Stem cells are undifferentiated, division-competent cells that reside in a given tissue, and that function to regenerate and/or replace all of the cell types that compose that tissue. Cancer is presently viewed as a clonal disease that depends on genetic mutation; and undifferentiated, division-competent cells are seen as prime targets for mutations that will be propagated to their progeny. Therefore, the attention of cancer biologists has turned to the stem cell compartments of all tissues. Generally, the biological properties of stem cells include self renewal, ie.: the ability to regenerate new stem cells, as well as the potential to generate all of the cell types of the tissue in which they exist (1, 2). Even tissues previously believed to consist wholely of terminally differentiated cells, such as adult brain have recently been shown to harbor a small population of pluripotent division-competent cells (3). Stem cells are important medically both because of the risk they pose in carcinogenesis, and for the potential they offer for organ regeneration or replacement. They are economically important in agriculture for the implications they have for increasing milk production in dairy cows. Recently we have been able to provide a morphological description of three division-competent cells in the murine mammary gland. The purpose of this mini-review is to make this information more available to mammary gland and tumor biologists.
Morphologically, it is difficult or impossible to distinguish the stem cells from the terminally differentiated cells of a particular tissue. For instance, the hematopoietic stem cell is morphologically identical to many of its differentiated progeny, because most of their differentiated features are specialized membrane receptors. For this reason positive stem cell markers such as c-kit (the receptor for stem cell factor) are used to distinquish them, isolate them, and study their function. For solid tissues such as skin, however, no specific markers for stem cells have been found, but the increased expression, of b1 integrins can be used to isolate epidermal cells with an unlimited capacity for self renewal and that also produce daughters that can amplify up to five divisions before differentiating (4).
The mammary gland epithelium is derived from the embryonic epidermis during fetal development, and is generally considered to be composed of two cell types: secretory or "lumenal" cells and myoepithelial cells. Since early in the century cell biologists and pathologists have recognized that there are structurally undifferentiated cells in the mammary epithelium, that do not take up any of the commonly used cytological stains (5). To qualify as structurally undifferentiated a cell must show a lack of polarity and contain few organelles in its cytoplasm. In contrast, a structurally differentiated cell contains a well developed system of specialized, membrane-bound organelles that may contain proteins whose identities can be demonstrated via immunocytochemistry. Further, if it resides in a solid tissue, it will also demonstate a physical polarization relative to its immediate connective tissue and either the parenchymal lumen or blood vascular elements of that tissue.
Since the undifferentiated cells in the mammary epithelium do not stain, they have been referred to in the literature as "pale cells" or "light cells" (5, 6). They are unpolarized and smaller than the lumenal cells, and argument about whether they are "wandering lymphocytes" (7) or stem cells (6) has been ongoing between pathologists and mammary stem cell biologists since they were first described.
Using light and electron microscopy we have recently been able to identify
morphological features that permit the differential description of two
pale-staining cellular morphotypes ("morpho" - shape, "type"
- kind) in murine mammary epithelium (8), and morphometric analysis has
shown that the less differentited of the morphotypes represents two functionally
different division-competent compartments. This means there are three division-competent
cell types in the mammary epithelium. We have used the presence of condensed
chromosomes in these cells as an indicator of division-competence, and
therefore of putative stem or progenitor cell function. Both morphotypes
are present in human (9), mouse, and rat (8). The following is a review
of what is known at present of stem cells in mammary epithelium. Nineteen
time points within the four stages of mammary epithelial growth, function
and involution were examined: nulliparous (animals were sampled without
regard to stage of estrus cycle); pregnancy: 2, 7, 13, 17, 19, and 21 days
(about 4 hours before parturition); parturition (after delivery of the
first pup); lactation: 3, 24, and 48 hours, and 4, 6, 11 and 12 days; involution:
24 and 48 hours, 6 and 9 days (pups were removed from their mothers on
day 4 of lactation to initiate involution).
Mammary epithelial morphotypes can be characterized ultrastructurally on the basis of eight structural features: 1) number of cells per unit area of epithelium 2) cell size, 3) location of the cell relative to the basement membrane and glandular lumen, 4) cell shape, 5) staining of the nuclear and cytoplasmic matrix, 6) nuclear morphology, 7) cytoplasmic morphology, and 8) grouping relative to each other and other epithelial morphotypes. Two categories of light-staining cells can be classified on the basis of both size and differentiated characteristics: "small light cells" (SLC) and "large light cells" (LLC). However, the LLC show variance in cytoplasmic differentiation and have been further classified into "undifferentiated large light cells" (ULLC) and "differentiated large light cells" (DLLC) (8).
Both large and small light cells occur in homogeneous (like) pairs suggesting that they are daughter cells of a recent mitosis, and hence can produce self-same daughters. They also occur in heterogeneous (mixed) pairs consisting of one of each type. Homogeneous pairing is defined as two undifferentiated cells of similar morphology surrounded by a morphologically uniform population of cells dissimilar to themselves. Heterogeneous pairing is one ULLC and one SLC. For consistency we have designated the differentiated lumenal cells as "large dark cells" (LDC). Myoepithelial cells remain "myoepithelial cells".
Differential Staining Characteristics Among Cell Morphotypes. The cause of the staining difference between the light cells and the dark cells, can be seen when the cytoplasm of all cellular morphotypes is examined at magnifications between 13,000X and 110,000X (SLIDE 1). At these magnifications an extremely fine fibrillar component of the cytoplasmic matrix is visible in the LDC that takes up osmium tetroxide as shown in SLIDE 1. This material is most easily visualized between the lamellae of the RER and in the cytoplasmic cortex. It is very scant in the SLC and ULLC, more densely distributed in DLLC, and very dense in LDC. It is assumed that this material also accounts for the greater staining density of LDC when paraffin sections are stained with toluidine blue. This feature varies with the relative abundance of membranous organelles in SLC and LLC and with the abundance of actin-myosin tonofilaments in myoepithelial cells,. The more developed and organized the organelles, the denser the dark-staining filamentous material. Similarly, the more condensed the chromatin in the nucleus, the denser the organization of the filamentous material in the nucleoplasm.
Small Light Cells . Small Light Cells (SLC) never contact the lumen, and they rest on the basement membrane or the suprabasal surface of a myoepithelial cell. SLC are irregularly shaped, often with numerous cytoplasmic processes. The nucleoplasm accepts neither osmium nor toluidine blue and the nucleus contains dense clumps of heterochromatin, and is sometimes indented (SLIDE 2 ). The cytoplasm is also characteristically pale-staining. It has a sparse complement of organelles clustered close to the nucleus and usually shows no structural evidence of specialized function. SLC are division-competent ( SLIDE 2 ), and although they are usually free of specialized membrane contacts with neighboring cells, they occasionally form nonjunctional focal contacts between each other and neighboring LDC (SLIDE 3). In very rare cases SLC (one pair out of 111 SLC observed in the rat) may share desmosomes between each other (not shown). The characteristic lack of cytoplasmic differentiation was consistent in all but one of 111. In this case the SLC contained 3 lipid droplets and an apically located nucleus (SLIDE 5 ). This cell is a transitional form between the SLC and ULLC morphotypes, and suggests that some SLC have differentiative potential. SLC occur as single cells scattered throughout the epithelium (SLIDE 2), in like pairs with another SLC (SLIDE 3), or as one member of a mixed pair whose other member is a ULLC (SLIDE 6).
Large Light Cells. The LLC population is actually comprised of two structurally distinct morphotypes with secretory potential. These are: the ULLC, which sometimes contacts the lumen and sometimes does not, and a well differentiated large light cell, the DLLC, with characteristics intermediate between ULLC and LDC. DLLC have very well developed secretory organelles and milk products and an apical surface in contact with the lumen.
Undifferentiated Large Light Cells. In cross section, ULLC are two to three times larger than SLC (Slide 6). They sometimes, but not always, contact the lumen. Some ULLC that do not contact the lumen have an apically located nucleus and a "trailing tail", that makes them gourd-shaped (SLIDE 6 ). Others, including those that do contact the lumen, are very round cells. Like the nucleoplasm, their cytoplasmic matrix accepts neither osmium or toluidine blue stains. Although their membrane systems are more abundant, they are only slightly more developed than those of SLC. Even so, ULLC can contain small secretory granules in their Golgi and lipid droplets in their cytoplasm. ULLC containing condensed mitotic chromosomes (SLIDE 7 ) appear in nulliparous, 13 days pregnant and 2 hours lactating rat mammary epithelium with about the same frequency. ULLC are present in the alveoli of nulliparous, mid-pregnant and newly lactating rat mammary glands as 1) scattered single cells, 2) one member of a mixed pair consisting of itself and an SLC, 3) like pairs, or 4) in clusters of ULLC and SLC. ULLC are the division-competent cells previously described in the mouse (6) and human (9, 10), and in the mouse they are immunopositive for casein.
Differentiated Large Light Cells. (SLIDE 9 ) DLLC are present in the rat mammary epithelium as singletons and pairs. Between mid-pregnancy and 4 days lactation they are also present in groups and large arrays which comprise up to half the cells visible in an alveolar or ductular cross section. Although there is no structural evidence of apoptosis occurring in large arrays of cells during lactation, these large groups of DLLC are conspicuously absent by 12 days lactation (8). We believe that they undergo terminal differentiation to LDC.
Population Sizes of Light Cell Morphotypes. SLC comprise about 3% (SLIDE 10 )of the epithelial cell population in all stages of development and regression in rat mammary gland, and are characteristically smaller than other epithelial cells. LLC comprise about 5% (SLIDE 10 )of rat mammary epithelium during both renewal and involution, with DLLC as a very prominent element of this population during pregnancy and early lactation. However, in the nulliparous rat all LLC are ULLC and comprise about 9% of the total cell population, and are three times more densely distributed than in other developmental stages (SLIDE 11 ). Morphometric analysis of the rat mammary epithelium at 4 of the 19 stages in which cell counts were performed showed that population density (number of cells/mm2) of SLC did not change from nulliparity through involution (Slide 10 ). This agreed with percentage determinations of SLC: in all 19 stages of rat mammary gland growth from nulliparity through 9 days postlactation the proportion of SLC remained at 3%. This means that although the number of mammary epithelial cells increases in the mouse by 27 fold (11) to 38 fold (Smith-unpublished observations), the percent that are SLC does not change. Therefore SLC increase and decrease in absolute number along with the rest of the cells. If the SLC were purely a stem cell population, they would be expected to show a decrease in population density during epithelial growth, and a concomitant increase with mammary involution, maintaining their absolute cell number throughout growth and regression. The single function of a stem cell is division, and they are expected to neither increase nor die in detectable numbers. Therefore the growth-related fluctuation of SLCs (which are defined solely by morphological criteria) indicates that they are very likely a combined population of stem cells and primary progenitor cells. In the mammary epithelium these two cells are morphologically indistinguishable, and the constancy of the population density of SLC is due to amplification of primary progenitors and their subsequent differentiation during pregnancy and lactation. Further, any remaining primary progenitors may undergo apoptosis during involution. However, the stem cell component of the SLC remain at a fixed (smaller than 3%) population size.
Cell Division in SLC and ULLC. In the mouse we observed ultrastructural images that suggested that both symmetric and asymmetric cell division occured in light cells (8). These images consisted of two structurally similar ULLC either side-by-side (suggesting a recent symmetric mitosis) or one above the other with the lower one containing small bundles of myofibrils and the upper withsome Golgi development. This second image suggested a recent asymmetric mitosis with the production of a presumptive myoepithelial cell below a differentiating secretory cell. It is unknown whether the asymmetric division occurs in SLC, ULLC or both. Although the daughter cells of the asymmetric divisions were all very undifferentiated, suggesting that the mother cell was an SLC, myoepithelial cells in the rat epithelium frequently contain a "rogue"lipid droplet. This suggests that some of them may arise from asymmetric division of ULLC. Myoepithelial cell genesis from asymmetric mitosis of light cells has also been described in the human (10), and we have not mitosis observed in myoepithelial cells in either the rat or the mouse (8). Based on our results we propose a model (SLIDE 12 ) for cell proliferation and differentiation in rodent mammary epithelium based on morphological markers. Briefly, SLC (stem cells), devoid of differentiated characteristics, generate daughter cells morphologically similar to the mother cell. One of the daughters retains its undifferentiated nature and remains a stem cell. The sister cell, although remaining morphologically identical to the stem cell, becomes a primary progenitor cell with the capacity to undergo multiple divisions. All daughters of a primary progenitor cell divide, and, depending on the orientation of the cleavage furrow at the time of determination, differentiate into secondary progenitor cells (ULLC) that are committed either to secretory or to myoepithelial differentiation. ULLC can divide multiple times to produce large numbers of ULLC which differentiate into DLLC and then LDC. Alternatively, they can produce myoepithelial cells, but these are not produced in very large numbers. Our scheme of myoepithelial cell genesis is based on the images we analyzed,so we have represented it in our model as if they arise from SLC.
1. Hall , P.A., and Watt, F.M. (1989). Stem Cells: the generation and maintenance of cellular diversity. Development 106, 619-633.
2. Potten, C.S. and Loeffler, M. (1990). Stem cells: attributes, cycles, spirals, pitfalls and uncertainties. Lessons for and from the crypt. Development 110, 1001-1020.
3. McKay, R.(1997) Stem Cells in the Central Nervous System. Science, 276, 66-71.
4. Jones, P.H. and Watt, F.M. (1993). Separation of human epidermal stem cells from transit amplyfying cells on the baxsis of differences in integrin function and expression. Cell 73 (4) 713-724.
5. Emmel, V.E., Weatherford, H.l., and Streicher, M.H. (1926). Leucocytes and lactation. Am. J. Anat. 38, 1-39.
6. Smith, G. H. and Medina, D. (1988). A morphologically distinct candidate for an epithelial stem cell in mouse mammary gland. J Cell Sci. 90, 173-83.
7. Ferguson, D.J.P. (1985) Intraepithelial lymphocytes and macrophages in the normal breast. Virchows Arch. [Pathol Anat] 407, 369-378.
8. Chepko,G. and Smith, G.H. (1997). Three division-competent, structurally distinct cell populations contribute to murine mammary epithelial renewal.Tissue and Cell 29, 239-253.
9. Ferguson, D. J. (1985). Ultrastructural characterisation of the proliferative (stem?) cells within the parenchyma of the normal resting breast. Virchows Arch A Pathol Anat Histopathol. 407, 379-85.
10. Ferguson, D. J., (1988). An ultrastructural study of mitosis and cytokinesis in normal "resting" human breast. Cell Tissue Res. 252, 581-7.
11. Nicoll, C. S. and Tucker, H. A. (1965). Estimates of parenchymal, stromal, and lymph node deoxyribonucleic acid In mammary glands of C3H/Crgl/2 mice. Life Sciences. 4, 993-1001.
Gloria Chepko
Lombardi Cancer Center
Georgetown University
Washington DC 20007
tel. 202-687-0340
e-mail: chepkog@gunet.georgetown.eduGilbert H. Smith
National Institutes of Health
Bethesda MD 20892
tel. 301-496-2385
last update: June 1998