A Brief History of HMEC Life, Aging, Death, and Immortality in our Culture Systems (chart 1)

(numbers in brackets refer to the sections in the Review which discuss these areas in greater detail)

Charts 1-4 in this section chart how the HMEC cultures we distribute were derived, provide the names of the cell types, and indicate the steps where the normal, finite lifespan HMEC overcome replicative senescence, acquire an indefinite proliferative potential and activate telomerase activity. Unless otherwise indicated, the finite lifespan HMEC we use and distribute are obtained from reduction mammoplasty tissues (with no abnormal pathology noted, or with mild fibrocystic disease) [I].

FINITE LIFESPAN HMEC (chart 2)

When grown in either our serum-containing MM, or serum-free MCDB 170 media [II.A.] active proliferation ceases after ~15-30 population doublings (PD)(3-5 passages in MM; 2-3 passages in MCDB170). These HMEC populations are called normal, pre-stasis, or pre-selection. We have limited amounts of MM-grown HMEC to distribute. As the HMEC approach this first barrier, the cobblestone-appearing cells become larger and flatter, and show increasing expression of the cyclin dependent kinase inhibitor (CKI) p16^INK4a. At growth arrest, cells stain positive for senescence-associated b-galactosidase (SA-b-gal) activity. Mean TRF (terminal restriction fragment) length is ~ 6-8 kb.

In MM-grown cells, there may be limited areas of proliferation that continue until around passage 10, but these are focal, with the proliferating cells soon acquiring a flatter, senescent morphology. In MCDB 170, a small number of cells are able to spontaneously overcome this first proliferation barrier; these cells show methylation of the p16 promoter and absence of p16 expression. This process is called self-selection, and the resulting p16(-) cells are called post-selection. We have large quantities of standardized post-selection HMEC to distribute. Post-selection HMEC grow actively for an additional ~30-70 PD without ever expressing any p16. The p16(-) HMEC express wild-type p53 which is present in a stabilized form. As they near a second proliferation barrier, post-selection HMEC become larger and vacuolated, with increasing expression of SA-b-gal. Mean TRF at growth arrest is ~5 kb.

Pre-selection HMEC, even MM-grown, can also overcome the first proliferation barrier, correlated with loss of p16 expression, following exposure to a chemical carcinogen (we used benzo(a)pyrene) [III.A.]. The resultant populations have been termed extended life (EL). We have very limited quantities of EL HMEC to distribute for collaborations. Since the extended life cultures are carcinogen exposed, they are not normal. For example, EL 184Aa, the precursor of our immortal cell line 184A1, contains mutations in the p16 alleles, and likely harbors additional errors.

Our current model of senescence in these HMEC postulates the following [II.B-C.].

1) The first proliferation barrier is mediated by the RB pathway, and is a response to accumulated stress, broadly defined as conditions that are not conducive to optimal metabolic function. This barrier can occur after varying numbers of PD in culture, depending upon stress levels. It can be overcome by alterations in the pathways governing RB. In cultured HMEC, the most common means by which this is effected appears to be the downregulation/loss of p16 expression. At this growth arrest, HMEC maintain normal karyotypes, are arrested in G1, have little DNA synthesis, and do not have critically short telomeres. We suggest that this barrier is most similar to what has generally been called as senescence or M1 in fibroblasts; however, some fibroblast strains, under some (low-stress) conditions, may not encounter this barrier. We also suggest that stress-induced growth inhibition occurs naturally both in vitro and in vivo, and encompasses what in vitro has been called “culture shock” or growth inhibition from “inadequate culture conditions”. In HMEC, oncogene-induced “premature senescence” represents a different type of proliferation barrier.

As a generic term, we are now calling this stress-associated barrier stasis. There may be significant cell type variability in how stasis occurs, e.g., in some cell types (not HMEC) p53-dependent p21 expression may play a major role.

2) The second proliferation barrier is determined by telomere length, and is extremely stringent. Mean TRF at this barrier is ~≤5 kb. Where p53 function is absent or abrogated, this barrier presents as crisis and cell death. Where wild-type p53 is present, this barrier has been recently termed agonescence, and produces mostly viable growth-arrested cells. Agonescence was not previously described because most human cells don’t spontaneously overcome the stasis barrier. Abrogation of p53 as well as RB function has been routinely used to get cells over stasis. Our post-selection and EL HMEC retain p53 function, and therefore mostly arrest at agonescence rather than die at crisis. As a consequence of telomere dysfunction, HMEC at agonescence and crisis show abnormal karyotypes (many telomeric associations) and mitotic failures. Most HMEC at agonescence remain viably arrested in both G1 and G2, have a moderate labeling index (~15%), and have never been observed to spontaneously immortalize. When normal p53-dependent checkpoints are mitigated by inactivation of p53 function, the cells are unable to growth arrest; there is an extended period of growth (labeling index >40%), followed by massive cell death.

The nature of the agonescence block can account for the observed stringent senescence in cultured HMEC; cells which fail to maintain a G1 or G2 arrest at agonescence will eventually die or become non-proliferative via mitotic catastrophe. Unlike an arrest based upon blocking cell cycle progression (e.g., elevated levels of CKIs), the agonescence barrier involving chromosomal derangements can not be readily overcome and is irreversible.

The means by which HMEC overcome agonescence or crisis have not yet been fully defined. Through the use of the multiple oncogenic agents described below we venture a model that postulates that multiple methods of repressing hTERT expression need to be inactivated, making overcoming senescence an extremely rare multi-step process in the absence of pre-existing errors and/or strong selection [IV.F.].

Note: Expression of a “senescent morphology” and SA-b-gal is seen at both these barriers, and thus these markers don’t distinguish between these molecularly distinct events.

Note: Rodent cells lack the second, stringent, telomere-dependent barrier.

INDEFINITE LIFESPAN (IMMORTAL) HMEC (chart 3 – chart 4 )

Note: all these cells are IMMORTALLY TRANSFORMED. Please do not refer to them as normal, untransformed, or non-transformed. They are TRANSFORMED from the normal condition of finite lifespan. Immortality is the most common transformation from normal seen in human carcinomas, and may be the rate-limiting step in human malignant progression. Our HMEC lines show obligate changes in some key signal transduction pathways as a consequence of immortal transformation.

Even after HMEC overcome both senescence barriers, the resultant cells with indefinite proliferative potential may still progress through further changes to reactivate telomerase and attain uniform good growth.

Overcoming agonescence is not coincident with the reactivation of telomerase activity. Overcoming agonescence means that these p53(+) cells can proliferate, without evidence of telomere dysfunction, even in the presence of critically short telomeres.

In HMEC that retain p53 function [184A1, 184AA4, 184B5 184ZN..., 184Aa-ZN..], telomerase reactivation occurs as part of a set of very gradual epigenetic changes, a process that we have termed conversion [IV.A-B]. In HMEC immortalized with loss of p53 function [184AA2, 184AA3, 184Aa-GSE22], the conversion process is greatly accelerated, and some aspects are not expressed. As a result, uniformly good-growing telomerase-expressing cells appear much more rapidly [IV.C.]. In HMEC immortalized via expression of ectopic hTERT [184-, 161-, 48R-hTERT], conversion does not occur [IV.D.].

Conversion in p53(+) post-agonescent HMEC appears to be inherently triggered by the extremely short telomeres that arise when the HMEC maintain proliferation in the absence of telomerase activity. Changes are first observed when the mean TRF declines to ~3 kb. If the HMEC immortalize when the mean TRF is still > 5 kb (prior to reaching agonescence; example: 1841), there is good uniform growth until the TRF declines to ~3 kb. These cells are pre-conversion. If the HMEC immortalize around the time agonescence begins, or during agonescence (examples: 184B5; 184AA4; 184-ZN..), when the immortal line appears it is already in the process of conversion. For both situations, the early passage populations have been called conditionally immortal, since the population as a whole, but not each individual cell, reproducibly expresses its immortal potential. Conditionally immortal cells undergo a prolonged period of slow heterogeneous growth, with the population gradually converting to become fully immortal.

The gradual conversion process in p53(+) HMEC is associated with the following changes:

1) p57^KIP2: p57 is not detected in our finite lifespan HMEC, but is seen during G0 arrest in conditionally immortal HMEC. When the mean TRF declines to ~3 kb. p57 remains expressed even after cells are released from G0. This change is associated with a sharp decrease in proliferative potential. Growth remains poor and heterogeneous while p57 is expressed at high levels. p57 levels gradual decrease as conversion progresses, associated with progressive increase in growth capacity.

2) Telomerase activity and mean TRF length: very low levels of telomerase activity become detectable once conversion starts, and the levels progressively increase. As telomerase is expressed, the faint <2.5 kb TRF signal gradual gets stronger and stabilizes at ~3-7 kb.

3) Growth in TGFb: the expression of hTERT in conditionally immortal HMEC (as well as in finite lifespan HMEC) leads to the gradual acquisition of resistance to TGFb-induced growth inhibition (increasing number of cells show progressively better growth in TGFb).

4) c-myc: In some immortal lines, the level of c-myc expression (mRNA and protein) during G0 arrest change from greatly reduced compared to cycling populations, to about the same level as seen in cycling cells.

5) Raf-1: as a consequence of conversion (not just overcoming agonescence or expressing hTERT) the ability of activated Raf-1 to induce growth inhibition is abrogated. Instead, Raf-1 induces increased malignancy-associated properties in fully immortal HMEC.

6) SA-b-gal: SA-b-gal is expressed in agonescent HMEC, and remains expressed in conditionally immortal HMEC, even those actively growing with mean TRF >3 kb. SA-b-gal expression is gradually lost in almost all cells as they convert to full immortality.

In HMEC immortalized with loss of p53 function, the main differences are that (1) p57 is never expressed, and the cells do not undergo a prolonged period of poor heterogeneous growth. (2) Telomerase activity is present soon after immortalization, and the mean TRF never declines to <3 kb.

In general, we have seen that different methods of producing immortal HMEC can yield cell lines with significantly different phenotypes. These methods may vary in the extent to which they model human malignant progression in vivo. Considering that human carcinomas generally have short telomeres, and human carcinoma cell lines have short regulated telomeres (mean TRF ~3-7 kb) we suggest that a conversion-like process likely occurs during human carcinogenesis. The attainment of short telomeres at agonescence/crisis, and during conversion, may play a pivotal role in malignant progression, including contributing to genomic instability through the perpetuation of bridge-fusion–breakage cycles [IV.F.].

Once the HMEC are immortally transformed, the introduction of one or two oncogenes can further transform these cells towards malignancy (anchorage-independent growth, growth factor independence, and/or tumorigenicity in nude mice). Truly non-transformed cells (i.e., finite lifespan) can not be rendered malignant by the same oncogenes. Thus we believe that it is the acquisition of the multiple derangements that produce immortality that may be rate-limiting for malignancy, and the importance we place in urging cancer researchers to refrain from inaccurately calling immortally transformed cells “normal, untransformed, or non-transformed”.