Experimental
MEASUREMENT SENESCENT
Cell Research 98 (1976) 298-302
OF DNA CONTENT HUMAN
MULTIPARAMETER
AND CELL
FIBROBLASTS SINGLE
VOLUME
UTILIZING
CELL
IN
FLOW
ANALYSIS
E. L. SCHNEIDER and B. .I. FOWLKES Laboratory of Cellular and Comparati\~e Physiology, Gerontology Research Center, National Institute on Aging, National Institutes of Health, PHS, US Department of Health, Education and Welfare, Bethesda, and Baltimore City Hospitals, Baltimore, MD 21224, and Department of Pathology, National Cancer Institute, National Institutes of Health, PHS, US Department of Health, Education and Welfare, Bethesda, MD 20014, USA
SUMMARY Cell cycle analysis of senescent cultured human fibroblasts by flow cytofluorometry reveals an increased proportion of cells in the G I period. An increased variation as well as a slight decrease in cellular DNA contents were observed in both G I and G 2+M senescent cells. Utilizing gated single parameter analysis, the increased cell volumes observed in these senescent fibroblasts were demonstrated to be present in G I as well as G2+M cells.
After a period of sustained rapid growth, cultured human fetal lung fibroblasts (WI38) enter a phase of slowed growth or “senescence” which culminates in the death of the cell culture [ll]. This characteristic of WI38 cell cultures has led to its utilization as an in vitro model system for cellular aging. The earliest changes observed in WI38 cell cultures during their in vitro lifespan are increased cell sizes [5, 13, 191 and an increased percent of non-replicating or slowly replicating cells [6, 121.Because of the difficulties in synchronizing diploid human fibroblasts, most of these studies were performed on cell populations in all phases of the cell cycle. The introduction of flow microfluorometry [2 l] permits the identification of cell cycle phase for individual cells by measurement of cellular DNA conExprlCrURes
98 (1976)
tents. Cells in G 1 with a diploid (2C) DNA content can be distinguished from cells in G 2 or M which have 4C DNA contents, and cells in S with intermediate DNA contents (2C-4C). Therefore, by employing this instrument, one can directly examine if the observed increase in non-dividing or slowly dividing cells is the result of an accumulation of cells in specific phases of the cell cycle. The recent development of multiparameter photometry permits the simultaneous measurement of volume and DNA content on individual cells [2, 3, 201. Utilizing gated single parameter analysis, electronic gates can be set for one parameter (2C or 4C DNA contents) and the second parameter, cell volume can be measured in populations of cells in specific cell cycle phases (G 1 or
Floes multiparameter
Fies I, 2. Ah&sat rel. DNA content: ordinrrfr: cell no: x 103. Fig. I. Distribution of DNA contents in senescent (- - -) and pre-senescent (-) WI38 cells in logarithmic growth. Pre-senescent WI38 cells at 27 cell population doublings (CPD) and senescent cells at 46 CPD were harvested 3 days after cell seeding and I day after media change to obtain maximal growth rates. The harvested cglls were fixed, stained-with mithramycin (Pfizer, Inc.) and DNA content auantified with a multiparameter cell sorter (Los Alamos Scientific Labs).
G 2+ M). This technique therefore enables us to discern whether the observed increased volumes of WI38 cells with increased in vitro age is due to specific enlargement of G I or G 2 cells or a generalized volume increase in all cell cycle phases.
analysis
of senescent
human Jibroblasts
299
becco’s calcium and magnesium-free phosphate buffered saline (CMFPBS). followed bv the addition of 0. I % pronase for I min. The pronase was neutralized immediately with media containing serum and the detached cells were collected. centrifuged at 200 g for IO mitt, washed with CMFPBS, recentrifuged at 200 g for IO min and resuspended in 70% ethanol fixative. The dispersed, fixed cells were stained for DNA content with IO0 pgiml mithramycin (Pfizer, Inc.) in 0.85% NaCl containing 15 mM MgCI, [4]. Total fluorescence and narrow angle light scatter [I61 were measured simultaneously on-eachcell using themultiparameter cell sorter (Los Alamos Scientific Lab.). A comprehensive description of this instrument has been presented elsewhere [20]. A Northern Scientific (Model 636) multichannel pulse height analyzer was used to display signals as pulse amplitude frequency distribution histograms. The pulse height analyzer was used in combination with a hard-wired signal processor to enable light scatter distributions to be displayed for only those cells having preselected ranges of fluorescent intensity. The instrument was calibrated with IO pm microspheres having a uniform fluorescence (Particle Technology, Inc.). During this study, the co-efficient of variation for the single amplitude distribution produced by these particles was maintained at approx. 4. I % for fluorescence and 7.14.4 % for narrow angle light scatter.
RESULTS The pulse amplitude distribution of DNA content in populations of pre-senescent and senescent WI38 cells is seen in figs 1 and 2.
4t
31
MATERIAL
AND METHODS
WI38 cells were provided by Dr L. Hayflick at early, I7 cell population doublines (CPD). and late. 37 CPD. passage. Cells were further subcultured (I :4 splits) in Eagle’s Minimal Essential Media supplemented with glutamine, 10% fetal bovine serum, and 50 &ml Aureomvcin fLederIe). Measurements of cell volume and DNA content determinations were performed on senescent (43 to 46 CPD) and presenescent (23 to 27 CPD) cultures on three separate occasions. To obtain cells in logarithmic growth phase, fresh medium was added to the culture on the second day after transfer (I :4 split) and the cells were harvested on the subsequent day. Growth arrested cells were obtained by harvesting the culture without media change two days after the monolayer became confluent. The cells were detached from their monolayer by washing with Dul-
Fig. 2. Distribution of DNA contents in senescent (- - -) and pre-senescent (--) WI38 cells at confluencv. Pre-senescent WI38 cells at 23 CPD and senescent ceils at 43 CPD were harvested 7 and I2 days after seeding without media change. Conditions similar to fig. I except that gain setting for DNA measurement changed. E.rpt/
Cell
Res 98 (1976)
300
Schneider and Fowlkes
Fig. 3. Abscissa: rel. DNA content. Illustration of the setting of electronic ure light scatter in G I or G 2+M cells.
gates to meas-
The large initial peak at 2 N DNA content represents cells in G 1. At twice the modal channel number of this G 1 peak, a smaller G2+M peak is seen. The cells between these peaks are in S. In both logarithmic growing (fig. 1) and confluent cultures (fig. 2), a greater proportion of senescent WI38 cells (dashed lines) are found in the G I phase of the cell cycle when compared with pre-senescent cells (solid lines). There is also a smaller proportion of ceils in S in logarithmically growing senescent cells. These senescent cells also have a wider distribution of DNA content values in both G I and G 2+M cells as indicated by an increased co-efficient of variation. This is most obvious in fig. 2 where the co-efficient of variation for confluent WI38 cells in G I increases from 3 5%in pre-senescent cells to 4.5 5%in senescent cell cultures. In addition, the modal DNA content of late passage or senescent cells in G 1 appears to be slightly less than in pre-senescent cells in both logarithmic growth (fig. I) and confluency (fig. 2). A similar shift to lower DNA contents is also observed in senescent G 2+ M cells. h/H/
Cdl RPS 98 (1976)
Despite the fact that the difference in modal DNA contents is only 1 or 2 channels, it was consistently seen on three separate comparisons of senescent and pre-senescent cultures. Fig. 3 demonstrates the setting of electronic gates for G 1 and G 2+M cells for the measurement of cell volume. Cell volume measurements in ungated logarithmically growing and confluent WI38 cells reveal a shift in distribution towards increased volume in both logarithmic growing and confluent senescent cells (figs 4, 5). The distribution of cell volumes in G I and G2+M cells (figs 4, 5) shows a clear increase in cell volume distribution in the senescent cell population in these phases of the cell cycle in both logarithmic growing UNGATED
G2 + M
Fig. 4. Abscissu: channel no.; ordinnte: cell no. x 103. Gated single parameter analysis of cell volume in senescent (- - -) and pre-senescent (-) WI38 cells in logarithmic growth. Electronic gates were set as illustrated in fig. 3 and cell volumes were measured by small angle light scatter in G I and G 2+M cells. Ungated cell volumes are presented for comparison to gated results.
Flow multiparameter UNGATED
Fip. 5. Abscissa: channel no.: ordinate: cell no. x IO*. bated single parameter analysis of cell volume in senescent (- - -) and pre-senescent (-_) WI38 cells as in fig. 4 except that cell cultures are confluent.
and confluent cultures. In addition to being twice the size of G 1 cells, G2+M cells in both senescent and pre-senescent cultures exhibited markedly increased cell volume variation. Cells in both confluent senescent and pre-senescent cultures (fig. 5) also had larger volumes as well as increased variation in cell volumes when compared with logarithmically growing cultures (fig. 4).
DISCUSSION Flow cytofluorometric measurement of DNA content has been shown to correlate well with classical techniques for cell cycle analysis [8]. Our results obtained with this new technique indicate that in senescent WI38 cells there is either an increased proportion of cells arrested in G 1, an increased length of the G 1 cell cycle phase or a combination of these changes. Yanishevsky et al. [23], employing a combination of DNA-cytophotometry and 20-761801
analysis of senescent
humanfibroblasts
301
autoradiography, have also found an increased proportion of G 1 cells in senescent WI38 cultures. They interpret their findings as indicating an accumulation of nondividing cells in G 1. Non-dividing cells are defined by these investigators as cells that have not incorporated [3H]thymidine into DNA during a 48-h labeling period. Other authors have also suggested that noncycling cells accumulate in senescent human fibroblast cultures [9, 10, 171. However, the recent observations of MacieiraCoelho [ 141as well as our own experimental data [ 151 indicates that with increased [3H]thymidine labeling times, over 90 % of senescent cell nuclei become labeled. Furthermore, by direct cinematographic measurement, Absher et al. [I] have demonstrated that although there is an increased proportion of non-dividing cells present in senescent cell populations, the dividing cells have increased division time. It is our interpretation that in senescent cell cultures there is an accumulation of both nondividing cells as well as cells with increased cell cycle times. Therefore the greater proportion of G 1 cells observed by cytofluorometry in senescent cultures is probably due to both an accumulation of non-dividing cells in G 1 as well as to a prolongation of the G 1 cell cycle phase in slowly dividing cells. The most likely explanation for the observed increased variation of DNA contents in senescent cells is the reported increase in aneuploidy [18]. This could also explain the shift to slightly lower DNA contents, since the aneuploidy observed in senescent cells tends to be hypodiploid. Other investigators have reported no increase in aneuploidy in WI38 cells as a function of in vitro aging [22]. These conflicting findings may be the result of the difficulty in obtaining adequate metaphase preparations in senescent culE.rpt/CellRrs
98 (1976)
tures. Since aneuploid cells are also known to have impaired replication [7], flow microfluorometry could be detecting aneuploidy that might be missed by conventional metaphase analyses. Alternate explanations for the increased variation and slight decrease in cellular DNA contents in senescent WI38 cultures include an increased frequency of chromosomal deletions, a decreased amplification of specific DNA genes, or a loss of cellular DNA. Altered binding of mithramycin to DNA is another possibility. However, Yanishevsky et al. [23] have also reported an increased variation in DNA contents in senescent WI38 cells utilizing Feulgen staining and cytophotometry. The increase in cell volumes demonstrated by small angle light scatter in ungated senescent cells can also be observed by measurement of cell diameter [5, 131and by Coulter volume [15]. Although ethanol fixation reduces cell volume, this reduction is relative and does not change the cell volume distribution [3]. Gated single parameter analysis clearly reveals that the increase in cell volume in senescent cells is not limited to any one phase of the cell cycle. It also confirms the increase in cell volume as cells progress through the cell cycle that had been suggested by dry mass measurements [24]. In conclusion, senescent cells appear to have increased cell sizes in both G I and G2+M cells in logarithmic growth as well as at confluency. Since the volume distribution of a cell culture appears to be closely related to its replication rate [1.5], the shift to larger cell volumes observed during the in vitro lifespan of WI38 cells may well be secondary to the observed increasing number of slow or non-dividing cells [l]. Similarly, increased variations in cell sizes may reflect increased heterogeneity in cell division times observed in senescent cells [l]. Exptl Cd RPS 98 (1976)
However. the remote possibility remains that a cell membrane alteration could be the primary event leading to both a loss of replicative ability and an increase in cell volume. WI38 cell cultures were obtained through contract no. I-HD-4-2828 from the National Institute of Child Health and Human Development to Dr Leonard Hayflick of Stanford University. We thank MS K. Braunschweiger for technical assistance.
REFERENCES 1. Absher. P M, Absher, R G & Barnes. W D, EXR cell res 88 (1974) 95. 2. Bonner, W A, Hulett, H R, Sweet, R G & Hertzenberg, L, Rev sci instr 43 (1972) 404. 3. Crissman, H A & Steinkamp, J A, J cell biol 59 (1973) 766. 4. Crissman. H A & Tobev. R A, Science 184(1974) 1297. 5. Cristofalo, V & Kritchevsky, D, Med exp I9 (1969) 313. 6. Cristofalo, V & Sharf, B, Exp cell res 76 (1973) 419. 7. Cure, S, Boue, A & Boue, J, Les accidents chromosomiques de la reproduction (ed A Boue & C Thibault) p. 95. Inserm, Paris (1973). 8. Dean, P N & Jett, J H, J cell biol 60 (1974) 523. 9. Gelfant, S & Smith, J G, Jr, Science I78 (1972) 357. IO. Good, P I & Smith, J R, Biophys j 14 (1974) 81 I. Il. Hayflick, L, Exp cell res 37 (1965) 614. 12. Macieira-Coelho, A, Ponttn, J & Philipson, L, Exp cell res 42 (1966) 673. 13. Macieira-Coelho, A & Ponttn, J, J cell biol 43 (1969) 374. 14. Macieira-Coelho. A. Nature 248 (1974) 42 I. 15. Mitsui, Y & Schneider, E L. Personal communication. 16. Mullaney, P F & Dean, P N, Biophys j IO (1970) 764. 17. Orgel, L E, Nature 243 (1973) 44 I. 18. Saskela, E & Moorhead, P S, Proc natl acad sci US 50 (1963) 390. 19. Simons, J W I M, Exp cell res 45 (1967) 336. 20. Steinkamp, J A, Fulwyler, M J, Coulter, J R, Hiebert, R H, Horney, J L & Mullaney, P F, Rev sci instr 44 (1973) 1301. 21. Van Dilla, M A, Trujillo, T T, Mullaney, P F & Coulter, J R, Science 163(1969) 1213. 22. Whitaker, A M, Gould, J & Sith, E M, Exp cell res 87 (1974) 55. 23. Yanishevsky, R, Mendelsohn, M L, Mayall, B H & Cristofalo, V J, J cell physiol84 (1974) 165. 24. Zetterberg, A & Killander, D, Exp cell res 39 (1965) 22. Received July 14, 1975
MEASUREMENT SENESCENT
Cell Research 98 (1976) 298-302
OF DNA CONTENT HUMAN
MULTIPARAMETER
AND CELL
FIBROBLASTS SINGLE
VOLUME
UTILIZING
CELL
IN
FLOW
ANALYSIS
E. L. SCHNEIDER and B. .I. FOWLKES Laboratory of Cellular and Comparati\~e Physiology, Gerontology Research Center, National Institute on Aging, National Institutes of Health, PHS, US Department of Health, Education and Welfare, Bethesda, and Baltimore City Hospitals, Baltimore, MD 21224, and Department of Pathology, National Cancer Institute, National Institutes of Health, PHS, US Department of Health, Education and Welfare, Bethesda, MD 20014, USA
SUMMARY Cell cycle analysis of senescent cultured human fibroblasts by flow cytofluorometry reveals an increased proportion of cells in the G I period. An increased variation as well as a slight decrease in cellular DNA contents were observed in both G I and G 2+M senescent cells. Utilizing gated single parameter analysis, the increased cell volumes observed in these senescent fibroblasts were demonstrated to be present in G I as well as G2+M cells.
After a period of sustained rapid growth, cultured human fetal lung fibroblasts (WI38) enter a phase of slowed growth or “senescence” which culminates in the death of the cell culture [ll]. This characteristic of WI38 cell cultures has led to its utilization as an in vitro model system for cellular aging. The earliest changes observed in WI38 cell cultures during their in vitro lifespan are increased cell sizes [5, 13, 191 and an increased percent of non-replicating or slowly replicating cells [6, 121.Because of the difficulties in synchronizing diploid human fibroblasts, most of these studies were performed on cell populations in all phases of the cell cycle. The introduction of flow microfluorometry [2 l] permits the identification of cell cycle phase for individual cells by measurement of cellular DNA conExprlCrURes
98 (1976)
tents. Cells in G 1 with a diploid (2C) DNA content can be distinguished from cells in G 2 or M which have 4C DNA contents, and cells in S with intermediate DNA contents (2C-4C). Therefore, by employing this instrument, one can directly examine if the observed increase in non-dividing or slowly dividing cells is the result of an accumulation of cells in specific phases of the cell cycle. The recent development of multiparameter photometry permits the simultaneous measurement of volume and DNA content on individual cells [2, 3, 201. Utilizing gated single parameter analysis, electronic gates can be set for one parameter (2C or 4C DNA contents) and the second parameter, cell volume can be measured in populations of cells in specific cell cycle phases (G 1 or
Floes multiparameter
Fies I, 2. Ah&sat rel. DNA content: ordinrrfr: cell no: x 103. Fig. I. Distribution of DNA contents in senescent (- - -) and pre-senescent (-) WI38 cells in logarithmic growth. Pre-senescent WI38 cells at 27 cell population doublings (CPD) and senescent cells at 46 CPD were harvested 3 days after cell seeding and I day after media change to obtain maximal growth rates. The harvested cglls were fixed, stained-with mithramycin (Pfizer, Inc.) and DNA content auantified with a multiparameter cell sorter (Los Alamos Scientific Labs).
G 2+ M). This technique therefore enables us to discern whether the observed increased volumes of WI38 cells with increased in vitro age is due to specific enlargement of G I or G 2 cells or a generalized volume increase in all cell cycle phases.
analysis
of senescent
human Jibroblasts
299
becco’s calcium and magnesium-free phosphate buffered saline (CMFPBS). followed bv the addition of 0. I % pronase for I min. The pronase was neutralized immediately with media containing serum and the detached cells were collected. centrifuged at 200 g for IO mitt, washed with CMFPBS, recentrifuged at 200 g for IO min and resuspended in 70% ethanol fixative. The dispersed, fixed cells were stained for DNA content with IO0 pgiml mithramycin (Pfizer, Inc.) in 0.85% NaCl containing 15 mM MgCI, [4]. Total fluorescence and narrow angle light scatter [I61 were measured simultaneously on-eachcell using themultiparameter cell sorter (Los Alamos Scientific Lab.). A comprehensive description of this instrument has been presented elsewhere [20]. A Northern Scientific (Model 636) multichannel pulse height analyzer was used to display signals as pulse amplitude frequency distribution histograms. The pulse height analyzer was used in combination with a hard-wired signal processor to enable light scatter distributions to be displayed for only those cells having preselected ranges of fluorescent intensity. The instrument was calibrated with IO pm microspheres having a uniform fluorescence (Particle Technology, Inc.). During this study, the co-efficient of variation for the single amplitude distribution produced by these particles was maintained at approx. 4. I % for fluorescence and 7.14.4 % for narrow angle light scatter.
RESULTS The pulse amplitude distribution of DNA content in populations of pre-senescent and senescent WI38 cells is seen in figs 1 and 2.
4t
31
MATERIAL
AND METHODS
WI38 cells were provided by Dr L. Hayflick at early, I7 cell population doublines (CPD). and late. 37 CPD. passage. Cells were further subcultured (I :4 splits) in Eagle’s Minimal Essential Media supplemented with glutamine, 10% fetal bovine serum, and 50 &ml Aureomvcin fLederIe). Measurements of cell volume and DNA content determinations were performed on senescent (43 to 46 CPD) and presenescent (23 to 27 CPD) cultures on three separate occasions. To obtain cells in logarithmic growth phase, fresh medium was added to the culture on the second day after transfer (I :4 split) and the cells were harvested on the subsequent day. Growth arrested cells were obtained by harvesting the culture without media change two days after the monolayer became confluent. The cells were detached from their monolayer by washing with Dul-
Fig. 2. Distribution of DNA contents in senescent (- - -) and pre-senescent (--) WI38 cells at confluencv. Pre-senescent WI38 cells at 23 CPD and senescent ceils at 43 CPD were harvested 7 and I2 days after seeding without media change. Conditions similar to fig. I except that gain setting for DNA measurement changed. E.rpt/
Cell
Res 98 (1976)
300
Schneider and Fowlkes
Fig. 3. Abscissa: rel. DNA content. Illustration of the setting of electronic ure light scatter in G I or G 2+M cells.
gates to meas-
The large initial peak at 2 N DNA content represents cells in G 1. At twice the modal channel number of this G 1 peak, a smaller G2+M peak is seen. The cells between these peaks are in S. In both logarithmic growing (fig. 1) and confluent cultures (fig. 2), a greater proportion of senescent WI38 cells (dashed lines) are found in the G I phase of the cell cycle when compared with pre-senescent cells (solid lines). There is also a smaller proportion of ceils in S in logarithmically growing senescent cells. These senescent cells also have a wider distribution of DNA content values in both G I and G 2+M cells as indicated by an increased co-efficient of variation. This is most obvious in fig. 2 where the co-efficient of variation for confluent WI38 cells in G I increases from 3 5%in pre-senescent cells to 4.5 5%in senescent cell cultures. In addition, the modal DNA content of late passage or senescent cells in G 1 appears to be slightly less than in pre-senescent cells in both logarithmic growth (fig. I) and confluency (fig. 2). A similar shift to lower DNA contents is also observed in senescent G 2+ M cells. h/H/
Cdl RPS 98 (1976)
Despite the fact that the difference in modal DNA contents is only 1 or 2 channels, it was consistently seen on three separate comparisons of senescent and pre-senescent cultures. Fig. 3 demonstrates the setting of electronic gates for G 1 and G 2+M cells for the measurement of cell volume. Cell volume measurements in ungated logarithmically growing and confluent WI38 cells reveal a shift in distribution towards increased volume in both logarithmic growing and confluent senescent cells (figs 4, 5). The distribution of cell volumes in G I and G2+M cells (figs 4, 5) shows a clear increase in cell volume distribution in the senescent cell population in these phases of the cell cycle in both logarithmic growing UNGATED
G2 + M
Fig. 4. Abscissu: channel no.; ordinnte: cell no. x 103. Gated single parameter analysis of cell volume in senescent (- - -) and pre-senescent (-) WI38 cells in logarithmic growth. Electronic gates were set as illustrated in fig. 3 and cell volumes were measured by small angle light scatter in G I and G 2+M cells. Ungated cell volumes are presented for comparison to gated results.
Flow multiparameter UNGATED
Fip. 5. Abscissa: channel no.: ordinate: cell no. x IO*. bated single parameter analysis of cell volume in senescent (- - -) and pre-senescent (-_) WI38 cells as in fig. 4 except that cell cultures are confluent.
and confluent cultures. In addition to being twice the size of G 1 cells, G2+M cells in both senescent and pre-senescent cultures exhibited markedly increased cell volume variation. Cells in both confluent senescent and pre-senescent cultures (fig. 5) also had larger volumes as well as increased variation in cell volumes when compared with logarithmically growing cultures (fig. 4).
DISCUSSION Flow cytofluorometric measurement of DNA content has been shown to correlate well with classical techniques for cell cycle analysis [8]. Our results obtained with this new technique indicate that in senescent WI38 cells there is either an increased proportion of cells arrested in G 1, an increased length of the G 1 cell cycle phase or a combination of these changes. Yanishevsky et al. [23], employing a combination of DNA-cytophotometry and 20-761801
analysis of senescent
humanfibroblasts
301
autoradiography, have also found an increased proportion of G 1 cells in senescent WI38 cultures. They interpret their findings as indicating an accumulation of nondividing cells in G 1. Non-dividing cells are defined by these investigators as cells that have not incorporated [3H]thymidine into DNA during a 48-h labeling period. Other authors have also suggested that noncycling cells accumulate in senescent human fibroblast cultures [9, 10, 171. However, the recent observations of MacieiraCoelho [ 141as well as our own experimental data [ 151 indicates that with increased [3H]thymidine labeling times, over 90 % of senescent cell nuclei become labeled. Furthermore, by direct cinematographic measurement, Absher et al. [I] have demonstrated that although there is an increased proportion of non-dividing cells present in senescent cell populations, the dividing cells have increased division time. It is our interpretation that in senescent cell cultures there is an accumulation of both nondividing cells as well as cells with increased cell cycle times. Therefore the greater proportion of G 1 cells observed by cytofluorometry in senescent cultures is probably due to both an accumulation of non-dividing cells in G 1 as well as to a prolongation of the G 1 cell cycle phase in slowly dividing cells. The most likely explanation for the observed increased variation of DNA contents in senescent cells is the reported increase in aneuploidy [18]. This could also explain the shift to slightly lower DNA contents, since the aneuploidy observed in senescent cells tends to be hypodiploid. Other investigators have reported no increase in aneuploidy in WI38 cells as a function of in vitro aging [22]. These conflicting findings may be the result of the difficulty in obtaining adequate metaphase preparations in senescent culE.rpt/CellRrs
98 (1976)
tures. Since aneuploid cells are also known to have impaired replication [7], flow microfluorometry could be detecting aneuploidy that might be missed by conventional metaphase analyses. Alternate explanations for the increased variation and slight decrease in cellular DNA contents in senescent WI38 cultures include an increased frequency of chromosomal deletions, a decreased amplification of specific DNA genes, or a loss of cellular DNA. Altered binding of mithramycin to DNA is another possibility. However, Yanishevsky et al. [23] have also reported an increased variation in DNA contents in senescent WI38 cells utilizing Feulgen staining and cytophotometry. The increase in cell volumes demonstrated by small angle light scatter in ungated senescent cells can also be observed by measurement of cell diameter [5, 131and by Coulter volume [15]. Although ethanol fixation reduces cell volume, this reduction is relative and does not change the cell volume distribution [3]. Gated single parameter analysis clearly reveals that the increase in cell volume in senescent cells is not limited to any one phase of the cell cycle. It also confirms the increase in cell volume as cells progress through the cell cycle that had been suggested by dry mass measurements [24]. In conclusion, senescent cells appear to have increased cell sizes in both G I and G2+M cells in logarithmic growth as well as at confluency. Since the volume distribution of a cell culture appears to be closely related to its replication rate [1.5], the shift to larger cell volumes observed during the in vitro lifespan of WI38 cells may well be secondary to the observed increasing number of slow or non-dividing cells [l]. Similarly, increased variations in cell sizes may reflect increased heterogeneity in cell division times observed in senescent cells [l]. Exptl Cd RPS 98 (1976)
However. the remote possibility remains that a cell membrane alteration could be the primary event leading to both a loss of replicative ability and an increase in cell volume. WI38 cell cultures were obtained through contract no. I-HD-4-2828 from the National Institute of Child Health and Human Development to Dr Leonard Hayflick of Stanford University. We thank MS K. Braunschweiger for technical assistance.
REFERENCES 1. Absher. P M, Absher, R G & Barnes. W D, EXR cell res 88 (1974) 95. 2. Bonner, W A, Hulett, H R, Sweet, R G & Hertzenberg, L, Rev sci instr 43 (1972) 404. 3. Crissman, H A & Steinkamp, J A, J cell biol 59 (1973) 766. 4. Crissman. H A & Tobev. R A, Science 184(1974) 1297. 5. Cristofalo, V & Kritchevsky, D, Med exp I9 (1969) 313. 6. Cristofalo, V & Sharf, B, Exp cell res 76 (1973) 419. 7. Cure, S, Boue, A & Boue, J, Les accidents chromosomiques de la reproduction (ed A Boue & C Thibault) p. 95. Inserm, Paris (1973). 8. Dean, P N & Jett, J H, J cell biol 60 (1974) 523. 9. Gelfant, S & Smith, J G, Jr, Science I78 (1972) 357. IO. Good, P I & Smith, J R, Biophys j 14 (1974) 81 I. Il. Hayflick, L, Exp cell res 37 (1965) 614. 12. Macieira-Coelho, A, Ponttn, J & Philipson, L, Exp cell res 42 (1966) 673. 13. Macieira-Coelho, A & Ponttn, J, J cell biol 43 (1969) 374. 14. Macieira-Coelho. A. Nature 248 (1974) 42 I. 15. Mitsui, Y & Schneider, E L. Personal communication. 16. Mullaney, P F & Dean, P N, Biophys j IO (1970) 764. 17. Orgel, L E, Nature 243 (1973) 44 I. 18. Saskela, E & Moorhead, P S, Proc natl acad sci US 50 (1963) 390. 19. Simons, J W I M, Exp cell res 45 (1967) 336. 20. Steinkamp, J A, Fulwyler, M J, Coulter, J R, Hiebert, R H, Horney, J L & Mullaney, P F, Rev sci instr 44 (1973) 1301. 21. Van Dilla, M A, Trujillo, T T, Mullaney, P F & Coulter, J R, Science 163(1969) 1213. 22. Whitaker, A M, Gould, J & Sith, E M, Exp cell res 87 (1974) 55. 23. Yanishevsky, R, Mendelsohn, M L, Mayall, B H & Cristofalo, V J, J cell physiol84 (1974) 165. 24. Zetterberg, A & Killander, D, Exp cell res 39 (1965) 22. Received July 14, 1975
