Mammary Epithelial Cell Hierarchy in the Dairy Cow Throughout Lactation.

The plasticity of the mammary gland relies on adult mammary stem cells (MaSCs) and their progenitors, which give rise to various populations of mammary epithelial cells (MECs). To face global challenges, an in-depth characterization of milk-producing animal mammary gland plasticity is required, to select more sustainable and robust dairy cows. The identification and characterization of MaSC and their progenitors will also provide innovative tools in veterinary/human medicine regarding mammary tissue damage (carcinogenesis, bacterial infections). This study aimed to determine the dynamics of mammary cell populations throughout a lactation cycle. Using mammary biopsies from primiparous lactating dairy cows at 30, 90, 150, and 250 days of lactation, we phenotyped cell populations by flow cytometry. To investigate cell lineages, we used specific cell-surface markers, including CD49f, CD24, EpCAM (epithelial cell adhesion molecule), and CD10. Two cell populations linked to milk production were identified: CD49f(+)/EpCAM(-) (y = 0.88x + 4.42, R(2) = 0.36, P < 0.05) and CD49f(-)/EpCAM(-) (y = -1.15x + 92.44, R(2) = 0.51, P < 0.05) cells. Combining immunostaining analysis, flow cytometry, daily milk production data, and statistical approaches, we defined a stem cell population (CD24(+)/CD49f(+)) and four progenitor cell populations that include bipotent luminal progenitors (CD24(-)/CD49f(+)), lumino-alveolar progenitors (CD24(-)/EpCAM(+)), myoepithelial progenitors (CD24(+)/CD10(-)), and lumino-ductal progenitors (CD49f(-)/EpCAM(+)). Interestingly, we found that the bipotent luminal progenitors (CD24(-)/CD49f(+)) decreased significantly (P < 0.05) during lactation. This study provides the first results of mammary cell lineage, allowing insight into mammary cell plasticity during lactation.


Mammary epithelial cell hierarchy in the dairy cow throughout lactation
Marie-Hélène Perruchot 1

INTRODUCTION
Milk yield (MY), and therefore the sustainability of livestock, is determined by breeding strategies and is finely regulated by cellular mechanisms determining the number and the activity of secretory cells [1] within the mammary gland. The mammary epithelium is composed of two cell layers: a luminal layer, lining the lumen of ducts and alveoli, and an external basal myoepithelial layer [2]. The stroma compartment that surrounds this epithelium is composed of several other cell types (adipocytes, fibroblasts and immune cells) that are responsible for mammary plasticity and homeostasis. During lactation in the dairy cow, the mammary tissue undergoes extensive morphological and functional remodeling, linked to cellular and architectural changes [3]. During early lactation, the rate of proliferating mammary cells overtakes the rate of apoptosis, thereby increasing the number of mammary cells. The decline in MY during late lactation is due to loss of mammary cells (apoptosis) and remodeling of mammary tissue. This results in alveolar regression during involution [4]. To date, our understanding of the phenotypic differentiation of mammary epithelial cells is limited. Current knowledge in mice and humans indicates the existence of stem cell and/or bipotent stem/progenitor cell populations that give rise to mature/differentiated cells [5]. The cleared mammary fat pads [8][9][10]. Furthermore, it was shown that epithelial cells isolated from virgin, pubertal, gestational, and lactating glands could all repopulate cleared fat pads [9]. As such a variety of parenchymal structures can regenerate a complete parenchymal framework, it is likely that progenitor or stem cells are present throughout the mammary epithelium. To date, MaSC studies in ruminants have been very sparse. Recently, Baratta et al. (2015) investigated different epithelial cell subpopulations in milk cells by flow cytometry [11]. They found an increase in CD49 f + cells and cytokeratin 14/18 cells at the end of lactation in dairy cows. In 2002, Ellis and Capuco were the first to provide data on the proliferation rate of mammary epithelial cells (MEC), using histological analysis of mammary gland explants from heifers, which were injected with BrdU before euthanasia [6]. They identified different staining intensities in mammary epithelial cell populations, and proposed that the lightly stained cells were the primary proliferative cell population. In 2010, Martignani et al. showed that a putative bovine mammary stem cell population was able to regenerate and maintain a complete mammary gland structure [12]. They hypothesized that these putative MaSC can Flow cytometry with specific cell-surface markers is a sensitive approach to lineage and quantify cell populations and subpopulations in the mammary gland [7]. The expression of cluster of differentiation (CD) molecules, such as heat stable antigen (CD24) and integrin alpha-6 (CD49 f ) have been widely used to identify epithelial cells in the mammary gland [13,14]. Recently, four populations have been found in the developing mammary gland subpopulations [15]. In the human mammary gland, the combination of the cell-surface markers EpCAM (Epithelial Cell Adhesion Molecule) and CD49 f has served to identify basal cell populations [16,17]. In lactating goats, EpCAM has been found by immunostaining in luminal populations of alveoli, and seems to be related to cell proliferation [18].

RT-qPCR Procedure
Total RNA was extracted from tissue samples after grinding them in liquid nitrogen with a mortar and pestle using Trizol (Invitrogen, Paris, France The amplification program consisted of an initial denaturation step at 95°C for 10 min followed by 40 cycles of denaturation at 95°C for 10 sec and combined primer-annealingextension at 60°C for 1 min, during which fluorescence was measured. A melting curve was   MACSQuantify analysis software (Miltenyi Biotec). Results were expressed in percentages (dot plot analysis).

Statistical analysis
Each experimental condition was repeated three times. Milk composition, flow cytometry and qPCR data were subjected to analysis of variance (one-way ANOVA) using the following model: y ij = µ + time i + ε ij (y=flow cytometry data or qPCR data; µ= mean; i= time of biopsies and ε=residuals). Post-hoc Tukey pairwise comparisons were used. Differences were considered significant at p<0.05. All statistical analyses were performed using R software (the R foundation). This article has been peer-reviewed and accepted for publication, but has yet to undergo copyediting and proof correction. The final published version may differ from this proof.

Milk yield of primiparous lactating dairy cows
Daily MY was recorded for the five primiparous cows, from calving to the drying-out period.
As expected, the MY was significantly different at each mammary gland biopsy timepoint of mammary biopsies (30, 90, 150 and 250 days) with 29.9 kg/day, 34.3 kg/day, 32.9 kg/day and 19 kg/day, respectively (p<0.05) ( Figure 1). The mean of daily MY described a classical lactation curve, with a peak of production at ≈ 60 days post-calving. Analysis of milk composition data showed that fat and protein yields varied according to MY, as expected during lactation in primiparous dairy cows (Table 2). Furthermore, somatic cell count (SCC) values were less than 250x10 3 cells / mL throughout lactation, showing the cows and their mammary glands were healthy. This prerequisite was the first step to validate our experimental model.

Mammary gland morphology and differentiation during lactation
The second step to validate our experimental approach was to examine the morphology of the mammary gland tissue using eosin/hematoxylin staining of tissue sections at each biopsy timepoint (

In situ localization of cell-surface markers in mammary gland tissue
To estimate the relevance of the four lineage markers used in the present study (CD10, CD49 f ,

Gating strategy for flow cytometry data analysis
To validate the multicolor panel of the four antibodies used in the present study, coupled with four different fluorochromes, we performed single cell staining with each antibody (Table 3) and Fluorescent Minor One. To develop this technique, essential to set up our multicolor assays, we used a manual compensation matrix based on the median fluorescence intensity procedure. We also ensured that only single living cells were analyzed, by eliminating cell aggregates and dead cells from our analysis. In order to eliminate aggregates, we performed doublet exclusion using FSC-H/A and SSC-H/A parameters. To discriminate between live cells and dead cells, we used PI staining. This technical strategy is essential to analyze each cell-surface marker in single living cells ( Figure 5A). The strategy to analyze the expression of the three others cellular markers used in the present study within a stained population (here, the CD49f positive cells) is illustrated in Figure 5B. In our example, we gated on CD49fpositive cells to analyze co-immunostained populations: CD49f and EpCAM, CD49f and CD10, CD49f and CD24. This procedure was applied for each cell-surface marker (CD49f, CD10, CD24 and EpCAM).

Cell hierarchy and lineage in the bovine mammary gland during throughout lactation
After tissue dissociation and single cell isolation, we determined the proportion of each  Figure 7B and 7C). Indeed, the correlation between CD49 f -/ EpCAMcells and milk production was negative (R 2 = 0.51, p<0.05, Figure 6B), whereas the proportion of CD49 f + / EpCAMcells positively correlated with milk production (R 2 = 0.35, p<0.05, Figure   7C).

DISCUSSION
During the development of the mammary gland (mammogenesis) and lactation (galactopoiesis) in dairy cows, each cell population has a specific feature involved in mammary gland plasticity. Like most glandular tissue, the adult mammary gland is composed of many cell types that interact to shape the organ and make it functional. In this study, we specifically focused on epithelial-like, myoepithelial and progenitor/stem cells implicated in mammary gland plasticity. Flow cytometry is a suitable method to analyze and to discriminate between single cell phenotypes on precious and small samples such as mammary gland biopsies. We used this method to analyze and to identify the lineage of cell populations based on the expression of cell-surface markers commonly used in the mammary gland field: CD49f, CD24, CD10 and EpCAM. The prerequisites for our study were to ensure that mammary explant samples were representative of a lactating mammary gland. Using five primiparous dairy cows, we first monitored their zootechnical data in order to ensure their This study provides a dynamic view of the mammary cell populations based on common markers used in mammary gland cell lineage identification. Our next goal will be to carry out the same analyses on mammary glands at different physiological stages: before and after puberty, during gestation, as well as at the peak of lactation and at the drying-out stage.
An analysis of functional proteins (cytokeratins, steroid receptors, aldehyde dehydrogenase activity) will be performed on sorted cells, isolated from the different populations described in this study and combined with mammosphere assays. In summary, this study not only contributes to a better understanding of the role of stem cells in the plasticity of the mammary gland, but also identifies new biomarkers for veterinary and human medicine.        Table 1: List of antibodies used in flow cytometry analysis and immunohistology.             been peer-reviewed and accepted for publication, but has yet to undergo copyediting and proof correction. The final published version may differ fr         been peer-reviewed and accepted for publication, but has yet to undergo copyediting and proof correction. The final published version may differ fr   This article has been peer-reviewed and accepted for publication, but has yet to undergo copyediting and proof correction. The final published version may differ from this proof.