The relationship of expression of statin, the nuclear protein of nonproliferating cells, to the...

Journal of Neuroscience Hesearch 26:l-15 (1990)

The Relationship of Expression of Statin, the Nuclear Protein of Nonproliferating Cells, to the Differentiation and Cell Cycle of Astroglia in Cultures and In Situ S. Fedoroff, I. Ahmed, and E. Wang Department of Anatomy, University of Saskatchewan, Saskatoon, Canada (S.F. , I .A.): The Bloomfield Centre for Research in Aging, Lady Davis Institute for Medical Research, The Sir Mortimer B. Davis Jewish General Hospital and McGill Univercity , Montreal. Quebec. Canada

Cells in the quiescent, nonproliferative state express a protein, statin, in their nuclei. When the cells re- enter the cell cycle, statin disappears and another protein, cyclin, appears. We have examined mouse astroglia at various stages of differentiation in cultures and astroglia in adult mouse brains for the presence of statin. In cultures initiated from the neopallium of newborn mice, the glial fibrillary acidic protein (GFAP) + stellate astrocytes were statin-negative (statin-) but cyclin-positive (cyclin+). In the same cultures, large flat cells (senescent cells) were statin+ but cyclin-. In frozen sections of the brains of adult mice and in brain smears, GFAP+ astrocytes were statin-. Neither stellate astrocytes grown in cultures for 30 or more days nor astrocytes in adult mouse brain were labeled when pulsed with bromode- oxyuridine (BudR). When astroglia were treated with dibutyryl cyclic adenosine monophosphate (dBcAMP), large stellate cells that closely resemble reactive astrocytes in situ formed. These cells were all statin + from 11-62 days in vitro; however, reactive astrocytes in mouse neopallium, 4-50 days after a stab wound, were statin-. In colony cultures, senescent cells became statin + , whereas stellate astrocytes and their precursor cells remained statin-. These ob- senations indicate that normal astrocytes both in cultures and in situ retain the potential to divide and probably progress through the cell cycle at a very slow rate.

Key words: statin, cyclin, astroglia, reactive astrocytes, colony cultures, GFAP

INTRODUCTION

The central nervous system (CNS) is composed of neurons and of glial cells, which include oligodendrog- lia, microglia, and astroglia. Neuronal precursor cells

divide until terminal mitosis and produce neurons with characteristic perikaryon. dendrites and axon, synapses, and the specific transmitters they synthesize. Oligoden- drocyte precursor cells also proliferate, enter the perma- nent nonproliferative state, and undergo well-defined stages of postmitotic differentiation (Imamoto et al., 19781, resulting in end cells that produce myelin and assume a specific relationship to neurons. Microglia are highly controversial cells; their origin has not yet been defined and whether they finally differentiate into mac- rophages has not yet been resolved.

Astrocyte precursor cells, proastroblasts and astroblasts, are proliferative cells; however, astrocytes are nonproliferative (Fedoroff et al., 1984a) or, possibly, have an extremely long proliferative cycle. When, or whether, astrocyte precursor cells enter terminal mitosis is not known, nor can the differentiated end cells of astrocytes be identified other than as process-bearing stellate cells that make contact with the basement mem- branes of pia and capillaries, connected by gap junctions to each other, to ependymal cells, and to oligodendrog- lia. There are no criteria that can be said categorically to define the terminally differentiated astrocyte.

In recent years two cell nuclear proteins, cyclin (Takasaki et al., 1984) and statin (Wang, 1985), have been found; cyclin is specific to cycling cells and statin to quiescent (Go phase) noncycling cells, and to senescent. nondividing cells (Takasaki et al . , 1984; Wang, 1985).

S-30 and S-44, mouse monoclonal antibodies to statin, were produced from hybridomas prepared from

Received August I , 1989; reviscd October 6 , 1989; accepted October 26, 1989.

Address reprint requests to Dr. S. Fedoroff, Department of Anatomy, University of Saskatchewan, Saskatoon. Saskatchewan S7N OWO, Canada.

0 1990 Wiley-Liss, Inc.

2 Federoff et al.

mice immunized with a cytoskeletal extract of an in vitro aged culture of human fibroblasts derived from a 66- year-old human donor (Wang, 1985). The antibodies to statin stain positively the nuclei of nonproliferating fibroblasts in cultures and differentiated cells in vivo in tissues such as the prickle cell layer of skin epidermis, esophageal epithelium, smooth muscle cells, chondro- cytes, and hepatocytes (Wang and Krueger, 1985). This paper describes an investigation of astroglia both in cultures and in situ, using antistatin antibodies to detect whether “mature” stellate astrocytes are noncycling cells.

MATERIALS AND METHODS Colony Cultures

The cerebral hemispheres of newborn C,H/HeJ mice were isolated aseptically and the meninges re- moved. The neopallia were dissected out and then gently forced through a sterile 75-pm Nitex mesh. The cclls were suspended in a modified Eagle’s minimum essential medium (MEM) containing 5% horse serum. For culturing, 5 X lo4 nigrosine-excluding cells were plated on 11 X 22-mm glass coverslips in 6O-mm Falcon culture Petri

dishes in a total volume of 3 ml growth medium and incubated at 37°C in a humidified atmospherc of 5% CO, in air. The cultures were then incubated for periods up to 60 days with the medium changed every 2-3 days until fixation.

Monolayer Cell Cultures Monolayer cultures were prepared in the same way

as colony cultures, except that 3 x los nigrosine-cx- cluding cells were plated per 60-mm Falcon culture Petri dish.

Reactive Astrocytes In Vitro Colony cultures werc prepared as described above.

After 12 days of incubation, the cultures were incubated up to another 50 days in medium containing 0.25 mM dibutyryl cyclic AMP (dBcAMPj and I % horse serum. The medium was changed every 2 days until fixation.

Reactive Astrocytes In situ C,H/HcJ mice, approximately 3 months old, were

anesthetized with ether. A puncture hole was made in the right hemisphere of the brain using a 25-gauge aseptic needle. The animals ”ere sacrificed 4, 7, 14, 21, 28. 35, 40, 50 and 60 days after the stab wound. and cryostat sections were made from the wounded rcgion of the cc- rebrdl hemispheres. For control, similar sections from normal brains were cut.

Cryostat Sections Whole brains of the C,H/HeJ mice were frozen

immediately after sacrifice in 2-methyl butanc and cooled with liquid nitrogen. Coronal sections of the brain, approximately 6 pm thick, were cut on a cryostat, placed on glass coverslips, and air dried. In case of stab- wounded animals, the sections were prepared from the wounded region. For controls, similar sections from brain tissue of normal animals were made.

Brain Tissue Smears Smears of fresh unfixed brain tissue from C,H/HeJ

mice were prepared according to the method of Olson and Ungerstedt (1970) and Bjorklund et al. (1984). In case of stab-wounded mice, tissue from the wound region was used. For controls, tissue from the correspond- ing area of normal brains was used.

Antisera Monoclonal antibodies to statin (S-30, and S-44

and S-41.1 j and cyclin-like protein (S- 132) were prepared as described by Wang (1985), Wang and Krueger (1985), and Connolly et al. (1988). S-30 in supernatant form was used in 1 : 100 dilution, S-44 in ascites form in 1 : 5,000 dilution, and 1 : 10,000, and 5-44.1 in ascites form in 1 : 25,000, while S-132 in supernatant form was used in 1 : 5 and in 1 : 10 dilution. All antibody dilutions were made in phosphate-buffered saline (PBS). Mouse monoclonal antibody to glial filament acidic protein (GFAP) and vimentin were purchased from Boehringer Mannheim GmbH (FRG). Rabbit polyclonal antibody to GFAP was prepared as described by Wang et al. (1984). Antibody to 5-bromodcoxyuridine (BudR) was purchased from AMAC, Inc. (Westbrook, ME, USA).

Secondary antibody, goat antimouse immunoglob- ulin (Ig) containing heavy and light chains, was from Jackson Immunoresearch Laboratories Inc. (Avondale, PA, USA). Fluorescent isothiocyanate (FITCj swine antigoat IgG was from Caltag Laboratories (San Francisco, CA, USA). FTTC goat antirabbit IgG and rhodamine goat antimouse IgG were from Miles-Yeda Ltd. (Reho- vat, Israel). Bovine TgG was purchased lrom Cappel Worthington (Malvern, PA, USA).

Immunocytochemistry: Statin or Cyclin and GFAP Double Staining

Cells in culture, brain sections, and smears were treated in the same way, except that cells in culture were rinsed briefly with Dulbecco’s PBS prior to fixation. The preparations were fixed for 10 min in a 1 : 1 mixture of methanol and acetone at -20°C and air-dried for about 30 min. They were then rehydrated and extracted with Triton X-100 for 10 min, washed, and incubated with

Statin in Astroglia in Cultures and In Situ 3

rescein. The statin-positive (statin ' ) or cyclin-positive (cyclin+) cells were counted and cells clascified as described above. A similar procedure was used for deter- mining the frequency of BudR positive cells among the GFAP ' and GFAP- cells.

statin or cyclin antibody for 16 hr at room temperature in a humidified chamber. The slides were washed, incubated with goat antimouse IgG (H + L) antibody diluted 1 : 100 for 30 min, washed, and incubated with FITC swine antigoat antibody diluted I : 100, then briefly washed and incubated for 30 min with normal goat serum (1 : 20). After blotting off the goat serum the slides were incubated for I hr at room temperature with mouse monoclonal antibody to GFAP diluted 1 : 25. washed, incubated for 30 min with rhodamine goat antimouse antibody diluted 1 : 100, and then washed. The preparations were then stained with the nuclear dye, Hoechst 33258, for 5 min, washed, and mounted in 50% glycerol in PBS (pH 7.8). For controls, primary antisera were omitted.

BudR and GFAP Double Staining For BudR and GFAP double staining, the method

described by Yong and Kim 1987) was used. Cells on coverslips were incubated in culture medium containing 10 pM BudR (Boehringer Mannheim GmbH, FRG) for 24 hr at 37°C in S% CO,. Cells were then washed thor- oughly with PBS containing 10 mM HEPES buffer and 2% horse serum, fixed and immunostained for GFAP as described previously, except that polyclonal GFAP antibody was used. Then the native DNA was denatured with 1 M HCI for 10 min followed by a wash and 0.1 M sodium borate buffer at pH 9 for 10 min. After rinsing, mouse monoclonal antibody to BudR (1 : 25 or 1 : 50 dilution) was applied for 30 min followed by three washes and a 30-min incubation in rhodatnine goat antimouse antibody. The coverslips were then washed again and stained with Hoechst dye 33258 as described above. For control, antibody to BrdU was omitted.

Immunofluorescence Microscopy and Cell Counts The preparations were examined in a Zciss Photo-

microscope I1 equipped with a mercury vapor lamp, epi- fluorescence optics, and interference filters. The speci- mens were observed in randomly selected fields with an exciter-barrier filter combination for Hoechst dye 33258. and cell counts were done. Then the same field stained with anti-GFAP antibody was observed with an exciter-barrier filter combination for rhodaminc, and all the GFAP-positive (GFAP+) and GFAP-negative (GFAP-) cells were counted. The GFAP' cells were classified into four groups: small, 10-30 pm (SC); medium, 35-60 pm (MC); large, 60-120 Frn (LC); and small stellate astrocytes (SS-4). GFAP- cells were classified into three groups: small, 10-30 pm (SC); medium, 35-60 km (MC); and large, 60-300 pm (LC). (No GFAP- SSA were observed.) Then, the same field was observed, stained with antistatin or anticyclin antibody with an exciter-barrier filter combination for fluo-

RESULTS Stellate Astroglia in Cultures and I n Situ

Astroglia from newborn mouse (C,H/HeJ) neopallium were grown in colony or rnonolayer cultures for 8-60 days. A few rat astroglia cultures were grown for 90 days. At various intervals, cultures were fixed and double-stained with antibodies to GPAP and to statin (S-30 or S-44) or cyclin-like protein (S- 132). Cell nuclei were stained with Hoechst dye 33258. In all cultures, colony and monolayer, regardless of the duration in culture, the nuclei of GFAP+ stellate astrocytes wcre always negative for statin (S-30 or S-44) (Figs. 1,2,5,6) but positive for cyclin-like protein (S- 132) (Figs. 3,4). The same cultures, however, contained flat cells of various sizes that were either GFAP + or CFAP-; inany of these cells had statin+ nuclei (Figs. 2,6).

To determine whether the nuclei of GFAP+ astrocytes in situ are also negative for statin, cryostat sections of the cerebral hemispheres (Figs. 7,s) and cerebellums (not shown) were double-stained with antibodies to GFAP and statin, and in some cases, antibodies to GFAP and cyclin-like protein. In cryostat sections in which many GFAP+ astroglia processes could be visualized, no statin+ nuclei related to GFAP+ cell processes were seen (Figs. 7,8). The statin+ nuclei observed did not relate to the GFAP' cell processes and probably be- longed to neurons.

icult to be certain which nuclei actually belong to the GFAP+ cell processes; therefore, smears from adult mouse cerebral hemispheres were made and double-stained with antibodies to GFAP and statin. In such preparations it was easy to identify GFAP+ stellate astrocytes. These cells always had statin- nuclei. When statin+ nuclei were seen, they did not belong to GFAP+ cells (Figs. 9,lO).

Reactive Astrocyte-like Cells in Cultures and Reactive Astrocytes In Situ

Treatment of 10-day-old astroglia cultures with 0.25 mM dBcAMP for 7 days transforms the astroglia into large stellate astrocytes (LSA), which are usually GFAP' . These cells in many respects resemble reactive astrocytes more than any other cells in situ (Fedoroff et al., 1984b; Abd-El-Basset et al., 1989). When reactive astrocyte-like cells (LSA) in cultures were double- stained with antibodies to GFAP and statin or antibodies to GFAP and cyclin-like protein, it was observed that,

Using cryostat sections, it is d

Figs. 1,2. Double immunofluorescence staining of cells from newborn mouse neopallium in a 21-day culture by antibodies to GFAP and statin (S-30), respectively. X 635.

Fig. 1 , The GFAP-containing intermediate filaments can be seen in the stellate astrocyte. A number of statin+ cell nuclei can be seen in the cells that are GFAP-.

Fig. 2. The nuclcus of the GFAP+ stellate astrocytc is statin- (arrow).

Figs. 3,4. Double immunofluorescence staining of cells from newborn mouse neopallium in a 30-day culture by antibodies to GFAP and cyclin (S-132), respectively. X 635.

Fig. 3. GFAP-containing intermediate filaments can be seen in the stellate astrocyte.

Fig. 4. The nucleus of the W A P + astrocyte is cyclin' (arrow). A number of cyclin+ nuclei of GFAP- cells can be seen.


Figs. 5 and 6. Double immunofluorescence staining of cells from newborn mouse neopallium in a 60-day culture of antibodies to GFAP and statin (S44), respectively. X635.

Fig. 5. A number of GFAP' cells can be seen. Fig. 6. The nucleus of the GFAP ' stellate astrocyte is statin- (arrow). whereas the nucleus

of the flat GFAP' cell is statin

for a few days after the initiation of treatment with dB- CAMP, the GFAP+ LSA were statin- (Figs. 11,12) and positive for cyclin-like protein (Figs. 13,14). After 8-15 days of treatment with 0.25 mM dBcAMP, a large ma- jority of the nuclei of GFAP+ LSA became statin+ (Figs. 15,16) but negative for cyclin-like protein (Figs. 17,18).

Cryostat sections of cerebral cortex made at various intervals after a stab wound in a hemisphere were double-stained with antibodies to GFAP and statin. We observed many strongly GFAP-posilivc cells with extensive cell processes, but they were always negative for statin (Figs. 19,20). The same observation was made on smears of wounded hemispheres. The GFAPf , stellate, reactive astrocytes were always statin- (Figs. 21,22).

Cells in Colony Cultures Colony cultures are initiated by a small inoculuni

of cells. In such cultures, cells attach to the substratum and the proliferating cells form a progeny of cells that form colonies with a definite arrangement of cells. In the center of the colony are situated relatively small-sized, pleomorphic cells, closely apposed to each other and

sometimes having an epithelial-cell-like appearance. To- ward the periphery of the colony, the cells increase in size. The colonies are usually surrounded by very large, flat cells that increase in number and size with time in culture. Cells in the intermediate zone, i.e., between the small cells in the center and large cells in thc periphery of thc colony, arc mcdium-sized, pleomorphic cells (Fig. 23). GFAP ' , small, stellate astrocytes (SSA) usually appear in older cultures, and are on top of the SC and MC .

cells, among GFAP+ and CFAP- cells in colony culturcs grown for various periods of time, is shown in Table 1. Thc GFAP+ and GFAP- SC in the center of the colony and the GFAP+ SSA did not become statin ' during culturing periods of as long as 60 days. A large propor- tion of MC, GFAP' cells, as well as GFAP- cclls, became statin+ with time in culture and all the LC, both GFAP' and GFAP-. were statin ' .

The frequency of occurrence of statin '

Cell Proliferation in Colony Cultures Table 11 summarizes the data on the degree of pro-

liferation (DNA synthesis) of the various cell types in

6 Federoff et al.

Figs. 7,8. Frozen section of 3-month-old mouse cerebral hemisphere double immunofluorescence stained with antibodies to GFAP and statin (S44), respectively. X 635.

Fig. 7. A number of GFAP+ cells can be seen. Fig. 8. The same cells seen in Figure 1 arc statin-.

Figs. 9,lO. Double immunofluorescence staining of 3-month- old mouse cerebral hemisphere smear by antibodies to GFAP and statin (S44.1). respectively. X 635.

Fig. 9. CFAP+ stellate astrocytes can be seen. Fig. 10. The same cells seen in Figure 9 have statin- nuclei

(arrows).

Figs. 11,12. Double immunofluorescence staining of large stellate astrocytes (LSA: reactive astrocyte-like cells) after I8 days culture and for the last 10 days trcatcd with 0.25 mM dBcAMP, by antibodies to GFAP and statin (S44. I ) , respectively. X 635.

Fig. 11. The GFAP-containing intermediate filaments in the LSA can be seen.

Fig. 12. The same cells seen in kigure 11 have statin- nuclei.

Figs. 13,14. Double immunofluorescence staining of large stellate astrocyte (LSA: reactive astrocyte-like cells) after 22 days culture and for the last 14 days treated with 0.25 mM dBcAMP, by antibodies to GFAP and cyclin (S-132), respectively. X 635.

Fig. 13. The GFAP-containing intermediate filaments in the LSA can be seen.

Fig. 14. The same cell seen in Figure 13 has a cyclin+ nucleus.

Figs. 15.16. Doublc immunofluorescence staining of large stellate astrocyte (LSA: rcactive astrocytc-like ccll) after 43 days culture and for the last 35 days treated with 0.25 niM dBcAMP, by antibodies to GFAP and statin (S-44), respectively. X 635.

Fig. IS. The GFAP-containing intermediate filaments in the LSA can be seen.

Fig. 16. The nucleus of the same cell is statin+.

Figs. 17.18. Double immunofluorescence staining of large stellate astrocytes (LSA: rcactivc astrocytc-like cells) alter 28 days culture and for the last 20 days treated with 0.25 mM dBcAMP, by antibodies to GFAP and cyclin 6-1321, respectively. x 635.

Fig. 17. The GFAP-containing intermediate filamcnts in the LSA can be seen.

Fig. 18. The nuclei of the same cells are cyclin (arrows).

Figs. 19,20. Frozen section through a stab wound i n the ce- Figs. 21,22. Double imrnunofluorescence staining of a smear rebral hemisphere of a 3-month-old mousc, doublc immuno- of the cerebral hemisphere of a 3-month-old mouse with a fluorescence stained with antibodies to GFAP and statin 14-day-old stab wound, by antibodies to GFAP and statin, (S44. I ) , respectively. X 635. respectively. X 635.

seen. Fig. 19. A number of GFAP+ reactive astrocytes can be

Fig. 20. The same cells have statin- nuclei.

Fig. 21. (ZAPf reactive astrocytes can be seen. Fig. 22. The same cells have statin- nuclei (arrows).

Figs. 23-26.


TABLE I. Frequency of Statin+ Cells among GFAP+ and GFAP- Cells in Colonv Cultures Grown for Various Periods of Time

No. of statin+ cells/GFAP+ cells No. of statin ' cellsiGFAP cells (%) (%I Total

Days in no. cells culture counted LC MC SC SSA LC MC SSA

8 352 14/26 (54.0) 121127 (9.4) 0139 (0.0) 0:s (0.0) 13114 (93.0) 9/132 (7.0) 019 (0.0) 14 302 38/42 (90.4) 341100 (34.0) 0/18 (0.0) 0132 (0.0) 31/31 (100.0) 27/65 (41.5) 0114 (0.0)

21 264 45150 (90.0) 40184 (48.0) 2/18 (11.0) 0134 (0.0) 44/44 (100.0) 10132 (31.2) 0/2 (0.0)

30 280 52/53 (98.1) 48/97 (49.4) 014 (0.0) 0/45 (0.0) 41/41 (100.0) 28/40 (70.0) -

60 347 92/93 (99.0) 46/64 (72.0) 019 (0.0) 0/43 (0.0) 82/82 (100.0) 35/56 (59.0) ~

LC, large cells: MC, medium-sized cclls; SC, small cells; SSA, small stellate astrocytes

TABLE 11. Frequency of BudR+ Cells among GFAY+ and GFAP- Cells in Colony Cultures Grown for Various Periods of Time

No of BudR- cells/GFAP ' cells (%) No o f B u d R +

cellsiGFAP- cells (%I Days in Total no culture cell? LC MC sc SSA LC MC

8 346 3/21 (14.3) 631 19 (53.0) 26/33 (79.0) 4/8 (50.0) 6/17 (3.5.3) 681148 (46.0)

14 365 1/47 (2. I ) 26/106 (24.5) 12/29 (4 1.4) 22/44 (50.0) 9/44 (20.4) 25/95 (26.3)

21 349 2/49 (4. I) 1311 19 ( 1 1.0) 9/23 (39.1) 3146 (6.5) 3/36 (8.3) 9176 (12.0)

30 355 3137 (8.1) 81154 (5.2) 7/16 (44.0) 0140 (0.0) 3/47 (6.4) 5/61 (8.1)

LC. large cells; MC, medium-sizcd cells; SC. small cells; SSA, sinall stellate astrocytes.

colony cultures. In young cultures (8 days) SSA, SC. and MC had high frequencies of BudR-positive cells. With time in culture, however, the SSA stopped synthesizing DNA, and the number of MC synthesizing DNA decreased drastically (from 53% to 5.2%). The number of SC synthesizing DNA decreased with time but, even in the 30-day-old cultures, 44% of the cells were actively synthesizing DNA. There was no significant difference between the GFAP ' and GFAP- cells as far as the per-

Fig. 23. Phasc-contrast photograph of a colony culturc of cells from newborn mouze neopallium showing topographical arrangement of cells. The small cells (SC) are situated at the center of the colony (right lower corner). Next to the SC, toward the periphery of the colony, are the medium-sized cells (MC) and at the periphery o f the colony (left side) are the large cells (LC). (kor more details, see text.) X 240.

Figs. 24-26. Immunofluorescence staining of a colony of cells from newborn mouse neopallium alter 28 days in culture with antibody to statin (544.1 j .

Fig. 24. Only cells in thc periphery of the colony are statin-; cclls in the center of the colony are statin . x 98.

Fig. 25. Only cells in the periphery of the colony are statin+. X 319.

Fig. 26. Thc cells in the center of the colony are statin ,

x319.

centagc of cells synthesizing DNA at various intervals of culturing was concerned.

Effect of Serum Concentration on Statin Expression It has been previously reported that less than 2% of

human fibroblasts of the 001 1 cell line, when grown in medium containing 10% serum, express statin in their nuclei. However, when the serum concentration is reduced to 0.5%, cells stop proliferating and more than 80% of the cells become positive for statin. (Wang and Liu, 1986).

We attempted to find out whether thc concentration of serum in the medium has an effect on the expression of statin in mouse astroglia, and if so, whether all or only some types of astroglia in colony cultures respond to serum concentration. To do this, we grew colony cultures in medium containing 5% serum, then in some cultures, reduced the concentration of serum to 1 % for 3 days. In other cultures, the concentration of serum was changed to 1% for 3 days, then back to 5 % for 3 days. The results are summarized in Table 111.

These experiments indicate that SC and MC, whether GFAPf or GFAP- , respond to a low concentration of serum by expressing statin, as do human fibroblasts. The expression of statin can be reversed by re- storing the original serum concentration. Regardless of changes in serum concentration, however, most LC remain slatin + and interestingly, SSC remain statin-

12 Federoff et al.

TABLE I l l . Effect of Concentration of Serum on Frequency of Statin+ Cells among GFAP+ and GFAP- Cells in Colony Cultures

No. of statin+ cellsiGFAP' No. of stain ' cellsiCFAP- cells (%) cells (8) Concn. of

serum in medium Total no. cells % (days) counted LC MC SC SSA LC MC-SC

5%'o(IOdt.5%(3d) 534 26/30 (86.6) 421215 (19.5) 0/102 (0) 0i29 (0) 31/31 (100) 2V!l27 (22.8) 5%(10d+10/(3d) 533 38 i38 (100) 1401181 (77.3) 28/59 (47.5) 0.20 (0) 24/24 (100) 1501211 (71.1)

5 7 4 1 0 d b 1 %(3d+S8(3d) 488 29129 (100) 501159 (31.4) 281129 (20.1) 0121 (0) 36136 (100) 321104 (30.8) 54."c(IOd)~54."c(3d)~5~(3d) 5 I4 32132 (100) 441172 (25.6) 11112.5 (8.8) 0124 (0) 30130 100) 291131 (22.1)

LC, large cells; MC, medium-sired cells; SC, small cells; SSA. small stellate astrocytes.

DISCUSSION Differentiation of Astroglia in Colony Cultures

The advantage of colony cultures is that development of the cells can be easily observed (Fedoroff, 1987). The colonies develop from a single cell or a few proliferative cells that attach to the plastic substratum. (In our hands, approximately 25% of colonies begin from a single cell.) In the center of the colony are the most immature cells, the small cells (SC), which proliferate and give rise to larger cells, the mcdium-sized cells (MC). SC and MC seem to have at least three developmental options: (1 ) to differentiate into small stellate astrocytes (SSA); (2) to enlarge in size and form large cells (LC) (Figs. 23,34); and (3) under the influence of prolonged treatment with dBcAMP, to develop into LSA (Fig. 27). All three types of cells assume specific rela- tionships to the cell colony of origin. The small and large stellate astrocytes (SSA and LSA) take a position on top of thc SC and MC, and the LC are always found on the extreme periphery of the colony (Fig. 27).

Small cells. In the center of the cell colony, the SC are either GFAP+ or GFAP-, but both kinds are vimentin+ , statin-, cyclin' , and proliferative and are responsive to the nutritionalitrophic state of the microen- vironment. We have demonstrated that in medium containing a low concentration of serum (l%), a large number of SC express statin in their nuclei. The phenomenon can bc interpreted to mean that SC in the presence of low serum concentrations stop dividing and become arrested in the G,/G,, phase of the cell cycle. Thc arrest can be reversed by increasing the concentration of serum. Thcn the cells probably re-enter the G, phase of the cycle and at the same time downgrade statin and bcgin to synthesize cyclin, which is required for cell division (Murray et a]., 1989).

Proastroblasts, cells that occur early in the astroglia lineage, are GFAP- and vimentin+ ; they differentiate into astroblasts, which are GFAP+ and vimentin+ (Fed- oroff, 1987). It seems reasonable to assume that the GFAP-, vimentin+ SC in the centers of the colonies are proastroblasts and that the GFAP+, vimcntin+ SC are astroblasts that have developed from the CFAP-, vimentin + SC.

Medium-sized cells. The MC can be either GFAP+ or GFAP-, but all are vimentin+. Some are statin+ and some are statin-. In young cultures, only a few MC are statin+ (see Table I); they are proliferative and can be arrested in the G,/G, phase of thc ccll cycle if serum concentration in the medium is lowered and can be relcascd from the arrest by increasing the serum concentration. However. with increased time in culture, even in medium with high serum concentration, most MC become statin+ (see Table I). This indicates that, with time, MC arc leaving the cell cycle permanently, a sign of cell senescence.

Large cells. The LC are next to the MC and form the outer cells of a colony. Probably as the MC differentiate and move farther away from the ccntcr of the colony, they are no longer apposed from all sides by other cells; if they do not receive appropriate differentiation signals. they undergo senescence. They stop dividing, probably by becoming arrested in late G I phasc (Rittling ct al., 1986), then begin to express statin and to enlarge greatly. The basis for suggesting that LC are indeed senescent cells forming under culture conditions, is that the increase in their number is directly related to the length of time in culture but not to the age of the embryo from which the cultures were initiated. Similar cells have been observed in "aging" cultures of various non-neuronal cell types and have been considered scnes- cent cells (Bowman and David, 1975; Mitsui and Schnei- der, 1976; Christofalo, 1970).

Small stellate astrocytes. The SSA always appear late (after 8-10 days in culture), are GFAP+, vimentin+, and statin- and stop DNA synthesis as they mature (Tables I and TI). Morphologically, SSA resemble fibrous astrocytes in situ (Fedoroff et al., 1984a) or Raff's type I1 astrocytes (Raff et al., 1983). In the cell colony, they are usually located on top of SC and MC. We believe that SSA evolve from cells of the SC-MC cell pool, which are proliferative and are GFAP+, vimentin ' , and statin-.

Large stellate astrocytes. We found that LSA in cultures that resemble reactive astrocytes in situ were statin+ but the rcactive astrocytes were statin-. The rea- son for the discrepancy could be that culturcs treated

GFAP'

V IM+

STATIN+

GFAP+ MC T i - - - *

SSA'

Fig.

LSA GFAP+ V IM+

STATI N: dBcAMP \ t I 1

sc sc G FAP + G FAP - VIM+ VIM' STAT IN - STATIN-

STATIN-

Diagrammatic representation of developmental rela- semble senescent cells and are GFAP+, statin+; and, in ie tionships of cells in colony cultures initiated rrom neopallium presence of 0.25 mM dBcAMP, to rorm large stellate astro- of newborn C3H/HeJ mice. The small cells (SC) form me- cytes (LSA), reactive astrocytc-like cells, which are GFAP+ dium-sized cells (MC). GFAP' and GFAP- SC are statin-, and which at the beginning of development are statin- but on whereas some MC cells are statin ' and some statin-. Cells in culturing become statin+. The GFAP~ SC develop into MC the GFAP', statin- SC-MC cell pool have three options: to and LC (senescent) cells which are GFAP- and statin+. It is develop into small stellate astrocytes (SSA), which are possible that GFAP- SC give rise 10 GFAP+ SC. For more GFAP', statin ; to develop into large cells (LC), which re- details, see text.

14 FederotT et al.

with dBcAMP were incubated in medium containing only 1% serum. The low concentration of serum was used because higher concentrations (5% or 10%) inhibit the property of dBcAMP to promote the differentiation of astroglia into LSA. In a separate experiment, we demonstrated that astroglia (SC and MC) grown in medium containing 1 % serum enter the G,/G, phase of the cycle and become statin+. It will be of interest to find out whether the dBcAMP promotes the differentiation of astroglia into LSA only when they are in G,/G,, phase of the cycle.

dBcAMP is also a mitotic inhibitor (Goldman and Fung-Chow, 1984; Senenbrenner et al., 1980). There- fore. if LSA are still in the cell cycle, perhaps prolonged treatment with dBcAMP will inhibit the progression of LSA through the cycle and may lead to the expression of statin.

Astroglial Differentiation and the Cell Cycle The main significance of the observations reported

in this paper is that astrocytes in adult mice and their counterparts, the SSA, in cultures, are statin-, but at the same time, only a small fraction of astrocytes in situ divide (Korr, 1986; McCarthy and Leblond, 1988); with time, SSA in culture stop dividing (Table 11). This find- ing suggests either that astrocytes behave differently from other kinds of cells in the body in respect to the relationship of cell division to the expression of the statin-cyclin proteins in cell nuclei, or that astrocytes are, indeed, in the cell cycle but progressing at a very slow rate.

The LSA and LC in cultures express statin, and SC and MC express statin in medium having a low concentration o f serum but downgrade it when transferred into medium with high serum concentration, just as human fibroblasts do. So far, thcrefore, there is no evidence that astrocytes (astroglia) behave differently than other cell typcs in the body in respect to the relationship between cell division and expression of statin and cyclin nuclear proteins. By contrast. it has been pointed out that [3H]- TdR or BudR pulse-labeling methods for identification of proliferative cells and for estimation of the duration of the cell cycle are quite inadequate to distinguish cells that are cycling slowly from senescent cells arrested in late G, phasc or from quiescent cells in GI/G,, phase of the cell cycle because the cells are not exposed long enough to [3H1-TdR or BudR (Skehan, 1988).

We observed that in young cultures a large percent- age of SSA incorporatc BudR, but with time in culture the incorporation of BudR ceases (Table 11). This can bc explained in two ways. First, we can assume that SSA originally have a short cell cycle, and with time the duration of traversing the cycle increases considerably so that a 25-hr pulse with BudR is an insufricient length of

time to detect its incorporation into the slow cycling cells. The second explanation is to assume that in young cultures BudR is incorporated into the last cell cycle of proliferating precursor cells, before they become nondividing stellate cells.

After extensive studies of thc proliferation kinetics of glia cells, Korr (1966) proposed that astroglia in the adult brain are in two compartments. One is a proliferative compartment (growth fraction) in which astroglia progress through a cell cycle calculated to be 18.8 2 0.4 hr in situ and 23.6 k 1.5 hr in cultures. The other compartment is composed of nonproliferating astrocytes that can entcr the proliferative compartment and complete the cell cycle, at the end of which some daughter cells may die and others re-enter the nonproliferative compartment; i.e.. maturc astrocytes can enter and exit the cell cycle. According to Korr's model, the astroglia in the nonproliferative compartment should be statin+ and in the proliferative compartment, statin-. Wc observed, however, that all mature astrocytes in situ and in cultures were statin-.

Recently, McCarthy and Leblond (1988) rcportcd observations on adult mice infused with ["HI-TdR con- tinuously for 30 days. These workers found that a small number of astrocytes in the mouse brain tissue did divide; they calculated that in aging mouse brain the turnover of astrocytes is 0.4% per day. Korr (1966) also observed cell division in astrocytes in adult rodents brains. He estimated that 0.4% of astrocytes are present in the proliferative cell compartment. Miyake et al. (1988) observed that after 6 days of cumulative labeling with 13H]-TdR, after stabbing of the cerebral hemispheres of 3-month-old rats, only 17% of GFAP ' reactive astrocytes were labeled.

CONCLUSIONS From the foregoing, it would appear that most as-

trocytes do not divide. However, we find all astrocytes statin-. indicating that they are not in a quiescent or senescent state. A possible cxplanation for the various observations is that as astrocytes differentiate they cycle more and more slowly; thus, the number of cells in the S phase at any given time decreases. The notion that astrocytes are in an extended cell cycle would explain the existence of the small cell growth fraction detectable by tritium labeling techniques, as well as the lack of statin in all astrocytes. The notion does not necessarily include the assumption that astrocytes traverse the cell cycle at a constant rate. It seems possible that, at certain restriction points, the rate may slow or the cells stop cycling tem- porarily. The main point is that astrocytes are not quiescent cells but are indeed in the cell cycle. This may have significant implications for the understanding of astro-


lmamoto K, I’aterson JD, Leblond CP (1978): Radioautographic investigation of gliogenesis in the corpus callosuni of young rats. I . Sequential changcs in oligodendrocytes. J Comp Neurobiol 180: I 1 5-1 37.

Korr H ( I 986): Proliferation and cell cycle parameters ofastrocytcs. In Fedoroff S. Vcrnadakis A (cds): “Astrocytcs.” Vol. 111. Or- lando. FL: Academic Press, pp 77-1 19.

McCarthy GF, Leblond CP ( 1988): Radiographic evidence for slow astrocyte turnover and modest oligodcndrocyte production in the corpus callosuni of adult mice infused with ’H-thymidine. J Comp Neurol 271:589-603.

Mitsui Y, Schneider EL (1976): Increased nuclear sire in senescent human diploid fibroblast cultures. Exp Cell Res. 100: 147-152.

Miyake T. Hattori T, Fukuda M, Kitamura T, Fuyita S (1988): Quan- titative studies on proliferative changes of reactive astrocytes in mouse cerebal cortex. Brain Res 451:133~-138.

Munay AW, Solomon MJ, Kirschncr MW (1989): The role of cyclin synthesis and degradation in thc control of maturation promot- ing factor activity. Nature (Lond) 339:280-286.

Olson L, Ungerstcdt U (1970): Monoamine fluorescence in CNS smears: Sensitive and rapid visualization of nerve tcrminals without freeye-drying. Brain Rcs 17:343-347.

Raff MC. Abney ER, Cohen J , Lindsay R , Noble bl (1983): Two types of astrocytes in cultures of developing rat white matter: Differences in morphology. surface gangliositles, and growth characteristics. J Neurosci 3: 1289-1300.

Rittling SK, Brooks KM. Cristofalo VJ, Baserga R (1986): Expression of cell cyclc-dependent genes in young and senescent W1-38 fibroblasts. Proc Natl Acad Sci (USA 83:3316-3120.

Sensenbrenner M, Devilliern G. Bock E, Porte A ( I 980): Biochemical and ultrastructural studies of cultured rat astroglial cells. Effect of brain extract and dibutyryl cyclic AMP on glial fibrillary acidic protein and glial filaments. Differentiation 1 7 5 1-61,

Skehan P (1988): Control models of cell cycle transit, exit and arrest. Biochein Cell Biol 66:467- 477.

Takasaki J , Fischauld D, Tan EM (1984): Characterization of proliferating cell nuclear antigen recognized by autoantibodics in lupus sera. J Exp Mcd 159:981-992.

Wane E (1985): A 57.000-mol-wt protein uniquely present in nonpro- liferatinp cells and senescent human fibroblasts. J Cell Biol 100545 -55 1

Wang E, Kruegcr JG (1985): Application of a unique monoclonal antibody as a marker for nonproliferating subpopulations of cells of some tiqsues. J Histochzni Cytochern 33:587-595.

Wang E. Liu SL (1986): Disappearance of statin, a protein marker for nonproliferating and senescent cells, following serum-stimu- lated cell cycle entry. Exp Cell Res 167:135-143.

Wang E. Cairncross JG, Liem RKH (1984): Identification of glial filament protein and vimentin (decamin) in the same intermediate filament system in clonal lines of human glioma cells. Proc Natl Acad Sci LISA 81:2101-2106.

Young WW. Kim SU (1987): A new double labclling immunofluorescence technique for the determination of proliferation of human astrocytes in culture. J Neurol Methods 21 :9-16.

cyte reactions in brain injuries. because control of cell numbers would relate to control of the rate at which astrocytes traverse the cycle. More study of the relationship between astrocyte differentiation and the cell cycle is indicated.

ACKNOWLEDGMENTS

This work was supported by grant MT423.5 from Medical Research Council of Canada (S.F.) and by grant MA-10207 from thc MRC of Canada and grant (AG0744) from the National Institute of Aging (E.W.). We would like to thank Osman Kademoglu for the pho- tography, Shirley West for typing of thc manuscript and Dr. A. Richardson for artwork.

REFERENCES Abd-El-Basset EM, Kalnins V1, Ahmed I , Fedoroff. S (1989): A 48

kilodalton intermediate filament associated protein (IFAI’) in reactive-like astrocytes induced by dibutyryl cyclic AMP in culture and in reactive astrocytes in situ. J Neuropathol Exp Neurol 48:245-254.

Bjijrklund H, Eriksdotter-Nilsson M, Dahl D, Olson L (1984): Astro- cytes ir i smears of CNS tissues as visuali/ed by GFA and vimentin immunofluorescence. Med Biol 62:38-48.

Bowman PD, Daniel CU (1975): Characteristics of proliferative cells from young, old and transformed W1-38 cultures. In Cristofalo VJ, Holeckova E. (eds): “Cell Impairment in Aging and De- velopment.” New York: Plenum Press, pp 107-122.

Connolly JA, Sanabia VE, Kelvin DJ, Wang E (1988): The disappearance of cyclin-like protein and the appearance of statin arc correlated with the onset of differentialion during myogenesis in virro. Exp Cell Res 174:461-471.

Cristofalo VJ (1970): Metabolic aspects of aging in diploid human cells. In Holeckovi E, Cristofalo V J (eds): “hging in Cell and Tissue Culture.” New York: Plenum Press, pp 83--1 19.

Fedoroff S (1987) From neuroepithclium to mature astrocytes. In Al- thaus HH, Seifert W (eds): “Glial-Neuronal Communication in Development and Regeneration. ” Berlin: Springer-Verlag, pp. 5-25.

Fedoroff S, Neal .I, Opas M , Kalnins V I (1984a): Astrocyte cell lin- eagc. 111. ‘Ihe morphology of differentiating mouse astrocytes in coloiiy culture. J Neurocytol 13: 1-20.

Fedoroff S , McAuley WAJ. Houle JD, Devon RM (1984b): Astrocyte cell lineage. V Similarity of astrocytes that form in the presence of dRcAMP in cultures to reactive astrocytes in vivo. J Neu- rosci Res 12:14-27.

Goldman JE. Fung-Chow C (1984): Dihutyryl cyclic AMP causec intermediate filament accuniulation and actin reorganization in astrocytes. Brain Res 3063-95.

The relationship of expression of statin, the nuclear protein of nonproliferating cells, to the...

Documents

Transcript of The relationship of expression of statin, the nuclear protein of nonproliferating cells, to the...