| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Advances in Brief |
Biological Engineering Division, Biotechnology Process Engineering Center, and Center for Environmental Health Sciences, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
 |
ABSTRACT |
|---|
 Top ABSTRACT IntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
|
Introduction |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
illustrates the salient features of the hypothesis. During mitosis in other tissue cells, random chromosome segregation occurs (Fig. 1B)
. In contrast, adult stem cells continuously cosegregate to themselves a complement of chromosomes bearing old parental DNA strands (i.e., immortal DNA strands; Fig. 1A
). By retaining the same original DNA templates and continuously discarding all of the newly synthesized DNA strands to their differentiating sisters, stem cells could avoid all of the mutations from replication infidelity. The earliest suggestion of nonrandom chromosome segregation came from in vitro autoradiographic analyses of radiolabeled mitotic chromosomes in cells as diverse as murine embryonic fibroblasts and plant meristem cells (3
, 4)
. Later, in vivo studies suggested the existence of immortal DNA strand cosegregation in rodent and human tissues (2
, 5
, 6)
. More recently, chromosome cosegregation has been proposed as a key feature of developmental programs responsible for body plan asymmetry in mammals (7)
. Proving the existence of immortal DNA strand cosegregation in vivo has been problematic because of uncertainty in marking immortal DNA, uncertainty in identifying adult stem cells in situ, and pharmacokinetics limitations of in vivo labeling studies in general (2
, 5
, 6)
. By using cultured cells that model the asymmetric cell kinetics of adult stem cells (8, 9, 10)
, we were able to avoid these complications and accomplish a direct demonstration of immortal DNA strands and cosegregation of the chromosomes that bear them.
|
|
Materials and Methods |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
CsCl Gradient Density-Shift Analyses.
Con-3 and Ind-8 cells were cultured for 24 h in medium supplemented with 75 µM ZnCl2 to establish asymmetric cell kinetics for Ind-8 cells (10)
. BrdUrd was added to a concentration of 20 µM and culture continued for 15 h (0.75 cell GT) and 48 h (2.4 GT). High molecular weight chromosomal DNA was isolated by phenol:chloroform extraction with proteinase K and RNAase treatment. Ten µg of DNA was centrifuged to equilibrium in a CsCl solution that had an initial refractive index of 1.40. Approximately 0.2–0.4 µg of DNA isolated from Con-3 cells labeled with [3H]thymidine was included in gradients as an internal marker for LL-DNA. The DNA content of gradient fractions was quantified by picogreen dye fluorescence using procedures specified by the supplier (Molecular Probes, Leiden, The Netherlands).
Mitotic Cell Selection.
Mitotic cells for flow cytometry and BrdUrd-pulse analyses were selected by mitotic shake-off into the medium of the culture after centrifugation to remove debris. Using the same culture medium stabilized the asymmetric cell kinetics during the period of mitotic cell selection. To avoid debris, cells from the first two collections were discarded. Collections were then taken every 60 min for the next 20 h. The isolated mitotic cells were maintained on ice until use immediately after the final collection. They were >95% viable by trypan blue analysis.
Flow Cytometry Analysis.
Mitotic cells were replated in their original culture medium used for mitotic selection and cultured for 5 h before flow cytometry analysis. Detection of incorporated BrdUrd was performed with an antibody kit per the instructions of the supplier (PharMingen). BrdUrd-specific fluorescence was determined with a Becton-Dickinson FACScan flow cytometer and Cell Quest software.
Chromosome Hoechst Fluorescence Analysis.
Mitotic cells for chromosome spreads were isolated by a 16-h treatment with Colcemid at a final concentration of 0.1 µg/ml. Mitotic spreads were prepared as described previously (14)
. Spreads were stained with Hoechst dye 33258 at a final concentration of 0.5 µg/ml for 5 min and washed three times for 5 min in PBS just before analysis. Fluorescent images were captured using a Nikon Diaphot-TMD microscope and a MTI 3CCD camera, and were analyzed with the Zeiss KS400 software package.
Cytochalasin D Analyses.
For BrdUrd-pulse studies, mitotic cells were isolated as for flow cytometry except the collection medium was supplemented with 2 µM cytochalasin D. After collection on ice, cells were allowed to adhere to chamber slides, cultured for 5 h, and then fixed for in situ immunofluorescence with a FITC-conjugated anti-BrdUrd antibody (PharMingen) per specifications of the supplier. Fluorescence microscopy was performed with a Zeiss Axioskop 2 phase/epifluorescence microscope and Axiocam digital camera.
For continuous BrdUrd labeling studies with Zn-dependent Ind-8 cells, asymmetric cell kinetics were induced at a cell density of 500 cells per 1.7-cm2 chamber slide by switching to growth medium containing 65 µM ZnCl2 (10 , 13) . BrdUrd was added after asymmetric cell kinetics induction to a concentration of 5 µM. After a 72-h growth period, the cultures were arrested by the addition of cytochalasin D (2 µM) for 16 h. For experiments with temperature-dependent 1 h-3 cells, asymmetric cell kinetics were induced at a density of 7500 cells/6-cm diameter well by switching to 32.5°C. After 8 h of growth, BrdUrd was added to a concentration of 10 µM. After a 72-h growth period, the cultures were arrested by the addition of cytochalasin D (2 µM) for 4 h. Detection of incorporated BrdUrd was performed using primary antibody Bu1–75 (Harlan) and a biotinylated secondary antibody (Vector Laboratories) with avidin-fluorescein visualization. Images were captured using a Nikon TE300 microscope, Orca camera, and Openlab software (Improvision). Analysis of captured images was performed with 1D Image Analysis software (Eastman Kodak).
|
Results and Discussion |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
The main cell line used for these studies, Ind-8, has experimentally controlled asymmetric cell kinetics (10 , 12 , 13) . Ind-8 cells are spontaneously immortalized p53-null murine embryo fibroblasts stably transfected with a wild-type p53 cDNA controlled by a metal-responsive gene promoter. Under routine culture conditions, Ind-8 cells divide with exponential kinetics, producing two similar dividing sister cells at each division. However, after addition of ZnCl2 to the culture medium and the ensuing restoration of p53 protein expression, the cells switch to deterministic asymmetric cell kinetics. Under these conditions, one sister cell acts like an adult stem cell and continues to cycle with a GT similar to that under conditions of exponential kinetics. The other sister acts like an in vivo differentiating cell and undergoes an immediate viable cell cycle arrest or occasionally divides once to produce two viable arrested cells (10) .
The immortal strand hypothesis is based on semiconservative DNA replication (18)
. The semiconservative nature of DNA replication during asymmetric cell kinetics was confirmed by BrdUrd density-shift analyses in equilibrium CsCl density gradients (19)
. We observed DNA species with buoyant densities corresponding to hemi-substitution (HL) and complete substitution (HH), indicating semiconservative DNA replication (Fig. 2, B and D)
. The HL and HH peaks were discrete (Fig. 2D)
, indicating synthesis of continuous BrdUrd-substituted DNA strands without significant sister chromatid exchange or repair synthesis.
|
; LL) derived from noncycling sister cells that were produced by asymmetric divisions before addition of BrdUrd (Fig. 2, B and D
, 
). In Fig. 2D
, the greater HL peak area was because of the next generation of stable nondividing sisters (
) produced from division of asymmetrically cycling cells that replicated DNA after BrdUrd addition.
Detection of Cosegregated Unlabeled Immortal DNA Strands.
In Fig. 2C
, mitosis by exponentially cycling cells that have undergone two S phases in BrdUrd yields two cycling sisters that have, on average, equivalent BrdUrd content (faded circles) as a result of random segregation of paired HH
HL sister chromatids. In each new sister, about half of the chromosomes are HH and half are HL. In contrast, the immortal strand hypothesis predicts that an analogous asymmetrically cycling sister (
in Fig. 2D
) will cosegregate all HL chromosomes to itself. This maneuver accomplishes retention of nonsubstituted DNA strands that were specified to be immortal strands before the addition of BrdUrd. It follows that the nondividing sisters would receive only HH chromosomes (
in Fig. 2D
). Therefore, asymmetric sister cells should exhibit a 50% difference in nuclear BrdUrd content (all HH versus all HL chromosomes).
To test this prediction, cytochalasin D, an actin antagonist that permits nuclear division but prevents cytokinesis, was used to evaluate the relative BrdUrd content of sister nuclei. Cytochalasin D treatment allowed sister nuclei to be trapped in the same cell cytoplasm for direct comparison of the relative BrdUrd content. The BrdUrd content of sister nuclei was determined by in situ immunofluorescence with anti-BrdUrd antibodies (Fig. 3)
. Cells were arrested with cytochalasin D after continuous culture in BrdUrd for four PDCs. A PDC is equivalent to one cell GT but refers specifically to division by all of the cells present in an initial cohort of cycling cells.
|
. The difference in BrdUrd content was confirmed by reciprocal quenching of Hoechst-fluorescence (20
; see Fig. 3b
). The observed mean difference in anti-BrdUrd fluorescence intensity was 34% (95% CI, 24–43), which differed significantly from the mean value of 15% observed for binucleates of exponentially cycling cells (mean number of binucleates evaluated = 24; P = 0.008). Whereas 37% (n = 19) of binucleates of asymmetrically cycling cells showed 
40% difference in BrdUrd content, all of the binucleates of exponentially cycling control cells showed <40% difference (n = 27).
The same binucleated cell analysis was performed with an independent cell line called 1h-3. 1h-3 cells are p53-inducible murine mammary epithelial cells derived with a temperature-dependent expression system (8
, 9
, 11)
. They exhibit the same type of deterministic asymmetric cell kinetics as Ind-8 cells, although not as efficiently (8, 9, 10)
. In 1h-3 analyses, 26% (n = 31) of binucleates of asymmetrically cycling cells versus 5% (n = 44) of binucleates of exponentially cycling cells showed 40% difference in BrdUrd content (P < 0.05). These results support the existence of immortal DNA cosegregation in cultured asymmetrically cycling cells, independent of their tissue of origin.
Detection of Mitotic Chromosomes Bearing Unlabeled Immortal DNA Strands.
Mitotic chromosome analyses were performed to detect immortal DNA strands without cytochalasin D treatment. Cells were cultured continuously in BrdUrd-containing medium for 6 PDC. Labeled cultures were then treated with Colcemid to prepare mitotic cells for chromosome spread analyses. To detect chromosomes with nonsubstituted DNA strands, mitotic chromosome spreads were stained with the fluorescent DNA dye Hoechst 33258. Incorporated BrdUrd partially quenches the fluorescence of DNA-bound Hoechst dyes (20)
. Mitotic chromosomes that are only hemi-substituted with BrdUrd exhibit one-half the degree of Hoechst-fluorescence quenching observed for bisubstituted chromosomes. Therefore, as a measure of BrdUrd content, the fluorescence intensity per pixel area was determined for individual anaphase chromosomes in single-cell spreads.
Chromosomes from cells cultured without BrdUrd exhibited the greatest Hoechst-fluorescence. Chromosomes from exponentially cycling cells cultured in BrdUrd showed a 24% reduction in median Hoechst-fluorescence (data not shown), indicative of quenching because of bisubstituted BrdUrd. Given this difference, immortal strand cosegregation is predicted to yield an 18% reduction in median Hoechst-fluorescence [i.e., equivalent to the average of complete reduction (24%) and one-half reduction (12%)]. An 18% reduction corresponds to a median fluorescence intensity/pixel area of 125. The observed value was 123 (95% CI, 122–124), indicating the persistence of chromosomes with unlabeled immortal DNA strands in asymmetrically cycling cells after six cell divisions.
Cosegregation of Chromosomes Bearing Marked Immortal DNA Strands.
The previous experiments provided indirect evidence for cosegregation of BrdUrd-free immortal DNA strands. Subsequent experiments had the goal of chemically marking immortal DNA strands and directly observing their cosegregation. To accomplish this goal, Ind-8 cells were first cultured with BrdUrd for 1 GT (producing HL chromosomes) under conditions of exponential kinetics, and then after removal of the BrdUrd (with or without a thymidine chase) the labeled cells were shifted to asymmetric cell kinetics for 5–6 PDC. There were several possible outcomes to this experiment. If only DNA strands that existed before the labeling period were selected to become immortal, then chromosomes with immortal strands would be unlabeled. If immortal strand selection by asymmetrically cycling cells were a random process, then on average half of the chromosomes with immortal strands might be labeled and half unlabeled. Finally, if only the more recently synthesized DNA strands were selected, then all of the chromosomes with immortal strands would contain BrdUrd. As described below, it was possible to mark the immortal DNA strands, and the data obtained were more consistent with the third strand selection mechanism.
To focus the BrdUrd-pulse analyses to cycling cells, mitotic cells were isolated by mitotic shake-off. Collected mitotic cells were returned to culture for 5 h. During this time, >95% of the cells attached and divided to produce new G1 cells. To inspect segregation patterns of BrdUrd-containing chromosomes between sister nuclei, isolated mitotic cells were cultured in the presence of cytochalasin D and then analyzed for anti-BrdUrd immunofluorescence. As predicted, exponentially cycling cells arrested by cytochalasin D showed very little anti-BrdUrd fluorescence (Fig. 4, a and b)
. Random segregation of labeled HL chromosomes in exponentially cycling cells causes their geometric dilution among newly synthesized unlabeled chromosomes. In flow cytometry analyses, new G1 cells produced from exponentially cycling mitotic cells isolated at 5.8 PDC after the BrdUrd-pulse showed a 6.8-fold reduction in BrdUrd content as compared with cells immediately after the BrdUrd pulse. The background fluorescence of BrdUrd-negative cells prevented determination of the predicted 56-fold reduction (i.e., 25.8-fold).
|
. In fact, flow cytometry analysis of G1 cells produced from asymmetrically cycling mitotic cells showed two populations of cells (data not shown). One cell population had a background level of fluorescence, corresponding to nondividing sisters that received no chromosomes with BrdUrd-substituted immortal DNA strands. The other cell population had a BrdUrd content equivalent to that of cells immediately after the BrdUrd-pulse, corresponding to asymmetrically cycling sisters that cosegregated chromosomes containing "old" BrdUrd-substituted DNA strands to themselves. In three separate flow cytometry experiments (mean PDC = 5.1), the BrdUrd-retaining population showed a mean dilution factor of 0.82 (95% CI, 0.4–1.2). This factor was not significantly different from 1.0, the value for complete retention of initially marked immortal DNA strands. However, 0.82 does differ significantly from 2.0 (P < 0.031), the factor for random selection. Therefore, our working hypothesis is that immortal DNA strands are not chosen randomly. Instead, on initiation of the asymmetric cell kinetics program the most recently synthesized DNA strands are selected.
Examination of the distribution of anti-BrdUrd immunofluorescence in binucleates from asymmetrically cycling cells revealed that 22% (n = 262) showed complete localization to only one sister nucleus (see examples in Fig. 4, e–j
). These findings strongly indicate retention and cosegregation of chromosomes bearing old BrdUrd-substituted immortal DNA strands. Moreover, the majority of arrested binucleated cells showed a high degree of unequal localization of BrdUrd immunofluorescence between sister nuclei. In experiments with continuous BrdUrd labeling (e.g., see Fig. 3
), which also permitted examination of BrdUrd localization in exponentially cycling binucleates, such highly unequal localization was never observed. Experiments performed with and without a thymidine chase gave similar results (data not shown).
The Biological Significance of Immortal DNA Strand Cosegregation.
By several independent measures, we provide evidence for immortal DNA strand cosegregation in cells that cycle with asymmetric cell kinetics. The molecular mechanisms by which immortal DNA strands are selected and cosegregated remain to be elucidated. Because p53 is responsible for asymmetric cell kinetics, it may also be involved in immortal strand mechanisms. However, the presented experiments do not address this possibility.
Two aspects of this study may have relevance to adult stem cells in vivo. First, the selection of immortal strands at initiation of asymmetric cell kinetics in culture may be related to the establishment of immortal strands in somatic stem cells in late fetal development. The finding that it is the most recently synthesized DNA strands that appear to be selected may be relevant to processes that establish adult stem cells in newly forming somatic tissues. The same establishment mechanism may recur in adult stem cells under special circumstances (e.g., after exponential divisions for repair of tissue damage). Secondly, immortal strand cosegregation was found to be specific for asymmetric cell kinetics, a property that distinguishes rare adult stem cells from their abundant differentiating progeny. As such, some factors required for immortal strand selection and cosegregation may uniquely identify adult stem cells.
By virtue of demonstrating immortal DNA strand cosegregation in cultured mammalian cells that cycle with asymmetric stem cell kinetics, this study supports the hypothesis of similar processes in adult stem cells in vivo. This increased confidence leads us to consider new ideas regarding the nature of aging mechanisms. Some changes that occur in tissues with advancing age are likely to reflect alterations in the number and function of adult stem cells (10) . Adult stem cells performing immortal strand cosegregation will retain a set of the same DNA molecules for long periods. Alterations that accumulate in these stable immortal DNA strands over time may compromise adult stem cell function and viability, precipitating a decline in tissue function. Possible alterations include stable covalent base modifications (e.g., methylation) and poorly repaired products of chemical reaction (e.g., oxidative damage and deamination). Thus, although immortal DNA strand cosegregation may be a key mechanism by which mammalian evolution has limited prereproductive death from cancer (1) , it may also be an important determinant of tissue aging and life span.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 J. R. M. and J. R. T. were supported by a research grant from the Defense Advanced Research Projects Agency (N0014–98-1–0760). J. A. L. was supported by Training Grant T32-ES07020 from the National Institute for Environmental Health Sciences and an Anna Fuller Fund fellowship administered by the Massachusetts Institute of Technology Center for Cancer Research. 
2 Present address: M. D. Anderson Cancer Center, Department of Molecular Pathology, Houston, TX 77030.
3 To whom requests for reprints should be addressed, at Biological Engineering Division, Massachusetts Institute of Technology, Room 16-743b, 77 Massachusetts Avenue, Cambridge, MA 02139.
4 The abbreviations used are: BrdUrd, bromodeoxyuridine; PDC, population division cycle; GT, generation time; LL, light:light; HL, heavy:light; HH, heavy:heavy; CI, confidence interval; DAPI, 4',6-diamidino-2-phenylindole.
Received 8/21/02. Accepted 10/10/02.
|
REFERENCES |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Hu, S. Liu, D. R. Breiter, F. Wang, Y. Tang, and S. Sun Octamer 4 Small Interfering RNA Results in Cancer Stem Cell-Like Cell Apoptosis Cancer Res., August 15, 2008; 68(16): 6533 - 6540. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kimoto, Y. Yura, M. Kishino, S. Toyosawa, and Y. Ogawa Label-retaining Cells in the Rat Submandibular Gland J. Histochem. Cytochem., January 1, 2008; 56(1): 15 - 24. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Weissmann Mortal and Immortal DNA: Craig Venter and Keats' "Lamia" FASEB J, December 1, 2007; 21(14): 3773 - 3776. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Cairns Cancer and the Immortal Strand Hypothesis Genetics, November 1, 2006; 174(3): 1069 - 1072. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Armakolas and A. J. S. Klar Cell type regulates selective segregation of mouse chromosome 7 DNA strands in mitosis. Science, February 24, 2006; 311(5764): 1146 - 1149. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Muntoni and R. R. Reddel The first molecular details of ALT in human tumor cells Hum. Mol. Genet., October 15, 2005; 14(suppl_2): R191 - R196. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Karpowicz, C. Morshead, A. Kam, E. Jervis, J. Ramunas, V. Cheng, and D. van der Kooy Support for the immortal strand hypothesis: neural stem cells partition DNA asymmetrically in vitro J. Cell Biol., August 29, 2005; 170(5): 721 - 732. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Goriely, G. A. T. McVean, A. M. M. van Pelt, A. W. O'Rourke, S. A. Wall, D. G. de Rooij, and A. O. M. Wilkie Gain-of-function amino acid substitutions drive positive selection of FGFR2 mutations in human spermatogonia PNAS, April 26, 2005; 102(17): 6051 - 6056. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Rambhatla, S. Ram-Mohan, J. J. Cheng, and J. L. Sherley Immortal DNA Strand Cosegregation Requires p53/IMPDH-Dependent Asymmetric Self-renewal Associated with Adult Stem Cells Cancer Res., April 15, 2005; 65(8): 3155 - 3161. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Hong and P. J. Stambrook Restoration of an absent G1 arrest and protection from apoptosis in embryonic stem cells after ionizing radiation PNAS, October 5, 2004; 101(40): 14443 - 14448. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. S. Klar A Genetic Mechanism Implicates Chromosome 11 in Schizophrenia and Bipolar Diseases Genetics, August 1, 2004; 167(4): 1833 - 1840. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. R. Williams and C.-t. Wu Does Random X-Inactivation in Mammals Reflect a Random Choice Between Two X Chromosomes? Genetics, July 1, 2004; 167(3): 1525 - 1528. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Maeshima, S. Yamashita, and Y. Nojima Identification of Renal Progenitor-Like Tubular Cells that Participate in the Regeneration Processes of the Kidney J. Am. Soc. Nephrol., December 1, 2003; 14(12): 3138 - 3146. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
, cycling cell. 
6 PDC (144 h) after the BrdUrd-labeling period, mitotic cells were selected in cytochalasin D-containing medium and returned to culture for 5 h to allow the formation of binucleated G1 cells. Cells were fixed and examined by in situ immunofluorescence with fluorescent anti-BrdUrd antibodies. Shown are low magnification fields of nuclear DNA DAPI fluorescence (a and c) and corresponding anti-BrdUrd fluorescence (b and d) for exponentially (a and b) and asymmetrically (c and d) cycling cells. e–g, examples of binucleates with unequal localization of anti-BrdUrd nuclear fluorescence at high power. Linear arrays correspond to phase contrast, DAPI fluorescence, and anti-BrdUrd fluorescence images. h–j, overlays of DAPI fluorescence and anti-BrdUrd fluorescence images for e–g, respectively.