Mammary adipocyte flow cytometry as a tool to study mammary gland biology

The mammary gland is a vital exocrine organ that has evolved in mammals to secrete milk and provide nutrition to ensure the growth and survival of the neonate The mouse mammary gland displays extraordinary plasticity each time the female undergoes pregnancy and lactation, including a sophisticated process of tertiary branching and alveologenesis to form a branched epithelial tree and subsequently milk‐producing alveoli. Upon the cessation of lactation, the gland remodels back to a simple ductal architecture via highly regulated involution processes. At the cellular level, the plasticity is characterised by proliferation of mammary cell populations, differentiation and apoptosis, accompanied by major changes in cell function and morphology. The mammary epithelium requires a specific stromal environment to grow, known as the mammary fat pad. Mammary adipocytes are one of the most prominent cell types in the fat pad, but despite their vast proportion in the tissue and their crucial interaction with epithelial cells, their physiology remains largely unknown. Over the past decade, the need to understand the properties and contribution of mammary adipocytes has become more recognised. However, the development of adequate methods and protocols to study this cellular niche is still lagging, partially due to their fragile nature, the difficulty of isolating them, the lack of reliable cell surface markers and the heterogenous environment in this tissue, which differs from other adipocyte depots. Here, we describe a new rapid and simple flow cytometry protocol specifically designed for the analysis and isolation of mouse mammary adipocytes across mammary gland developmental stages.

The mammary gland is a vital exocrine organ that has evolved in mammals to secrete milk and provide nutrition to ensure the growth and survival of the neonate The mouse mammary gland displays extraordinary plasticity each time the female undergoes pregnancy and lactation, including a sophisticated process of tertiary branching and alveologenesis to form a branched epithelial tree and subsequently milk-producing alveoli. Upon the cessation of lactation, the gland remodels back to a simple ductal architecture via highly regulated involution processes. At the cellular level, the plasticity is characterised by proliferation of mammary cell populations, differentiation and apoptosis, accompanied by major changes in cell function and morphology. The mammary epithelium requires a specific stromal environment to grow, known as the mammary fat pad. Mammary adipocytes are one of the most prominent cell types in the fat pad, but despite their vast proportion in the tissue and their crucial interaction with epithelial cells, their physiology remains largely unknown. Over the past decade, the need to understand the properties and contribution of mammary adipocytes has become more recognised. However, the development of adequate methods and protocols to study this cellular niche is still lagging, partially due to their fragile nature, the difficulty of isolating them, the lack of reliable cell surface markers and the heterogenous environment in this tissue, which differs from other adipocyte depots. Here, we describe a new rapid and simple flow cytometry protocol specifically designed for the analysis and isolation of mouse mammary adipocytes across mammary gland developmental stages.
The mammary gland is a unique organ that evolved in mammals to provide postnatal nutrition to the offspring via lactation [1,2]. Mouse mammary gland development is a highly regulated cyclical process which is repeated for each pregnancy. During puberty, it is manifested by morphological changes, including the formation of terminal end buds, mammary epithelial ductal tree growth and bifurcation. During gestation, due to exposure to hormones such as oestrogen, progesterone and prolactin [3], the mammary epithelium undergoes additional tertiary branching and alveologenesis to form alveoli which will become functional and produce milk to nourish the offspring after birth [1].
The processes of normal mammary gland development require a specific environment, the mammary fat pad, which mostly comprises mammary adipocytes [4,5]. Research has shown that the fat pad not only provides an appropriate stroma for the development of the epithelial tree but may also serve as an endocrine organ to facilitate adipokine and cytokine signalling, similar to other adipose depots [6].
The nulliparous mammary gland contains mostly white adipocytes. Lactation stimulates mammary adipocytes to change their physical properties and lose most of the lipids which provide an initial source of lipids for milk fat production [4]. During involution, a two-stage process of programmed cell death and extensive signalling leads to the first stage of involution, characterised by the activation of lysosomal pathway of cell death, known as lysoptosis. The first phase of involution lasts up to 48 h in the mouse and is triggered by milk stasis in the alveolar lumens. Key proteins regulate this phase, including Stat3, JAK1 and LIF. This process is reversible if suckling resumes at this stage. Subsequently, activation of the second irreversible phase of involution occurs, characterised by the drop in circulating prolactin levels [7]. Throughout involution, mammary adipocytes increase in size and expand to repopulate the tissue in a process that is not fully understood.
In recent years, there has been increased awareness of the unique roles of different cell types within this tissue, which present distinct transcriptional profiles, dynamics and mechanobiology [8][9][10].
Despite the vast proportion of mammary adipocytes in the tissue, most of the studies in the field focus on the mammary epithelium, while the physiology and molecular mechanisms of mammary adipocytes remain largely unknown. In the last three decades, the function of adipose tissue has become recognised, with growing knowledge revealing the cellular diversity and the roles of different types of adipose tissue, namely white, brown and beige adipose. It is now clear that adipose is not solely a lipid storage depot but also an endocrine and thermogenic organ [11], which includes adipose stromal and stem cells with distinct markers related to the anatomical location and pathophysiology of metabolic diseases [12].
At the end of pregnancy and during lactation, mammary adipocytes show unique characteristics which are not observed in normal white or brown adipocytes. Those include unique morphology, compartmentalised lipid droplets, cytoplasmic septa, large mitochondria, cytoplasmic projections, numerous peroxisome and abundant rough ER. Some researchers suggest that during lactation, mammary adipocytes undergo transdifferentiation or adipoepithelial conversion and refer to them as pink adipocytes [13,14]. However, further evidence is required to fully understand the identity of mammary adipocytes during lactation.
New techniques to study adipose were implemented, including immortalised preadipocyte and adipocyte cultures and new mouse models such as Cre-loxP mediated lineage. However, many available methods are appropriate for a largely homogeneous adipose tissue but have not been not explicitly adapted to the mammary gland.
Exploring mammary adipocytes poses various challenges because these cells reside in a highly plastic environment which changes dramatically with development. For example, mammary adipocytes enlarge during midgestation, change their lipogenic capacity and either deplete or de-differentiate after parturition [5,14].
During lactation, their physical properties change dramatically, and their numbers drop, making it more difficult to localise them [15].
In the last few years, several groups successfully isolated fresh adipocytes using flow cytometry [16,17], overcoming challenges such as their large cell size, fragility and buoyancy. However, this was performed in adipose tissue, a very different cellular context which is not directly transferable to the mammary gland.
Here, we developed a protocol designed to address the specific challenges of sorting mammary adipocytes. We mitigate mammary cellular heterogeneity and adjusted our protocol to fit nulliparous tissue as well as adult developmental stages, including gestation, lactation and involution, allowing single-cell exploration of mammary adipocytes.
This protocol works efficiently for all adult developmental stages and allows the isolation of over 10 000 cells in all tested stages. Typical numbers of isolated cells that were obtained from the cell sorter are shown in Fig. 4A.

Materials
Animal handling and tissue preparation

Solutions preparation
Prepare the following solutions as described below: 1 Krebs-Ringer Modified Buffer (KRB) 1 Select an appropriate animal and euthanise it using a CO 2 or isoflurane chamber. If possible, avoid using cervical dislocation as it causes blood accumulation around the cervical mammary glands. For this experiment, use 10-12week-old females (or older) who may be from nulliparous animals, or pregnancy, lactation or involution stages.
2 Place the mouse on its back on a polystyrene surface wrapped with aluminium foil, spread its limbs and pin using 20 G needles. 3 Spray the animal with 70% ethanol. 4 Using serrated forceps, pull the skin above the urethral orifice and make a small incision using fine sharp tip scissors. 5 Using blunt-tip Lexer-baby scissors, perform a blunt dissection to gradually separate the peritoneum from the subcutaneous facia and make a medial cranial incision to the jawline. 6 Make four additional incisions from the midline towards the limbs (Fig. 1A). 7 Dip a cotton bud in PBS 91, slide in the medial incision and roll in both caudal and cranial directions. Perform this bilaterally. 8 Spread the skin and pin to the polystyrene using 20 G needles. 9 Dissect and remove the lymph node from the inguinal mammary gland (#4) using serrated forceps and fine-tip scissors (Fig. 1B). 10 Dissect mammary gland tissue using fine-tip scissors and tissue forceps (Fig. 1B).

Digestion and generation of single cells
1 Place the dissected tissue on ice and add some drops of PBS 91. When dissecting tissue from lactation or early involution stages, wash the tissue several times in PBS 91. Mince the tissue using a scalpel and forceps into very fine pieces of approximately 1 mm.
2 Transfer the tissue fragments to a 50-mL falcon tube with a 10 mL digestion buffer. 3 Incubate the suspension on an orbital shaker incubator for 1 h at 37°C at 100 r.p.m. and make sure the tube rolls freely on the shaker. 4 Pass through a 200 lm mesh PluriStrainer filter and wash the cells by pipetting 10 mL wash buffer. 5 Centrifuge the filtered suspension at room temperature, 150 g, for 5 min. Following the centrifugation, mammary adipocytes will float and form a supernatant layer. 6 Using a p1000 lL tip, pipette the supernatant mammary adipocytes layer and move into a 15 mL falcon tube. 7 Wash the cells by gently resuspending them in 5 mL wash buffer. 8 Harvest the cells by centrifugation at room temperature, 150 g for 5 min. Transfer the adipocyte layer to a 1.5 mL tube ( Fig. 2A).

Cell staining
1 Resuspend the adipocytes in blocking buffer and block the sample for 15 min at room temperature. 2 Centrifuge at room temperature, 150 g, for 5 min. 3 Following centrifugation, mammary adipocytes will float.
Collect the supernatant layer and add wash buffer to 500 lL. 4 Aliquot your sample into 1.5 mL tubes and spare 10% (50 lL) for negative controls.Controls shall include unstained cells and single-channel staining (in this case, AF488 and LipidTox Deep Red). Add 450 lL wash buffer to each control tube to reach an equal volume for the staining. *A smaller volume can be spared for negative controls and pooled together from multiple samples when sorting mammary adipocytes from several animals. 5 The rest of the sample (450 lL) will be used for full staining. 6 Add LipidTox Deep Red to the adipocyte suspension at a 1 : 125 dilution (4 lL into 500 lL) and mix gently by inverting the tube. 7 Add fluorescent antibodies of interest (BP-1-Bio, Ter119-Bio, CD45-Bio, CD31-Bio, CD49f-AF488 and add 1 lL of each antibody into 500 lL). 8 Incubate the cells at room temperature for 15 min and mix every 5 min. 9 Centrifuge cell suspension at 150 g for 5 min. Remove the subnatant from the layer of floating mammary adipocytes. *At this stage, we remove the subnatant rather than transferring the adipocyte layer (supernatant) between tubes to limit the handling and lysis of the fragile mammary adipocytes.    Fig. 3A, using a nulliparous animal. 6 Keep the collection tubes at room temperature and sort the cells into 100 lL wash buffer.

Tips & Tricks
Sample selection and processing Dissecting mammary glands differs between nulliparous mice and gestation/lactation stages. During gestation and lactation, the gland will become larger and thicker, and blood vessels will be more apparent. During gestation, the glands will become pinker in their colour, and during lactation and early involution, become white due to the presence of milk. This does not affect the way the glands are identified and removed. Representative haematoxylin and eosin-stained sections of some of the sorted samples show a significant difference in the histology of the tissue (Fig. 2C).
Identifying the lymph nodes in the inguinal mammary gland (#4) can be challenging during lactation and early involution due to milk engorgement. We recommend practising on nulliparous mammary glands before moving to the latter stages.
The preparation of the single-cell suspension is a key factor in the efficiency and reproducibility of this protocol. Reduced digestion time or temperature will result in a lower yield of cells.
During the digestion, it is important to ensure that the falcon tube containing the tissue and digestion buffer can move freely on the orbital shaker.
Avoid vortexing the cells, as mammary adipocytes tend to be fragile, and vortexing may result in cell lysis.
Versene is used to minimise the sticky nature of the sample during the sorting stage at the flow cytometer; however, heparin may be used as an alternative. Using Versene instead of FBS provides a better signal-to-noise ratio in flow cytometry.

Staining and flow cytometry
This protocol uses antibodies for cell surface Alexa Fluor 488 staining which here labels CD45, CD31, CD49f, Ter119 and BP-1 that were chosen to exclude immune, endothelial and epithelial cells, erythrocytes and tumours based on well-established protocols [18]. LipidTox Deep Red fluorescent dye was chosen for labelling adipocytes and neutral lipid droplets based on a previously published flow cytometry protocol for adipose tissue [16]. Further optimisation may be required to explore different subtypes of mammary adipocytes.
In some flow cytometers, using a neutral density (1.5 ND) filter added in front of the FSC detector improved the detectability of mammary adipocytes. ND filtering allows the detection of large particles, events that may be off the scale on the FSC axis. By decreasing the FSC signal, the events corresponding to mammary adipocytes are kept on the scale.
In addition to the ND filter, setting up a high FSC threshold allows the detection of larger events. Both options may be used to improve the detection of mammary adipocytes.
Mammary adipocytes may have very high autofluorescence. Therefore, it is essential to use unstained cells and single-channel staining to set up the gates. Using fluorescently tagged beads will result in unreliable cell sorting.
The buoyancy of mammary adipocytes will result in phase separation of the cells and the Versene as soon as the sample is still. As soon as phase separation occurs or a drop in detected events is observed, pause the cell sorting, close the 1.5 mL tube and flick it until the sample becomes homogenous. On average, this is required every 5 min.
When sorting mammary adipocytes from lactation or early involution stages, the sample may be cloudy due to the presence of milk or milk fat globules in the sample. These events may be gated out as they appear in the lower end of the FSC signal. Additionally, the number of events per second is expected to be very high, and the efficiency would be lower than usual. Figure 3B shows typical flow cytometry plots for gestation Day 9.5, lactation Day 5 and involution Day 1.

Considerations of cell numbers
Throughout the development of the mammary gland, the tissue undergoes substantial morphological changes, affecting the size of the tissue and the number of cells. Specifically, mammary adipocytes make up a significantly smaller proportion of mammary tissue during gestation and lactation, when milk-producing alveoli populate the tissue. During involution, mammary adipocytes become visible and gradually contribute more to the cellular composition of the gland. These changes affect the number of sorted mammary adipocytes at those developmental stages and the efficiency of the cellsorting process. Since the tissue becomes larger during gestation and lactation and remains relatively large at early involution, the number of sorted adipocytes from all glands is at least 10 000 cells, even when fewer mammary adipocytes are visible in the tissue. Figure 4A shows the number of sorted mammary adipocytes for each time point.

Validation of sorted populations
It is highly advisable to use several independent validations for this protocol: ○ Use fluorescent microscopy to visualise the stained cells before and after sorting them (Fig. 4B, C, respectively). You may wish to add DAPI to a small aliquot of the sorted cells to ensure that you have sorted cells rather than lipid droplets (Fig. 4C). ○ Extract RNA and perform qPCR validations on the sorted fractions, including negative cell populations, to ensure that the sorted mammary adipocytes express adequate markers, such as Adipoq, and are not expressing epithelial markers, such as Krt5. Figure 4D shows a typical expression profile of the different populations, demonstrating relevant markers highly expressed in sorted mammary adipocytes. Adiponectin was expressed 54-fold higher in adipocytes compared with Alexa Fluor 488 + cells, while MCAM, also known as CD146, was expressed 10-fold higher. We suggest the following primer pairs in Table 1 for qPCR validation.

Advantages and limitations
The current protocol provides an efficient way to isolate mammary adipocytes. To date, adipocyte cell sorting is not a common practice in this field due to technical difficulties, the physical properties of these cells and non-standard cell-sorting parameters. Existing protocols to isolate adipocytes were developed for adipose tissue rather than for mammary glands, which differ substantially and present a challenging cellular heterogeneity. This unique protocol has been developed specifically for the mammary gland. It addresses the challenges this tissue presents, including isolating a relatively small population of mammary adipocytes during lactation and involution.
The experimental design proposed in this protocol can be further optimised for culturing mammary adipocytes following cell sorting. This will require the addition of live/dead staining, a standard practice when culturing isolated cells. In our hands, live/dead staining such as DRAQ5 or propidium iodide caused artefacts and autofluorescence, which compromised the compensation in the cell sorter. Therefore, we chose not to use live/dead staining. Further adaptations of this protocol will be required for subsequent cell culturing.
Further optimisations may be required to characterise the isolated population of mature adipocytes in flow cytometry, for example using cell surface markers such as CD36, a marker of preadipocytes and mature adipocytes and Sca-1/CD34, a marker of stem cell niche.