEA) exact in GeneSpring (Agilent, USA). The significance of the association between the dataset and the GO terms were measured using the ratio of the number of genes from the dataset that map to the pathway divided by the total number of genes in the test. A threshold of p-value≤ 0.05 was used to indicate biological processes, molecular functions, and signaling pathways that are significantly represented by genes in an annotated gene list. To interpret the results, genes' functions and interactions were analyzed based on the literature, and genes' interactions networks were developed using Pathway Studio software (Ariadne Genomics, Inc.).

Identification and quantification of mammary stem/progenitor cells
To identify putative mammary stem/progenitor cells, we applied double labeling with previously used anti-Sca-1-FITCconjugated antibody (Motyl et al. 2011) and antibody against FNDC3B, which is considered a putative marker of mammary stem/progenitor cells. Previous studies have shown that FNDC3B can be used as a potential marker of LREC in the bovine mammary gland (Choudhary and Capuco 2012). Cytometric analysis of the number of cells expressing both FNDC3B and Sca-1 antigens revealed their significantly higher value in HF (2.94 ± 0.35 %) than in LM (1.72 ± 0.20 %) heifers. Their undifferentiated status was confirmed by the lack of ERα expression (data unpublished).
Transcriptomic profiles of the mammary tissue of dairy and beef breed heifers The transcriptomic background of a higher mammogenic potential in HF heifers was investigated by comparing the differences between gene expression profiles in the mammary tissue of HF and LM heifers. Fold change analysis followed by Student's t-test and Benjamini-Hochberg FDR correction (p-value≤0.05) identified 3,153 probes representing 1,987 genes that were significantly altered by at least 1.3-fold between the two examined breeds. The classification of genes according to their molecular function revealed four major groups of up-regulated genes in HF heifers: transcription regulator activity (53 genes), cytokine activity (17 genes), growth factor activity (12 genes), and chemokine activity (9 genes). Among genes with significant changes in expression, we selected those representing tissue hormones, growth factors, and cytokines, which expression was linked with the function of the mammary gland (Table 1). Special attention was focused on the expression of genes whose products could affect mammary stem/progenitor cells (Table 2). Among them were genes involved in cell maintenance, stem cell renewal, and stem cell development.
On the basis of the functions of genes presented in Tables 1 and 2, and the available literature, we were able to predict and discuss the role of the identified genes in the mechanisms controlling stem/progenitor cells function in outgrowth and development of the mammary tissue.

Discussion
There is a great deal of evidence showing the importance of pubertal mammary development as a determinant of heifers' milk yield potential. Mammary growth during puberty is affected by endocrine regulation, mainly ovarian and somatotropic axis hormones, as well as nutrition, including feeding level and specific nutrients (Sejrsen 1994). The results of the present study revealed that, apart from central endocrine and nutritional mechanisms controlling mammogenesis, the intra-mammary potential of the stem/progenitor cells population may influence better mammary tissue outgrowth and development in dairy heifers. We were able to show a higher number of stem/progenitor cells and up-regulation of genes involved in the formation of favorable niche for the maintenance, renewal, proliferation, and differentiation of stem/progenitor cells in HF than LM heifers (Tables 1 and  2). The higher number of stem/progenitor cells, as well as the elevated expression of genes involved in the growth and differentiation of the glandular tissue, can make the postpubertal development of mammary tissue more advanced, and the gland better prepared for pregnancy and lactation in HF heifers.

Estimation of the mammary stem/progenitor cell number in different breed heifers
Many potential molecular markers have been investigated in search of a universal stem cell marker, or a set of markers that would allow assessment of the number of stem cells within the mammary gland. The approach used in the identification and isolation of MaSC from murine and human mammary glands was based on a combination of several surface markers: CD24, CD29 (β1 integrin), and CD49f (α6 integrin) Shackleton et al. 2006;Eirew et al. 2008).
Recently, Rauner and Barash (2012) used FACS to determine the expression of CD24 and CD49f in mammary epithelial cells (MECs) isolated from the bovine mammary gland. On the basis of the expression levels of both surface markers, the authors distinguished four different subpopulations of bovine mammary epithelial cells, of which the putative stem cells population (puStm) was characterized by moderate levels of CD24 and high levels of CD49f (CD24 med CD49f high ). The measurement of aldehyde dehydrogenase (ALDH) activity in bovine MECs was also used as a marker, allowing for the separation of luminal and basal compartments, and it was shown that the basal compartment was enriched in stem celllike activity (Martignani et al. 2010). Sca-1 belongs to the group of surface markers which have been linked with the stemness of cells (Welm et al. 2002;Deugnier et al. 2006). In our previous study (Motyl et al. 2011), we found that only a small percentage of cells (2.14 %) expressed Sca-1 in the mammary tissue of 2-year-old HF heifers. The Sca-1 pos cells did not express estrogen receptor, but showed up-regulation of genes characteristic for cells of hematopoietic origin. Since Sca-1 is expressed not only by epithelial stem/progenitor cells but also by non-differentiated cells of bone marrow origin, it cannot be used as a single marker, defining a stem cell population. Therefore, in the present study, we decided to use an [NM_175050] −1.3 0.027 Controls the proliferation and differentiation of hematopoietic stem and progenitor cells additional stem/progenitor cell marker, FNDC3B, to determine the epithelial stem/progenitor cell population. FNDC3B has been recently identified by microarray analysis of LREC as a potential marker of bovine MaSC (Choudhary et al. 2010). FNDC3B-positive cells were demonstrated to have increased telomerase activity, which is a feature characteristic of stem and progenitor cells (Choudhary and Capuco 2012). Cells double-labeled with antibodies against Sca-1 and FNDC3B, which were conjugated with fluorescent dyes, were counted using scanning cytometry. It is a good method for stem/progenitor cell analysis, since it allows direct examination of the cells in the tissue, without interference of its structure. Scanning cytometry enables to relocate cells with a specific phenotype in a heterogeneous tissue structure (Godlewski et al. 2008). It also gives the possibility of imaging counted cells, which not only allows to assess the accuracy of allocating cells to specific gates, but also enables to compare gated cells between the samples. A comparison of cytograms from the mammary tissues of both examined breeds revealed a significantly higher number of Sca-1 pos FNDC3B pos cells in the glandular tissue of HF than in LM heifers. We assume that the difference in the number of these cells, which can constitute a putative population of stem/progenitor cells, can belong to the factors affecting the different ability of outgrowth and development of the mammary gland, and, as a consequence, may reflect the differences in milk production between dairy and beef breed cattle.
Identification of genes potentially regulating mammary stem cell niche in the mammary gland of heifers Transcriptomic analyses carried out as a part of this study identified a number of biological processes, molecular functions, and signaling pathways which may pose a dairy potential in HF heifers compared to LM heifers. The literature data show that the entire functional outgrowth of mammary epithelium may comprise the progeny from a single cell (Kordon and Smith 1998). In most tissues of adult organisms, microenvironments, also called niches, are present and serve as regions of stem cell activity ). Niches protect the stem cells from their excessive expansion, regulate their activity, and keep them in maintenance by homeostasis caused by the surrounding tissues (Bussard and Smith 2011). All niche structures (basement membrane, extracellular matrix) and cells of the mammary gland (progenitor, myoepithelial, fibroblasts, adipocytes, and cells of the immune system) are maintained in a state of dynamic equilibrium by adjusting auto-and paracrine regulators, including growth factors, hormones, and cytokines activity (McCave et al. 2010). The dominance of the microenvironment over the stem cell's autonomous phenotype has been demonstrated in several reports involving cells crossing lineage "boundaries" to regenerate "foreign" tissues (Booth et al. 2008). Some very interesting data were provided by Boulanger and coworkers (2007), who observed cooperation of cells isolated from the seminiferous tubules of the mature testis into the murine mammary fat pads, contributing robust numbers of epithelial progeny to normally growing mammary glands. The same group of researchers transplanted neuronal stem cells (NSCs) from WAP-Cre/Rosa26R mice together with wildtype mammary epithelial cells into epithelium-divested mammary fat pads of prepubertal female mice, and observed that NSCs interact with MECs on transplantation, and contributed mammary epithelial-specific progeny to normal mammary outgrowths (Booth et al. 2008). The complex interactions between mammary stem/progenitor cells, stromal cells, and other components of the glandular environment are still poorly understood. However, based on the results obtained in our microarray experiment, ontological analysis, genes interaction networks, and the available literature, we have proposed a model of the mammary gland niche controlling stem cells activity (Fig. 1). The microarray data contained a number of genes associated with the activity of the MaSC and HSCs (Table 1). We revealed changes in the expression of genes involved in the maintenance, differentiation, and renewal of stem cells (Table 2).
The higher degree of development of the mammary gland in HF heifers was accompanied by the up-regulation of many genes representing factors related to stem cell maintenance and mammary tissue remodeling (Table 1). Among the upregulated genes, we identified those encoding: Janus kinase 2 (JAK2), responsible for the regulation of alveolar cells differentiation and their maintenance during differentiation (Wagner et al. 2004); colony-stimulating factors: CSF1, CSF2, associated with the regulation of MaSC and macrophages activity, as well as stimulation of the outgrowth potential and regenerative abilities of the mammary gland (Gyorki et al. 2009); neuregulin 1 (NGR1), involved in the promotion of growth, differentiation, and stimulation of branching morphogenesis, lobulo-alveolar budding, and milk proteins production (Yang et al. 1995); transcription factor FOSL1, that takes part in the promotion of vasculogenic and angiogenic processes in the epithelium and forming tube-like structures (Evellin et al. 2013).
Among the transcripts up-regulated in the mammary gland of HF heifers were also: fibroblast growth factor 2 (FGF2), which was shown to play an important role in the differentiation of stem cells to mesodermal lineages (Sharpe et al. 2011); betacellulin (BTC), linked with the development of a lactating-like phenotype in the mammary gland of virgin female mice (Dahlhoff et al. 2011); nerve growth factor (NGF), involved in mammary tumor stem cell metastasis, proliferation, and survival (Adriaenssens et al. 2008) (Table 1). Products of the above-mentioned genes (FGF2, BTC, and JAK2) are involved in the EGFR signaling pathway, whose activity differs significantly between the two compared cattle breeds. The EGF-EGFR-mediated signaling pathway plays a key role in mammary gland development, autophagy regulation, cell cycle progression, estrogen metabolism, proliferation, cell survival, angiogenesis, and cell migration (Lurje and Lenz 2009;Sobolewska et al. 2009). Our study also revealed a group of genes differentially expressed between HF and LM breeds, which are involved in the ID signaling pathway: NGF and ID1, ID2. ID signaling is important in the regulation of maintenance, migration, and differentiation of stem cells. Increased level of proteins taking part in this pathway may result in the loss of differentiation and gain of migration and proliferative abilities (ID1), and activation of mammary differentiation and alveologenesis during pregnancy (ID2) (Itahana et al. 2003;Dong et al. 2011).
Another group of genes exerting higher expression in the mammary tissue of HF than LM comprised factors essential for the maintenance of hematopoietic stem cells' pools, and for mammary gland involution after lactation (LIF), as well as maintenance of cellular homeostasis, mammary epithelium growth, and stem cells activity (BMI1) (Pietersen et al. 2008;Mathieu et al. 2012). We observed a lower expression of genes encoding proteins responsible for the growth arrest of mammary epithelial cells and maintenance of mammary gland progenitor cells in the quiescent state (FABP3, LAMA1) (Bionaz and Loor 2008a; Bussard and Smith 2011) ( Table 2).
Interestingly, the mammary gland of HF heifers also showed a higher expression of genes involved in the regulation of stem cells renewal. We observed changes in the expression of genes which regulate the renewal of hematopoietic stem cells (HSCs), e.g., HOXB4, MLL5, TAC1, and MYB. HSCs are particularly important in the context of their penetration into the mammary gland and their potential effect on glandular development (Niku et al. 2004). HOXB4 (showing up-regulated expression in this study) activates Myc transcription factor, which also showed a higher expression. It has been demonstrated that proteins from this family of transcription factors prevent hematopoietic stem cells from differentiating, Fig. 1 Proposed scheme of molecular regulation of the mammary stem cell microenvironment. The scheme illustrates specific dynamic regulations leading to increased potential of milk production. The stem cell niche in the mammary gland comprises mammary stem cells, as well as epithelial and myoepithelial cells, which are separated by basement membrane from adipocytes, fibroblasts, and blood vessels. The scheme was developed based on the comparison of transcriptomic profiles of mammary glands from Holstein-Friesian (HF) heifers and Limousin (LM) heifers. The names of genes regulating stem cell activity are shown in bold. The arrows indicate the direction of gene expression in dairy heifers in relation to beef heifers while, at the same time, allowing their self-renewal (Stoelzle et al. 2009;Lee et al. 2013). Another identified gene, MLL5, encodes a protein (myeloid/lymphoid or mixed-lineage leukemia 5), which is a key regulator of hematopoiesis involved in terminal myeloid differentiation and in the regulation of HSCs self-renewal (Heuser et al. 2009).
This study revealed also a significantly higher expression of several cytokines and growth factors in the mammary tissue of dairy heifers (HF) when compared to beef breed heifers (LM) ( Table 1). IL6 and IL8 are associated with multiple processes important for mammary gland development and function, including: mammary tissue remodeling during involution, regulation of immune response, insulin signaling pathway, negative regulation of fat cell differentiation, positive regulation of epithelial cell proliferation, and promotion of angiogenesis (Zhao et al. 2002;Sansone et al. 2007;Liu et al. 2011;Singh et al. 2013). The significant difference in the expression of genes encoding these cytokines between HF and LM indicates their important role in the regulation of development and physiological function of the mammary gland.
The results of the microarray experiment also showed highly significant differences in the expression of genes involved in adipogenesis and adipose tissue activity between the examined breeds. In LM mammary glands, a higher expression of genes involved in fatty acid metabolic process, lipid storage, and lipid biosynthesis was observed. Adipocytes are essential for the development of the mammary gland. It has been shown that mammary epithelium lacking paracrine signals from stromal adipocytes forms rudimentary glandular structures with no signs of ductal branching (Landskroner-Eiger et al. 2010). Moreover, it has been suggested that white adipocytes can transdifferentiate into alveolar epithelial cell during pregnancy, whereas the alveolar epithelial cells can transdifferentiate into white adipocytes during mammary gland involution (Morroni et al. 2004). These results provide evidence that adipocyte-to-epithelium transdifferentiation constitutes another mechanism contributing to mammary gland development during pregnancy and lactation.
It is worth noting that our microarray analysis revealed a higher expression of genes involved in the oxidation/ reduction process in LM than HF. The production of reactive oxygen species increases during adipogenic differentiation (Reid et al. 2013;Ogasawara et al. 2009). Differences between HF and LM in the expression of genes involved in redox balance are probably related to the fact that beef heifers have more developed mammary gland adipose tissue. It should be noted that the metabolism of beef cattle is generally more directed towards fat deposition and intramuscular lipid synthesis. The results obtained may also indicate an important effect of fatty tissue and lipid metabolism on dairy potential, but this topic is still poorly understood. The literature data indicate a relationship between adiposity and dairy potential, since fat accumulation and increased growth rate, caused by high feeding level before puberty, can lead to reduced pubertal mammary growth and reduced milk yield potential (Sejrsen et al. 2000). We revealed a higher expression of transcription regulator peroxisome proliferator-activated receptor gamma (PPARG) in HF, which plays an important role in fatty tissue activity and has many regulatory interactions with other significantly regulated genes and processes. Other reports have demonstrated that the expression of the PPARG gene in the mammary gland is correlated with the abundance of adipocytes at several stages of pregnancy, i.e., smaller number of adipocytes in late pregnancy results in lower PPARG expression (Bionaz and Loor 2008b). During lactation, the expression of PPARG is up-regulated, suggesting an important role of this nuclear receptor in milk fat synthesis (Bionaz and Loor 2008b;Kadegowda et al. 2009). PPARG was shown to regulate the expression of genes involved in triacylglyceride synthesis (LPIN1), fatty acids synthesis (ACACA, FASN, SREBF1), metabolism (SCD), and import (CD36) in bovine mammary epithelial cells (Kadegowda et al. 2009). Furthermore, several studies have shown that PPARG regulates signaling pathways, controlling and improving insulin sensitivity, cell proliferation, fatty acid β-oxidation, glucose utilization, adipocytes differentiation, and improves HSC maintenance (Ito et al. 2012;Yuan et al. 2012).

Conclusions
In conclusion, we assume that the higher mammogenic potential in postpubertal dairy heifers in comparison with beef heifers not only depends on central endocrine regulations but also on local intramammary factors, including higher mammary stem cells (MaSC) number and auto-and paracrine regulations of the MaSC environment, forming a favorable niche for their maintenance, self-renewal, and development.