In Vitro Cell. Dev. Biol. 27A:557-561. July 1991 © 1991 Tissue Culture Association 0883-8364/91 $01.50+0.00

EFFECTS OF DIFFERENTIATION OF EMBRYONAL CARCINOMA CELLS (P19) ON MITOCHONDRIAL DNA CONTENT IN VITRO

GURMIT SINGH ~D KAREN L. VELTRI

Ontario Cancer Foundation, Hamilton Regional Cancer Centre and McMaster University, Department of Pathology, 1200 Main Street West, Hamilton, Ontario, LSN 3Z5 Canada

(Received 12 November 1990; accepted 26 February 1991)

SUMMARY

The embryonal carcinoma cell line P19 is derived from mouse teratocarcinomas. These pluripotent cells can be induced to differentiate into a variety of cell types by exposure to various drugs. We used retinoic acid to induce embryonal carcinoma cells to differentiate into neuronlike cells. In this study, we show that changes occur in mitochondria during differentiation of embryonat carcinoma cells to neuronlike cells. We found that various morphologic parameters such as mitochondrial fractional area and mitochondrial size decrease as embryonal carcinoma cells differentiate into neuronlike cells. Similar changes were also observed in mitochondrial DNA content. Stereologic analysis of cell preparations provided a measure of mitochondrial fractional area per cell and mtDNA content was assessed by radiolabeled mtDNA probe. This study establishes that mitochondria are regulated as cells differentiate.

Key words: mitochondrial DNA; differentiation; P19-carcinoma cells.

INTRODUCTION

The purpose of this study was to examine whether in vitro we can observe changes in mitochondrial DNA content as cells differen- tiate. This was prompted by our observation that in vivo differen- tiated organs have distinct mtDNA genomic copies (15). Studies of differentiation occurring in the developing mammalian embryo have been facilitated through the use of P19, a clonai line of embryonal carcinoma (EC) cells. EC cells are undifferentiated stem cells de- rived from mouse teratocarcinomas (7). Subsequent to differentia- tion of EC cells in vitro, EC cell development proceeds in a similar manner to that of the normal mouse embryo (10). This similarity renders the P19 cell line a useful model system for early mamma- lian developmental events. Retinoic acid is derived from Vitamin A (retinol) (12) and affects differentiation of EC cells. After exposure to micromolar amounts of retinoic acid, the pattern of differentiation of EC cells into neuronlike cells closely resembles that of brain tissue of embryonal origin (4,5,8,11).

Differences in mitochondrial DNA (mtDNA) content in various cell types were shown in earlier investigations (1,9). More recently, mtDNA content per mitochondrial density was found to be distinct in different mouse organs (15). In the current study, we assessed mtDNA content and mitochondrial fractional area in neuronlike and undifferentiated cells.

Mitochondrial fractional area was determined by stereologic anal- yses of cell preparations using electron microscopy, mtDNA content per cell was estimated using a specific mtDNA probe. We report here that mtDNA levels parallel mitochondrial fractional area in neuronlike cells differentiated in vitro. In undifferentiated cells, mtDNA content per mitochondrial density was about 3 times higher

than in differentiated cells. This characteristic may be used to distin- guish between differentiated and undifferentiated P19 cells.

MATERIALS AND METHODS

Cells. P19 is an embryonal carcinoma cell line isolated from C3H/He mice (6). Cells were provided by Dr. M. W. McBurney (Department of Medical Biology, University of Ottawa). Cells were grown in Dulbecco's modified Eagle's medium with 10% fetal bo- vine serum (GIBCO, Grand Island, NY). The cell suspension was seeded at a density of l 0 s cells/ml supplemented with 100 #g/ml of antibiotic-antimycotic [containing penicillin G sodium (10 000 U/ml); amphotericin B as Fungizone (25 #g/ml); streptomycin sul- fate (10 000 ttg/ml)].

Aggregates form in bacteriologic grade petri dishes after 4 days, an an aliquot of aggregate was added to tissue culture dishes con- taining medium supplemented with 10% fetal bovine serum and 100 ttg/ml antibiotic-antimycotic solution. The medium was re- placed at 48-h intervals. These cells were used as controls.

In vitro differentiation of EC cells with retinoic acid. Embryonal carcinoma cells (P19) were seeded at a density of l 0 s cells/ml in alpha medium supplemented with 10% fetal bovine serum and 0.3 ~tM retinoic acid (all trans-retinoic acid) prepared as a stock (1 mM in 95 % ethanol) (14). These cells aggregated sponta- neously. Because aggregates were formed, retinoic acid-supple- mented medium was replaced with medium lacking retinoic acid. The non-neuronal cells were eliminated by adding cytosine arabino- side (5 #g/ml) (11) and a population rich in neuronlike cells was obtained. The neuronlike characteristics were visible on Day 8 of

557

5 5 8 SINGH AND VELTRI

the experiments. Cells were harvested on Day 9 for mitoehondrial DNA quantitation and electronmicroscopic morphometry.

Electron microscopy. Embryonal carcinoma cells were fixed in 2% glutaraldehyde (in 0.1 M sodium cacodylate, pH 7.4) and incu- bated on ice for 1 h. The cells were postfixed in 1% osmium tetrox- ide (in 0.1 M sodium caeodylate, pH 7.4) for 1 h at 4 ° C, and dehydration in graded ethanol followed. The final pellet was embed- ded in Spurr's resin (13). Thin sections of 70 nm were stained with uranyl acetate and lead citrate and transferred to grids. Using the JEOL-Biosystem electron microscope attached to a video camera, electron mierographs (scape magnification X 12Ki were videotaped. The videotape was then scored by another individual so that ob- server bias was not encountered. Random samples of the electron microscope images were used until a minimum of 5100 #m 2 of reference areas was analyzed. From these images, stereologic analy- ses were done using the Standard Program Image Analysis System of the Kontron MOP-VIDEOPLAN (Zeiss). This program allowed us to measure mitochondrial surface area within a frame of refer- ence. A total area of approximately 5100 #m 2 was analyzed for each cell type.

mtDNA isolation from cell cultures. After harvesting the cells, the pellet was reconstituted in medium. One-milliliter aliquots of various cell concentrations were lysed for 2 h at 60 ° C with 10% sodium dodecyl sulfate (SDS) and proteinase K (0.2 mg/ml) to inhibit nucleases. A known quantity of a smaller circular plasmid DNA (100 ng of PBR322) was added to each aliquot of cells before extraction of total DNA. We ensured that the aliquots were rela- tively translucent before extraction. DNA was extracted using phenol:chloroform [1:1] and ch]oroform:isoamyl alcohol [24:1] and precipitated with ethanol in the presence of 3 M sodium acetate. The final pellet was reconstituted in buffer (containing 10 mM Tris- C1, 1 mM EDTA pH 7.4). An aliquot of DNA was then removed and electrophoresed on a 0.7% agarose gel. This ensured precipitation of plasmid DNA, which was then hybridized to estimate the recovery of DNA from the extraction. Similar amounts of DNA were electro- phoresed in control wells for comparison of radioactivity. This pro- cedure usually yielded a recovery of 85 to 90% of plasmid DNA, with a similar recovery for mtDNA. DNA was denatured in 3 M NaOH for 1 h in a 60 ° C water bath and neutralized in 2 N ammo- nium acetate before DNA blotting directly onto a nitrocellulose filter (Schleieher and Schuell, BA 85, 0.45#m size).

mtDNA hybridization. Nitrocellulose filters were prehybridized for 6 h at 68 ° C in buffer [20)< standard saline citrate (SSC), 50)< Denhardt's yeast transfer RNA (20 mg/ml)]. Filters containing bound mtDNA were hybridized with [a2 P]dATP-labeled mtDNA probe for 18 h at 68 ° C in 20 )< SSC. Vector PSP64 contains the mouse mtDNA genome inserted into the Sac I restriction site (pro- vided by Dr. D. Clayton, Stanford University).

Autoradiography and densitometric analysis. The filters were exposed to X-ray film (Kodak Diagnostic Film). Quantitation of mtDNA on slot blots were done by using the BIO-RAD model 620 Video Densitometer. Standard curves with various concentrations of mtDNA in duplicate were used to estimate mtDNA in the various cell types. DNA dilutions were done to estimate the content in the linear range.

RrSULTS

Phase contrast microscopy. The neuronlike cell population was characterized by the presence of elongated processes radiating from

FIG. 1. Phase contrast photomicrographs of llve EC cultures. A, con- trois: EC cells aggregated for 4 days in medium lacking any drug. After transfer to tissue culture dish, most cells remain undifferentiated. Bar = 2 mm. B, retinoic acid-treated (neuronlike): Embryonal carcinoma cells ag- gregated for 4 days in medium supplemented with 0.3 #M retinoic acid. After transfer to tissue culture dish, a population of >90% neuronlike cells was formed. Bar = 2 mm.

the cell body. Figure 1 A illustrates the morphologic appearance of undifferentiated P19 ceils (controls). Figure 1 B illustrates the characteristics of the neuronlike cell population observed with phase contrast optics. Aggregates formed in the absence of retinoic acid (controls) retained the characteristic appearance of undifferen- tiated cells lacking neuronal processes.

Electron microscopy. Neuronlike cells were identified by the presence of numerous and highly developed organelles with particu- larly prominent Golgi apparatus. Figure 2 B illustrates these char- acteristics. Characteristics of undifferentiated cells (controls) were evident in Fig. 2 A. These include few undeveloped organelles.

Stereologic analysis. Sterelogic measurements from videotaped electron micrographs (scope magnification )<12K) allowed us to obtain the total mitochondrial area per reference area (mitochon- drial fractional area). Values of mitochondrial fractional area are

M1TOCHONDRIAL DNA CONTENT 559

FiG. 2. Electron micrographs of EC cells. A, controls: Aggregates not exposed to retinoic acid show characteristics of undifferen- tiated cells: fewer undeveloped organelles, m = mitochondria. Bar = 1 #m. B, retinoic acid-treated (neuronlike): Retinoic acid-treated aggregates show characteristics of developing neurons: more abundant organelles with prominent Golgi apparatus, m = mitochondria. Bar = 1 gm. C, controls: Low magnification electron micrograph illustrating undifferentiated cell size. Bar = 1 #m. D, retinoic acid-treated (neuronhke): Low magnification electron micrograph illustrating neuronlike cell size. Bar = 1 gm.

listed in Table 1. Mitochondrial a rea was obtained by measu r ing the

cross-sect ional a rea of each mitochondrion within each micrograph;

reference area was obtained by measu r ing the cross-sect ional a rea

of each micrograph. The cumulat ive mitochondrial a rea for a ser ies

of micrographs is the " total mitochondrial a rea . " The "total refer-

ence a rea" is the cumulat ive cross-sect ional a rea of a ser ies of

micrographs . To ensu re statistical significance, a total re ference

area of approximately 5 1 0 0 # m 2 was used. The mitochondrial frac-

tional a rea in neuronl ike cells was about 5 0 % of that of control

cells. The average mitochondrial cross-sect ional a rea in neuronl ike

cells was about 5 0 % of that of controls (Table 1). However, neuron-

like cells were about twice the size of contrals (Table 1). The differ-

ence in cell size between control and neuronl ike ceils is i l lustrated

in Fig. 2 (C,D). Using the two-sample t test, significant differences at P < 0 .0 5

were found in mitochondrial size between neuronl ike ceils and con-

TABLE 1

STEREOLOGIC ANALYSIS OF EC CELL SECTIONS.

AreaMT," MT Area, Mean Area/MT, b Mean Area/Cell, b CeU Type /~ln 2 Fractional Area #m 2 #m 2

Control 590 0.11 1.12 _+ 0.74 c 61 -+ 22 c Retinoic acid (neuronlike) 316 0.06 0.48 + 0.41 c 151 + 89 c

~" ~ Area~T" refers to total mitochondrial cross-sectional area expressed per ~ 51 O0 gm 2. b Values for Mean Area per MT and Mean Area per Cell are mean + or. ~ Significant difference at P < 0.05.

560 SINGH AND VELTRI

TABLE 2

MtDNA QUANTITATION IN EC CELLS

Ratio of MtDNA/CelP MtDNA Content/Cell, Copies" MT Area/Cell, to Mt Area/CelL

Cell Type ttg MtDNA/Cell /I/m 2 copies/ttm 2

Control 10 × 10 -s + 1.1 × 10 -so 6667 6.7 995 Retinoic Acid

(neuron-like) 5 X 10 -s +_ 1.9 × 10 -s" 3333 9.3 358

- 1 copy corresponds to 1.5 × 10-n ttg DNA. b Values for MtDNA content per cell are mean + a. ~ Significant difference at P < 0.05.

trois. There was also a significant difference at P < 0.05 in cell size between neuronlike and control cells.

mtDNA content. Densitometric analysis of autoradiographs re- vealed an approximate twofold difference in mtDNA content per cell. Control cells contained a higher mtDNA content (Table 2). By multiplying mitochondrial fractional area values by mean area per cell, we established a value for the proportion of mitochondrial area per cell cross-sectional area or "MT area/cell ." When expressed per cell, the ratio of mtDNA content (neuronlike:control) was ap- proximately 1:2 whereas the ratio of mitochondrial area (neuron- like:control) was approximately 1.5:1. The mtDNA content per mi- tochondrial area was then expressed as a ratio of neuronlike:control and was about 1:3 indicating a threefold decrease in mtDNA con- tent per mitochondrial fractional area in neuronlike cells (Table 2). In both neuronlike and control cells, the mtDNA content was pro- portional to mitochondrial fractional area. Using the two-sample t test, significant differences at P < 0.05 were found in mtDNA content per cell between neuronlike ceils and controls.

DISCUSSION

Embryonal carcinoma cells have been used in differentiation stud- ies by various investigators (3,4,6,8,11). Results in the present study show coordinate changes in mtDNA and mitochondrial frac- tional area exhibited by neuronlike cells differentiating in vitro and in undifferentiated cells.

This study shows a concomitant decrease in mtDNA content and mitochondrial fractional area as EC cells differentiate into neuron- like cells. A study on HL-60 promyelocytic cells, which had differ- entiated into granulocytes, used Rhodamine 123 fluorescence to determine the concentration of the mitochondrial enzyme cy- tochiome c oxidase and electron microscopy to determine mito- chondrial density (2). Collins and Foster (2) reported that differen- tiated cells exhibit decreased mitochondrial density simultaneous with a decrease in cytochrome c oxidase. Williams (16,17) found a direct relationship between mtDNA level and mitochondrial gene copy number in mammalian striated muscle. Also, studies using electrically stimulated muscle have found that mitochondrial num- ber increases proportionately to an increase in mtDNA content (16,17).

Morphologic similarities exist between cell types present in em- bryonic tissue and neuronlike cells derived from P19 cells differen- tiated in vitro (11). We thus identify embryonal carcinoma-derived cell types on the basis of resemblance to embryonic tissue. We confirmed our neuronhke and control cell population by compari-

son of phase contrast micrographs to those previously reported (4,5,11). Electron microscopic analysis enabled us to further char- acterize neuronlike and control cells. Neuronlike cells have previ- ously been identified by the prevalence of numerous and highly developed organelles with conspicuous Golgi apparatus. Control cells lacked these neuronlike qualities manifested by fewer and less-developed Ürgane••es (8). We confirmed our neuronlike and control cell populations on the basis of resemblance of electron micrographs to those of the previous study (8). At the molecular level, developing neurons in embryonic tissue and their counter- parts differentiated in vitro do not express the genes encoding pro- teins found in terminally differentiated adult tissue (11).

The data from this study support the conclusion that mtDNA content, mitochondrial fractional area, and mitochondrial size di- minish concurrently as neurons develop in vitro. Thus it shows a chemical stimuli that leads to differentiation has a similar effect as an electrical stimuh on mtDNA content (15,16). The significance of a change in mtDNA as ce~s differentiate is unknown at present. However, one could speculate that mtDNA may be related to the function of the differentiated cells. Inasmuch as mtDNA is under the regulatory control of the nucleus, the state of differentiation also determines the amount of mtDNA in the differentiated cell.

In summary, this study demonstrates that mtDNA content changes as cells differentiate but it does not show a distinct change in copy number per mitochondria as seen in vivo. Further studies are required to resolve this issue.

REFERENCES

1. Bogenhagen, D.; Clayton, D. A. The number of mitoehondfial deoxyri- bonucleic acid genomes in mouse L and human HeLa cells. J. Biol. Chem. 249:7991-7995; 1974.

2. Collins, J. M.; Foster, K. A. Differentiation of promyelocytic (HL-60) cells into mature granulocytes; mitochondrial-specific Rhodamine 123 fluorescence. J. Cell Biol. 96:94-99; 1983.

3. Goto, K.; Hayashi, S.; Shirayoshi, Y., et al. Exogenous 5-crystallin gene expression as probe for differentiation of teratocarcinoma stem cells. Cell Differ. 24:139-148; 1988.

4. Jones-Villeneuve, E. M. V.; McBurney, M. W.; Rogers, K. A., et al. Retinoic acid induces embryonal carcinoma cells to differentiate into neurons and glial ceils. J. Cell Biol. 94:253-262; 1982.

5. Jones-Villeneuve, E. M. V.; Rudnicki, M. A.; Harris, J. F., et al. Retinoic acid-induced neural differentiation of embryonal carci- noma ceils. Mol. Cell. Biol. 3:2271-2279; 1983.

6. McBumey, M. W.; Jones-Villeneuve, E. M. V.; Edwards, M. K. S., et al. Control of muscle and neuronal differentiation in a cultured em- bryonal carcinoma cell line. Nature 299:165-167; 1982.

7. McBurney, M. W.; Rogers, B. J. Isolation of male embryonal carci-

MITOCHONDRIAL DNA CONTENT 561

noma cells and their chromosome replication patterns. Dev. Biol. 89:503-508; 1982.

8. McBurney, M. W.; Reubl, K. R,; Ally, A. I., et al. Differentiation and maturation of embryonal carcinoma-derived neurons in cell culture. J. Neurosci. 8:1063-1073; 1988.

9. Nass, M. M. K. Mitochondrial DNA: advances, problems and goals. Science 165:25-35: 1969.

10. Papaioannou, V. E. Interactions between mouse embryos and terato- carcinomas. INSERM (Int. Nat. Sante Rech. Med.) Symp. 10:141- 155; 1979.

11. Rudnicki, M. A.; McBurney, M. W. Cell culture methods and induction of differentiation of embryonal carcinoma cell lines. In: Robertson, E. J., ed. Teratocarcinomas and embryonic stem cells: a practical approach. Oxford. England: IRL Press; 1987:19-49.

12. Sporn, M. B.; Roberts, A. B.; Goodman, D. S. The retinoids, vols. 1,2. New York: Academic Press; 1984.

13. Spurr, A. R. A low-viscosity epoxy resin embedding medium for elec- tron microscopy. J. Ultrastruct. Res. 26:31; 1969.

14. Strickland, S.; Mahdavi, V. The induction of differentiation in teratocar- cinoma stem cells by retinoic acid. Cell 15:393-403; 1978.

15. Vehri, K. L.; Espiritu. M.; Singh, G. Distinct genomic copy number in mitochondria of different mammalian organs. J. Cell. Physiol. 143:160-164; 1990.

16. Williams, R. S.; Salmons, S.; Newholme, E. A., et at. Regulation of nuclear and mitochondrial gene expression by contractile activity in skeletal muscle, J. Biol. Chem. 261:376-380; 1986.

17. Williams, R, S. Mitochondrial gene expression in mammalian striated muscle. J. Biol. Chem. 261:12390-12394; 1986.

We thank Dr. Michael W. McBurney (Department of Medical Biology, University of Ottawa) for providing us with P19 embryonal carcinoma cells; Dr. D. Clayton (Stanford University) for providing us with vector mtDNA probe; and Mr. Ernie Spitzer and Dr. G. Simon (Director, McMaster University Medical Centre) for their guidance in electron microscopic morphometric analysis. We also acknowledge Dr. M. Richardson and Dr. R. M. K. Lee for discussion regarding morpho- metric analysis of our data.

This study was financially supported by the Medical Research Council of Canada.