EXPERIMENTAL CELL RESEARCH 236, 355–358 (1997)ARTICLE NO. EX973760

SHORT NOTE

Different Kinetics of Senescence in Human Fibroblastsand Peritoneal Mesothelial Cells

Emma Thomas, Eman Al-Baker,* Simona Dropcova, Stephen Denyer, Nasser Ostad,Andrew Lloyd, Ian R. Kill,* and Richard G. A. Faragher1

Department of Pharmacy, University of Brighton, Brighton BN2 4GJ, United Kingdom; and *Department of Biological Sciences,The University, Dundee DD1 4HN, United Kingdom

INTRODUCTIONSenescence has been reported for a wide variety of

The classic work of Hayflick and Moorehead [1] dem-human cell types. In cultures of human fibroblasts theonstrated that the growth of human fibroblasts is char-process is due to a percentage of the cells becomingacterized by a finite and predictable life span. Furthersenescent at each passage rather than all the cells en-work demonstrated that this limited life span was duetering senescence simultaneously at the end of the lifeto a progressively declining fraction of cells capablespan. By measuring the percentage of fibroblasts whichof undergoing division rather than all the cells thatare still cycling at each passage, a rate of decline in thecomposed the culture simultaneously entering senes-growth fraction, which mirrors the rate of senescence,cence at the end of the culture life span [2, 3]. Measure-can be obtained. However, such an analysis has never

been undertaken in multiple cell types using the same ment of the percentage of fibroblasts which are stillmethod to identify cycling cells. It is thus unknown if traversing the cell cycle at each passage allows the ratethe rate of senescence is the same or different in cul- at which this subpopulation declines (known as the celltures of different human cell types. To answer this ques- kinetics) to be determined [4–6]. Senescent fibroblasts,tion the rates of decline in the cycling fractions were although incapable of dividing, show continued viabil-simultaneously measured in two cultures of human ity for long periods after growth has ceased, indicatingcells (AGO7086A, peritoneal mesothelial cells; and 2DD, that the failure of the culture to grow is not due to cellhuman dermal fibroblasts) which have practically iden- death [7, 8].tical in vitro life spans. 2DD fibroblasts showed a rate Senescence has also been reported for a range ofof decline of 0.89% cycling cells per population doubling other cell types, including keratinocytes, adrenocorticalwhen the data obtained were fitted to a simple linear cells, and T-lymphocytes [9–11]. Unfortunately veryequation. However, AGO7086A gave a decline of ap- few of these studies have directly related any observa-proximately 2.2% per population doubling. Thus meso- tions to the fibroblast literature. None have measuredthelial cells enter senescence significantly faster than

the cell kinetics in a manner which allows direct com-fibroblasts (P õ 0.001). This decline in the growth frac-parison with other cell types. It is thus unknown if thetion was accompanied by an increasing fraction ofrate of senescence is the same or different in culturesmesothelial cells which retained detectable endoge-of different human cell types.nous b-galactosidase activity at pH 6. Such activity has

This study compares the cell kinetics of cultures ofpreviously been shown to be associated with senescenthuman fibroblasts (2DD) and peritoneal mesothelialhuman fibroblasts. These findings suggest that the pro-cells (AGO7086A) [12]. Both cultures undergo senes-cess of senescence has common features in differentcence at Ç50 population doublings (PD), which makescell lineages but that the rate of the process can differthem an ideal model system in which to study cell ki-markedly between them. q 1997 Academic Press

netics independent of differences in culture life span.The percentage of growing cells was determined usingan antiserum to pKi67, a heavily studied and routinelyused nuclear marker of proliferation in both normal1 To whom correspondence and reprint requests should be ad-

dressed. Fax: /44 (1273) 679333. and neoplastic materials [13]. In addition, senescence-

355 0014-4827/97 $25.00Copyright q 1997 by Academic Press

All rights of reproduction in any form reserved.

AID ECR 3760 / 6i27$$1041 09-19-97 03:09:05 ecl

356 SHORT NOTE

mide (Live/Dead kit, Molecular Probes) showed thatthe postmitotic endothelial cells remained alive. Suchcontinual viability after replicative failure is a criterionof senescence [16].

Figure 1 shows the fraction of Ki67-positive nuclei inhuman dermal fibroblasts and AGO7086A mesothelialcells during culture. In both cases the fraction of cyclingcells was seen to decline continuously and graduallyrather than to undergo an abrupt collapse. To obtaina numerical estimate of the rate of decline, simplestraight lines were fitted to the data by least-squaresregression analysis. The proliferating fraction of 2DDfibroblasts showed a decline of 00.88 { 0.13% of Ki67-positive cells per population doubling (PCPD). By con-trast the culture of mesothelial cells started with aconsiderably highly dividing fraction but entered se-nescence at a rate of 02.20 { 0.20 PCPD. StatisticalFIG. 1. The kinetics of decline in the growth fraction of human

diploid fibroblast strain DD1 (squares) compared with those of hu- analysis based upon a modified Student t test for differ-man peritoneal mesothelial cell strain AGO7086A (circles). ence of slope showed this difference in rate to be highly

significant (P õ 0.001).It has been reported [15] that senescent fibroblasts

can be demonstrated both in vitro and in vivo throughspecific b-galactosidase staining was applied to thethe use of a high pH modification of the standard tech-AGO7086A cells as a marker of the senescent state.nique designed to detect b-galactosidase activity. Fig-ure 2 shows the results of a similar analysis in whichMATERIALS AND METHODS

Cell culture. Human dermal fibroblasts (strain 2DD) were de-rived from an explant culture of juvenile foreskin. The cells werecultured as previously described [5].

Human peritoneal mesothelial cells (strain AG07086A) derivedfrom a 27-year-old female were purchased from the NIA (CoriellInstitute, Camden, NJ) at 19 PD. The cells were grown in a 1:1mixture of M199 and MCDB104 supplemented with 10% (v/v) FCS,EGF, and hydrocortisone at 6000 cells cm02 as previously described[12]. At each passage cells from both strains were plated onto 13-mm-diameter glass coverslips for the cytochemical studies detailedbelow.

Detection of proliferation-associated antigen Ki67. Cells were in-cubated with a 1:50 dilution of the antiserum raised against theproliferation-associated antigen Ki67 (Dako) as described in [8]. Toassess proliferation 1000 total or 200 positive nuclei were scored inrandomly selected fields.

Demonstration of normal and senescence-associated b-galactosi-dase activity. b-Galactosidase activity was demonstrated in cellsgrown on coverslips by a variation of the indigogenic method of Lojda[14]. To demonstrate total b-galactosidase activity the incubationbuffer was kept at pH 5, whereas to demonstrate senescence-associ-ated b-galactosidase (Sen-bgal) activity the pH was shifted to pH 6[15]. The proportion of Sen-bgal-positive cells was determined bycounting 200 positive or 1000 total cells.

RESULTS

AGO7086A underwent a total of 46 population dou-blings before entering senescence, which is comparableto previously reported data for this cell strain. Simi-larly 2DD human diploid fibroblasts attained 53 PD FIG. 2. Fraction of mesothelial cells displaying detectable b-ga-before the culture became senescent. Staining using a lactosidase activity at pH 6 correlated with the Ki67-positive percent-

age at that passage.commercial mixture of calcein AM and ethidium bro-

AID ECR 3760 / 6i27$$1042 09-19-97 03:09:05 ecl

357SHORT NOTE

the percentage of AGO7086A mesothelial cells with de- should be borne in mind by workers utilizing the Sen-b galactosidase technique until its mechanistic basistectable b-galactosidase activity at pH 6 is plotted

against the growth fraction of the culture as measured can be defined in greater detail.It may be that each tissue in the body shows distinc-by Ki67 staining. A nonlinear inverse relationship ap-

pears to exist between the percentage of cycling cells tive cell kinetics, as distinct from a different maximumlife span. Further comparative studies of the rates ofin a mesothelial cell culture and the fraction of cells

which are Sen-bgal positive. senescence using identical markers in cultures fromdifferent tissues appear to be warranted to determineif this hypothesis is correct. Interestingly, peritonealDISCUSSIONwounds show rates of closure which are affected bypatient age much more clearly than dermal woundsThese data show that mesothelial cells in culture

undergo a clear senescence in culture. This observation [19]. Different rates of accumulation of senescent cellsboth during the aging process and during wound heal-has been reported previously but no further character-

ization was undertaken [12]. In fibroblast cultures se- ing could explain this observation. This model wouldpredict the detection of elevated numbers of Sen-b ga-nescence is due to a gradual rate of decline in the per-

centage of cells which are able to proliferate (the lactosidase-positive cells in aged and wounded perito-neal mesothelium compared to dermis.growth fraction) rather than a sudden complete col-

lapse of proliferative ability at the end of the life span It has been suggested that telomere shortening is acounting mechanism controlling the senescence of cellsas was originally implied by Hayflick [17]. Unfortu-

nately, comparable data on the rates of senescence of in culture, and this shortening has been demonstratedin a number of different cell types, including T-cells,other cell types are not available. We have compared

the rates of decline in the growth fractions of two hu- fibroblasts, and mammary epithelia [20]. However, thekinetics of senescence using proliferation markers haveman cell types with identical life spans in culture and

found, unexpectedly, that the rate of mesothelial senes- not been assayed simultaneously with the telomereshortening rate. On the basis of the data presentedcence was approximately 2.5-fold faster than that ob-

served in fibroblasts. Although some variability in the here we hypothesize that, on a simple model, the over-all telomere shortening rate should be approximatelyrate of senescence is observable in fibroblast cultures

derived from different normal donors, this is not statis- 2.5-fold faster in mesothelial cells than in dermal fi-broblasts. Measurement of telomere shortening ratestically significant [5, 18]. Both 2DD and AGO7086A are

standard cell strains previously characterized by other in AGO4677A cells would allow a new test of the tel-omeric shortening theory.workers [12]. Kinetic findings obtained with these

strains may thus provisionally be considered typical.It was recently reported that a novel histochemical

REFERENCEStechnique based upon the detection of endogenous b-galactosidase activity could be used to specifically visu-

1. Hayflick, L., and Moorehead, P. S. (1961) Exp. Cell Res. 25,alize senescent cells [15]. We have applied this tech-585–621.nique to mesothelial cells and found an increase in

2. Cristafalo, V. J., and Scharf, B. (1973) Exp. Cell Res. 76, 419–b-galactosidase-positive cells with declining growth427.

fraction analogous to that seen in fibroblasts. However,3. Merz, G. S., and Ross, J. D. (1969) J. Cell Physiol. 82, 219–221.the percentage of cycling cells labeled using anti-Ki674. Ponten, J., Stein, W. D., and Shall, S. (1983) J. Cell Physiol.antiserum and the percentage of cells stained using the

117, 342–352.b-galactosidase staining technique do not total 100% of

5. Faragher, R. G. A., Kill, I. R., Hunter, J. A. A., Pope, F. M., Tan-the observed cells at each passage. The b-galactosidase nock, C., and Shall, S. (1993) Proc. Natl. Acad. Sci. USA 90,staining technique is known to be dependent on the 12030–12034.absolute amount of enzyme present, which is usually 6. Norsgaard, H., Clark, B. F. C., and Rattan, S. I. S. (1996) Exp.a function of cell size [15]. Large cell size correlates Gerontol. 31, 563–570.with senescence in both fibroblasts and mesothelial 7. Macieira-Coelho, A., Ponten, J., and Philipson, L. (1966) Exp.cells, so the correlation of b-galactosidase staining at Cell Res. 42, 673.high pH and senescence may be indirect and dependent 8. Kill, I. R., and Shall, S. (1990) J. Cell Sci. 97, 473.on the propensity of senescent cells to be larger than 9. Willie, J. J., Pittelkow, M. R., Shipley, G. D., and Scott, R. E.their growing counterparts in the system under study. (1984) J. Cell Physiol. 121, 31–44.An alternative explanation could be that there is sim- 10. Hornsby, P. J., Ryan, R. F., and Cheng, C. Y. (1989) Exp. Geron-ply a delay between the timing of loss of Ki67 staining tol. 24, 539–558.in each senescent cell and the accumulation of Sen-b 11. Perillo, N. L., Walford, R. L., Newman, M. A., and Effros, R. B.

(1989) Exp. Gerontol. 24, 177–187.galactosidase activity to a detectable level. This caveat

AID ECR 3760 / 6i27$$1042 09-19-97 03:09:05 ecl

358 SHORT NOTE

12. Rheinwald, J. G. (1989) Cell Growth and Division: A Practical 16. Shall, S. (1987) in Perspectives on Mammalian Cell Death(C. S. Potten, Ed.), pp. 184–201, Oxford Univ. Press, London.Approach (R. Baserga, Ed.), pp. 81–95, IRL Press, Oxford.

13. Heidebrecht, H. J., Buck, F., Haas, K., Wacker, H. H., and 17. Hayflick, L. (1965) Exp. Cell Res. 37, 614–636.Parwaresch, R. (1996) Cell Proliferation 29, 413–425. 18. Kill, I. R., Faragher, R. G. A., Lawrence, K., and Shall, S. (1994)

J. Cell Sci. 107, 571–579.14. Dropcova, S., Cejkova, J., and Lojda, Z. (1992) Acta Histochem.Cytochem. 25(1–2), 299. 19. Ashcroft, G. S., Horan, M. A., and Ferguson, M. J. W. (1995) J.

Anat. 187, 1–26.15. Dimri, G. P., Lee, X. H., Basile, G., Acosta, M., Scott, C., Roskel-ley, C., Medrano, E. E., Rubel, J. I., Linskins, M., Pereira- 20. Vaziri, H., Dragowska, W., Allsopp, R. C., Thomas, T. E., Har-

ley, C. B., and Lansdorp, P. M. (1994) Proc. Natl. Acad. Sci.Smith, O., Peacocke, M., and Campisi, J. (1995) Proc. Natl.Acad. Sci. USA 92, 9363–9369. USA 91, 9857–9860.

Received June 11, 1997

AID ECR 3760 / 6i27$$1042 09-19-97 03:09:05 ecl