Reduction in DNA repair capacity following differentiation of murine proadipocytes

Experimental

Cell Research 174 (1988) 502-5 IO

Reduction in DNA Repair Capacity following Differentiation of Murine Proadipocytes PHILIP J. TOFILON’ Department

and RAYMOND

E. MEYN

of Experimental Radiotherapy, The University of Texas M. D. Anderson Hospital and Tumor Institute at Houston, Houston, Texas 77030

It has been suggested that terminally differentiated mammalian cells have a decreased DNA repair capacity, compared with proliferating stem cells. To investigate this hypothesis, we have examined y-ray-induced DNA strand breaks and their repair in the murine proadipocyte stem cell line 3T3-T. By exposure to human plasma, 3T3-T cells can be induced to undergo nonterminal and then terminal differentiation. DNA strand breaks were evaluated using the technique of alkaline elution. No difference was detected among stem, nonterminally differentiated, and terminally differentiated cells in the initial levels of radiation-induced DNA strand breaks. Each of the strand break dose responses increased as a linear function of y-ray dose. The strand breaks induced by 4 Gy rejoined following biphasic kinetics for each cell type. At each time point examined after irradiation, however, the percentage of strand breaks that had not rejoined in terminally differentiated cells was three to six times greater than in stem cells. The rate of strand break rejoining in nonterminally differentiated cells was of an intermediate value between that of the stem and of the terminally differentiated cells. These results indicate that, at least for 3T3-T cells, differentiated cells have a reduced capacity for DNA repair. 0 1987 Academic PXSS, II-C.

A definitive characteristic of a terminally differentiated ceil is its inability to initiate DNA replication. Although the reasons for this inability to synthesize DNA have not been clearly defined, several studies have noted that the activity of DNA synthetic enzymes are decreased in nonproliferating, differentiated cells, presumably because the genes coding for these enzymes are suppressed during differentiation [l-3]. Because of the putative commonality among some of the enzymes and processes involved in DNA synthesis and the repair of DNA, it has been suggested that the ability of a differentiated cell to repair DNA damage may also be compromised [4]. Indeed, Hahn et al. found that as rat muscle cells (myoblasts) differentiate into myotubules in vitro, the repair of DNA damage induced by methylmethane sulfonate is significantly reduced [5]. Other investigators using this same model system reported that compared with myoblasts, myotubules are also deficient in repairing the DNA damage induced by 4nitroquinoline-l-oxide [6] and uv light [7]. In contrast to the reports of reduced repair of the DNA lesions induced by these agents, Chan et al. found no difference between myoblasts and the differentiated myotubules in the repair of X-ray-induced DNA strand breaks [6]. However, more recent studies using in uiuo differentiation models have suggested that as ’ To whom reprint requests should be addressed at Department of Experimental Radiotherapy, Box 66, M. D. Anderson Hospital and Tumor Institute, 1515 Holcombe, Houston, TX 77030. Copyright @ 1988 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/88 $03.G1l

502

DNA repair during differentiation

503

cells differentiate, their ability to repair the DNA strand breaks induced by ionizing radiation is compromised [8, 91. These studies compared DNA repair in rat brain tumor cells to rat cerebellar neurons and in cycling to noncycling cells in the gut and bone marrow of mice. In each case, the rate and extent of repair of radiation-induced strand breaks was decreased in noncycling differentiated cells when compared with the cycling cells. In an attempt to better understand the relationship, if any, between the state of cellular differentiation and the repair of DNA damage induced by ionizing radiation, we have examined a model system recently developed by Scott and colleagues, the murine 3T3-T proadipocyte cell culture system [lO-12, 151.In this system, differentiation is a well-controlled and defined process. Proliferating stem cells are induced by exposure to human plasma to undergo growth arrest at a stage in G, specific for subsequent differentiation (Gn). They then proceed to a state of nonterminal differentiation (NTD) and finally, upon further stimulation, become terminally differentiated (TD) adipocytes. Each of these states of differentiation is easily identified morphologically by the presence of lipid droplets within the cells, and relatively pure populations of each cell type can be obtained. We present in this report the results of studies in which we have used the alkaline elution assay to determine the effects of 3T3-T cell differentiation on y-rayinduced DNA damage and its repair.

MATERIALS

AND METHODS

Cell culture. The murine 3T3-T proadipocyte cell line, which was generously provided by Dr. Robert Scott, Mayo Clinic, Rochester, Minnesota, was used in these studies. Cultures were maintained at 37°C in a humidified 9 % CO* : 91% air atmosphere. Stock culture growth medium consisted of Dulbecco’s modified essential medium (DMEM) (high glucose) containing 10% heat-inactivated fetal calf serum (FCS). Preparation of human plasma and the CEPH plasma fraction. Human plasma (HP) was prepared as described by Krawisz and Scott [l 11.In brief, human plasma purchased from the M. D. Anderson Blood Bank was made platelet free by centrifugation at 25,OOOgfor 30 min at 22°C. Platelet-poor plasma was stored at -70°C overnight and thawed slowly over a 24-h period at 4°C. It was then centrifuged at 1OOOgfor 10 min at 4°C and the supematant was passed through a 0.45 urn Millipore filter. To induce growth arrest (G,) and terminal differentiation (TD), 3T3-T cells were exposed to DMEM containing 20 u/ml heparin and 25% HP. Differentiated 3T3-T cells are easily identified in that they contain large lipid droplets [ 111. The CEPH fraction of HP was used to induce nonterminal differentiation (NTD) and was prepared as described by Wier and Scott [12]. Briefly, platelet-poor HP was adjusted to pH 8.6 and BaCI, (1 M) was added to a final concentration of 0.1 M. It was then mixed for 45 min at 4°C and centrifuged at 3600g for 30 min. The barium chloride precipitate was partially dissolved in 0.9% NaCU0.02 M sodium citrate, mixed for 45 min. and recentrifuged at 3600s for 30 min. The resulting supernatant (CEPH) was dialyzed for 24 h against 6.5 mM sodium citrate (pH 7.4) and filtered prior to use. To induce nonterminal differentiation, DMEM containing 25% CEPH, 20 u/ml heparin, and 1 uM biotin was used. NTD cells are identified by the numerous small lipid droplets within the cells [12]. Induction of nonterminal (NTD) and terminal (TD) diff erentiation. NTD was induced by the bacteriological plate procedure described by Wier and Scott [12]. Rapidly growing 3T3-T cells, which were previously labeled with 0.01 $X/ml [‘4C]thymidine for 2 days, were nonenzymatically removed from the tissue culture surface (2 mM EDTA, 20 min), centrifuged at 5OOgfor 5 min, and resuspended in DMEMCEPH. Cell suspensions (5x 10’) were plated into loo-mm Lab-Tek (Miles Laboratories, Inc., Naperville, IL) bacteriological petri dishes containing DMEMCEPH and 0.01 uCi/ml [‘4C]thymidine. With this method, cells reached maximal differentiation (>80%) at Day 5. In studies of TD, rapidly growing 3T3-T cells were seeded into loo-mm tissue culture dishes

504

Tofilon and Meyn

(Coming, Glass Works, Coming, NY) containing normal growth medium and 0.01 $X/ml [‘4C]thymidine. Two days later, dishes were rinsed twice with phosphate-buffered saline [PBS] and DMEM/HP containing 0.01 &i/ml [“‘Clthymidine was added. In this procedure, cultures reached the growth arrest stage (Go) after 5 days of exposure to HP and maximal TD (>70%) at Day 12. Cultures were fed on Days 4 and 8 with DMEM/HP that did not contain [‘4C]thymidine. Isolation of purified populations of NTD and TD cells. In our hands, the maximum induction of NTD and TD reached approximately 80 and 70%, respectively, of the total cells in culture. To obtain relatively pure populations (>90%) of adipocytes (NTD or TD), cultures were trypsinized (0.05% trypsin/l mM EDTA) on ice for 3 min (NTD) or 10 min (TD) and collected in 5 ml of DMEM/lO% FCS. Suspensions were then centrifuged at 500g for 5 min at 4°C. Since differentiated cells (NTD or TD) contained lipid droplets, they remained in the supernatant while cells than were not differentiated were pelleted. This procedure is similar to that described by Dugail et al. for the isolation of adipose cells from in uiuo preparations [13]. Alkaline elurion. The alkaline elution technique used was that described by Kohn et al. [14]. Cells (8-9x 10’) in complete growth medium were layered onto a 25 mm diameter, 0.8~urn polycarbonate filter (Nuclepore Corp., Pleasanton, CA) and lysed with 5 ml with 2 M NaCl, 0.04 M EDTA, 1% SDS, and 0.5 mg/ml proteinase K (pH 10.0). After 30 min, the lysate was washed with 5 ml of 0.02 M EDTA, and the DNA was eluted in the dark with 0.1 M tetrapropylammonium hydroxide, 0.02 M EDTA (free acid), and 0.1% SDS, pH 12.2, at a constant flow rate of 0.04 mhmin. Fractions were collected every 90 min for 15 h. Any DNA retained on the filter at the end of the elution time was recovered by heating to 60°C for 60 min in HCl and vortexing extensively. DNA remaining in the filter holder was recovered by flushing with 0.4 N NaOH. The DNA in each eluted fraction was assayed by liquid scintillation counting as described [14]. The strand-scission factor (SSF), which reflects the level of DNA strand breaks present, was calculated using the equation SSF = log (f,if,), where& andf, are the proportion of DNA retained on the filter after an eluted volume of 17.5 ml for the unirradiated and irradiated samples, respectively [9]. Data are the average of at least three independent experiments. Irradiations. Monolayer cultures were irradiated on ice in 2 mM EDTA using a 13’Cssource (5 Gyimin). For repair experiments, cultures were irradiated at room temperature in their appropriate medium, returned to the incubator for a specified period of time (I5 to 90 mitt), and then trypsinized on ice and the alkaline elution assay was performed. Plating efficiency. The plating efficiency of 3T3-T stem, NTD, and TD cells was determined by seeding 100cells into 60-mm tissue culture dishes containing DMEM/30% FCS/SOug/ml insulin. This medium has been shown to reverse the NTD phenotype [12] and, in our hands, has no negative effect on the plating efficiency of 3T3-T stem cells. After incubation for 10 days plates were stained with crystal violet and the number of colonies was counted. Data represent the mean f SE of six culture dishes.

RESULTS To investigate the effects of differentiation on initial levels of radiation-induced DNA damage, the strand breaks induced by y-irradiation of 3T3-T stem, NTD, and TD cells were determined using the technique of alkaline elution. In Fig. 1, the SSF calculated from individual alkaline elution profiles are shown as a function of y-ray dose. No differences were detected between the three cell types with respect to SSF, and each response increased as a linear function of dose. These results indicate that, at least as detected by alkaline elution, differentiation has no effect on the initial level of radiation-induced DNA strand breaks. The effects of differentiation on the rate of rejoining of y-ray-induced DNA strand breaks were determined by plotting the percentage of strand breaks remaining as a function of time after y-irradiation (Fig. 2). Each cell type exhibited biphasic rejoining kinetics, which corresponds to previously published studies on the repair of radiation-induced DNA damage [14]. At each time point examined, however, the percentage of strand breaks not rejoined in TD cells was three to six times greater than in the undifferentiated stem cells. When compared

DNA repair during differentiation

505

Dose (Gy) Fig. 1. Dose-response curves indicating the level of DNA strand breaks induced (SSF) as a function of y-ray dose for stem (O), NTD (+), and TD (m) cells. The three cell types were irradiated in monolayer on ice, and the SSF was calculated as described under Materials and Methods. Values represent the mean + SE for three to four independent determinations.

with the stem cells, the fast phase of rejoining in the TD cells was considerably slower, consequently, more of the strand breaks were subjected to the slow phase of repair. NTD cells had an intermediate strand break rejoining rate, but as time after irradiation increased, they appeared to be approaching an extent of rejoining similar to that of the stem cells. In the development of the 3T3-T model, Wier and Scott noted that in some cases, depending on the particular plasma preparation, a population of NTD cells may actually be a mixture of NTD and TD cells [ 121.Thus, in order to ensure that in our hands the NTD cells were truly in a reversible and not a terminal state of differentiation, NTD cultures were prepared in the same manner as those subjected to DNA repair analysis and induced to dedifferentiate [ 121, and their plating efficiencies were determined. When these NTD cultures were trypsinized and plated in tissue culture dishes containing growth medium with 30% FCS plus insulin (50 ug/ml) the plating efficiency was 60+5 % compared with 61+4% for 3T3-T stem cells. When a similar procedure was performed using TD cells, a 6.0f0.6% plating efficiency was obtained. These data indicate that the state of differentiation of the NTD cultures analyzed in the alkaline elution experiments is indeed reversible and is not likely to be a mixture of NTD and TD cells. Since dfferentiated cells are also growth arrested, it was necessary to at least attempt to distinguish the relative influence of this process on repair. This was accomplished by placing 3T3-T cells in a state of growth arrest. Five days after HP exposure 3T3-T cells stopped dividing, which is in agreement with previous studies [15], and became arrested in a substage of G, specific for subsequent differentiation (Fig. 3) [15]. At this time, there was no evidence of differentiation (Le., lipid droplets) and cells were referred to as being in Gn. The SSF of Gn cells as a function of y-ray dose is the same as the other cell types (data not shown).

506

Tofilon and Meyn

0

60 30 Repair time (mix.)

90

Fig. 2. Percentage of DNA strand breaks not rejoined as a function of time after irradiation (4 Gy) for stem (+), Gd (A), NTD (O), and TD (m) cells. Cells were irradiated in the appropriate medium at room temperature and placed back into the 37°C incubator for specified times before being assayed for strand breaks. Values represent the mean f SE of three to four independent determinations.

However, the results in Fig. 2 illustrate that Gn cells rejoin strand breaks at essentially the same rate and to the same extent as the undifferentiated stem cells. Thus, simple growth arrest does not appear to be responsible for the compromised repair capabilities of NTD and TD cells. DISCUSSION Reports based on experimental in vivo models have shown that proliferating cells are able to repair DNA strand breaks induced by ionizing radiation to a greater extent than nonproliferating cells [8, 91. These studies were interpreted as indicating that as a result of differentiation, cells decrease their repair capacity. However, owing to the nature of the in vivo systems used, it was not possible to determine the exact clonal relationship between the proliferating and nonproliferating populations examined. In addition, the possibility that merely growth arrest, and not differentiation, played a role in DNA repair could not be ruled out. Thus, in contrast to the in vitro myoblast/myotubule studies using chemical carcinogens and uv light [5-71, the effects of differentiation on the repair of DNA strand breaks induced by ionizing radiation remained unclear. The data presented

DNA repair during differentiation

507

60. 50. 40.

10. :7. 6543-

2-

/ 1C 0

1

2

4

3

5

6

7

Days

Fig. 3. The number of cells per colony expressed as a function of time after plating. The arrow indicates the time of human plasma (HP) addition. Values represent the mean for 50 colonies. Stem cells (-) before and (---) after addition of HP.

here clearly indicate that as 3T3-T stem cells differentiate to form end-stage adipocytes, their ability to rejoin y-ray-induced DNA strand breaks is decreased; this decrease in repair capabilities is not detected in growth-arrested (Gn) 3T3-T cells. These results thus support the hypothesis that differentiated cells have a reduced capacity for DNA repair. In the 3T3-T system, Scott and colleagues have shown that the formation of TD cells requires that cells first reach a state of NTD [ 101.NTD cells are nonproliferative and contain lipid droplets, although smaller than those in TD cells. The most obvious functional difference between NTD and TD cells is that upon the appropriate stimulus NTD cells can reenter the cell cycle, reinitiate DNA synthesis, and eventually divide. Although the rejoining of DNA strand breaks is decreased in NTD 3T3-T cells compared with stem cells or growth-arrested cells, it remains faster and more complete than in TD cells. This difference in repair between these states of differentiation may be related to the ability of NTD cells to reinitiate DNA synthesis, which suggests the continued availability of DNA replicative and presumably repair enzymes. It is also noteworthy that the apparent stepwise decrease in the repair of DNA strand breaks as cells move from GD to NTD to TD appears to correlate with the progressive nature of the differentiation process as defined in the 3T3-T cells [lo].

508

Tofilon and Meyn

The molecular mechanisms responsible for the decrease in repair of y-rayinduced strand breaks as 3T3-T cells differentiate remains to be determined. It is possible that as a result of changes in gene expression that accompany differentiation, the synthesis of DNA replicative enzymes, which theoretically are not necessary in nonproliferative cells, is suppressed. This particular mechanism was suggested in many of the early studies comparing DNA repair in rat muscle myotubules and myoblasts [5-71. An alternative explanation is that the actual levels of repair enzymes are not altered, but their accessibility to damaged DNA is. Wheeler and Wierowski, in comparing DNA repair in rat brain tumor cells and cerebellar neurons, hypothesized that the more compact chromatin of differentiated cells prevents repair enzymes from reaching the damaged DNA [16]. In another model system, chromatin was shown to be more compact in differentiated cells than in stem cells [17]. Although chromatin structure and repair enzyme levels have not been characterized in the 3T3-T cell differentiation system, it is not unreasonable to speculate that one or a combination of these factors may be involved in the differentiation-induced decrease in DNA repair capability. The inability of a cycling cell to repair DNA damage is usually associated with cell death, i.e., loss of reproductive capacity. The functional consequence of the reduced ability of a nonproliferating end-stage (hence reproductively dead) cell to repair DNA damage is more difficult to assess. The radiation doses used in our studies of 3T3-T cells induce DNA damage but are not large enough to result in membrane damage, inhibit transcription, or result in any other identifiable cellular injury. Thus, while the results of these studies clearly show that as 3T3-T cells differentiate, their ability to repair DNA is compromised; the biological relevance of this effect remains in question. The reduced repair capability of differentiated cells could lead to an increase in the fixation of DNA damage resulting in the retention of lesions over long periods of time, which may indeed play a role in cell function. In experiments designed to investigate the functional consequences of DNA damage in terminally differentiated cells, Lett et al. examined both DNA damage and cell loss in the retina of rabbits [18]. They reported that as the age of unirradiated control animals increased degradative type changes in their retinal DNA occurred in an age-dependent manner and before the actual loss of retinal cells. These investigators then found that when the rabbits were optically irradiated at 6 weeks of age, retinal cells exhibited a similar time-dependent DNA degradation but at younger ages than in unit-radiated animals; the larger the radiation dose, the earlier the onset of cell damage. The appearance of retinal DNA damage in these irradiated rabbits also preceded the loss of retinal cells. Thus, these results suggest that DNA damage is not without deleterious sequelae in nonproliferating cells. Studies using cultured cells obtained from patients with diseases involving primary neuronal degeneration have also suggested, although somewhat indirectly, that DNA repair may be crucial for maintaining the function of differentiated cells. In studies of patients with xeroderma pigmentosum (XP), a genetic disorder characterized by uv hypersensitivity and a defect in DNA repair ability, a

DNA repair during differentiation

509

correlation was shown between the sensitivity of fibroblasts cultured from these patients to uv radiation and their respective age of onset and severity of neurological symptoms [ 191.In addition, hypersensitivity to the lethal effects of X-rays and radiomimetic chemicals have been reported for cells obtained from patients with ataxia telangiectasia [20], Huntington’s disease [20], Parkinson’s disease [21], and Alzheimer’s disease [21], all of which are characterized by primary neuronal degeneration. These studies are consistent with the neuronal DNA integrity theory of Robbins et al., which postulates that neurons undergo primary degeneration as a result of the accumulation of unrepaired DNA damage [22]. The condition of neuronal DNA in these disease states, however, has yet to be reported. Regarding the 3T3-T system, there is no functional assay for the terminally differentiated adipocyte cells and thus the biological consequences of a decrease in DNA repair ability, as for the neurons, are subject to speculation. However, because it is one of the few well-characterized and well-controlled models of cellular differentiation, 3T3-T cells represent a unique in vitro system for the study of the mechanisms and processes involved in the regulation of DNA repair. We thank Dr. Robert E. Scott of the Departments of Pathology and Cell Biology, Mayo Clinic and Foundation, Rochester, Minnesota, for kindly providing us with the 3T3-T system. This work was supported by Grant CA-06294 from the National Institutes of Health. We thank Deborah Thomas Elum and Donna McGhee for preparing this manuscript.

REFERENCES 1. O’Neill, M., and Strohman, R. (1969) J. Cell. Physiol. 73, 61. 2. Stockdale, F. E. (1970) Deu. Biol. 21, 462. 3. Shainberg, A., Yogil, G., and Yaffe, D. (1971) Deu. Biol. 25, I. 4. Alexander, P. (1967) Symp. Sot. Exp. Biol. 21, 29. 5. Hahn, G. M., King, D., and Yang, S.-J. (1971) Nature (New Biol.) 230, 242. 6. Chan, A. C., Ng, S. K. C., and Walker, I. G. (1976) 1. Cell. Physiol. 70, 685. 7. Stockdale, F. E. (1971) Science 171, 1145. 8. Wheeler, K. T., and Wierowski, J. V. (1983) Radiat. Enuiron. Biophys. 22, 3. 9. Murray, D., and Meyn, R. E. (1987) Radiat. Res. 109, 153. 10. Scott, R. E., Wille, J. J., Pittelkow, M. R., and Sparks, R. L. (1985) Carcinogenesis (Barrett, J. C., and Tennant, R. W., Eds.), Vol. 9, p. 67, Raven Press, NY. 11. Krawisz, B. R., and Scott, R. E. (1982) J. Cell. Physiol. 94, 394. 12. Wier, M. L., and Scott, R. E. (1985) Cancer Res. 45, 3339. 13. Dugail, I., Quignard-Boulange, A.. Ardouin, B., and Brigant, L. (1986) In vitro Cell Dev. Biol. 22, 375.

14. Kohn, K. W., Ewig, R. A. G., Erickson, L. C., and Zwelling, L. A. (1981) DNA Repair: A Laboratory Manual of Research Procedures (Friedberg, E., and Hanawalt, P., Eds.), p. 379, Dekker, NY. 15. Wille, J. J., and Scott, R. E. (1982) J. Cell. Physiol. 112, 115. 16. Wheeler, K. T., and Wierowski, J. V. (1983) Radiar. Res. 93, 312. 17. Szabo, G., Damjanovich, S., Sumegi, J., and Klein, G. (1987) Exp. Cell Res. 169, 158. 18. Lett, J. T., Bergtold, D. S., and Keng, P. C. (1986) Mechanisms of DNA Damage and Repair, Basic Life Sciences (Simic, M. G., Grossman, L., and Upton, A. C., Eds.), Vol. 38, p. 139, Plenum, NY. 19. Andrews, A. D., Barrett, S. F., and Robbins, J. H. (1978) Proc. N&l. Acad. Sci. USA 75, 1984. 20. Moshell, A. N., Tarone, R. E., Barrett, S. F.. and Robbins, J. H. (1980) Lancer 1, 9.

510

Tofilon and Meyn

21. Robbins, J. H., Otsuka, F., Tarone, R. E., Polinsky, R. J., Brumback, R. A., and Nee, L. E. (1985) J. Neurol. Neurosurg. Psychol. 48, 916. 22. Robbins, J. H. (1983) Cellular Responses to DNA Damage (Friedberg, E. L., and Budges, B. A., Eds.), p. 671, A. R. Liss, NY. Received July 15, 1987

Printed

in Sweden