CELLULAR AGING: BASIC PHENOMENA

Development proceeds through many steps that ultimately produce germ (reproductive cells) and somatic (nonreproductive) cell lineages. The germ cell lineage population is potentially immortal, because the genes it carries can be passed on indefinitely. The somatic cells will ultimately age and die. Early studies by Carrel and coworkers suggested that, when isolated from organisms, individual somatic cells were immortal. This view remained prominent for many years, though subsequent studies showed that it was actually untrue.

Studies conducted in many laboratories have shown that after the migration of cells from tissue pieces into culture vessels there is a phase of rapid proliferation, followed by one of declining proliferative capacity, and ultimately a total loss of mitotic activity. Cultures that have entered this permanently quiescent state are termed "senescent." During this later nonmitotic phase, the cells change in size and morphology, become granular in appearance, and accumulate debris. The loss of proliferative capacity in human diploid cell culture populations is a well-regulated phenomenon. As populations of cells divide successive times, the generation time of the culture increases, and there is also failure of an increasing fraction of the population to replicate at all. There is strong evidence that substantial heterogeneity develops in populations of cells as they proliferate, with some members of the population exhibiting a high proliferative capacity while others are capable of only a small number of divisions.

Cellular senescence has been observed in a variety of cell types. Although the proliferative capacity differs between various types of cells, for any given cell type the proliferative capacity tends to be relatively constant. The inability of cell cultures to proliferate indefinitely is neither the result of technical difficulties nor of the depletion of essential metabolites. Hayflick and Moorhead (1961) concluded that the limited lifespan phenomenon observed in human diploid cells might be programmed, and/or that accumulation of genetic damage might limit proliferative capacity. They suggested that the limited proliferative lifespan of cell cultures was an expression of cellular aging in vitro.

Changes in cell morphology and contact

The steady loss of replicative potential is accompanied by a greater heterogeneity of cell sizes; cells near the end of their proliferative capacity tend to be much larger than those that have doubled a small number of times. A number of subcellular structures also change with age (see Figure 1). An increase in nuclear size, in nucleolar size, and in the number of multinucleated cells parallels the increase in cell size seen in slowly replicating or nonreplicating cells. Prominent Golgi apparati, evacuated endoplasmic reticula, increased numbers of cytoplasmic microfilaments, vacuolated cytoplasm, and large lysosomal bodies have been observed in senescent human fibroblasts.

Cultures near the end of their proliferative potential display a reduced harvest density and a lowered saturation density at the plateau phase of growth; that is, the number of cells per unit area is lower in near-senescent cultures that exhibit inhibition of growth due to contact with neighboring cells. At the end of their life span in vitro, substantial cell death occurs; however, a stable population emerges that can exist in a viable, nonproliferative state for many months. This stable population is capable of maintaining only an extremely low saturation density, equivalent to less than 5 percent of that reached by cultures that have doubled a small number of times. The observed decrease in saturation densities probably reflects an increased sensitivity to intercellular contact, rather than increased numbers of larger cells. Changes at the level of extracellular matrix (ECM; proteins secreted by the cell that remain attached to the cell), specific secretory proteins not connected with the ECM, and membrane-associated molecules occur in cells near the end of their proliferative life, but it is unclear whether they account for alterations in the nature of cell contact among these cells.

Senescence and cell-cycle progression

The growth arrest observed in senescent cultures is not random; senescent cells appear to stop dividing at a unique point in the cell cycle. Treatment of cells that have divided a relatively small number of times with growth stimulators (generically referred to as mitogens ) is followed by a so-called gap phase (G₁), which is a period during which cells appear quiescent (see Figure 2). In fact, many biochemical changes occur during G₁, but these are not apparent from gross cell morphology. After G₁, cells enter a synthesis (S) phase in which chromosomes are replicated. A second pause or gap occurs after the S phase (G₂) and before mitosis (M). When essential growth factors are removed or decreased, young cells may enter a distinct quiescent phase called G₀. The cells in living tissues are generally thought to be in a G₀ state of growth. Much evidence supports the view that cultures of senescent cells are blocked in late G₁ near the G₁/S boundary in the cell cycle. It is clear, though, that the irreversibly arrested condition ultimately reached by senescent cells is distinct from the G₀ phase young cells enter or any other definable stage of the cell cycle. When senescent cultures are maintained with reduced levels of mitogens that induce a quiescent state (G₀) in early-passage cells, the pattern of gene expression exhibited by senescent cells is different from that of a functional G₀ state. Furthermore, chromatin condensation patterns are consistent with arrest in late G₁. The fact that cells near the end of their proliferative life can express some of the gene characteristics of the G₁/S boundary suggests an abortive attempt to initiate DNA synthesis and arrest that growth in a unique state.

The reason that senescent cells fail to enter the S phase is unclear. Both young and senescent cells appear to respond to fresh mitogens by carrying out some of the same cell-cycle processes in roughly the same time frame. However, the ability to complete the mitogen-initiated cascade of signal transduction pathways and to synthesize DNA is lost in senescent cells. In fact, the hallmark of senescence in culture is the inability of cells to replicate their DNA following stimulation with mitogens.

In general, the synthesis rate of macromolecules decreases as cells approach the end of their proliferative life span in vitro, while the cellular content of macromolecules (except DNA) increases. This observation seems to indicate that, in senescent cells, there is a general dysregulation of coordinated processes, which uncouples DNA synthesis from the synthesis of other macromolecules. The synthetic rates of DNA, RNA, and protein decrease in cultures nearing the end of their proliferative potential, which may be related to altered chromatin template activity in senescent cells. Gorman and coworkers tested the possibility that the replicative enzymes themselves and/or replicationassociated processes such as control of DNA hierarchical structural orders, were reduced or altered during senescence. They observed that, when treated with simian virus 40 (SV40), senescent cells can initiate an additional round of semiconservative DNA synthesis in old cells, indicating that the proteins necessary for DNA synthesis can still function in senescent cells following at least some types of stimuli.

Growth signals and senescence

If the cellular machinery that replicates DNA is intact in senescent cells, then it is possible that the proteins that carry some of the signals that stimulate DNA replication are diminished in senescent cells. Human diploid cells, near the beginning of their in vitro replicative life span, vigorously respond to stimulation with serum or a combination of certain growth factors by initiating DNA synthesis and mitosis. As these cells approach the end of their proliferative potential in culture, they become increasingly resistant to mitogenic signals. The basis for this loss of responsiveness cannot be attributed to any dramatic reductions in the number of cell-surface growth-factor receptors, nor to changes in the binding affinities with which these receptors bind their ligands (the generic name given to substances that bind to receptors). Instead, the intracellular proteins that convey signals from receptors to the nucleus decrease in number or become inactive.

Cells respond to mitogens through the intracellular actions of secondary events, including phospholipid turner, protein kinase C activation, and calcium mobilization. Alterations in post-receptor transduction pathways have been documented for each of these pathways. Repression of c-fos transcription in senescent cells is evidence for an early block in one or more pathways potentially required for DNA synthesis. However, this alone does not account for the cessation of growth in senescent cells, since overexpression of c-fos fails to prevent senescence. Interestingly, in skin fibroblast cultures derived from individuals with Werner syndrome (a disease of precocious aging), c-fos expression in response to serum is equal in both young and senescent cultures. The expression of some late-acting cell-cycleregulated genes, is also lower in senescent cells. Thus, it is the attrition of multiple signaling pathways associated with cell growth, rather than a decrease in any one pathway, that appears to govern the appearance of a senescent phenotype.

It is also known that the expression of certain genes that suppress growth, such as p53, p21, and p16, increases in senescent cells. These changes in expression may exert profound effects on the proliferative life of cultures. Treatment of normal cells with antisense p53 and retinoblastoma protein (pRB) extends the life span of senescent cells in a cooperative manner. Disruption of tumor/growth suppressor genes such as p21, a gene controlled by p53, can extend proliferative life span. Additionally, transfection of human tumor cells that lack p53 with the wildtype gene can induce a senescent phenotype; yet elimination of p53, and hence p21, fails to prevent senescence in human cells. Loss of p16, an inhibitor of cell cycle regulatory proteins, has also been reported to extend proliferative life span. Constitutive activation of the cellular-signaling protein MEK (a component of the MAP Kinase cascade) induces both p53 and p16 and results in permanent growth arrest in primary mouse fibroblasts. In contrast, overexpression of constitutively active MEK causes uncontrolled mitogenesis and transformation in cells lacking either p53 or p16. Although these changes might at first appear to suggest that signaling proteins decline while growth-inhibitory proteins increase—as part of a coordinated mechanism— this may not be the case. At least one of the growth inhibitory genes (p16) appears to increase, because the family of proteins (Id proteins) that block its transcription in young cells decline in senescent cells. Thus, the senescence-associated increase in p16 results partially from loss of an inhibitor, rather than from direct stimulation of transcription.

The Genetics of cellular senescence

Skin fibroblasts from pairs of monozygotic twins show no significant difference in replicative life span within each pair, but do show such differences among pairs. In heterokaryons, which are formed by the fusion of two different cells, the nonproliferative phenotype of senescent cells in culture is dominant over normal proliferative cells and over immortalized cells, but not over immortalized variants having high levels of DNA polymerase α(DNA pol α) or those transformed by DNA tumor viruses. There are numerous reports documenting the presence of inhibitors of DNA synthesis in senescent cells, although the nature of these inhibitors is poorly understood, and whether any of them plays a causal role in senescence is unclear. The idea that senescent cells actively make an inhibitor of DNA synthesis, however, is supported further by the observation that expressible RNA derived from senescent fibroblasts, when microinjected into proliferation-competent cells, can inhibit their entry into DNA synthesis.

Some of the most compelling evidence in support of a genetic component for cellular senescence has been the finding that introduction of particular chromosomes into immortalized cells causes them to acquire a senescent, or at least a nongrowing, phenotype. Evidence that the cessation of proliferation is a programmed phenomenon is seen in studies that used SV40 large T antigen to produce a reversible escape from cellular senescence. This evidence suggests that senescence confers at least two distinct mortality states.

Genetic influences over the process of cellular senescence would necessarily be reflected in reproducible changes in gene expression. The list of molecular markers of senescence in culture has dramatically increased as the result of examining genes isolated from selective libraries and monoclonal antibody pools. Remarkably, the number of genes isolated by these methods includes those involved with the ECM, secretory proteins involved in growth-factor-mediated function, differentiation and shock proteins, inhibitors of DNA synthesis, and genes of unknown function.

Cellular senescence and aging in organisms

Aging changes in organisms involve different kinds of cells and tissues and may result from multiple mechanisms. Thus, the aging of postmitotic cells, such as neurons, may proceed by a different set of mechanisms than that of proliferating tissues such as skin, the lining of the gut, or the blood-forming elements. Extracellular matrix macromolecules, such as collagen and elastin, also change with age, and presumably through mechanisms that are unique to them. There are also potentially profound effects resulting from interactions among these components during aging. Questions that have long been pondered by researchers studying proliferative senescence is whether the phenomenon occurs in vivo and whether it reflects changes that occur in organisms as they age. It is generally accepted that aging occurs at a cellular level. If this is true, then the mechanisms that control aging operate, and should be quantifiable, at a cellular level. But do the same mechanisms lead to cellular senescence? There is no simple answer to this question.

One of the strongest supports for the use of the cell-culture senescence model in studies of aging was the demonstration of a relationship between species' maximal life span and proliferative life span. There is also a donor-age-associated decline in mitotic activity and proliferation rates in a wide variety of human and rodent tissues in vivo. Both of these observations seem to suggest that the factors controlling the life span of intact organisms also control the growth characteristics and proliferative potential of cells in a culture environment. It has been observed that cells from patients with accelerated aging syndromes, such as Werner syndrome, have a reduced proliferative capacity compared with control cells from normal donors of the same ages. Additionally, this disease is associated with decreased mitotic activity, DNA synthesis, and cloning efficiency. Other types of diseases can also strongly modulate proliferative life expectancy. The effects of diabetes are particularly pronounced. All of these observations would appear to suggest that the physiological condition of the donor is reflected to some extent by the rate of cellular senescence observed in vitro.

On the other hand, studies by Robbins et al. (1970) were unable to identify any cells in the skin of older individuals with the phenotype that emerges in vitro. This raises the question of whether or not proliferative senescence actually occurs in the tissues of intact organisms? One of the strengths of the cell-culture model is that it has permitted the study of a single cell type; however, this can also be a limitation. Most studies of cell proliferation are unable to elucidate the relative contribution of intrinsic versus extrinsic factors. For example, do cells only senesce when separated from other types of cells and extracellular matrices, as in a culture environment? A group of studies bearing on this point involved the serial transplantation of normal somatic tissues to new, young, inbred hosts each time the recipient approached old age. In general, normal cells serially transplanted to inbred hosts exhibit a decline in proliferative capacity and probably cannot survive indefinitely. Additionally, it has been reported that proliferative capacity is diminished in spleen cells derived from old animals and transplanted into young irradiate hosts, and that mouse epidermis from old donors retain an age-associated increase in susceptibility to carcinogens regardless of whether they were transplanted into young or old recipients.

Donor age and proliferative life span

What has generally been accepted as the strongest evidence supporting the usefulness of the cell senescence model of aging has been reports from several laboratories of a negative correlation between donor age and proliferative life span in vitro. On the basis of these observations, it follows that the physiological effects of aging in vivo are reflected by the life span of cells maintained in vitro. Other laboratories report that the colony-forming capacity of individual cells also declines as a function of donor age. These observations, and the relationship between growth potential and maximal species life span is generally regarded as clear evidence of a direct relationship between aging in vivo and in vitro.

In spite of the fact that the relationship between donor age and proliferative lifespan became widely accepted, the correlations reported were always quite low. Furthermore, the studies were never standardized for biopsy site or culture conditions, which further added to individual variation and further decreased the correlation. Probably the most important problem with human studies was that so few of them actually assessed the health status of the donors from whom the cell lines were established. Many of the cell lines used were established from cadavers. It is known that diseases such as diabetes exert a profound affect on the proliferative potential of cell cultures. It is probable that other diseases can also affect in vitro proliferative capacity. The failure to screen for donor health probably skewed the results of many studies. These factors were seldom discussed when considering the correlations between donor age and proliferative lifespan. Instead, the relatively low correlations were seen as the result of individual variations in a highly outbred and otherwise uncontrolled population, and it was generally accepted that stronger correlations would be found in a longitudinal study where cell lines were established from a small group of individuals throughout their life. Due to the relatively large numbers of lines examined, the correlations reported found general acceptance, even though they were weak. Nevertheless, it remained difficult to assess whether the reported correlations between donor age and replicative life span indicated any compromise of physiology or proliferative homeostasis in vivo.

In order to address some of the problems with standardization and health status, 124 cell lines, established from 116 donors who participated in the Baltimore Longitudinal Study of Aging (BLSA), and 8 samples of fetal skin were examined by Cristofalo and coworkers (Cristofalo, et al., 1998). In this study, all of the donors were medically evaluated and determined to be healthy (by the criteria of the BLSA). None of the donors had diabetes when the cell lines were established. The results of this study showed that in healthy donors there was no relationship between donor age and proliferative life span. A longitudinal study was also performed to evaluate the effects of age on in vitro life span within individuals. Multiple cell lines established from the same individuals at time intervals spanning as long as fifteen years were compared. Surprisingly, this longitudinal study also failed to reveal any relationship between donor age and proliferative capacity in the five individuals studied. In fact, the proliferative potentials of cultures established when donors were older was found to be greater than that in lines established at younger donor ages in four of the five donors.

In spite of these findings, the fact that the colony-forming capacity of individual cells had also been reported to decline with donor age still seemed to support a relationship between donor age and proliferative potential in culture. However, the colony-forming assay is strongly influenced by variations in growth rates. Both the initial growth rates and the rate of thymidine incorporation into cells is dramatically higher in fetal fibroblast lines than in adult lines, which indicates that cell lines established from fetal skin initially divided more frequently, even though the replicative life span of these cell lines do not exceed that observed in the postnatal cell lines. It would seem equally relevant that the rate of thymidine incorporation and initial growth rates of large groups of lines established from adults do not vary significantly with respect to age. In view of differences in initial growth rates and intraclonal variations in the proliferative potential of single cells, it seems probable that the clone-size distribution method of estimating proliferative life span and an actual determination of replicative life span measure different things.

It should be noted that studies supporting the inverse relationship between donor age and proliferative life span are not limited to human cells. Studies of rodent skin fibroblasts also appeared to support the existence of a small, though significant, inverse correlation between donor age and replicative life span. Furthermore, it was observed that treatment of hamster skin fibroblasts with growth promoters could extend the proliferative life of cultures established from young donors but had negligible effects on cultures established from older donors. However, even in rodents, the relationship between donor age and proliferative potential is not entirely clear. For example, an examination of hamster skin fibroblast cultures established from the same donors at different ages reveals no age-associated changes in proliferative potential in animals older than twelve months.

In vitro studies on aging have used two kinds of cell cultures. The predominant one has been fetal- or neonatal-derived cultures that show aging changes when subcultured at regular intervals; some of these alterations parallel aging changes in vivo. The other related paradigm is that of cells derived from donors of different ages and studied after only one, or a few, subcultivations. In either case, attempts to relate changes in these individual cells to changes occurring in organisms as they age forms one basis of the use of the in vitro model for studies of in vivo aging.

Biomarkers of cellular aging in vitro and in vivo

The failure to confirm the relationship between donor age and proliferative life span does not invalidate cell culture as a model for aging studies, because other significant alterations occur during senescence in vitro that also occur during aging in vivo. These changes include an increased number of copies of genetic material, increased cell size, decreased mitogenic response to growth signals, a decline in the expression of genes involved in growth control, increased expression of genes of the extracellular matrix, and various changes in cell morphology. In addition, a correlation between replicative capacity and initial telomere length of cells derived from donors up to ninety-three years of age has been reported.

However, not all changes that occur during senescence in vitro occur during the aging of organisms. For example, the induction of the c-fos gene following stimulation with serum is lost in senescent cells but is undiminished in old individuals. Even markers that have gained relatively widespread use have been found to detect different phenomena when they occur both in vivo and in vitro. One marker thought to be potentially universal was reported by Dimri et al. (1995), who observed increases in cytochemically detectable β-galactosidase activity (SA β-gal) at pH 6.0—both in cell cultures and in tissue sections obtained from old donors. They also observed that immortal cells exhibited no SA β-gal staining under identical culture conditions. Additionally, they interpreted their results as providing a link between replicative senescence and aging in vivo. These observations appear to be supported by studies in rhesus monkeys. This model has gained widespread use as a crude measure of senescence; however, a number of subsequent studies have shown that it is actually not as specific or universal as first claimed. For example, SA β-gal positive cells have been observed in quiescent cultures of mouse cells as well as some types of human cancer cells that were chemically stimulated to differentiate, even though none these cell types can be classified as senescent. Furthermore, one biochemical analysis demonstrated that β-gal activity was present at both pH 6.0 and pH 4.5 in a number of different tumor cells. Other studies failed to detect any correlation with donor age either in tissue sections or in skin fibroblast cultures established from donors of different ages. It was also observed that staining in skin sections occurred only in the lumen of sebaceous glands and shafts of hair follicles, which is consistent with detection of microbial invasion, rather than actual changes in the biochemistry of the human cells in tissue sections (Severino, et al., 2000). The source of these discrepancies cannot be entirely elucidated; however, it is clear that this type of staining cannot be used as a marker for all types of aging under all conditions.

An important biomarker that is linked to the cessation of mitotic activity associated with cellular senescence is the progressive loss of chromosomal telomeric repeats. The telomeres of human chromosomes are composed of several kilobases of simple repeats: (TTAGGG)n. Telomeres protect chromosomes from degradation, rearrangements, end-to-end fusions, and chromosome loss. During replication, DNA polymerases require an RNA primer for initiation. The terminal RNA primer required for DNA replication cannot be replaced with DNA, which results in a loss of telomeric sequences with each mitotic cycle of normal cells. The observation that telomere shortening correlates with senescence provides an attractive model for the way in which cells might "count" divisions. Telomere shortening has been observed in aging models both in vitro and in vivo.

An examination of some biomarkers of senescence lends at least some support to the hypothesis that the phenomenon occurs in vivo. For example, the messenger RNA (mRNA) for the angiogenesis inhibitor EPC-1 declines more than one-hundred-fold with proliferative aging in culture. Although the EPC-1 mRNA is readily detected in tissue sections from both young and old donors, the total amount of EPC-1 mRNA is also lower in older individuals. Furthermore, it is present in a mosaic pattern in the skin sections from older individuals. Thus, the decrease occurs only in some cells. Similarly, the average length of telomeres, which decreases progressively with increasing numbers of mitoses, exhibits large variations in multiple subclones established from one individual, again suggesting different telomere lengths are associated with different tissue regions in vivo. These observations suggest that loss of proliferative potential, in vivo, occurs in a mosaic pattern, and that both long- and short-lived cells may lie in close proximity in tissues. In a culture environment, the cells with the greatest growth capacity will ultimately become predominant in the culture. Hence, the selection during the procedure to establish cultures probably obscures differences in proliferative capacity that exist in vivo.

Senescence and differentiation

Although most of the studies of cellular senescence have focused on similarities to aging, some have considered homologies with developmental processes. A commonly held view is that cellular aging may follow a differentiation lineage model. In fact, it has been suggested that the finite life span of cells may represent a differentiation of cell types, and that the process of diploid cell growth may have an in vivo counterpart in hyperplastic processes (uncontrolled proliferation). Chronological time appears to be significantly less important than division events in the progression of the phenomenon. Although the state of differentiation may change in cells that senesce in vitro, there is, in fact, no evidence that the changes in gene expression observed in fetal cells as they senesce, in vitro, are tantamount to differentiation, in vivo. While analogous changes can be found, they are greatly outnumbered by the dissimilar features that characterize these two distinct phenomena. Hence, a comparison of senescence-associated changes and differences that exist between fetal and postnatal cells reveals little similarity.

At least some analogous similarities exist between senescence in fetal fibroblasts and developmental changes that occur in vivo. For example, it has been observed that addition of platelet-derived growth factor-BB (PDGF-BB) stimulates an increased mRNA abundance of the transcript encoding PDGF-A chain in fetal and newborns and a relatively small response in adult cells. Senescence in vitro of newborn fibroblasts appears to result in the acquisition of the adult phenotype by decreasing their response to PDGF-BB. On the other hand, there are a number of differences reported between cell lines established from fetal and adult tissues that are related to growth factor requirements for proliferation and migration that remain disparate, even as these cultures become senescent. For example, fetal dermal fibroblasts will proliferate in either plasma or serum, while adult dermal fibroblasts require serum. It is also noteworthy that the expression of some genes, such as SOD-2, increases during proliferative senescence but only in some types of fibroblasts; in other types of fibroblasts no change is observed (see Allen, et al., 1999). It might be expected that cells placed in culture will be deprived of those signals that direct the normal sequence of developmental pathways, and that differentiation, if it occurs, is to an aberrant state. Alternatively, fetal cell lines may arise from different precursor cells than adult fibroblasts, and thus merely differentiate to a different fibroblast type. What is apparent from studies of senescence is that similarities as well as discrepancies are observed when it is compared to either aging or development. As a result of this, the model is obviously less useful for predicting aging or development changes in vivo than it is for studying known changes in an isolated cell system.

Vincent J. Cristofalo R. G. Allen

See also Cellular Aging; Cellular Aging: Cell Death; Cellular Aging: Telomeres; Physiological Changes, Fibroblast Cells; Physiological Changes: Stem Cells; Theories of Biological Aging.

BIBLIOGRAPHY

Allen, R. G.; Tresini, M.; Keogh, B. P.; Doggett, D. L.; and Cristofalo, V. J. "Differences in Electron Transport Potential Antioxidant Defenses and Oxidant Generation in Young and Senescent Fetal Lung Fibroblasts (WI-38)." Journal of Cellular Physiology 180 (1999): 114–122.

Bruce, S. A.; and Deamond, S. F. "Longitudinal Study of In Vivo Wound Repair and In Vitro Cellular Senescence of Dermal Fibroblasts." Experimental Gerontology 26 (1991): 17–27.

Carrel, A., and Ebling, A. H. "Age and Multiplication of Fibroblasts." Journal of Experimental Medicine 34 (1921): 599–523.

Cristofalo, V. J.; Volker, C.; Francis, M. K.; and Tresini, M. "Age-Dependent Modifications of Gene Expression in Human Fibroblasts." Critical Reviews in Eukaryotic Gene Expression 8 (1998): 43–80.

Cristofalo, V. J., and Pignolo, R. J. "Cell Culture As a Model." In Handbook of Physiology: Aging. Edited by E. J. Masoro. New York: Oxford University Press, 1984. Pages 53–82.

Cristofalo, V. J.; Pignolo, R. P.; Cianciarulo, F. L.; Dipaolo, B. R.; and Rotenberg, M. O. "Changes in Gene Expression During Senesence in Culture." Experimental Gerontology 27 (1992): 429–432.

Cristofalo, V. J.; Doggett, D. L.; Brooks-Frederich, K. M. and Phillips, P. D. "Growth Factors as Probes of Cell Aging." Experimental Gerontology 24 (1989): 367–374.

Cristofalo, V. J.; Allen, R. G.; Pignolo, R. P.; Martin, B. M.; and Beck, J. C. "Relationship between Donor Age and the Replicative Life Spans of Human Cells in Culture: A Reevaluation." Proceedings of the National Academy of Sciences of the United States of America 95 (1998): 10614–10619.

Cristofalo, V. J.; Volker, C.; and Allen, R. G. "Use of the Fibroblast Model in the Study of Cellular Senescence." In Aging Methods and Protocols. Edited by Y. Barnett and C. R. Barnett. Totowa, N.J.: Humana Press, 1984.

Dimri, G. P.; Lee, X.; Basile, G.; Acosta, M.; Scott, G.; Roskelley, C.; Medrano, E. E.; Linskens, M.; Rubelj, I.; Pereira-Smith, O.; Peacocke, M.; and Campisi, J. "A Biomarker that Identifies Senescent Human Cells in Culture and in Aging Skin In Vivo." Proceedings of the National Academy of Sciences of the United States of America 92 (1995): 9363–9367.

Feng, J.; Funk, W. D.; Wang, S.-S.; Weinrich, S. L.; Avilion, A. A.; Chiu, C.-C.; Adams, R. R.; Chang, E.; Allsopp, R. C.; Yu, J.; Le, S.; West, M. D.; Harley, C. B.; Andrews, W. H.; Greider, C. W.; and Villeponteau, B. "The RNA Component of Human Telomerase." Science 269 (1995): 1236–1241.

Gorman, S. D., and Cristofalo, V. J. "Reinitiation of Cellular DNA Synthesis in BrdUSelected Nondividing Senescent WI-38 Cells by Simian Virus 40 Infection." Journal of Cellular Physiology 125 (1985): 122–126.

Hayflick, L., and Moorhead, P. S. "The Serial Cultivation of Human Diploid Cell Strains." Experimental Cell Research 25 (1961): 585–621.

Lin, A. W.; Barradas, M.; STone, J. C.; van Aelst, L.; Serrano, M.; and Lowe, S. W. "Premature Senescence Involving p53 and p16 Is Activated in Response to Constitutive MEK/MAPK Mitogenic Signaling." Genes and Development 12 (1998): 3008–3019.

Lumpkin, C. K., Jr.; McClung, J. K.; Pereira-Smith, O. M.; and Smith, J. R. "Existence of High Abundance Antiproliferative mRNA's in Senescent Human Diploid Fibroblasts." Science 232 (1986): 393–395.

Medcalf, A. S. C.; Klein-Szanto, A. J. P.; and Cristofalo, V. J. "Expression of p21 Is Not Required for Senescence of Human Fibroblasts." Cancer Research 56 (1996): 4582–4585.

Ohtani, N.; Zebedee, Z.; Huot, T. J. G.; Stinson, J. A.; Sugimoto, M.; Ohashi, Y.; Sharrocks, A. D.; Peters, G.; and Hara, E. "Opposing Effects of Ets and Id Proteins on p16^INK4a Expression During Cellular Senescence." Nature 409 (2001): 1067–1070.

Oshima, J.; Campisi, J.; Tannock, C. A.; and Martin, G. M. "Regulation of c-fos Expression in Senescing Werner Syndrome Fibroblasts Differs from That Observed in Senescing Fibroblasts from Normal Donors." Journal of Cellular Physiology 162 (1995): 277–283.

Pignolo, R. J.; Rotenberg, M. O.; and Cristofalo, V. J. "Alterations in Contact and Density-Dependent Arrest State in Senescent WI-38 Cells." In Vitro Cellular and Developmental Biology 30A (1994): 471–476.

Robbins, E.; Levine, E. M.; and Eagle, H. "Morphologic Changes Accompanying Senescence of Cultured Human Diploid Cells." Journal of Experimental Medicine 131 (1970): 1211–1222.

Rubin, H. "Cell Aging In Vivo and In Vitro. " Mechanisms of Ageing and Development 98 (1997): 1-35.

Severino, J.; Allen, R. G.; Balin, S.; Balin, A.; and Cristofalo, V. J. "Is β-Galactosidase Staining a Marker of Senescence In Vitro and In Vivo?" Experimental Cell Research 257 (2000): 162–171.

Shelton, D. N.; Chang, E.; Whittier, P. S.; Choi, D.; and Funk, W. D. "Microarray Analysis of Replicative Senescence." Current Biology 9 (1999): 939–945.

Smith, J. R., and Whitney, R. G. "Intraclonal Variation in the Proliferative Potential of Human Diploid Fibroblasts: Stochastic Mechanisms for Cellular Aging." Science 207 (1980): 82–84.