| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Immunology; |
Program in Cell Biology, Department of Pediatrics, National Jewish Medical & Research Center, Denver, Colorado 80206
ABSTRACT
Following the cessation of lactation, the mammary gland undergoes a physiologic process of tissue remodeling called involution in which glandular structures are lost, leaving an adipose tissue compartment that takes up a much larger proportion of the tissue. A quantitative morphometric analysis was undertaken to determine the mechanisms for clearance of the epithelial cells during this process. The involution process was set in motion by removal of pups from 14-day lactating C57BL/6 mice. Within hours, milk-secreting epithelial cells were shed into the glandular lumen. These cells became apoptotic, exhibiting exposure of phosphatidylserine residues on their surfaces, activation of effector caspase-3, staining for caspase-cleaved keratin 18, loss of internal organellar structure, and nuclear breakdown, but minimal blebbing or generation of apoptotic bodies. Clearance of residual milk and the shed epithelial cells was rapid, with most of the removal occurring in the first 72 h. Intact apoptotic epithelial cells were engulfed in large numbers by residual viable epithelial cells into spacious efferosomes. This process led to essentially complete involution within 4 days, at which point estrous cycling recommenced. Macrophages and other inflammatory cells did not contribute to the clearance of either residual milk or apoptotic cells, which appeared to be due entirely to the epithelium itself.
apoptosis, efferocytosis, immunology, involution, mammary glands, mammary involution, phagocytosis
Apoptosis is important for tissue remodeling throughout the body, both during development and in the adult. Normally, apoptotic cells are rapidly removed in situ with minimal local tissue response and without inflammation. Recent evidence indicates that the apoptotic cells are recognized by neighboring cells and engulfed by what appear to be unique ingestion mechanisms, recognition molecules, receptors, and signaling pathways [1, 2]. These processes are highly conserved in multicellular organisms and have led us to coin the term efferocytosis (from effero, to carry to the grave, to bury) to describe them [3]. Despite significant evidence to the contrary, including apoptotic cell removal in organisms that do not have macrophages and during embryonic development before macrophages are present, there is a prevailing perception that apoptotic cell removal in vivo is largely mediated by macrophages. Studies in vitro have demonstrated that a wide variety of cells can recognize and engulf apoptotic cells, and that candidate receptors for apoptotic cell recognition are very widely distributed. Whether the mechanisms of uptake are different between so-called professional phagocytes (mononuclear phagocytes and immature dendritic cells) and tissue cells is not completely clear. Evidence for differences has been presented [4], but could largely reflect rates of uptake rather than qualitative differences. Other studies, including those from our own laboratory, suggest that uptake mechanisms, receptors, and signaling are conserved, not only evolutionally but also between different cell types.
The involuting mammary gland offers a unique system for examination of physiologic tissue remodeling and removal of apoptotic cells in an adult mammal because the process can be synchronized by weaning of the pups. Previously, the process of involution has been suggested to occur in three stages [5]: 1) cessation of milk secretion; 2) collapse of the alveoli with apoptosis of the milk-secreting alveolar cells and clearance of both residual milk and the apoptotic cells; and 3) regrowth of the stromal adipose tissue and return to a structure similar to the postpubertal, virgin gland, i.e., a bare, ductal tree with regressed end buds. However, the processes mediating the clearance phase are poorly understood.
Following weaning, it has previously been shown that epithelial cells undergo apoptosis [6, 7], and both macrophages and epithelial cells have been implicated in their clearance (see [8]). In vitro, both primary mammary epithelial cells and mammary epithelial cell lines have been shown able to act as phagocytes, in the sense that they bind and engulf apoptotic cells [9]. These cells express many of the receptors employed by so-called professional phagocytes, including structures that recognize phosphatidylserine (PtdSer), CD36, integrin alpha V (ITGAV), and LRP1 (low-density lipoprotein related receptor/CD91/alpha-2-macroglobulin receptor/TβRV). A likely important stimulus for LRP1-mediated uptake, calreticulin (CALR) [10], is present in the milk, [11] as is MFGE8 (milk fat globule-EGF factor 8 protein, also known as lactadherin). This molecule has been shown to participate in apoptotic cell uptake as a bridge molecule between PtdSer on the apoptotic cell and integrins on the engulfing cell [12]. Thrombospondin (THBS1) has also been implicated as a bridge molecule, in this case to CD36, as have ITGAV and/or LRP1. When mammary epithelial cells in vitro were exposed to apoptotic cells, they exhibited membrane ruffling, rapid cytoskeletal actin reorganization, and macropinocytosis [9], as described previously for macrophages and fibroblasts [13, 14]. In addition, following interaction with apoptotic cells, mammary epithelial cells secreted transforming growth factor β (TGFB1) and suppressed proinflammatory cytokine release, again a feature reported for both macrophages and other tissue cells responding to apoptotic cells. Accordingly, we hypothesize that the mammary epithelial cells likely play a key role in removal of cells during involution.
To address this hypothesis, we undertook a detailed and quantitative study of mammary gland involution and the processes by which the epithelial cells die, are shed into the lumen, and are subsequently cleared. Remarkably, these highly efficient processes were found to be essentially complete within the first 4 days after weaning and appeared to be mediated by the epithelial cells themselves without evidence of inflammation or involvement of macrophages.
Mouse Care and Induction of Mammary Involution
C57BL/6 males and females were ordered from Harlan Laboratories for all studies except the macrophage counts, which were C57BL/6 and FVB from Taconic Farms. All mice were housed in micro-isolator cages with filter tops and subjected to 12L:12D cycles. Autoclaved, aspen-chip bedding was provided, and LabDiet-5020 (PMI, Richmond, IN) was fed ad libitum. The Institutional Animal Care and Use Committee of National Jewish Medical and Research Center approved all procedures.
The estrous cycles of six C57BL/6 females per week, age 10–12 wk, were synchronized by exposure to male mouse urine. Approximately 60 h later, females were placed with stud males until a vaginal plug was observed. The females were then housed singly and given cornhusk and cotton nesting material. Females birthed 19 days after the observation of a plug. Within 2 days of parturition, pups were culled or cross-fostered to standardize the litters to eight pups. Litters were weighed and inspected daily, Monday through Friday, for lost pups and evidence of lactation failure. Females with lost litters were allowed 21 days of involution, remated, and used on the second lactation.
Utilizing this strategy, 15% of the mice had failed first litters. Another 23% lost one or more pups during lactation, and at least 18% had evidence of premature involution of one or more glands at the time of dissection, illustrating the difficulty of using C57BL/6 mice for mammary studies.
To force mammary involution, all pups were removed in the morning of Lactation Day 14. Tissues were collected at 12-h time points after weaning.
Mice were anesthetized with sodium pentobarbital (0.1 mg/g body weight) and perfused intracardially with 20 ml ice-cold PBS (with Ca++/Mg++) followed by 20 ml each of 2% and 4% paraformaldehyde (PFA) in PBS (with Ca++/Mg++). Both #4 mammary glands (inguinal mammary glands) were removed and weighed. Both #3 glands were removed. All were further fixed in 4% PFA for 4 h. One #4 gland was sent to our histology lab for routine paraffin embedding. The other three glands were frozen between sheets of aluminum foil, in isopentane cooled with liquid nitrogen, and stored at –80°C until sections could be cut.
Frozen, fixed mammary tissue was sectioned and collected onto Cell Tak (BD Biosciences)-coated coverslips (coated using the blood-smear technique given by the manufacturer). The sections were brought to room temperature inside a humid chamber containing a 4% PFA-soaked paper towel. They were further vapor-fixed in this fashion for 20 min. Sections were gently wetted with PBS and then permeabilized with 0.2% Triton X-100 in PBS (no Ca++) for 5 min. Sections were washed with PBS and blocked with 10% donkey serum in PBS (sterifiltered) for 20 min. They were stained with FITC-conjugated M30 CytoDEATH (1:100; Roche Biochemicals), rabbit anti-active caspase 3 (1:100; Cell Signaling), rat anti-mouse EMR1 (EGF-like module containing, mucin-like, hormone receptor-like sequence 1, also known as F4/80; 1:100; Caltag), guinea pig anti-adipophilin (1:100; Research Diagnostics, Inc.), or rabbit anti-perilipin (1:100; Affinity Bioreagents, Inc.) in PBS for 1 h and then washed with PBS, six changes, 5 min each. Primary staining was visualized or enhanced by incubation with secondary antibodies: Alexa 488-conjugated, anti-FITC antibody (1:100; Molecular Probes), fluorescently conjugated donkey anti-rabbit IgG, donkey anti-rat IgG2, or donkey anti-guinea pig IgG (1:100; Jackson ImmunoResearch). Sections were counterstained with Rhodamine- or FITC-phalloidin (1:100; Molecular Probes) and Hoechst 33258 (0.5 µg/ml) for 1 h. The sections were soaked overnight in PBS for the final wash and then mounted with a glycerol-based mounting medium containing o-phenylenediamine antifade. Coverslips were sealed with nail polish and stored at –20°C until able to view.
Images were collected on a Marianas system (Intelligent Imaging Innovations, Inc., Denver, Co) using SlideBook Software version 4.0 (Intelligent Imaging Innovations). The hardware components of the system included a Zeiss Axiovert 200M microscope stand, a Sutter DG-4 excitation source and Sutter Lambda 10–2 filter wheels, an x-y stage controller from ASI, and a Cooke Sensicam, cooled CCD camera. The filters were from Chroma and included standard DAPI, FITC, and Cy3 sets. The Zeiss objectives used included a x10 Fluar (N.A. 0.3), x40 Plan-Neofluar (N.A. 1.3), x63 Plan-Apochromat (N.A. 1.4) and x100 Plan-Apochromat (N.A. 1.4).
Apoptotic cell counts. Two nonadjacent sections from one #3 or #4 gland from each mouse were stained. Tissues stained with an antibody to active caspase-3 were used to ascertain the number of apoptotic cells in the gland in the days following forced-weaning involution. Five fields from each section were imaged using a x10 objective. Sections were chosen that spanned the entire breadth of the section. The imaging was biased toward the middle of the gland in the direction of sectioning only and toward portions of the gland that were intact and unfolded. A combination of positive staining and morphology was used in the cell counting: only whole cells (diameter >8 µm) were counted. Fragmented cells were not counted. Because the tissue was never dried and there was no difference between the wet weights of the fresh and perfusion-fixed tissue, no adjustments for tissue shrinkage were made.
Actual section thickness (14.6 ± 2.5 µm) was measured with the SlideBook software using the x100 oil objective, correcting for the numerical aperture of the lens. The actual size of the shed, apoptotic cells (median 15.2 µm, range 8.5–25.6 µm) was determined by imaging cells isolated from the milk (see method below). Since the shed cells themselves are regular in shape, the correction method of Abercrombie was used [15]. The number of apoptotic cells per mm3 tissue was calculated from the number observed in each field, correcting for the section thickness and the size of the cell: number = count x (section thickness)/(section thickness + cell height).
Adipose and milk volume. Two sections from the #3 or #4 gland from each mouse were stained for adipophilin and perilipin. Fields spanning the whole section were captured. One field per section was randomly chosen for analysis using a random number table. Masks delineating luminal space, epithelium, and adipose tissue were created by hand on each field using the histology of the above immunostains. Area percent (=volume percent) of each compartment was calculated, and, using the measured volume of the whole gland, the volume of each compartment was determined. Two methods were used to show that perfusion fixation did not cause shrinkage or swelling of the tissue. First, the density of apoptotic cells in the milk space as measured by morphometry was comparable to the concentration of cells in freshly collected milk: 10.2 million (±3.1 SD, n = 3), and 8.7 million per ml (±8.0 SD, n = 11), respectively. Second, the wet weights of the fresh and perfusion-fixed glands were comparable at every developmental time point (data not shown).
Fresh tissue volumes of both #4 glands were measured by the water displacement technique from four mice per time point [16].
Mouse Milking and Milk Processing
A female mouse 1 or 2 days postwean was anesthetized with an i.p. injection of Avertin (Tribromoethanol, Sigma-Aldrich #T48402, 640 µg/g body weight). Upon loss of toe-pinch reflex, the mouse was laid on her back and injected intraperitoneally with oxytocin (0.01 USP units per g body weight = 0.2 ml of a 1:10 dilution). Upon let-down, the teat plug was cleared by gentle manual pressure and milk was collected into a small tube on ice in a side-arm flask using the cut tip of a 10-µl pipette tip, with the house vacuum line to provide suction.
Mouse milk was mixed 1:1 with 10% sucrose in saline and then layered under x4 volume of sterile saline. The milk was centrifuged at 1500 x g at 15°C for 20 min. The milk fat globules were removed with a spatula, the saline layer by gentle suction. The apoptotic cells were gently resuspended in PBS and centrifuged at 1500 x g for 5 min.
C57BL/6 females were injected intraperitoneally with 4% thioglycolate aged 1 yr [17]. Twenty-four hours later, the mice were killed with CO2, and the peritoneal cavity was lavaged with 10 ml Hanks Balanced Salt Solution. The cells were collected by low-speed centrifugation. Staining was performed on peritoneal cells and mammary cells isolated from the milk. FITC-conjugated anti-EMR1 (Clone: BM8, Caltag) and FITC-conjugated rat IgG2b, as well as Alexa 488-conjugated anti-LY-6G and Alexa 488-conjugated rat IgG2a (BD Pharmingen), were each used at 1 µg/ml to stain 1 000 000 cells.
Quantitation of Tissue Macrophages
Frozen sections collected as previously described were permeabilized with 0.2% Triton-X100 and blocked with 5% donkey serum. FITC-conjugated rat anti-mouse EMR1 (clone: A3–1, Serotec) was used at 1:50. Co-staining was done with Alexa 546-conjugated phalloidin and Hoechst 33258. Imaging was performed on an epifluorescent, multifluor imaging, deconvolution, and analysis system with Slidebook software. Three C57BL/6 and three FVB mice were used for each time point. Ten random images were taken of each slide at x10 magnification. Images were all set at the same fluorescent intensity for macrophage stains and converted to TIFF files with Adobe Photoshop, and the number of macrophages per x10 field was counted.
Determination of Mouse Estrous Cycling
Vaginal smears were performed according to published methods [18]. Briefly, sterile saline was used for vaginal lavage. The lavage fluid was subjected to cytospin, fixed, and stained with Hema 3 (Fisher, a methanol fix, and a modified Wright-Giemsa stain). Estrous stage (metestrus, diestrus 1, diestrus 2, proestrus, estrus) was subjectively determined by the number of cornified epithelial cells, nucleated epithelial cells, ovoid cells, and neutrophils. Serial cytology was performed daily for 21 days postwean.
Histological Changes in the Mammary Gland
Involution was initiated by forced weaning at midlactation, leading to abrupt cessation of milk secretion and coordinated regression of the proximal and distal parts of the gland. Specifically C57BL/6 mice were subjected to timed matings, litters were standardized to eight pups, and the pups were weaned on Day 14 postpartum to initiate the involution. As shown in Figure 1A, regression did not occur immediately, as evidenced by the continued weight gain of the gland during the first 24 h. Others have noted that milk secretion continues for a period of time and that involution in the mouse is reversible for up to 24–36 h, possibly depending on the strain of mouse (C. Palmer, personal communication). After 24 h there was a rapid loss of gland weight, reaching a nadir by Day 4 postwean, at which time the gland was back to its basal, prelactation weight. Histological analysis of tissue sections spanning this clearance phase showed a loss of luminal contents and epithelial elements, as well as a reappearance of stromal adipose tissue (Fig. 1B). Frozen sections of the mammary gland stained with an antibody to perilipin, the major protein coating the cytoplasmic lipid droplets in adipose cells (but not those in the mammary epithelial cells), showed a large accumulation of lipid between Days 1 and 3 postwean, i.e., the redevelopment of adipose tissue (Fig. 2A). Also obvious was the concomitant loss of luminal contents along with staining for adipophilin, which is present on lipid droplets in the milk-secreting epithelial cells [19]. However, standard examination of histological sections provided only a limited view of the changes occurring in the whole gland.
|
|
Using morphometric analysis combined with tissue volume measurements, we determined the volume of the residual milk in the gland, as well as the total volume of adipose tissue stained with anti-perilipin antibody at various time points postweaning. As seen in Figure 2B, at 24 h postwean (Involution Day 1 [I-1]), the #4 mammary gland of C57BL/6 mice, weighing on average 0.9 g (Fig. 1A and Table 1), contained more than 400 µl of residual milk, and 60% of the gland volume was luminal space. This milk space is a significant increase from the prewean status of the gland, representing the continuing milk secretion noted above. However, the luminal volume dropped rapidly to less than 0.5% of the gland by Day 4. Conversely, the amount of adipose tissue in the gland increased from 6% of the total gland volume on Lactation Day 14, when the pups were removed, to 60% by Day 4 postweaning. Since the total volume of the gland was dropping during this period of time, the rise in the total volume of adipose tissue in the gland was actually only about 2-fold. This fact is not evident from the histological data.
|
Apoptosis of the Epithelial Cells
Histologic examination showed that cells with the appearance of apoptotic epithelial cells accumulated within the lumina of the gland during the involution process (Figs. 1B and 3B). Electron micrographs of the cells during shedding (Fig. 3A) amplify this point. At 24 h these cells already have characteristics of early apoptosis, including disrupted attachment, loss of secretory pathway organelles including RER, Golgi and secretory vesicles, and condensation of nuclei. They also stain positive for caspase-cleaved keratin 18 (Fig. 3C), demonstrating their epithelial origin and supporting an apoptotic, as opposed to necrotic, cell death. By contrast, cytoplasmic lipid droplets, both small and very large, were maintained in the shed epithelial cells at this time. Some cells isolated from the milk at Involution Day 1 were still able to discharge these lipid droplets (data not shown), although whether this release mimics active milk fat secretion is not known.
|
Another hallmark of apoptosis is the externalization of PtdSer on the plasma membrane. Accordingly, cells isolated from the milk 24 h after weaning were examined for annexin V binding. As shown in Figure 3D, cells with condensed, apoptotic nuclei showed high levels of annexin V binding, whereas cells in the same population with normal-appearing nuclei did not. Approximately 90% of the cells isolated from the milk 24 h postweaning had high annexin V binding (Fig. 3E). The presence of a few apparently "normal" cells with low PtdSer exposure in this population suggests that the epithelial cells may acquire characteristics of apoptosis after being shed, and may progress to late stages of apoptosis within the lumen prior to removal. At Day 1 postwean, approximately 35% of the shed cells were anucleate, suggesting destruction by caspase-activated DNase, a late-stage marker of apoptosis. A similar process of epithelial cell shedding and cell death has been described for MDCK cells and chick epithelium undergoing apoptosis [20].
The appearance of apoptotic cells in the involuting gland was mapped using morphometric analysis after immunostaining. Figure 4A shows sections of mammary glands stained with an antibody against cleaved (active) caspase-3 (red). On Day 1, cells shed into the luminal space were positively stained, but most of the cells remaining in the alveolar wall were not, and there was little evidence of ingested, stained material within them. On the other hand after 3 days, there was a marked increase in positive cells in the lumen, along with significant accumulation of active caspase-staining material in efferosomes (specialized phagolysosomes) within remaining alveolar wall epithelial cells. The cells were also positive for caspase-cleaved keratin 18, an apoptosis marker for epithelial cells (see Fig. 3C). Counting the number of apoptotic cells per microscope field suggested that the peak of apoptosis was between Days 2.5 and 3 (Fig. 4B). However, these numbers really represent a peak in the density of apoptotic cells. When the total number of apoptotic cells in the whole gland was calculated (using the total volume of the gland and the analyzed volume), the number of such cells was shown to be high even at Day 1 postwean (Fig. 4C). The increased appearance or higher density of apoptotic cells in the gland at Day 3 results from the clearance of milk from the gland.
|
Clearance of Milk and Epithelial Cells During Involution
As noted, luminal milk was cleared from the gland within 3 days. Uptake of milk fat globules by mammary epithelial cells can be demonstrated in vitro (Monks, unpublished results). It seems likely that this is also a major mechanism for their clearance in vivo and globules were observed within remaining viable epithelial cells, though whether due to uptake or continuing intracellular production could not be determined. On the other hand, large numbers of intact apoptotic epithelial cells were seen within these remaining viable epithelial cells during the clearance phase (Fig. 3B). Many rounded, detached cells with condensed nuclei were seen within the lumen, and nearly every remaining cell within the epithelium contained an efferosome with a whole apoptotic cell or cellular material within it. The shed and ingested cells were verified to be apoptotic alveolar epithelial cells by staining with an antibody to caspase-cleaved keratin 18 (Fig. 3C). These data show definitive evidence of apoptotic epithelial cell clearance by epithelial cells.
In order to determine the period of most rapid cell clearance, the fraction of whole apoptotic cells contained within efferosomes in the epithelial layer was counted. The highest percentage of ingested apoptotic cells was observed on Day 4 postwean (Fig. 5). However, again extrapolating to the whole gland, the greatest number of cells was shown to be engulfed on Days 2 and 3 postwean. In addition, these calculations cannot take into account the rate at which engulfed apoptotic epithelial cells (or milk fat globules) are digested, shown to be complete by 8–10 h in the macrophage [21].
|
Lack of Evidence of Clearance by Inflammatory Cells
It has been suggested that both epithelial cells and macrophages play a role in removal of apoptotic cells during involution of the mammary gland [22, 23], and luminal contents during involution have been reported to contain both macrophages and neutrophils [22, 24–27]. In contrast, in the studies reported here, clearance of apoptotic cells appeared to be mediated solely by remaining epithelial cells. To examine this discrepancy more carefully, frozen sections were stained with an antibody against EMR1 (Fig. 6A), and macrophages were counted in sections. During the clearance phase of the involution period, fewer than five macrophages per x10 field were detected, and no macrophages were seen on the luminal side of the alveolar basement membrane, i.e., where all the clearance of cells and cell fragments was occurring. To verify this conclusion, cells in the milk were analyzed by flow cytometry and microscopy using antibodies to EMR1 to stain macrophages and morphologic criteria to identify neutrophils, again with no evidence for either of these cell types (Fig. 6B, and data not shown). A slight increase in the number of macrophages was seen on Day 4 postwean (Fig. 6C), after the bulk of apoptotic cell and milk clearance. To determine if this increase was strain specific, we also examined the glands of FVB mice. Again, no macrophages were seen during the clearance phase of involution. Preliminary data suggest that macrophages do accumulate in the gland even later in involution in this strain (not shown). However, the absence of these cells in the luminal contents at any point during the clearance phase of involution suggests that, in this system, weaning alone does not cause an inflammatory infiltration into the milk space and that inflammatory cells are not involved in the clearance of either milk or epithelial cells during this period.
|
Analysis of In Situ Efferocytosis by Epithelial Cells
Figure 7 shows an electron micrograph of a shed epithelial cell in the presumptive early stages of engulfment (efferocytosis) by a secretory epithelial cell still within the monolayer. Although the condensed nucleus is not seen in this section, the apoptotic cell (colored red in the inset) has lost attachment and polarity as well as secretory organelles, yet still contains several small cytoplasmic lipid droplets and one very large, attached droplet. A cell process extended by the engulfing cell (colored green) can be seen to be in direct contact with the membrane covering the droplet. Similar processes appear on the sides and top of the apoptotic cell, presumably part of the same extension or an extension from a neighboring cell. These membrane processes appear similar to the actin-containing ruffles induced in mammary epithelial cells exposed to apoptotic cells in vitro (data not shown). The engulfing cell still exhibits an active nucleus and a polarized phenotype, including tight junctions and apical microvilli. However, it appears to have a disrupted secretory pathway with reduced rough endoplasmic reticulum and secretory vesicles, presumably a mechanism for halting milk secretion. Neighboring epithelial cells on either side exhibit efferosomes containing engulfed and partially digested material (colored blue).
|
Return to Normal Hormonal Cycling
It has been previously suggested that there may be a second round of apoptosis during involution induced by forced weaning [28]. However, as noted above, we found a return to near-prepregnancy size and structure within 4 days after weaning. An alternative hypothesis might be that this second wave of cell death could represent a return to hormonal cycling, as the estrous cycle is known to cause waves of proliferation and cell death in the mammary gland [29, 30]. Vaginal smear cytology was performed on eight dams for 21 days after forced weaning of their litters. Four out of eight entered metestrus on Day 4 postwean and the other four on Day 5. Earlier smears showed very few cells, most closely matching the diestrus-2 smears reported by Schedin et al. [30]. The mice maintained their respective 4- and 5-day cycles, half entering metestrus again on Day 8 and half on Day 10. After this, their cycles became more sporadic (60%, 4 day; 28%, 5 day; and 11% other intervals), resulting in a distribution of estrus stages at any particular time point postwean. Reexamining the mammary weight data (Fig. 1A), we notice slight increases in the gland weight on Days 5 and 10, which is possible evidence of 5-day cycling in these mice. We conclude from these observations that involution is effectively over at 4 to 5 days after weaning in this system and that subsequent epithelial changes are more a remodeling effect of estrus cycling than a continued involutional process.
These studies show that during mammary gland involution in the mouse, epithelial cells themselves are responsible for clearance of both residual milk and the majority of the epithelial cells. Quantitative analysis also demonstrates that most of the clearance occurred during the first 2 to 3 days after forced weaning and was complete by 4 days, and that the gland returned to normal estrus cycling within 4 to 5 days.
As reported in the literature [31], gland weight increased over the first 24 h because of continued milk production, but then declined rapidly over the next 3 days. Apoptotic epithelial cells were detected shortly after weaning and also peaked in total numbers at 24–30 h. Evidence that the epithelial cells were apoptotic included demonstration of active caspase 3, generation of caspase-cleaved keratin 18, altered nuclear morphology, and exposure of surface PtdSer. Milking the gland allowed analysis of the apoptotic cells and showed that, though most exhibited surface-exposed PtdSer (annexin V staining), the few that did not also lacked apoptotic nuclear morphology. This observation is consistent with the finding that the epithelial cells are released from the monolayer before they expose PtdSer, as noted in in vitro studies of epithelial apoptosis by Rosenblatt et al. [20]. Attempts to examine milked cells at earlier time points did not reveal a higher incidence of PtdSer-negative cells (data not shown), an observation that supports a fairly rapid development of this feature after the cells are released into the lumen.
The stimuli that induce epithelial apoptosis in this forced-weaning system are not known (for review see [32]). The observations reported above are consistent with a prior extrusion of the cells from the epithelium (i.e., anoikosis), but this could still result from even earlier apoptotic signals driving the extrusion [20]. Hormonal changes could contribute to the induction of involution and apoptosis, although teat-sealing one gland during normal suckling of the others still induces apoptosis and involution. Mechanisms and kinetics of the loss of epithelium in other model systems, including natural involution, teat-sealing involution, and concurrent-pregnancy involution, will help elucidate the signals.
Importantly, in this system, no inflammatory cells could be detected within the glandular lumen. A few interstitial macrophages were observed, but did not seem to change in numbers or distribution over the clearance phase of the involution process and they remained outside the alveolar basement membrane. We did observe a small increase in macrophages at the latest stage examined (Day 4 postwean); this increase correlates with the timing of macrophage infiltration predicted from microarray studies of the gland [31]. Other investigators have suggested participation of inflammatory cells in apoptotic cell clearance during involution [22–27]. Whether these differences reflect strain/species effects, possible concurrent mastitis, or simply technical differences in the timing and detection methods, the data reported herein suggest that the clearance process can occur without macrophages. In addition to the specific case of tissue remodeling in the mammary gland, most emphasis in the literature on clearance of apoptotic cells has focused on macrophages as engulfing cells—despite evidence for the ability of so-called "non-professional" phagocytes to recognize and ingest apoptotic cells in vitro [9, 33]. On the other hand, tissue remodeling involving apoptotic cell removal occurs during development, before the appearance of macrophages or, in fact, in their absence in Sfpi1 knockout (a.k.a. PU-1-deficient) mice [34]. Involution of the mammary gland represents a classic case of physiologic tissue remodeling in the adult, and we believe it is not inflammatory in nature. Continued investigation into the role of macrophages later in mammary involution is underway in collaboration with other laboratories.
The studies bring up a number of key questions. One of these is what determines whether an apoptotic cell becomes a target for extrusion, apoptosis, and engulfment rather than becoming a phagocytic (efferocytic) cell. Since the alveolus gradually becomes smaller and smaller, it is also possible that individual epithelial cells may first act as efferocytes and then later undergo apoptosis and themselves be consumed. Is there a subpopulation of epithelial cells that is predestined to remain after involution is complete? Could ingestion of apoptotic cells signal the efferocyte to become such a residual cell? Can clearance of milk occur in the absence of apoptotic cells? Studies addressing these questions are currently underway.
A second set of questions addresses the ligands on the apoptotic epithelial cells and their cognate receptors on the engulfing cells that mediate the uptake. As noted in the introduction, many potential ligands and bridge molecules are known to be present in milk and/or in the involuting mammary gland, including CALR, MFGE8, and THBS1. We showed the presence of PtdSer on the luminal apoptotic cells and in a previous study [9] showed that mammary epithelial cells express the PtdSer recognition structures involved in apoptotic cell clearance. However, the relative roles of the different receptor systems in vivo remain to be determined.
Thirdly, one may question whether the uptake mechanism seen for cannibalistic ingestion of apoptotic epithelial cells in vivo is the same as that seen for uptake into macrophages or cultured epithelial cells in vitro. Answering this question will require interventional examination of uptake in vivo. However, it should be noted that the ingested apoptotic cells (Fig. 3) appear to reside in spacious phagosomes and cause actin ruffling, which are hallmarks of efferocytic processes in vitro [3].
The initiation of normal estrus cycling almost immediately after the apoptotic cells have been cleared (4 or 5 days) presumably imposes a new round of epithelial cell division, apoptotic cell death, and removal [29, 30], although on a smaller scale than remodeling of the lactating gland. Clearance of apoptotic cells during the normal estrous cycle has not been studied in either the parous or nulliparous mammary gland, but may occur through processes similar to those described herein for full-scale involution. Alternatively, macrophage infiltration may occur, contributing to apoptotic cell clearance late in involution and during the normal estrous cycle. Mammary epithelial cells acting as phagocytes may indeed be a differentiated-cell phenotype seen only during the early clearance phase of postweaning mammary involution. The final functioning of the milk-secreting cells may be to clear residual milk and their dying neighbors, and then to die themselves.
ACKNOWLEDGMENTS
Special thanks to Colin Monks for expert microscopy consultation, to Dallas Hyde for morphometric and stereology consultation, to Margaret Neville for critical discussion, and to Jess Cramer, Gail Smith, Lynn Cunningham, and Bill Townend for excellent technical assistance.
FOOTNOTES
1Supported by NIH grants PO1 HD 38129 and RO1 GM 061031. 
Correspondence: 2Jenifer Monks, Reproductive Sciences, Ob/Gyn, MS 8309, University of Colorado Health Sciences Center, Anschutz Medical Campus, P.O. Box 6511, 12800 E. 19th Ave., Aurora, CO 80045. FAX: 303 724 3512; e-mail: Jenifer.Monks{at}UCHSC.edu
Received: 13 August 2007.
First decision: 11 September 2007.
Accepted: 19 November 2007.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  C. J. Watson Immune cell regulators in mouse mammary development and involution J Anim Sci, April 1, 2009; 87(13_suppl): 35 - 42. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
