Maier, A.B.
Citation
Maier, A. B. (2008, December 11). Cellular senescence in vitro and organismal ageing. Retrieved from https://hdl.handle.net/1887/13335
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded
from: https://hdl.handle.net/1887/13335
2
Persistence of high-replicative capacity in cultured fibroblasts from nonagenarians
Andrea B. Maier, Saskia le Cessie, Corine de Koning-Treurniet, Joke Blom, Rudi G.J. Westendorp, Diana van Heemst
Ageing Cell 2007; 6, 27-33
Summary
.
Earlier studies on human fibroblast cultures have demonstrated an inverse relationship between the total number of population doublings (PDs) and donor age. As more recent studies were unable to replicate these findings, we set out to analyze growth characteristics of fibroblast cultures from nonagenarians who represent the extreme of human lifespan. Therefore, we obtained skin biopsies from 68 participants of the Leiden 85-plus Study, all aged 90 years. None of the 68 strains failed to proliferate and all were easily cultured under highly standardized conditions. Within a time window of 30 months, all strains displayed a high and reproducible replicative capacity that was maintained for at least 50 PDs. A decline in mitotic activity was observed between 26 and 81 PDs. Out of the 68 cell strains, 57 strains reached the post-mitotic phase with an onset between 51 and 108 PDs.
The growth pattern of each senescent strain was fitted by a piecewise linear model, which allowed calculation of the transition towards the phase of decreased growth speed, as well as by a nonlinear continuous model; goodness-of-fit was high and not different between the models (both > 0.99). Growth characteristics were not associated with morbidity or mortality of the donors. We conclude that fibroblasts from nonagenarians maintain a high-replicative capacity despite a huge variability in the onset of senescence. These results cast further doubt on the association between in vitro growth characteristics of fibroblast cultures and the length of human lifespan.
Introduction
It has generally become accepted that human diploid fibroblasts divide a finite number of times, but the significance of this phenomenon has not been elucidated yet. The original description of the course of a fibroblast culture ex vivo distinguishes three phases (Swim and Parker, 1957; Hayflick and Moorhead, 1961). Phase I consists of taking a skin biopsy and transferring the cells from the in vivo environment to in vitro conditions to initiate the culture. Following the establishment of the culture, cells undergo a long period of steady proliferation (phase IIa). Hereafter, a period follows in which the proliferative capacity, measured by the time required to reach confluence, is decreased (phase IIb).
Cultures then degenerate and cell division ceases, resulting in growth arrest (phase III or senescence). Phase III has been shown to occur at about 50 cumulative population doublings (PDs) in human fibroblast cultures (Hayflick and Moorhead, 1961), regardless of the passage level at which strains are frozen (Hayflick, 1965).
More recently, the transition between the growth phases has been described to be less abrupt, resulting in a more continuous decline in proliferative growth (Serra and von Zglinicki, 2002).
Numerous studies using various designs have been carried out to address the question whether growth kinetics in vitro can be linked to the donor’s chronological age in humans. Inverse relationships between the age of the donor and proliferative capacity have been found by some authors (Hayflick, 1965;
Martinet al., 1970; Schneider and Mitsui, 1976; Smith et al., 1978; Allsopp et al., 1992; Serra and von Zglinicki, 2002), but others failed to find such a correlation (Cristofaloet al., 1998; Tesco et al., 1998; Smith et al., 2002). One study found a trend towards lower replicative potential when subjects were followed longitudinally and repeated biopsies were taken (Smith et al., 2002). Remarkably, in most of these studies a large interindividual variation in the proliferative capacity independent of the age of the donors was observed.
To further address the clinical significance of growth characteristics of fibroblast cell cultures, here we analyzed in vitro growth kinetics of fibroblasts from 68 nonagenarian donors who represent the extreme of human lifespan. To account for all possible sources of variability in growth kinetics, we have made great efforts to standardize procedures as much as possible.
Material and Methods
Subjects
Subjects were recruited from the Leiden 85-plus Study, a large population based study in which all inhabitants of Leiden, the Netherlands, aged 85 years were invited to take part. Between September 1997 and September 1999, 599 members of the 1912 to 1914 birth cohort were enrolled in the study (der Wiel et al., 2002).
Participants were followed up to five years and annually undertook comprehensive series of physical and psychological tests. The medical history was obtained from the general practitioner. All participants were followed for mortality up until February 2006.
During the period December 2003 up to May 2004, biopsies were taken from 68 surviving 90-year old participants, 42 women and 26 men. These participants were in good physical and mental condition and were able to come to the research institute, where the same qualified physician carried out the procedures. In May and June 2005 a repeated biopsy was taken from nine participants out of 68 participants; these participants were randomly chosen. All subjects gave informed consent. The medical ethical committee of Leiden University Medical Center approved the study.
Skin biopsy and cell culture
A three millimeter full thickness punch biopsy was taken from the sun unexposed medial site of the mid-upper arm in a standardized way. In case of a second biopsy, skin was biopsied on the other arm. Biopsies were collected in medium
supplemented with 20% fetal calf serum (FCS, lot no. 40G4932F) and stored at room temperature for maximally 24 h before establishment of the primary fibroblast culture. The biopsy was diced into six equally small pieces and split into two 25-cm2 culture flasks, each containing three fragments. The flasks with the biopsy pieces were put aside in an upright manner without cap to dry for 15 min.
Afterwards the adhered fragments were covered with a thin layer of medium (1.5 mL). When the first fibroblasts grew out of the fragments the amount of medium was increased to 5 mL. Twice a week, the pieces of tissue were gently refed with fresh medium until a couple of hundred fibroblasts had appeared around each biopsy fragment. The primary culture was defined as passage zero and PD zero (P0
= 0 PD), and the first subcultivation labelled as passage one. Cellular growth from passage zero to passage one was estimated as six PDs (P1 = 6 PDs). The second subcultivation was carried out when cells covered the entire bottom of the flask with a confluent monolayer of fibroblasts, corresponding to an estimated 11 PDs (P2 = 11 PDs).
Fibroblasts were incubated at 37°C in a dark 5% CO2 and 100% humidity and grown in two 25-cm2 flasks for each cell strain with D-MEM:F-12 (1:1) medium supplemented with FCS, 1 mM sodium pyruvate, 10 mM HEPES, 2 mM Glutamax I and antibiotics (100 Units per mL penicillin, 100 ȝg per mL streptomycin and 0.25–2.5 ȝg per mL amphotericin B). All reagents were obtained from Gibco, Breda, the Netherlands. Cells were continuously fed in a standardized manner by replacing one-half of the medium with fresh medium twice a week. Cells at passage zero were cultured with medium containing 20% FCS. Hereafter cells were fed with medium containing 10% FCS. After passage two, cells were transferred and expanded using two 75-cm2 flasks for each cell strain to freeze cells at different passages. Hereafter, cells were cultured in 25-cm2 flasks to the end of their replicative lifespan. Each time when the culture reached 95% to 100%
confluence, medium was removed and cell monolayers were rinsed with phosphate buffered saline. A trypsin solution was added to each culture flask (5 mg trypsin
per mL). Progression of digestion was checked by microscopy. Separated cells were diluted into 5 mL medium and plated for serial passaging in a 1:3.3 split ratio.
To avoid contamination, cells were seeded in new flasks every 2 weeks or, in case of slower doublings capacity, every subcultivation. If a strain was not subcultured for over a period of 35 days, the split ratio was changed to 1:1. Cells were defined post-mitotic when the cell density was stable or decreasing within a period of at least 45 days after the last subcultivation with two changes of medium per week and at least 75 days without subculturing. The onset of phase III was defined as the passage level and the day of the final subcultivation. Cell populations were maintained as stationary post-mitotic fibroblast populations with every week two medium changes and a microscopic control.
The cumulative PD level was calculated by tracking the increase in cell number in sequential passages and therefore the sum of all previous PDs.
Because of increasing susceptibility to fungal infection, we lost cultures at higher replicative stages. A secondary culture was restarted at least four passages lower than the passage number of the lost primary culture to avoid influence of the thawing procedure. Cultures were monitored every two months for Mycoplasma contamination and all were found to be negative.
Reproducibility
Twenty-four out of 68 cell strains were partially grown in parallel cultures. At various time points, secondary cultures were established by thawing fibroblasts of different passage numbers. Growth kinetics of the primary and secondary culture were compared at the time when both entered a specified passage number. The minimal period of parallel growth was nine passages with a mean of 16 passages.
The first three passages of the secondary culture were excluded from the analysis because of influence of the thawing procedure. Fibroblasts from the nine repeated biopsies were cultured under the same conditions as the cultures established out of the first biopsy. Growth kinetics from the initial and the repeated biopsy were compared over a time period of 284–353 days.
Statistical analyses
The statistical analysis was started at passage three, after the initiation of the culture. Growth kinetics from cell strains that entered the senescent phase (57 out of 68 cell strains) were modelled per person using a piecewise linear change-point model with a change-point during the growth phase II and a nonlinear continuous Chapman model. For each senescent cell strain the fit of each model was calculated by using a goodness-of-fit analysis.
According to the change-point model, the expected PD for a person at a certain day is PD = Į1 + ȕ1 x day, for day < t1 and Į2 + ȕ2 x day, for t1 < day, with restrictions on Į1 and Į2 such that the function is continuous in the point t1, i.e., (Į1
+ ȕ1 x t1) = (Į2 + ȕ2 x t1). The model was fitted using software for nonlinear regression models. The change-point (t1) corresponds to the onset of phase IIb, which is characterized by a decrease in mitotic activity. The slope of the line between passage three and the change-point (ȕ1) reflects the growth speed during phase IIa. The second slope of the model (ȕ2) reflects the growth speed during phase IIb, the line between the change-point and the onset of phase III.
According to the Chapman model: growth curves were fitted with the equation PDs = y0 + a(1 – e–bt)cas described by Serra and Zglinicki (2002). In this model the maximal PD value is y0 + a. If c > 1, the maximal growth speed occurs at day t + ln(c)/b and is equal to a x b x (1 – 1/c)(cí PD per day. If c < 1, the maximum growth speed occurs at t = 0 and is equal to infinity.
The growth speed of primary and secondary cultures, and cultures out of the first and second biopsies were compared using the co-efficient of variation, which was calculated as 100 x (standard deviation/mean).
Growth characteristics of the senescent cell strains were correlated with health characteristics of the donors by using the Pearson correlation. P values < 0.05 were defined as statistically significant. Analyses were performed with the computer program SPSS 12.00 (SPSS Inc., Chicago, IL).
Results
From December 2003 to May 2004, we established fibroblast cultures from skin biopsies taken from 68 elderly women and men, aged 90 years, who were in relatively good physical and mental condition and able to come to the research institute to donate a skin biopsy. We serially cultured these up to June 2006 (see Figure 1). Twelve to 16 months after taking the initial biopsy, a subsequent biopsy was taken from nine out of 68 participants to test the reproducibility of the growth patterns. Reproducibility was tested by comparing the growth speed of the initial and the repeated fibroblast cultures. The coefficient of variation was 5.1% (± 3.1).
The growth speed of primary fibroblast cultures and secondary cultures that were started by using frozen fibroblasts from primary cultures were also compared.
Twenty-four cell strains were grown as secondary cultures for at least nine passages. The coefficient of variation for the growth speed of primary and secondary cultures at the same PD level was 7.9% (± 7.9).
Figure 1. Flow chart of the cell cultures included in the study.
Initial biopsies Repeated biopsies
2003 December
2004 May
2005 January
May
2006 June
Fibroblast strains No. = 68
Senescent state No. = 57
Fibroblast strains No. = 9
Growth kinetics
None of 68 initial strains failed to proliferate and all were easily cultured. The growth curves of all cell strains were characterized by the typical growth phases:
initiation (phase I), growth (phase II), and senescence (phase III). Figure 2A shows the growth characteristics of the 57 senescent fibroblast strains. Figure 2B shows the 11 established strains, which did not reach the senescent state within the time window of 30 months.
The growth patterns of all 57 senescent cell strains were analyzed by fitting a piecewise linear change-point model to the observed growth data to obtain the change-point in growth speed that marks the onset of phase IIb and is characterized by a lower growth speed when compared to phase IIa. The data were also fitted by a nonlinear continuous Chapman model which allowed for estimating the maximal number of PDs. Two representative examples of the observed growth and the model fitting are presented in Figure 3. Below, we describe the growth characteristics of the strains during the three growth phases in more detail.
Phase I (passage 0–2): After initiation of the primary fibroblast cultures, the mean time to reach passage one, defined as the migration of fibroblasts out of the adhered fragments resulting in a fibroblast film of a couple of hundred cells around each fragment, was 17.9 days (± 2.8, range 11–25). Cells reached passage two, the first confluent cell layer in a 25-cm2 flask, after a mean time of 25.9 days (± 2.9, range 20–33).
Phase II: All 68 cell strains showed a similar growth pattern during the first period of phase II, indicated by nearly parallel growth lines as outlined in the insets of Figure 2. The onset of phase IIb, as obtained by the change-point model, was very heterogeneous, ranging from 87 to 280 days (± 42 days, average 173), respectively, and the corresponding PDs ranged from 26 to 81. The growth speed between passage three and the onset of phase IIb of each cell strain was very similar with an average 0.29 PD per day (± 0.04, Table 1). Thereafter, the average speed of growth during phase IIb was 0.09 PD per day (± 0.03, Table 1).
Figure 2. Observed growth kinetics of the 68 fibroblast strains up to 30 months of culturing. (A) 57 fibroblast strains became senescent. fibroblast strains did not reach the senescent state yet. The growth patterns of all established cell strains are initially similar (see inlets), where after the growth curves display a fan-shaped distribution because of increasing heterogeneity in cell growth.
Figure 3. Goodness-of-fit for the continuous nonlinear model and piecewise linear model in two representative cell cultures. The continuous lines display the predicted values calculated by the change-point model and the dotted lines by the nonlinear continuous model while the dots indicate the individually observed data points. (A) Fibroblast strain S597. (B) Fibroblast strain S613.
Phase III: The observed onset of the postmitotic phase of the 57 senescent strains differed widely with a range from 294 to 590 days, respectively, and corresponding cumulative PD values ranging from 51 to 108, the average being 73 (± 10.0). The maximal PDs were also estimated with the Chapman model and overestimated with an average of 85 PDs (± 21.7, Table 1).
The accuracy of the piecewise linear change-point model and the nonlinear continuous Chapman model to describe the growth kinetics of the senescent cell strains was estimated by using a goodness-of-fit analysis. Goodness-of-fit was high and not different between the models (both > 0.99, Table 1).
Table 1. Observed and predicted growth characteristics of 57 fibroblast cell cultures that reached senescence.
Growth characteristic Average (SD) Range
Change-point model
Growth speed phase IIa (PD per day) 0.29 (0.04) 0.20–0.36 Growth speed phase IIb (PD per day) 0.09 (0.03) 0.03–0.19
Onset phase IIb (PD) 54 (10.4) 26–81
Goodness-of-fit 0.995 (0.003) 0.985–0.999
Chapman model
Growth speed (PD per day) 0.32 (0.06) 0.06–0.42
Maximum PD (number predicted) 85 (21.7) 51–158
Goodness-of-fit 0.997 (0.003) 0.979–0.999
Replicative capacity
Maximum PD (number observed) 73 (10.0) 51–108
The growth characteristics of the fibroblast strains were analyzed mathematically by using a piecewise linear change-point model as well as a nonlinear continuous Chapman model.
Relationship between phase I and the onset of phase IIb and phase III
To analyze whether the observed interindividual differences in growth pattern were already present in phase I, we tested for correlation of duration of phase I with the onset of phase IIb or phase III. The days required for the initiation of the culture (phase I) were not correlated with the onset of phase IIb, either expressed in days or in PDs (phase I – onset phase IIb in days: Pearson correlation 0.23, p = 0.09;
phase I – PD value at onset phase IIb: Pearson correlation 0.24, p = 0.07). Phase I was also not correlated to the onset of phase III (phase I – onset phase III in days:
Pearson correlation 0.11, p = 0.43; phase II – PD value at onset phase III: Pearson correlation 0.16, p = 0.22).
Relationship between growth characteristics and health status of donors
The correlation of the growth characteristics of the 57 senescent cell strains with the health characteristics of the donors from which the biopsies were taken was also tested. Nine out of 57 donors suffered from diabetes mellitus at the age of 90 years. Malignancy (at present or in the past) was documented in nine out of 57
Observed and predicted growth characteristics of 57 fibroblast cell cultures according to health characteristics. Change-point modelChapman modelReplicative capacit Phase IIa (PD per day)Phase IIb (PD per day)Onset phase IIb (PD)Growth speed (PD per day)PD max (n)PD max (n) 0.31 (± 0.3) 0.09 (± 0.02)57 (± 6.4) 0.47 (± 0.12) 80 (± 7.4) 75 (± 6.6) 0.29 (± 0.04)0.09 (± 0.04)53 (± 10.4)0.5 (± 0.2)85 (± 23.5) 72 (± 10.4) value0.1070.8080.2640.6130.541 0.439 0.31 (± 0.04)0.09 (± 0.04)52 (± 11.4)0.53 (± 0.16)88 (± 27.3) 78 (± 7.5) 0.29 (± 0.04) 0.09 (± 0.03)54 (± 10.3) 0.49 (± 0.2)84 (± 20.7) 72 (± 10.2) value0.1460.9840.5150.580.552 0.119 0.31 (± 0.03) 0.08 (± 0.02)55 (± 10.1) 0.53 (± 0.15)73 (± 10.6) 70 (± 9.8) 0.29 (± 0.04) 0.09 (± 0.04) 54 (± 10.6) 0.49 (± 0.2) 87 (± 22.5) 73 (± 10.1) value0.1420.1810.7150.6280.11 0.427
donors. Within 30 months after the biopsies were taken, eight out of 57 donors died. As shown in Table 2, none of these characteristics were associated with any of the growth characteristics.
Discussion
To address the significance of growth kinetics in vitro, we started to culture fibroblasts obtained from nonagenarians that represent the extreme of human lifespan. If indeed a relation exists between fibroblast growth characteristics and organismal lifespan, we would have expected a high failure in the initiation of the cultures, a low remaining replicative capacity and a limited interindividual variation in the growth characteristics of these fibroblast strains. In contrast, all biopsies were easily cultured, and maintained a high-replicative capacity.
Interindividual differences in growth characteristics were large whereas intraindividual differences were limited.
Earlier studies found an extreme variability in the maximal cumulative PDs of parallel cultures arising from a single biopsy (Holliday et al., 1977; Thompson and Holliday, 1983), whereas Cristofalo et al. reported a high reproducibility of the replicative lifespan of fibroblasts in repetitive studies (1998). After standardization of the procedures used in taking and processing the biopsy, we found a remarkably high reproducibility of the growth speed tested over limited time periods, when primary cultures were compared to cultures established from frozen primary fibroblast cultures or from second biopsies.
After a 30-month period of serial culturing, we were able to formally quantify in 57 senescent strains the growth phases described by Hayflick in 1961. The first significant interstrain differences in growth kinetics were found during the transition from phase IIa to IIb, the onset of decreased growth speed. During phase IIb, growth was maintained at a lower and again, in most cases, constant uniform level, followed by a period of further decrease in mitotic activity ending with the
post-mitotic, metabolically active state. The cell strains displayed a huge variation in the onset of senescence.
In contrast to previously published studies with a relatively high percentage (above 15%) of growth failure of primary cultures established from human skin biopsies (Smith et al., 2002), no skin biopsies in our study failed to proliferate. The first confluent monolayers in our experiments (25-cm2) were reached after approximately 26 days with little interbiopsy variation, which is comparable to findings published by Tesco et al. (1998). Therefore, we can conclude that, even in the skin of nonagenarians, sufficient fibroblasts with high mitotic outgrowth capacity are present.
The first period of phase II was characterized by an exponential increase in cell number, which was followed by a decrease in growth speed at highly variable time points. A more abrupt transition towards the phase of decreased growth speed has been described in the commitment theory of cellular aging (Holliday et al., 1977).
The commitment theory predicts that an uncommitted cell population doubles in size until the first cells die, thereafter cells grow at a reduced but steady rate.
Cells out of a biopsy form a mixture of clones of different growth potential (Bayreuther et al., 1988). The decrease in growth speed during phase II may therefore be due to a decrease or exhaustion of mitotic activity of one of the dominant cell clones with high mitotic activity. On the other hand it has been suggested that a stochastic process with a random probability at the cell cycle is responsible for determining the limited proliferative capacity of strains (Smith and Whitney, 1980), which results in a continuous nonlinear growth curve without a change-point during phase II. To find arguments for one of these assumptions, the growth patterns of the senescent strains were fitted by a piecewise linear change- point model as well as a nonlinear continuous Chapman model to estimate the abruptness of the transition towards a decreased growth speed during phase II. As the goodness-of-fit was very high and similar for both models, neither of the models is favored.
The differences in growth speed in phase IIa and IIb were small; however, the onset of phase IIb and III differed substantially, which was not correlated with the duration of phase I. Therefore, our described characteristics of replication cannot be explained by differences during phase I.
We observed a high replicative capacity of fibroblasts from nonagenarians with huge variation in the onset of phase III. In earlier studies, fibroblasts reached the post-mitotic phase after about 50–55 cumulative PDs (Hayflick and Moorhead, 1961; Wagner et al., 2001). However, variation in the definition of phase III, differences in culture techniques, culture medium, FCS and skin explants make direct comparisons difficult. Despite these limitations, the data on the replicative capacity of fibroblasts of nonagenarians exceeded all our initial expectations.
Within our study population the growth characteristics of the fibroblast cultures were not associated with the presence of diabetes mellitus or malignancy in the donors. Growth characteristics were also not different between those who died and those who survived during the 30 months of culturing. These data stand in opposition to findings of earlier studies (Goldstein et al., 1978; Azzarone et al., 1981). Our study was not designed to test the influence of specific pathology on growth, the prevalence of disease was low and therefore the power to detect differences in growth characteristics between those with good and ill health was limited. The relatively good health status of subjects is reflected in the low incidence (7% per year) of death during follow-up.
Consciously, we chose a study population of the same chronological age to address the relation between fibroblast growth characteristics and organismal lifespan. Our data support the theory of individually defined attributes and impressively show that even in the very elderly a crucial number of cells with high mitotic capacity are left to give rise to fibroblast strains with the capacity for more than 100 PDs. Both observations contrast with earlier reported associations between in vitro growth characteristics of fibroblast cultures and the length of human lifespan. Our future research will focus on the underlying mechanisms of
cellular aging, particularly those that determine the onset of the different growth phases.
References
Allsopp, R.C., Vaziri, H., Patterson, C., Goldstein, S., Younglai, E.V., Futcher, A.B., Greider, C.W., Harley, C.B., 1992. Telomere length predicts replicative capacity of human fibroblasts. Proc. Natl. Acad. Sci. U. S. A. 89, 10114–10118.
Azzarone, B., Romanzi, C.A., Macieira-Coelho, A., 1981. Human skin fibroblasts from patients with mammary tumours: differences in growth properties. Cancer Detect.
Prev. 4, 249–259.
Bayreuther, K., Rodemann, H.P., Francz, P.I., Maier, K., 1988. Differentiation of fibroblast stem cells. J. Cell Sci. Suppl. 10, 115–130.
Cristofalo, V.J., Allen, R.G., Pignolo, R.J., Martin, B.G., Beck, J.C., 1998. Relationship between donor age and the replicative lifespan of human cells in culture: a reevaluation. Proc. Natl. Acad. Sci. U. S. A. 95, 10614–10619.
Goldstein, S., Moerman, E.J., Soeldner, J.S., Gleason, R.E., Barnett, D.M., 1978.
Chronologic and physiologic age affect replicative life-span of fibroblasts from diabetic, prediabetic, and normal donors. Science 199, 871–872.
Hayflick, L., 1965. The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res. 37, 614–636.
Hayflick, L., Moorhead, P.S., 1961. The serial cultivation of human diploid cell strains.
Exp. Cell Res. 25, 585–621.
Holliday, R., Huschtscha, L.I., Tarrant, G.M., Kirkwood, T.B., 1977. Testing the commitment theory of cellular aging. Science 198, 366–372.
Martin, G.M., Sprague, C.A., Epstein, C.J., 1970. Replicative life-span of cultivated human cells. Effects of donor’s age, tissue, and genotype. Lab. Invest. 23, 86–92.
Schneider, E.L., Mitsui, Y., 1976. The relationship between in vitro cellular aging and in vivo human age. Proc. Natl. Acad. Sci. U. S. A. 73, 3584–3588.
Serra, V., von Zglinicki, T., 2002. Human fibroblasts in vitro senesce with a donor-specific telomere length. FEBS Lett. 516, 71–74.
Smith, J.R., Pereria-Smith, O.M., Schneider, E.L., 1978. Colony size distribution as a measure of in vivo and in vitro aging. Proc. Natl. Acad. Sci. U. S. A. 75, 1353–1356.
Smith, J.R., Venable, S., Roberts, T.W., Metter, E.J., Monticone, R., Schneider, E.L., 2002.
Relationship between in vivo age and in vitro aging: assessment of 669 cell cultures derived from members of the Baltimore Longitudinal Study of Aging. J. Gerontol. A Biol. Sci. Med. Sci. 57, B239–B246.
Smith, J.R., Whitney, R.G., 1980. Intraclonal variation in proliferative potential of human diploid fibroblasts: stochastic mechanism for cellular aging. Science 207, 82–84.
Swim, H.E., Parker, R.F., 1957. Culture characteristics of human fibroblasts propagated serially. Am. J. Hyg. 66, 235–243.
Tesco, G., Vergelli, M., Grassilli, E., Salomoni, P., Bellesia, E., Sikora, E., Radziszewska, E., Barbieri, D., Latorraca, S., Fagiolo, U., Santacaterina, S., Amaducci, L., Tiozzo, R., Franceschi, C., Sorbi, S., 1998. Growth properties and growth factor responsiveness in skin fibroblasts from centinarians. Biochem. Biophys. Res. Commun. 244, 912–916.
Thompson, K.V., Holliday, R., 1983. Genetic effects on the longevity of cultured human fibroblasts. I. Werner’s Syndrome. Gerontology 29, 73–82.
Wagner, M., Hampel, B., Bernhard, D., Hala, M., Zwerschke, W., Jansen-Durr, P., 2001.
Replicative senescence of human endothelial cells in vitro involves G1 arrest, polyploidization and senescence-associated apoptosis. Exp. Gerontol. 36, 1327–1347.
der Wiel, A.B., van Exel, E., de Craen, A.J., Gussekloo, J., Lagaay, A.M., Knook, D.L., Westendorp, R.G., 2002. A high response is not essential to prevent selection bias:
Results from the Leiden 85-plus study. J. Clin. Epidemiol. 55, 1119–1125.
