You are here

| Mammary Stem Cells

Keeping abreast of cell fate hierarchy: mammary tissue can bounce both ways



By Carla Mellough

The dogma that stem cells exist at the apex of a cellular hierarchy and divide to self-renew as well as generate more differentiated progeny in a unidirectional fashion may now be under question following a recent report published in PNAS. The paper, by Christine Chaffer et al.1, originates from various academic centres in Massachusetts and provides evidence that a certain type of human mammary epithelial cell displays an unexpected degree of plasticity, and can revert from a more differentiated cell towards a stem cell-like state. Observations that normal and neoplastic cells can naturally co-exist in populations of epithelial cells grown in vitro led the authors to study the biology of subpopulations of primary human mammary epithelial cells (HMECs) and HMECs from a human breast cancer cell line. And what they discovered in fact is that breast cells can swing either way.

Working with breast cancer cell line cultures, Chaffer et al.1 observed that a population of floating cells existed above adherent cultures of HMECs under normal growth conditions. This HME-floating population of cells (HME-flopcs) were collected and generated adherent cultures after transfer into new culture dishes. Although both HMECs and HME-flopcs expressed epithelial markers and formed polarised epithelial acinar structures after clonal plating on Matrigel, HME-flopcs did not exhibit the typical epithelial monolayer morphology nor form tight junctions between cells as seen in HMECs. Flow cytometric analysis revealed that 2.5% of HME-flopcs were enriched for markers usually associated with the bipotent progenitor/stem cell fraction of mammary epithelial tissue (CD44hiCD24loESA-), compared to only 0.2% of HMECs. Further, they found that HME-flopcs CD44hi cells were most similar to progenitor/stem cells whereas HME-flopcs CD44lo, HMEC- CD44lo cells and more differentiated mammary cell types were more alike. The authors purified CD44hi and CD44lo cells through single cell cloning which revealed that both cell types existed within clonal colonies. By assessing 1) purified isolated FACS populations and 2) heterogenous populations of CD44hi cells expressing tomato fluorescent protein mixed with purified unlabelled HME-flopcs CD44lo cells, the authors went on to demonstrate that in fact HME-flopcs CD44lo cells can spontaneously regenerate CD44hi cells in 2D culture. Interestingly, in contrast to HME-flopcs populations, this was not observed in HME cultures; HME CD44lo cells were poorly able to spontaneously generate CD44hi cells. In vitro mammosphere forming assays confirmed the stem-like quality of CD44hi cells; bulk HMECs did not form mammospheres whereas HME-flopcs cells did, with the most efficient rate of mammosphere formation from the HME-flopcs CD44hi fragment. Further experiments indicated that CD44hi cells were able to maintain a stem-like state in 2D culture, but that 3D culture drove them to form more differentiated cell types.

But what about primary HMEC tissue, would this give the same result? The authors identified a CD44lo CD24+ESA basal population that was similar to HME-flopc-CD44lo cells, indicating that a biological equivalent of this cell fraction existed in the human mammary gland. So could these CD44lo cells convert back to a CD44hiCD24loESA stem-like state? Basal and luminal cells were FACS purified from primary HMECs and then both populations monitored in vitro for 12 days. What Chaffer et al.1 found was that a proportion of primary basal CD44lo cells were indeed able to convert back into CD44hi stem-like cells, whereas in luminal populations CD44lo cells did not readily make this conversion. These results show that primary mammary basal epithelial cells are capable of reversion towards a more multipotent phenotype.

Chaffer et al.1 then turned their attention to the implications of these results for cancer cell biology. Many human solid tumor types house a population of stem-like cells, called cancer stem cells (CSCs). CSCs are demonstrated to spawn new tumors by giving rise to more differentiated cell types within a tumor (which lack tumor-initiating ability) as well as replenishing the CSC compartment. Both normal stem cells and their neoplastic counterparts are self-renewing and, until now, were understood to make unidirectional cell fate decisions down the cellular hierarchy. The authors transformed HME and HME-flopc cells by introducing the SV40 early region (SV40-ER) and H-ras oncogenes. Transformed HME-flopc-CD44lo cells converted to CD44hi cells fivefold more efficiently than untransformed HME-flopc-CD44lo cells. Interestingly, even after transformation luminal HME-CD44lo cells were inefficient at converting to a CD44hi stem-like state, just like their untransformed equivalents.

To determine whether transformed HME cells can go on to give rise to CSCs in vivo, the authors injected transformed cells from HME-CD44lo (with poor conversion capability), HME-flopc-CD44lo (which efficiently convert to CD44hi cells) and mixed populations of CD44hi+lo cells (with mixed conversion ability) into immunodeficient mice. Importantly, they observed that their in vitro results were not mirrored in vivo. Transformed HME-flopc-CD44lo-derived tumors were found to generate an inflammatory response in vivo resulting in largely cystic tumors containing 0.2% of CD44hi cells, whereas populations of transformed HME-CD44lo cells - previously observed to be inefficient at producing CD44hi cells – readily underwent conversion in vivo, with 16% of these tumors comprising of CD44hi cells. However, tumor forming ability of the various transformed cell subtypes in vitro was found to be related to the differentiation state of the cell-of-origin, oncogenic transformation of stem-like cells generated more proliferative tumor spheres that could be serially passaged than did the oncogenic transformation of more differentiated mammary cell types.

Transformed mammary progenitors and stem cells are already implicated in the pathogenesis of breast cancer in humans. One study2 has shown that breast tissue from women carrying mutations in the BRCA1 gene (encoding the tumor suppressor protein BRCA1) harbours a luminal HMEC progenitor population which undergoes factor-independent growth in vitro and is transformed in BRCA1-associated tumors. Results from another study3 consolidate the observations of Chaffer et al.1 regarding tumor forming ability and also indicate that the differentiation state of the cell-of-origin influences the aggressiveness of resulting neoplastic cell populations and their metastatic capability.

That cells already appearing committed to a differentiated cell fate were then able to revert to a stem cell-like state without genetic manipulation is a particularly interesting and important result given the ongoing efforts to achieve induced pluripotency from various differentiated somatic cell types. As the pluripotent nature of these spontaneously arising de novo stem-like cells was not addressed in this study (evidence for teratoma formation or incorporation into the germline was not determined), the implications of this work currently holds the greatest value for the development of cancer therapeutics. Certainly, a deeper understanding of CSC biology will shed new light on more suitable and therefore hopefully more successful ways to prevent tumor progression using strategies directed against the cell-of-origin, and maybe even prevent tumor initiation in individuals under high risk of breast cancer. Moreover - could further investigation into the mechanisms underlying the conversion of mammary cells back towards a pluripotent state then also potentially lead to the development of manipulation-free methods for the reversion of somatic cell types back to pluripotency?



1 Chaffer et al. Normal and neoplastic nonstem cells can spontaneously convert to a stem-like state PNAS 2011 108(19):7950-7955; published ahead of print April 15, 2011, doi:10.1073/pnas.1102454108.

2. Lim et al. Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 2009, 15:907–913.

3. Ince et al. Transformation of different human breast epithelial cell types leads to distinct tumor phenotypes. Cancer Cell. 2007 Aug;12(2):160-70.