An Overview of Growth, Aging, Senescence, and Immortality in our HMEC Culture System
The goal of our studies since 1976 has been to understand the normal processes governing growth, aging, and senescence of HMEC, and how these normal processes are altered during immortal and malignant transformation. To address these goals, our labs have generated a large variety of HMEC types, ranging from primary organoid material and normal HMEC strains, to isogenic series of cultured cells at various stages of transformation (Chart 1). Examination of these cultures has elucidated the many significant differences between normal and abnormal HMEC, and produced a new, molecularly defined model of the senescence barriers encountered by cultured HMEC (Figure 1, Table 1). These senescence barriers function to suppress tumorigenesis, thus understanding how they are overcome as normal cells transform to immortality and malignancy can provide insight into the mechanism of human malignant progression in vivo, and into possible therapeutic interventions. Our HMEC culture system has been shown to accurately model many aspects of early stage breast carcinogenesis in vivo, and can serve as an experimentally tractable model to examine factors that influence human cellular aging and carcinogenesis. Importantly, our ability to achieve long-term growth of normal finite HMEC of multiple lineages makes possible extensive examination of normal HMEC biology, differentiation, and aging. We believe that understanding aberrant biological process requires prior knowledge of the normal state.
We postulate that cultured finite lifespan HMEC encounter two mechanistically distinct barriers to indefinite proliferation, stasis (stress-associated senescence) and replicative senescence (telomere dysfunction due to telomere attrition). Finite HMEC are also vulnerable to oncogene-induced senescence (OIS). Errors are needed to bypass or overcome the stasis and replicative senescence barriers. Once cells become immortal (telomerase expressing) they are no longer vulnerable to OIS, and a gain-of-function oncogenic error can be sufficient to confer malignant properties.
Fig.1. Model of senescence barriers encountered by cultured HMEC
Stasis is a stress-associated barrier, mediated by the retinoblastoma (RB) pathway, that is independent of telomere length and extent of replication. The onset of stasis in cultured HMEC correlates with increased expression of p16, but not p21 (1-4). The number of population doublings (PD) achieved prior to stasis depends upon culture conditions; we have observed a range of ~10-60 PD (3-6). Molecular correlates that can identify stasis, in addition to p16 expression, include arrest in G1, low labeling index (LI), non-critically short telomeres and normal karyotypes (2-4). These parameters are consistent with an RB-mediated arrest and the absence of a significant DNA damage response (DDR). Cells at stasis express senescence-associated β-galactosidase (SA-β-Gal) activity and a senescent morphology. Stasis can be bypassed or overcome in cultured HMEC by multiple types of single alterations (genetic and/or epigenetic) in pathways governing RB, and does not require loss of p53 function (1,6-9). Overcoming stasis may correlate with hyperplasia/atypical hyperplasia in vivo, which commonly display clonal growth and errors in the RB pathway (e.g., loss of p16 expression, mutated RB, overexpressed cyclin D1) (1,10-13). Gross genomic aberrations are not common at this stage in vivo (14), and are not associated with overcoming stasis in vitro (2,4).
We postulate that stasis can also be enforced by p53-dependent p21 in response to DNA damaging stresses such as oxidative damage or radiation. Although neither cultured HMEC or their isogenic mammary fibroblasts express p21 at stasis, other cell types may be more vulnerable to DNA damage inducing stresses in culture, express p21, and show greater evidence of a DDR at stasis, including telomeric foci. HMEC in vivo may also experience p53-inducing stresses. The p53-dependent type of stasis arrest does not require critically short telomeres or genomic instability, and inactivation of p53 (or p21) function may facilitate overcoming this arrest (15,16). Reactivation of telomerase is neither necessary nor sufficient to overcome stasis.
Replicative senescence occurs in post-stasis HMEC (cells that have bypassed or overcome stasis) due to ongoing proliferation producing progressively shortened telomeres, in the absence of sufficient telomerase activity. When telomeres become critically short (mean TRF ≤ 5 kb), genomic instability and a DDR is elicited. Where wild-type p53 is present, most cells show a viable arrest; this barrier has been termed agonescence (2,3,17). Karyotypic analysis of HMEC at agonescence has shown that virtually all metaphases exhibit gross chromosomal abnormalities, predominantly telomere associations (2). This result is not consistent with a hypothesis that a p53-dependent senescence arrest due to telomere attrition occurs as soon as one uncapped telomere is present (18,19). When p53 is non-functional a viable arrest is not possible, and crisis-associated massive cell death occurs (3). Agonescence can be distinguished from stasis in HMEC by the presence of critically short telomeres and genomic instability, higher LI (~15%), arrest at all phases of the cell cycle, and presence of a DDR (Table 1). HMEC at agonescence as well as at stasis display a senescent morphology and SA-b-Gal, so these properties do not readily distinguishable between these two molecularly distinct senescence barriers. Crisis can be distinguished from agonescence in HMEC by a higher LI (~40%) and the absence of a viable arrest. Since most human epithelial and fibroblast cells induced to transform to immortality in culture had inactivation of p53 function to facilitate overcoming stasis (e.g., using viral oncogenes or inhibitors of p53 function), only crisis was observed in such cultures at replicative senescence.
The barrier due to telomere attrition can be overcome by the expression of sufficient telomerase to maintain stable telomere lengths. Overcoming replicative senescence may correlate with DCIS in vivo, which commonly display short telomeres, genomic instability, and telomerase reactivation (14,20). That most DCIS contain critically shortened telomeres indicates that the precursor cells did not express sufficient telomerase for telomere maintenance.
Cultured finite lifespan HMEC are vulnerable to oncogene-induced senescence (OIS) (21,22). HMEC that have attained immortality via reactivation of endogenous telomerase are no longer vulnerable to OIS, and show gain of malignancy-associated properties when exposed to oncogenes such as Raf-1, Ras or ErbB2 (21,23,24). HMEC immortalized by exogenous hTERT transduction may appear initially sensitive to OIS, but can maintain growth (25). The molecular correlates of OIS in HMEC differ from those seen in cells at stasis or telomere dysfunction (Table 1). OIS in HMEC does not require p16 or p53 function, and is independent of telomere length; its molecular properties are consistent with a DDR (21,22).
HMEC derived from reduction mammoplasties, milk, benign tumors, and non-tumor mastectomy tissues have been grown in either serum-containing (MM, M85, M87A) or serum-free (MCDB 170) media (Chart 1)(4-6,26,27). Depending upon the media and culture conditions, active proliferation has ceased after ~10-60 PD (Figure 2). In media that support fewer PD, levels of p16 expression increase earlier; virtually all cells express p16 at stasis in all media used (1,4). The molecular profile of the HMEC at stasis is similar regardless of their PD potential or growth media (Table 1), with one noticeable difference. HMEC grown in serum-containing media have a typical senescent morphology of large flat vacuolated cells, whereas HMEC that had been grown in serum-free MCDB170 exhibit a more elongated morphology showing abundant stress fibers (4,6,27). We believe this difference is due to the serum-free medium being more stressful for cultured HMEC, consistent with the early rise of p16 and the low PD potential of HMEC initiated in MCDB170 (commercial MEGM [Lonza] and M171 [Life Technologies] are based on our original MCDB170) (1,6). This difference in morphology may have led other investigators to consider this stasis arrest distinct, and refer to it as “M0” (28,29).
Pre-stasis in M87A: Our most recently developed medium, M87A, supports long-term growth of the normal pre-stasis HMEC (Figures 2,3) (4). The pre-stasis HMEC we now distribute were grown in this medium. Populations contain a mixture of cells with markers of myoepithelial, luminal, and progenitor lineages; later passage cultures show fewer luminal cells (30,31). While not yet carefully examined, cells with luminal markers may also cease proliferation as a consequence of terminal differentiation rather than p16(+) stasis. Senescent cells remain genomically stable. We have examined gene transcript profiles, global promoter methylation, and DDRs as a function of passage from several individual’s HMEC (4,9). As expected, gene expression changes significantly with passage, while no obvious differences were seen for promoter methylation. Some interindividual differences could be detected in gene expression and extent of DDRs. Previous studies have shown interindividual differences in carcinogen metabolism (32), leading us to recommend that at least 2 individuals be examined to determine normal HMEC properties. We have also observed differences in lineage markers and differentiation correlated with the age of the specimen donor (14,30). With increasing age, HMEC from 4th passage pre-stasis strains and from uncultured dissociated organoids showed a decline of myoepithelial cells, and an increase of luminal cells that exhibited molecular features usually ascribed to myoepithelial cells (increased expression of integrin a6 and keratin (K)14). The proportion of c-Kit expressing cells (putative progenitors) also increased with age, and exhibited impaired microenvironment-directed differentiation and lineage specificity, consistent with alterations in Hippo pathway activation. These data suggest that the observed age-associated increase in luminal breast cancer could be connected to changes that occur normally with aging in the human breast. Myoepithelial cells are thought to be tumor-suppressive and progenitors are putative etiological roots of some breast cancers. Thus during the aging process, the potential target cell population may increase while there is a simultaneous decrease in the cells thought to suppress tumorigenic activity.
Figure 2. Population doubling potential of pre-stasis HMEC in different media. Primary cultures from reduction mammoplasty specimen 184 were initiated from organoids in different media. The number of PD in primary culture cannot be accurately determined; growth is shown starting from passage 2. All proliferation stopped in HMEC grown in serum containing media (MM and M87A with oxytocin (X)) after 10-50 PD beyond passage 2. The extensive proliferative potential in M87A+X supports generation of large batches of early passage pre-stasis HMEC from individual donors. HMEC initiated in serum-free MCDB 170 (commercial MEGM) show rapid induction of p16 and cessation of growth. When cultures are allowed to sit without subculture for 2-3 weeks, post-selection post-stasis HMEC emerge and maintain growth to replicative senescence. If cultures are repeatedly subcultured, fewer to no post-selection cells may emerge.
Figure 3. Characterization of pre-stasis HMEC grown in M85 with oxytocin. A. Expression of markers associated with proliferation (LI) and senescence (p16, SA-β-Gal) in pre-stasis 184D HMEC with increasing passage; stasis was at passage 15. Note the reciprocal relationship between the small cells with a positive LI, and the larger, often vacuolated cells (senescent morphology) that are positive for p16 and SA-β-Gal, and negative for LI. B. Immunohistochemistry expression of luminal lineage marker K19 in pre-stasis 184D HMEC. C. Immunofluorescence expression of luminal lineage markers EpCam and Muc1 in pre-stasis 48RT HMEC. Size marker = 200 microns.
Pre-stasis and post-stasis in MM: We have cultured pre-stasis HMEC from over 150 individuals in serum-containing media and have not observed even a single instance of a cell spontaneously overcoming the stasis barrier. However, early experiments that exposed primary cultures of specimen 184 HMEC grown in MM to the chemical carcinogen benzo(a)pyrene (BaP) resulted in the emergence of HMEC colonies that maintained growth after the bulk of the cultures ceased proliferation at stasis (7,33). These BaP post-stasis populations (originally called Extended Life), ceased growth after an additional 10-40 PD, with the exceptional of very rare cells that became immortal cell lines (see below). All three BaP post-stasis cultures that have been examined showed loss of p16 expression, associated either with mutation or promoter silencing (1,13). Whole exome analysis (34) indicates that these cultures contain ~100-200 BaP-induced point mutations, the majority consistent with the known mutation spectrum of BaP. We have very limited quantities of BaP cultures available for distribution on a collaborative basis.
Pre-stasis and post-stasis in MCDB170: When the HMEC are cultured in the highly stressful serum-free MCDB170 medium, a small number of cells are able to overcome stasis in the absence of additional oncogenic exposures (6) (Figure 2). These post-stasis cells show methylation of the p16 promoter and absence of p16 expression, as well as nearly 200 other mostly cancer-associated changes in promoter methylation (in contrast to the BaP post-stasis cultures, which display only ~10 changes) (1,9). We originally called the emergence of these post-stasis cells “selection” and this class of post-stasis HMEC “post-selection”. We now recognize that selection (what other labs latter termed “M0”) is a stasis arrest. Although the pre-stasis populations may be heterogeneous with respect to a cell’s ease in silencing p16 to become post-selection (35), we believe the post-selection cells are induced by growth in the stressful (oncogenic) serum-free MCDB170 medium, i.e., post-stasis cells are not present in the starting normal pre-stasis cultures. This is based on the total absence in over 30 years of our work of any post-stasis cell emerging from normal pre-stasis HMEC grown in any of our serum-containing media, as well as the absence or reduction of post-selection HMEC emerging from pre-stasis HMEC grown in MCDB170 when there are small changes in media composition (e.g., absence of a cAMP stimulator) or methodology (e.g., sub-culturing cells approaching stasis/selection rather than waiting 2-3 weeks at stasis without subculture for the post-selection cells to emerge; we presume the induction of the p16(-) cells is occurring during this time when no cell divisions are observed). It is likely that post-stasis cells exist in some breast tissues; p16(-) HMEC have been seen in apparently normal breast tissues in vivo (35). These rare cells have been called vHMEC; the nature of the error(s) leading to the silencing of p16 in vHMEC in vivo is not known; the term vHMEC has also been used by others to refer to p16(-) post-stasis cells in culture that are specifically post-selection. Of note, it has been suggested that the aberrant post-selection post-stasis HMEC (which are sold commercially as normal primary HMEC, e.g., Lonza CC-2551 and Life Technologies A10565) may be on a pathway to metaplastic cancer (36).
Post-selection post-stasis p16(-) HMEC grow actively for an additional ~30-70 PD, depending on the individual. They express wild-type p53 that is present in a stable form (3,37,38). As they near agonescence, they exhibit a senescent morphology, SA-β-Gal, a DDR, and genomic instability (2,3). If p53 function is inactivated (e.g., using the genetic suppressor element GSE22 (3W9)) cells continue to proliferate for an additional ~2-4 passages, with increasing evidence of cell death and debris (i.e., crisis) (Figure 4)(3). The telomere dysfunction barrier is very stringent in post-selection post-stasis HMEC. We have never seen any unperturbed cell at agonescence spontaneously immortalize. We have also never seen any immortalization at crisis in post-selection HMEC, but rare immortalization at crisis using DN-p53 constructs has been reported by others (40,41). This stringency is due in part to the molecular nature of this barrier; cells that fail to maintain a G1 or G2 arrest with critically short telomeres will eventually die or become non-proliferative as a consequence of the genomic instability and mitotic catastrophes. Overcoming replicative senescence requires induction of sufficient telomerase; in post-selection post-stasis cells (see more below), more than one error is needed to reactivate telomerase. The probability of this occurring in one p53(+) cell during telomere dysfunction is exceedingly low, accounting for the lack of immortalization in the post-selection cultures at agonescence.
We have large supplies of post-selection HMEC available for distribution from women of various ages. It’s important to recognize that these cells are not normal, and acquire genomic instability as they are propagated in culture. In the past, due to the inability to attain long-term culture of normal pre-stasis HMEC, we provided post-selection HMEC for studies on finite HMEC. Since it is now possible to grow large quantities of normal pre-stasis HMEC, we recommend that studies aiming to understand normal HMEC behavior use normal HMEC and not the aberrant post-selection HMEC. For some experimental purposes, post-selection or other post-stasis HMEC may be preferable, e.g., examining the requirements for and mechanisms of overcoming the telomere dysfunction barrier, or assaying cells at different stages in progression.
Figure 4. Growth and morphology of post-stasis post-selection 184 with and without functional p53. 184B HMEC were transduced with GSE22-containing or control (Babe) vectors at passage 5. (A) growth curves of 184B-Babe and 184B-GSE22. Note the additional PD in the cultures lacking functional p53. We believe growth rates are similar ± p53, but the absence of p53-mediated growth inhibition allows more cells to continue to proliferate to crisis, leading to apparent faster growth of the population as cells near telomere dysfunction. (B) 184B-Babe at agonescence, 2 months after plating at passage 15, contains mostly large, flat cells with some vacuolization; the cell population can retain this morphology and viability for over a year. (C) 184-GSE22, two weeks after plating at passage 15, shows areas of small proliferating cells and many very large flat cells (arrows). (D) 184B-GSE22, four months after plating at passage 15, shows mostly large multi-nucleated, vacuolated cells and abundant cell debris. All photographs are at the same magnification. (3)
Post-stasis in M85 or M87A: More recently, we have generated additional types of post-stasis HMEC using shRNA to p16 (p16sh) (42) and transduction of a cyclin D1/CDK2 construct (D1) (43). As expected, in early passage HMEC grown in M85 or M87A, direct inhibition of p16 using p16sh, or transduction of D1, led to widespread bypass of stasis. Post-stasis p16sh HMEC show few differences in their pattern of promoter methylation or gene transcripts compared to their precursor pre-stasis cultures ((1,9)unpublished). They grow for an additional ~20-30 PD until agonescence (Figure 5). Unlike post-selection post-stasis, and similar to BaP post-stasis cells, p16sh post-stasis HMEC have generated rare clonal immortal lines during the period of genomic instability at agonescence. This difference (see more below) is presumably due to the ability of just one error to immortalize these cultures (42). We can provide limited amounts of the p16sh post-stasis HMEC for specific requests.
I want to add a few comments about nomenclature since this issue has frequently come up in discussions. It’s my general experience in science that functionally distinct molecules or molecular processes are given distinct names. Confusion could result if there were not distinct names for different, though closely related family members (e.g., growth factors and their receptors) or related mechanisms (e.g., apoptosis, anoikis, mitotic catastrophe). One of our overall goals is to try to model the many different in vivo pathways a normal cell can take to become malignant. Such information may assist individualized clinical interventions. Our data thus far indicate that molecular properties differ among different pathways, starting with early stage carcinogenesis. Pre-stasis HMEC differ from post-stasis HMEC, although both are finite, and as discussed above, there are significant differences among the various post-stasis types. Since these different post-stasis cultures are functionally different, we have given them distinct names. Similarly, since agonescence is molecularly and morphologically distinct from crisis, although both result from telomere attrition, we believe it important that there be distinct names. Confusion may also arise if similar mechanisms are given distinct names, e.g., we view what we are defining as stasis as having also been referred to as M0, M1, MINT, M1.5, premature senescence, replicative senescence, and culture shock. Telomere dysfunction due to telomere attrition (replicative senescence displaying as agonescence or crisis) has been called replicative senescence, crisis, M2, and M1. Indeed, it was this situation that prompted our initial efforts to generate molecularly defined nomenclature for the senescence barriers (2). I encourage everyone to employ the molecularly defined nomenclature we have presented here for our HMEC culture system.
The replicative senescence barrier can be overcome or bypassed by the expression of sufficient telomerase to maintain stable telomere lengths. However, unlike bypassing an arrest based upon keeping an active RB (e.g., abrogating the p16 expression that enforces stasis), the widespread chromosomal derangements caused by telomere dysfunction are not reversible. Consequently, in cultured HMEC, overcoming replicative senescence differs from bypassing/overcoming stasis in that the escaped immortal cells will contain the genomic abnormalities accumulated to that point, and may retain some degree of genomic instability (14). We hypothesized that the inherent genomic instability consequent to telomere attrition can give rise to the errors permissive for telomerase reactivation, as well as many of the “passenger” errors seen in breast carcinomas, and that the generation of breakage-fusion-bridge (BFB) cycles prior to immortalization may underlie some of cancer-associated genomic instability (3,44). Thus many errors that can contribute to the ultimate cancer cell phenotype, including level of aggressiveness, may arise prior to immortalization and malignancy. This hypothesis is consistent with publications indicating that many properties of invasive tumors are already present in their pre-invasive DCIS lesions, such as tumor markers, gene expression profiles, gene methylation, PIK3CA mutations, and genomic errors (45-49).
Our recent studies on immortalization support these hypotheses, as well as our model of the HMEC senescence barriers (34,42). Reactivation of sufficient telomerase in cultured finite lifespan HMEC has been difficult to achieve using pathological relevant agents (i.e., not ectopic hTERT or viral oncogenes). Reported immortal lines have been rare clonal events (3,7,8,40-42,50-52). This likely reflects the fact that large long-lived animals such as humans have evolved mechanisms for stringent repression of telomerase in normal adult non-stem cells, presumably for tumor suppression. In contrast, cells from small short-lived animals such as mice do not show such stringent telomerase repression, and, lacking the replicative senescence barrier, readily immortalize once they overcome stasis (53-55). We believe that immortalization with telomerase reactivation is a rate-limiting step in human epithelial carcinogenesis, and thus great caution should be exercised in extrapolating mechanisms of rodent malignant progression to humans. One of the goals of our long-term program in developing an HMEC model system of transformation has been to make available experimentally tractable human cells for examination of this crucial step in human malignant progression, since it cannot be accurately modeled in mice (42). As described further below, our initial work generated rare clonally immortalized lines containing multiple genomic errors, making it difficult to examine the immortalization process. Our recent work has shown that efficient non-clonal immortalization is possible when the two main senescence barriers (stasis and replicative senescence) are directly targeted; resulting immortalized lines display normal karyotypes. These new methods should allow experimental examination of the immortalization process in HMEC.
We have generated a variety of immortally transformed lines using various oncogenic agents (see: Chart 1 and Cell Types Generated) (3,7-9,21,33,42,50,56-59). Most of these lines were derived from post-stasis cultures, although in a few instances (involving hTERT or c-Myc transduction) lines emerged from perturbations of pre-stasis populations. Our first immortal lines were obtained from the BaP post-stasis cultures, 184Aa, 184Bd, and 184Be (7,8,33,56). Extremely rare immortal lines have appeared at agonescence (184A1, 184AA4, 184AA8, 184B5, 184BE1, 184BE2). The BaP post-stasis cultures harbor small mutations from their BaP exposure, as well as loss of p16 expression (34). Large genomic errors such as copy-number variations (CNV) were seen in examined immortal lines but not the precursor BaP post-stasis cultures (34,60). Presumably, the genomic instability at agonescence produced the CNV and the rare errors that allowed telomerase reactivation and immortalization. More frequent but still rare clonal lines appeared at agonescence following transduction of the breast cancer–associated oncogene ZNF217 into the 184Aa population (184AaZN1-3 (50)). More frequent immortal clonal outgrowths at crisis were seen when p53 was inactivated in 184Aa using GSE22 (184AaGS1-2). All examined (karyology and/or aCGH) clonal lines exhibited numerous large genomic errors. Uniform immortalization was obtained following transduction of c-Myc into three different BaP post-stasis cultures (184AaMY1-5, 184BeMY, 184CeMY (42), indicating that one error is sufficient to immortalize these BaP post-stasis cultures.
Figure 5. Effect of c-MYC on post-stasis HMEC growth and TRAP activity. Top panel: schematic representation of protocol to directly target senescence barriers to achieve non-clonal immortalization. Middle panel: Pre-stasis 184D and 240L HMEC grown in M87A+CT+X were transduced at 3p with a p16sh-expressing retrovirus (MSCV, blue) or empty vector (black). At 4p cultures ±p16sh were transduced with c-Myc (BH2)(red +p16sh; purple -p16sh). c-Myc-transduced p16sh post-stasis HMEC maintained active growth indefinitely, associated with increased TRAP activity. The continuous exponential growth following c-Myc transduction of the 4p p16sh-post-stasis populations reflects the observed non-clonal immortalization. Cells transduced with p16sh alone bypassed stasis and ceased net growth at agonescence, with rare clonal immortalization at agonescence. Cells transduced with c-Myc alone ceased growth at stasis, with rare clonal escape from stasis leading to immortalized lines. Control cultures transduced with empty vectors ceased growth at stasis. In some TRAP assays, heat-treated controls (+) were run next to unheated (-) samples. Positive TRAP control samples are indicted by “+”. E.
Bottom panel: Pre-stasis 122L and 805P HMEC grown in M87A+CT+X were transduced at 3 or 4p respectively with a p16sh-expressing retrovirus (MSCV, blue), D1/CDK2 vector (blue dotted line) or empty vector (black). At 4p or 5p cultures ±p16sh or D1 were transduced with c-Myc (BH2)(red +p16sh; red dotted +D1). c-Myc-transduced post-stasis HMEC maintained active growth indefinitely. Each point indicates a passage.
No post-selection post-stasis HMEC has been observed to spontaneously immortalize, presumably due to the need for two errors for immortalization. Rare immortal lines have appeared at agonescence following overexpression of the breast cancer associated oncogene ZNF217 (184ZN4-7) (42,50). Transduction of c-Myc alone produced 1 clonal line in 10 independent experiments (184SMY1). We hypothesize that rare errors generated by the genomic instability at agonescence may complement ZNF217 or c-Myc to allow telomerase reactivation. Overexpression of both c-Myc and ZNF217 in post-selection HMEC was able to produce clonal immortal lines in repeat experiments (184ZNMY1-4, unpublished); some of these clonal lines immortalized early, prior to agonescence.
More recently, we have tested our hypothesis about the two main senescence barriers by targeting these barriers in early passage pre-stasis cells grown in M85 or M87A (42). If we were correct in postulating that the genomic errors seen in clonal immortalized cells (in vitro and in vivo cancer) were required to bypass/overcome these barriers, but genomic instability was not needed per se, then directly targeting these barriers should produce non-clonal immortalized lines lacking gross genomic errors. That is indeed what we have observed, using primary HMEC from 4 different specimen donors ranging in age from 19 to 91 (Figure 5). The stasis barrier was targeted using p16sh or D1/CDK2, and the replicative senescence barrier was targeted by transduction of c-Myc. Efficient immortalization has been seen with all specimens in all experiments (currently 8 independent). Karyotypes obtained from 6 different immortalized lines (184Dp16sMY, 240Lp16sMY, 240LD1, 122Lp16sMY, 122LD1, 805Pp16sMY) were all normal diploid at early passage. From visual observation, there could be <100% transition to post-stasis using either p16sh or D1 alone, while the transition to immortality from the post-stasis cultures appeared close to 100%. c-Myc transduction of the BaP post-stasis cultures also appeared to give close to 100% immortalization efficiency. This transduction was done in 184Aa and 184Be after the start of telomere dysfunction, so the immortalized lines do not exhibit a normal karyotype, while 184Ce, transduced earlier, generated 184CeMY with a normal karyotype (tho with many BaP-induced point mutations). The ability of the p16sh, D1, and BaP post-stasis cells to be immortalized by one error (i.e., c-Myc) likely accounts for the rare “spontaneous” immortalization seen in these cultures at agonescence (184Fp16s, 184Dp16s, 240Lp16s, 805Pp16s and the BaP-exposed lines). Amplification at the c-Myc region was observed in some of these clonal lines. We do not know why the post-selection post-stasis cells differ in this respect. We hypothesize that their prior experience of high stress may play a role. Transduction of c-Myc alone into pre-stasis HMEC in low stress-media gave rare clonal escape from stasis; these “post-stasis” cells maintained growth to became clonal immortal lines (184FMY2, 184DMY3, 240LMY, 122LMY). All clonal lines examined showed multiple CNV by aCGH. These new lines are still being characterized and are available upon specific request.
Our studies have led us to propose the existence of a previously unreported step in HMEC immortalization, termed conversion, based on observing the need for multiple steps to attain a fully immortal potential (i.e., synthesizing sufficient telomerase to maintain stable telomere lengths) (3,8,21,42,56,59) (8,21,50,56-59,61,62). Even after HMEC have acquired the errors allowing them to bypass/overcome both stasis and replicative senescence, and express hTERT, the resultant cells with indefinite proliferative potential still progress through further changes. Conversion is most prominent in cells that immortalize while retaining functional p53. Consequently, this process has not been widely studied, as most in vitro immortalized human epithelial and fibroblast cells had p53 inactivated prior to immortalization.
Conversion has been most extensively studied in the immortal 184A1 line, which first appeared ~passage 8 in the 184Aa BaP post-stasis population, and had a mean TRF value of ~5 kb when first examined at passage 11 (Figure 6). We noted that cells that overcame agonescence gained the potential to express telomerase, but initially displayed little telomerase activity, and had ongoing telomere erosion with proliferation. When telomeres got extremely short (<3 kb), the conversion process ensued. Expression of the CKI p57Kip2 initially abruptly increased and then slowly declined, associated with initial slow-heterogeneous growth and then gradual re-attaining of uniform good growth. Telomerase activity gradually increased, and the faint very short telomeres seen during conversion gradually became stabilized with a mean TRF of ~3-7 kb. As telomerase activity increased, the immortal lines gradually developed the ability to maintain growth in the presence of TGFb; this change is a direct consequence of the hTERT expression, as transduction of hTERT into post-selection HMEC confers the ability to maintain growth in TGFb in addition to producing uniform immortalization (58). A significant change that is associated with conversion and telomerase expression but not initial immortal potential is the loss of vulnerability to OIS (21), likely related to a role of telomerase in providing resistance to OIS (63).
Figure 6. Conversion of newly immortal p53(+) HMEC lines is associated with changes in many key properties. (A.B.) The p53(+) 184A1 line undergoing conversion exhibits changes in growth capacity (CFE) and expression of p57, expression of telomerase activity and mean TRF length, gains the ability to maintain growth in the presence of TGFb, and becomes resistant to OIS (8,21,50,56,61). When pre-conversion 184A1 is transduced with GSE22, there is a rapid increase in telomerase activity associated with stabilization of TRF length (8).
When we obtained immortal HMEC lines that lacked functional p53 (184AA2, 184AA3) we noted that they showed some initial telomerase activity, no p57 expression, and quickly attained good uniform growth ± TGFb. Their mean TRF length stabilized at ~4-5 kb and never declined to the very low levels seen in the p53(+) lines. The role of p53 in repressing telomerase activity in newly immortal lines was then demonstrated by inactivating p53 (using GSE22) in pre-conversion 184A1 (Figure 6) (8). Endogenous telomerase activity was quickly expressed and mean TRF lengths stabilized; existing p57 expression was rapidly reduced. GSE22 transduction into the finite lifespan precursors of the immortal lines did not induce significant telomerase activity indicating that abrogation of p53 function alone is not sufficient for telomerase reactivation in post-stasis HMEC. These results suggest that the newly immortal p53(+) lines have the potential to express telomerase, but expression is low due to a p53-mediated repression (unpublished data have also indicated that newly immortal p53(+) lines express low telomerase activity which can be inhibited). We believe that fully immortal p53(-) lines are expressing an accelerated but molecularly similar conversion process as occurs in the p53(+) lines. The resulting cultures express similar properties, e.g., short stable telomeres, resistance to OIS, and hundreds of promoter methylation changes (9), but the p53(-) lines never encountered the extremely short telomeres and p57 expression shown by the p53(+) lines. Newly immortal 184AA2 and 184AA3 did both initially briefly show slower growth and lower TRAP activity than at later passages.
While we have gained much information about the molecular properties associated with conversion, much about this process remains unknown. Our current speculation is that conversion may reflect a need to alter chromosome conformation at the telomeres when cells transition from a finite state (no stable telomere length maintenance) to one where sufficient telomerase maintains the short stable telomeres. As well studied in yeast, immortal cells can have “counting” mechanisms to maintain telomeres within a limited size range (64). Since most human carcinoma cells, as well as our immortal HMEC lines, maintain telomeres within a short range (mean TRF ~3-7 kb) (56,65,66), some type of “counting” mechanism likely is involved. Short stable telomeres are not seen in normal telomerase expressing human cells such as stem cells and lymphocytes (67), suggesting that active processes may be required for conversion to the distinct telomeric state seen in the immortalized and cancer-derived cells. Functional p53 may present a partial barrier to the conversion process until very short telomeres provoke a structural change at the telomeric ends. Since the majority of breast cancers express wild-type p53, it is possible that the slower p53(+) version of the conversion process may be relevant to early-stage breast carcinogenesis in vivo. We have speculated that the low levels of telomerase expression coupled with extremely short telomeres could make newly immortal p53(+) breast cancers particularly vulnerable to therapeutic interventions targeting telomere dynamics.
We, and others, have also immortalized finite HMEC by ectopic overexpression of the hTERT gene (58,68,69). Initial lines came from transducing hTERT into post-stasis post-selection HMEC grown in serum-free medium (184BTERT, 48RTERT, 161TERT); consequently, these immortalized lines contain the many cancer-associated and other aberrancies found in this p16(-) metaplastic post-stasis type. We and others were unable to immortalize pre-stasis HMEC grown in serum-free media with transduced hTERT, and only rare clonal immortalization occurred when hTERT was transduced into pre-stasis HMEC grown in MM (184FTERT) (58). Hypothesizing that high stress exposure might prevent hTERT immortalization of pre-stasis HMEC, we transduced hTERT into passage 3 pre-stasis 184D grown in low-stress M87A, once that medium was available. Nearly uniform immortalization was observed, generating the non-clonal line 184DTERT1 (42). Caution: TERT reactivation during in vivo malignant progression is not caused by ectopic introduction. The errors and mechanisms responsible for TERT reactivation are key to understanding human tumorigenesis (and its potential prevention), and they cannot be investigated or recapitulated in lines created by TERT transduction. Further, we have seen that ectopic TERT overexpression creates lines with properties not found in normal or abnormal HMEC in vivo. TRAP activity is quite high [unpublished], compared to more modest levels in most in vitro immortalized and cancer-derived lines. Where examined, mean TRF length is longer than the typical 2-8 kb range of immortalized and cancer-derived lines (56,58,66), suggesting that these TERT-immortalized lines did not undergo conversion and therefore have telomere dynamics distinct from most carcinoma cells in vivo. Telomerase expression can enable cells to become resistant to OIS and maintain growth in the presence of TGFβ (21,25,58,63); other signaling pathways are also affected by transduced TERT (70). 184DTERT did not express and was not inducible for p16 [unpublished]. Based on such data, I do not think that TERT-immortalized HMEC represent accurate models for normal, aberrant, or cancer (including cancer stem) HMEC in vivo; rather their expression of in vitro induced artifacts can potentially obscure understanding of in vivo biology. Since these immortalized lines are aberrant compared to normal HMEC (which are finite, with no or low TRAP activity, and with intact stasis, OIS, and replicative senescence tumor suppressor barriers), they cannot be accurate normal controls.
In general, we have seen that different methods of producing immortal HMEC, e.g., culture medium, oncogenic agent, agent used for direct bypass of stasis, can yield cell lines with significantly different phenotypes. The age of the specimen donor may also influence phenotype. Immortalized lines generated from younger women tend to express a phenotype most similar to the basal subtype of human breast cancers, while more luminal phenotypes can be seen in lines from older women (43). The use of the cyclin D1/CDK2 construct to bypass stasis promoted a more luminal phenotype, even in younger specimens (43). Immortalization of the post-selection post-stasis types may favor a metaplastic phenotype (36). Possibly, the prior prevalence of immortalized lines with only a basal or metaplastic phenotype was a consequence of using cells from mostly younger women combined with stressful culture medium that did not support growth of normal HMEC with luminal or progenitor properties. Additionally, in vivo human malignant progression is influenced by interactions with the microenvironment, which our current culture conditions do not recapitulate.
Once the HMEC are immortally transformed and no longer vulnerable to OIS, 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) (21,23,24). Finite lifespan, OIS-sensitive HMEC cannot be rendered malignant by the same oncogenes. Thus many reports in the literature show that one error can confer malignancy-associated properties to abnormal, immortally transformed lines such as MCF10A and 184B5 – because these aberrant lines have abrogated all the major tumor suppressor mechanisms! Comparisons of non-malignant immortal lines with oncogene-exposed derivatives that had gained anchorage-independent growth did not show major differences in gene transcript profiling or global promoter methylation, in contrast to the major differences seen between all finite and all immortalized cultures (9,71). These data are consistent with the acquisition of immortality, rather than the acquisition of malignancy, as the step in human carcinogenesis most associated with molecular alterations.
Altogether, these studies validate our model of the senescence barriers encountered by cultured HMEC, indicate that genomic instability per se is not needed for immortalization, and support the hypothesis that carcinoma-associated genomic instability (along with many “passenger” errors) may have its origin in the inherent instability induced by telomere dysfunction at replicative senescence. The reproducible ability to generate non-malignant (OIS resistant) immortalized lines lacking gross genomic errors can facilitate further investigation of the molecular underpinnings of the immortalization process, and provides substrates to examine the effects of additional oncogenes and in-vivo identified genomic alterations on malignant progression.
Note: Immortal HMEC have been actively transformed to immortality. Normal human somatic cells are finite, and vulnerable to multiple tumor suppressor barriers (stasis, replicative senescence, OIS). The stringent replicative senescence barrier to immortality has evolved in large-long-lived animals like humans to suppress tumorigenesis. Immortality (expression of sufficient telomerase to maintain stable telomere lengths) is the most common alteration from normal associated with human solid cancers. We believe that attaining immortality is likely the most rate-limiting step in human carcinogenesis – immortally transformed lines such as 184A1 and MCF10A have acquired the errors that allowed them to overcome all tumor suppressor barriers, so that the overexpression of one oncogene can confer malignancy. Our immortal lines cluster with tumor derived lines and not finite HMEC in properties such as gene expression, promoter methylation, and resistance to OIS and TGFb growth inhibition. Immortal lines may be non-malignant, but they are NOT normal, “normal”, or untransformed. Please do not refer to immortal HMEC (or any immortalized human cells) as normal or untransformed. I sometimes despair about how we will be able to understand and develop therapeutics for early stage human epithelial carcinogenesis when immortal cell lines such as 184A1 and MCF10A, or TERT-immortalized post-selection post-stasis HMEC, are routinely referred to in the literature as normal or untransformed, employed as “normal” controls, or used as a starting point to study “early stage” carcinogenesis. I view this as similar to calling telomerase(+) DCIS “normal”, and using it as a normal control for cancer or the starting point for studying early stage carcinogenesis. There is already much in the literature to indicate that many, if not most of the significant alterations seen in breast and other carcinomas are already present in the pre-malignant stage.
1. Brenner AJ, Stampfer MR, Aldaz CM. Increased p16INK4a expression with onset of senescence of human mammary epithelial cells and extended growth capacity with inactivation. Oncogene 1998;17:199-205.
2. Romanov SR, Kozakiewicz BK, Holst CR, Stampfer MR, Haupt LM, Tlsty TD. Normal human mammary epithelial cells spontaneously escape senescence and acquire genomic changes. Nature 2001;409(6820):633-7.
3. Garbe JC, Holst CR, Bassett E, Tlsty T, Stampfer MR. Inactivation of p53 function in cultured human mammary epithelial cells turns the telomere-length dependent senescence barrier from agonescence into crisis. Cell Cycle 2007;6(15):1927-36.
4. Garbe JC, Bhattacharya S, Merchant B, Bassett E, Swisshelm K, Feiler HS, Wyrobek AJ, Stampfer MR. Molecular distinctions between stasis and telomere attrition senescence barriers shown by long-term culture of normal human mammary epithelial cells. Cancer Res 2009;69(19):7557-68.
6. Hammond SL, Ham RG, Stampfer MR. Serum-free growth of human mammary epithelial cells: Rapid clonal growth in defined medium and extended serial passage with pituitary extract. Proc Natl Acad Sci USA 1984;81:5435-39.
7. Stampfer MR, Bartley JC. Induction of transformation and continuous cell lines from normal human mammary epithelial cells after exposure to benzo[a]pyrene. Proc Natl Acad Sci U S A 1985;82(8):2394-8.
8. Stampfer MR, Garbe J, Nijjar T, Wigington D, Swisshelm K, Yaswen P. Loss of p53 function accelerates acquisition of telomerase activity in indefinite lifespan human mammary epithelial cell lines. Oncogene 2003;22(34):5238-51.
9. Novak P, Jensen TJ, Garbe JC, Stampfer MR, Futscher BW. Step-wise DNA methylation changes are linked to escape from defined proliferation barriers and mammary epithelial cell immortalization. Cancer Res 2009;67:5251-58.
11. Loughran O, Malliri A, Owens D, Gallimore PH, Stanley MA, Ozanne B, Frame MC, Parkinson EK. Association of CDKN2A/p16INK4A with human head and neck keratinocyte replicative senescence: relationship of dysfunction to immortality and neoplasia. Oncogene 1996;13(3):561-8.
14. Chin K, de Solorzano CO, Knowles D, Jones A, Chou W, Rodriguez EG, Kuo WL, Ljung BM, Chew K, Myambo K, Miranda M, Krig S, Garbe J, et al. In situ analyses of genome instability in breast cancer. Nat Genet 2004;36(9):984-8.
15. Rheinwald JG, Hahn WC, Ramsey MR, Wu JY, Guo Z, Tsao H, De Luca M, Catricala C, O'Toole KM. A two-stage, p16INK4a-and p53-dependent keratinocyte senescence mechanism that limits replicative potential independent of telomere status. Mol Cell Biol 2002;22(14):5157-72.
16. Evans RJ, Wyllie FS, Wynford-Thomas D, Kipling D, Jones CJ. A P53-dependent, telomere-independent proliferative life span barrier in human astrocytes consistent with the molecular genetics of glioma development. Cancer Res 2003;63(16):4854-61.
17. Tlsty TD, Romanov SR, Kozakiewicz BK, Holst CR, Haupt LM, Crawford YG. Loss of chromosomal integrity in human mammary epithelial cells subsequent to escape from senescence. J Mammary Gland Biol Neoplasia 2001;6(2):235-43.
19. Ouelette MM, Liao M, Herbert B, Johnson M, Holt SE, Liss HS, Shay JW, Wright WE. Subsenescent telomere lengths in fibroblasts immortalized by limiting amounts of telomerase. J Biol Chem 2000;275:10072-76.
20. Meeker AK, Argani P. Telomere shortening occurs early during breast tumorigenesis: a cause of chromosome destabilization underlying malignant transformation? J Mammary Gland Biol Neoplasia 2004;9(3):285-96.
21. Olsen CL, Gardie B, Yaswen P, Stampfer MR. Raf-1-induced growth arrest in human mammary epithelial cells is p16-independent and is overcome in immortal cells during conversion. Oncogene 2002;21:6328-39.
22. Cipriano R, Kan CE, Graham J, Danielpour D, Stampfer M, Jackson MW. TGF-beta signaling engages an ATM-CHK2-p53-independent RAS-induced senescence and prevents malignant transformation in human mammary epithelial cells. Proc Natl Acad Sci U S A 2011;108(21):8668-73.
25. Sherman MY, Meng L, Stampfer M, Gabai VL, Yaglom JA. Oncogenes induce senescence with incomplete growth arrest and suppress the DNA damage response in immortalized cells. Aging Cell 2011;10(6):949-61.
29. Ramirez RD, Morales CP, Herbert BS, Rohde JM, Passons C, Shay JW, Wright WE. Putative telomere-independent mechanisms of replicative aging reflect inadequate growth conditions. Genes Dev 2001;15(4):398-403.
30. Garbe JC, Pepin F, Pelissier FA, Sputova K, Fridriksdottir AJ, Guo DE, Villadsen R, Park M, Petersen OW, Borowsky AD, Stampfer MR, Labarge MA. Accumulation of multipotent progenitors with a basal differentiation bias during aging of human mammary epithelia. Cancer Res 2012;72(14):3687-701.
31. Pelissier FA, Garbe JC, Ananthanarayanan B, Miyano M, Lin C, Jokela T, Kumar S, Stampfer MR, Lorens JB, LaBarge MA. Age-related dysfunction in mechanotransduction impairs differentiation of human mammary epithelial progenitors. Cell Rep 2014;7(6):1926-39.
35. Holst CR, Nuovo GJ, Esteller M, Chew K, Baylin SB, Herman JG, Tlsty TD. Methylation of p16(INK4a) Promoters Occurs in Vivo in Histologically Normal Human Mammary Epithelia. Cancer Res 2003;63(7):1596-601.
36. Keller PJ, Arendt LM, Skibinski A, Logvinenko T, Klebba I, Dong S, Smith AE, Prat A, Perou CM, Gilmore H, Schnitt S, Naber SP, Garlick JA, et al. Defining the cellular precursors to human breast cancer. Proc Natl Acad Sci U S A 2012;109(8):2772-7.
37. Lehman T, Modali R, Boukamp P, Stanek J, Bennett W, Welsh J, Metcalf R, Stampfer M, Fusenig N, Rogan E, Reddel R, Harris C. p53 mutations in human immortalized epithelial cell lines. Carcinogenesis 1993;14(5):833-39.
39. Ossovskaya VS, Mazo IA, Chernov MV, Chernova OB, Strezoska Z, Kondratov R, Stark GR, Chumakov PM, Gudkov AV. Use of genetic suppressor elements to dissect distinct biological effects of separate p53 domains. Proc Natl Acad Sci USA 1996;93(19):10309-14.
41. Gollahon LS, Shay JW. Immortalization of human mammary epithelial cells transfected with mutant p53 (273his). Oncogene 1996;12:715-25.
42. Garbe JC, Vrba L, Sputova K, Fuchs L, Novak P, Brothman AR, Jackson MW, Chin K, LaBarge MA, Watts G, Futscher BW, Stampfer MR. Immortalization of normal human mammary epithelial cells in two steps by direct targeting of senescence barriers does not require gross genomic alterations. Cell Cycle 2014;13(21):3423-35.
43. Lee JK, Garbe JC, Vrba L, Miyano M, Futscher BW, Stampfer MR, Labarge MA. Age and the means of bypassing stasis influence the intrinsic subtype of immortalized human mammary epithelial cells. Frontiers in Cell and Developmental Biology 2015.;3:13
44. Stampfer MR, Garbe JC, Labarge MA. An Integrated Human Mammary Epithelial Cell Culture System for Studying Carcinogenesis and Aging. In: Schatten H, editor. Cell and Molecular Biology of Breast Cancer. New York: Springer; 2013. p 323-61.
46. Muggerud AA, Ronneberg JA, Warnberg F, Botling J, Busato F, Jovanovic J, Solvang H, Bukholm I, Borresen-Dale AL, Kristensen VN, Sorlie T, Tost J. Frequent aberrant DNA methylation of ABCB1, FOXC1, PPP2R2B and PTEN in ductal carcinoma in situ and early invasive breast cancer. Breast Cancer Res 2010;12(1):R3.
47. Miron A, Varadi M, Carrasco D, Li H, Luongo L, Kim HJ, Park SY, Cho EY, Lewis G, Kehoe S, Iglehart JD, Dillon D, Allred DC, et al. PIK3CA mutations in in situ and invasive breast carcinomas. Cancer Res 2010;70(14):5674-8.
48. Ma XJ, Salunga R, Tuggle JT, Gaudet J, Enright E, McQuary P, Payette T, Pistone M, Stecker K, Zhang BM, Zhou YX, Varnholt H, Smith B, et al. Gene expression profiles of human breast cancer progression. Proc Natl Acad Sci U S A 2003;100(10):5974-9.
51. Soule HD, Maloney TM, Wolman SR, Peterson WD, Brenz R, McGrath CM, Russo J, Pauley RJ, Jones RF, Brooks SC. Isolation and characterization of a spontaneously immortalized human breast epithelial cell line, MCF-10. Cancer Res 1990;50(September 15, 1990):6075-86.
54. Seluanov A, Hine C, Bozzella M, Hall A, Sasahara TH, Ribeiro AA, Catania KC, Presgraves DC, Gorbunova V. Distinct tumor suppressor mechanisms evolve in rodent species that differ in size and lifespan. Aging Cell 2008;7(6):813-23.
55. Yasaei H, Gilham E, Pickles JC, Roberts TP, O'Donovan M, Newbold RF. Carcinogen-specific mutational and epigenetic alterations in INK4A, INK4B and p53 tumour-suppressor genes drive induced senescence bypass in normal diploid mammalian cells. Oncogene 2013;32(2):171-9.
56. Stampfer MR, Bodnar A, Garbe J, Wong M, Pan A, Villeponteau B, Yaswen P. Gradual phenotypic conversion associated with immortalization of cultured human mammary epithelial cells. Mol Biol Cell 1997;8:2391-405.
57. Stampfer MR, Yaswen P. Immortal transformation and telomerase reactivation of human mammary epithelial cells in culture. In: Mattson M, Pandita T, editors. Telomerase, Aging and Disease. Volume 8, Advances in Cell Aging and Gerontology. Amsterdam: Elsevier; 2001. p 103-30.
58. Stampfer MR, Garbe J, Levine G, Lichtsteiner S, Vasserot AP, Yaswen P. Expression of the telomerase catalytic subunit, hTERT, induces resistance to transforming growth factor beta growth inhibition in p16INK4A(-) human mammary epithelial cells. Proc Natl Acad Sci U S A 2001;98(8):4498-503.
61. Nijjar T, Wigington D, Garbe JC, Waha A, Stampfer MR, Yaswen P. p57/KIP2 loss of heterozygosity and expression during immortal conversion of human mammary epithelial cells. Cancer Res 1999;59:5112-18.
62. Stampfer MR, Labarge MA, Garbe JC. An Integrated Human Mammary Epithelial Cell Culture System for Studying Carcinogenesis and Aging. In: Schatten H, editor. Cell and Molecular Biology of Breast Cancer: Springer; in press.
66. Listerman I, Sun J, Gazzaniga FS, Lukas JL, Blackburn EH. The major reverse transcriptase-incompetent splice variant of the human telomerase protein inhibits telomerase activity but protects from apoptosis. Cancer Res 2013;73(9):2817-28.
67. Aubert G, Baerlocher GM, Vulto I, Poon SS, Lansdorp PM. Collapse of Telomere Homeostasis in Hematopoietic Cells Caused by Heterozygous Mutations in Telomerase Genes. PLoS Genet 2012;8(5):e1002696.
68. Kiyono T, Foster SA, Koop JI, McDougall JK, Galloway DA, Klingelhutz AJ. Both Rb/p16INK4a inactivation and telomerase activity are required to immortalize human epithelial cells. Nature 1998;396(6706):84-8.
71. Li Y, Pan J, Li JL, Lee JH, Tunkey C, Saraf K, Garbe JC, Whitley MZ, Jelinsky SA, Stampfer MR, Haney SA. Transcriptional changes associated with breast cancer occur as normal human mammary epithelial cells overcome senescence barriers and become immortalized. Mol Cancer 2007;6:7.