Journal of Neuroscience Research 28:601406 ( 1991)
Rapid Communication (x Isoform of Smooth Muscle Actin Is Expressed in Astrocytes In Vitro and In Vivo E. Lecain, F. Alliot, M.C. Laine, B. Calas, and B. Pessac Centre De Biologie Cellulaire, UPR 3 10 I , C.N.R.S., Ivry Sur Seine CEDEX (E.L., F.A., M.C.L., B.P.); Centre De Recherche De Biochimie Macromoleculaire, UPR 8402. C.N.R.S.. Montpellier (B.C.), France We have previously reported that astroglial cell lines derived from spontaneously immortalized mouse cerebellar cultures as well as primary astrocyte cultures express the mRNA of the a isoform of smooth muscle actin. In this report, we have used an antiserum specific for the a smooth muscle actin protein to investigate the presence and the pattern of expression of a smooth muscle actin protein at the cellular level with immunocytochemical methods. The results show that an anti-smooth muscle vessels a actin antiserum labels a typical actin network in the D19 astroglial cell clone and in flat astrocytes of primary cultures derived from various CNS regions of embryonic and postnatal mice. Furthermore, this antiserum labels distinct populations of astrocytes in the adult mouse brain, in particular in the corpus callosum and the fornix. However, in the corpus callosum, astrocytic processes are strongly labeled by anti-SMV a actin antibodies only in parasagittal planes. Thus, a smooth muscle actin represents a new marker for subsets of astrocytes. Key words: mouse, corpus callosum, central nervous system INTRODUCTION Actin is a ubiquitous cytoskeletal protein. In mammals, actin is a multigene family as there are at least six different isoforms of actin, each isoform being coded by a distinct gene. The different isoforms of actin include two striated muscles isoforms (askeletal and ci cardiac), two smooth muscle isoforms (avascular and y enteric) and two cytoplasmic isoforms (p and y cytoplasmic) (Vanderkerckhove and Weber, 1978). The expression of the different actin genes appear to be tissue specific (Garrels and Gibson, 1976). For instance, smooth muscle vessels ci actin (SMV-a actin) has only been described in smooth muscle cells such as those around vessels (Fatigati and Murphy, 1984). 0 1991 Wiley-Liss, Inc.
We have recently reported that astroglial cell lines as well as primary cultures of astrocytes express the mRNA of the a isoform of SMV actin (Rhyner et a]., 1990). However, it was possible that this unexpected expression of SMV a actin in astroglial cells did not result in protein synthesis. We now report that a monospecific antiserum and a monoclonal antibody (MAB) directed to the 1-9 amino acid sequence of SMV 01 actin label a typical actin network in D19 astroglial cells and in astrocytes in brain primary cultures. These findings prompted further experiments to investigate the localization of SMV ci actin in brain sections of adult mice. Quite surprisingly, we have observed that, in addition to vessel walls, processes of a subset of astrocytes are clearly labeled by the specific anti-SMV a actin antibodies.
MATERIALS AND METHODS Cell Cultures The D19 cell line can be regarded as the in vitro counterpart of the velate protoplasmic astrocytes. Its culture conditions have been previously described (Alliot and Pessac, 1984; Alliot et a]., 1988). Primary cultures of astrocytes from ED 15 to PN2 forebrain, cerebellum, and brain stem were prepared according to the following procedure. After the meninges were peeled, brain cells were dissociated by trypsinization and gentle pipetting and plated at a density of 150,000-200,000 cells/cm’ in Eagle’s basal medium supplemented with 10%fetal bovine serum (FBS). These cultures were used 5-30 days later. Immunolabeling showed that such cultures are composed of large flat
Received November 20, 1990; revised January 30, 1991; accepted February 6 , 1991. Address reprint requests to Dr. Bernard PESSAC, C.B.C.-C.N.R.S. UPK 3101. 67, rue Maurice Gunsbourg. 94205 lvry sur Seine. Cedex. France.
602
Lecain et al.
intensely glial fibrillary acidic protein (GFAP) immunoreactive astrocytes.
Brain Sections Adult C57BI mice were killed by neck elongation. Their brains were rapidly dissected, snap frozen in isopentane cooled at -40°C in Dry Ice for 1 min, and immediately sectioned in a cryostat. 10-12 pm brain parasagittal and transversal sections were made. Frozen sections were kept at -20°C.
antiserum diluted 1/100followed by an incubation with a 1/20 solution of a FITC goat anti-rabbit IgG.
RESULTS Astroglial cultures The D19 astroglial cell line. Figure 1 shows a typical actin network in cultures of the D19 astroglial clone after labeling with a monospecific anti-SMV a actin antiserum. A similar result was obtained with a monoclonal antibody specific for SMV a actin (not Immunocytochemistry shown). These cultures were made of confluent nonCultures and sections were fixed in paraformalde- synchronous cells in which many, but not all, cells were hyde 4% in phosphate buffer saline (PBS) for 20 min, immunoreactive to either SMV a actin probe. However, then permeabilized with Triton X 100 1% for 5 min. All in double labeling experiments with phalloidin, which incubations were carried out at room temperature for 1 binds to all actin isoforms, all cells in these confluent hr . cultures were strongly labeled. Moreover, when D19 asSingle immunolabeling. The cultures or sections torocytes were seeded at low density and as long as they were first incubated with a 1/100 dilution of the anti- remained sparse, all cells were labeled by the SMV a SMV a actin antiserum (Hanin et al., 1989) or 1/50 actin antibodies (now shown). dilution of the monoclonal anti-SMV a actin antibody Brain primary cultures. Primary cultures from (PROGEN GmbH ASM-I) (Skalli et al., 1986), or a various brain regions at different developmental stages 1/100dilution of the monoclonal anti-GFAP (GA5, ICN, were double-immunolabeled with a GFAP specific Cat, no. 694-1 102) antibody; after extensive washings, monoclonal antibody (GA5) and the anti-SMV a actin they were incubated with a 1/20 FITC goat anti-rabbit Ig antiserum. Figure 2 shows a primary culture of 2-day (Southern Biotechnology Associates, Cat. no. 4010-02) postnatal forebrain after 14 days in vitro. Comparison of or a 1/20 FITC goat anti-mouse IgG2a (Southern Bio- Figure 2a and 2b shows that most, if not all, GAS-postechnology Associates, Cat. no. 1080-02) or a 1/20 FITC itive astrocytes (b) are immunoreactive to the anti-SMV goat anti-mouse (Southern Biotechnology Associates, a actin (a), although to different extents. In addition, Cat. no. 1070-02) solution, respectively. No immunor- there are cells which show a typical actin network but are eactivity was detected when the first antibody was omit- not labeled by GA5 (arrows). It is possible that these cells are not astrocytes or that they are immature astroted. As a control, the monospecific antiserum was ad- cytes that do not yet express GFAP. Figure 2c shows that sorbed overnight at 4°C with smooth muscle vascular a when the anti-SMV a actin antiserum was adsorbed by actin extracted from aorta (a gift from Dr. J . C. Cava- SMV a actin, no actin network was seen. Taken todore) under shaking. After centrifugation at 4,000g for gether, these results show unambiguously that in vitro 30 min at 4"C, the supernatant that contained the ad- astrocytes produce SMV a actin into a typical network. sorbed SMV a actin antibody was used. Double labeling. The anti-SMV a actin poly- Brain Sections This unexpected finding suggested that in vivo asclonal antiserum was mixed with the anti-GFAP monoclonal (GA5). After washings, a 1/20 dilution of a RITC trocytes, or a subset of astrocytes, might also be immugoat anti-mouse IgG 1 (Southern Biotechnology Associ- noreactive to monospecific SMV (Y actin antibodies. To ates, Cat. no. 1070-03) was mixed with the FITC goat investigate whether astrocytes express SMV a actin in anti-rabbit Ig antiserum. When the monoclonal anti- vivo, we then carried out immunocytochemistry experiSMV a actin was used. slides were incubated succes- ments in parasagittal and frontal sections throughout the sively with the anti-SMV 01 actin monoclonal, the FITC adult mouse brain. goat anti-mouse IgG2a, the monoclonal GA5 and the A frontal section through cerebellum and brainstem RITC goat anti-mouse IgG 1 antibodies. No immunore- is shown in Figure 3. Figure 3a shows that phalloidin activity to SMV a actin or GFAP was detected when the binds to actin in all cell types in this region of the CNS. corresponding antibody was omitted. By contrast, Figure 3b shows that the monospecific SMV For double-labeling with phalloidin and the anti- a actin only labels vessel walls. However, other sections clearly show immunoreSMV a actin antiserum, slides were first incubated for 5 min at 37°C in a 1/1000 dilution of a rhodamine-conju- activity of astrocytes to SMV a actin antibodies. Figure gated phalloidin solution, then with the anti-SMV a actin 4a,b shows a parasagittal section, through the third ven-
Smooth Muscle Vascular (Y Actin in Astrocytes
tricle and corpus callosum, labeled with GA5 antibody. Figure 4b is a field of Figure 4a at higher magnification. In Figure 4a,b astrocytes, very abundant in this region, are strongly immunoreactive to the anti-GFAP antibody. Figure 4b,c represents the same field labeled by antiGFAP (b) and anti-SMV a actin (c) antibodies, respectively. The striking result is that most, if not all astrocytes, in these parasagittal sections, are immunoreactive to the SMV a actin antiserum. Figure 5a-c shows frontal sections of forebrain through corpus callosum at the limit of the two brain hemispheres. Figure 5a,b is immunolabeled with the GA5 antibody; Figure 5b is a magnification of Figure 5a. Figure 5c shows a section labeled by the anti-SMV a actin antiserum, adjacent to Figure 5b. In this frontal section, few corpus callosum astrocytes are immunoreactive to SMV a actin although processes of other astrocytes are strongly immunoreactive to this antibody (arrows). Figure 6 represents another frontal section of corpus callosum which also shows that in that plane, few astrocyte processes are immunoreactive to SMV a actin. In an other parasagittal section through the fornix, astrocytes are also strongly immunoreactive to GA5 (not shown) and labeled by the monoclonal antibody antiSMV a actin (Fig. 7).
DISCUSSION The fact that D19 astrocytes display an SMV a actin network was not unexpected, since they express the corresponding mRNA. However, D19 cells are an immortal astroglial cell line and their expression of SMV a actin might have been related to the immortalization process or to their astroglial nature. This led us to investigate whether primary astrocytes in short term cultures also contain SMV a actin. Our data clearly show that, indeed, this is the case. The next step was to investigate whether astrocytes or a subset of astrocytes also express SMV a actin in vivo. Indeed, the most striking result of the present study is the demonstration of an immunoreactivity to SMV a actin antibodies in astrocytes of distinct brain regions. These data rely on the specificity of the anti-SMV a actin probes used in this study. First, the actin network labeled by the anti-SMV a actin antiserum was no longer visible when this antiserum was preadsorbed with SMV a actin. Second, experiments carried out with a monoclonal anti-SMV a actin antibody gave results identical to those obtained with the SMV a actin antiserum. Third, in many brain sections, only vessel walls which contain smooth muscle cells, are labeled by the antiSMV a actin antiserum. These data clearly show that the
603
monospecific antiserum and the monoclonal antibody used in this study only label the SMV a actin isoform. Although these results do not formally eliminate the possibility that the antibodies might recognize an other antigen sharing the amino-terminal of SMV a actin, two types of data indicate that this is very improbably. In Western blots of brain extracts, the anti-SMV a actin monoclonal antibody recognized a 43 kD band, which corresponds to the molecular weight of actin (Omri B., data not shown); furthermore, a search in NBRF protein data base shows that no protein has been reported to have the same amino-terminal as SMV a actin. The fact that processes of corpus callosum astrocytes are strongly immunoreactive to anti-SMV a actin antibodies only in sagittal sections came as a surprise. This suggests the existence of structural differences between distinct astrocyte processes. Indeed, it has been reported that different actin isoforms are not equivalent (Bcrnstein, 1990). It is possible, therefore, that SMV a actin, present in the astrocyte processes that follow the sagittal planes of the corpus callosum and fornix, play a specific physiological role. For instance, SMV a actin in these processes might act as a scaffold for long myelinated fibers. In this context, it will be interesting to investigate if the SMV a actin begins to be expressed when long myelinated fibers appear. Another possibility is that SMV a actin is expressed throughout the brain during development and becomes restricted to astrocytes in some regions of the adult CNS, as already described for skeletal muscles (Sawtell and Lessard, 1989). In conclusion, our results clearly show that distinct populations of astrocytes can be distinguished by a biochemical marker, SMV a actin. It will be of interest to determine whether SMV a actin is also present in human astrocytes and in tumors derived from astrocytes.
ACKNOWLEDGMENTS We thank J.C. Cavadore for the monospecific smooth muscle vascular (Y actin antiserum, for helpful suggestions and critical reading of the manuscript and B. Omri for the Western blot experiments. We are grateful to Michel Louette for the photographic work and Patricia Gaudoin and Isabelle Angelchic for typing the manuscript. This work was supported by the Centre National de la Recherche Scientifique, Ministere de la Recherche et de la Technologie (M.R.T., grant 87C0527) and Association de la Recherche contre le Cancer (A.R.C., grant 6777) and Fondation pour la Recherche Medicale Fransaise. Eric Lecain was supported by Ministere de la Recherche et de la Technologie and Association Fransaise contre les Myopathies (A.F.M.).
Fig. 1. Actin network visualized with a monospecific antiSMV a actin antiserum in the astroglial D19 cell line. x 250.
double labe!ed. b. With GAS antibody. x 350. c. With antiSMV a actin antiserum. x 250.
Fig. 2 . Primary cultures of a 2-day postnatal forebrain after 14 days in vitro. a,b. Represent the same field. a. Immunolabeled by GAS antibody. X 250. b. Labeled by anti-SMV a actin antiserum. X 250. c. Another field immunolabeled by antiSMV a actin antiserum adsorbed by smooth muscle a actin from aorta.
Fig. 5 . Frontal brain sections through corpus calloaum. a,b. Same field immunolabeled with the GAS antibody. a. X 100. b. x 250. c. An adjacent section immunolabeled with antiSMV a actin antiserum. x 250.
Fig.3 Frontal brain section through cerebellum and brain stem double-labeled. a. With phalloidin rhodamine. b. With antiSMV a actin antiserum. x SO. Fig. 4. Sagittal brain section through the third ventricle and corpus callosum. a. Immunolabeled with the GA5 antibody. X 100. b,c. Higher magnification of a field from Figure 4a
Fig. 6. Frontal sections show another field of the corpus callosum of the frontal sections shown in Figure 5.a. Immunolabeled with GA.5 antibody. X 250. b. Immunolabeled with antiSMV a actin antiserum. X 2.50. Fig. 7. Parasagittal brain section through the fornix immunolabeled with the monoclonal anti-SMV a actin antibody. x 250.
Figs. 4-7.
606
Lecain et al.
REFERENCES Alliot F. Pessac B (1984): Astrocytic cell line clones derived from established cultures of 8 day postnatal mouse cerebella. Brain Res 306:283-29 1. Alliot F, Delhaye-Bouchaud N. Geffard M, Pessac B (1988): Role of astroglial cell clones in the survival and differentiation of cerebellar embryonic neurons. Dev Brain Res 44247-257. Bernstein PA (1990): The functional importance of multiple actin isoforms. Bio Essays 12:309-315. Fatigati V, Murphy RA (1984): Actin and tropomyosin variants in smooth muscles. Dependence on tissue type. J Biol Chem 259: 14383-14388. Carrels JI. Gibson W (1976): Identification and characterization of multiple fornis of actin. Cell 9:793-805. Hanin V, Naharisoa H. Sarrade V. Calas B (1989): Production of oligoclonal antibodies directed to the N-terminus of smooth
muscle a-actin using peptidyl polyacrylic resin as direct immunogen. Peptide Res. 2367-372. Rhyner TA. Lecain E, Mallet J. Pessac B (1990): Isolation of cDNAs from a mouse astroglial cell line by a subtracted cDNA library. J Neurosci Res 27:145-152. Sawtell NM, Lessard J (1989):Cellular distribution of smooth muscle actin during mammalian embryogenesis: Expression of the avascular but not y-entcric isoform in differentiating striated myocyte. J Cell Biol 109:2929-2937. Skalli 0, Ropraz P, Traeciak A , Benzonana G , Gillessen D. Gabbiani G (1986): A monoclonal antibody against alpha-smooth muscle actin: A new probe for smooth muscle differentiation. J Cell Biol 1032787-2796. Vandekerckhove J, Weber K (1978): At least six different actins are expressed in higher mammal: An analysis based on the aminoacid sequence of the amino-terminal peptide. J Mol Biol 126:783-802.
Rapid Communication (x Isoform of Smooth Muscle Actin Is Expressed in Astrocytes In Vitro and In Vivo E. Lecain, F. Alliot, M.C. Laine, B. Calas, and B. Pessac Centre De Biologie Cellulaire, UPR 3 10 I , C.N.R.S., Ivry Sur Seine CEDEX (E.L., F.A., M.C.L., B.P.); Centre De Recherche De Biochimie Macromoleculaire, UPR 8402. C.N.R.S.. Montpellier (B.C.), France We have previously reported that astroglial cell lines derived from spontaneously immortalized mouse cerebellar cultures as well as primary astrocyte cultures express the mRNA of the a isoform of smooth muscle actin. In this report, we have used an antiserum specific for the a smooth muscle actin protein to investigate the presence and the pattern of expression of a smooth muscle actin protein at the cellular level with immunocytochemical methods. The results show that an anti-smooth muscle vessels a actin antiserum labels a typical actin network in the D19 astroglial cell clone and in flat astrocytes of primary cultures derived from various CNS regions of embryonic and postnatal mice. Furthermore, this antiserum labels distinct populations of astrocytes in the adult mouse brain, in particular in the corpus callosum and the fornix. However, in the corpus callosum, astrocytic processes are strongly labeled by anti-SMV a actin antibodies only in parasagittal planes. Thus, a smooth muscle actin represents a new marker for subsets of astrocytes. Key words: mouse, corpus callosum, central nervous system INTRODUCTION Actin is a ubiquitous cytoskeletal protein. In mammals, actin is a multigene family as there are at least six different isoforms of actin, each isoform being coded by a distinct gene. The different isoforms of actin include two striated muscles isoforms (askeletal and ci cardiac), two smooth muscle isoforms (avascular and y enteric) and two cytoplasmic isoforms (p and y cytoplasmic) (Vanderkerckhove and Weber, 1978). The expression of the different actin genes appear to be tissue specific (Garrels and Gibson, 1976). For instance, smooth muscle vessels ci actin (SMV-a actin) has only been described in smooth muscle cells such as those around vessels (Fatigati and Murphy, 1984). 0 1991 Wiley-Liss, Inc.
We have recently reported that astroglial cell lines as well as primary cultures of astrocytes express the mRNA of the a isoform of SMV actin (Rhyner et a]., 1990). However, it was possible that this unexpected expression of SMV a actin in astroglial cells did not result in protein synthesis. We now report that a monospecific antiserum and a monoclonal antibody (MAB) directed to the 1-9 amino acid sequence of SMV 01 actin label a typical actin network in D19 astroglial cells and in astrocytes in brain primary cultures. These findings prompted further experiments to investigate the localization of SMV ci actin in brain sections of adult mice. Quite surprisingly, we have observed that, in addition to vessel walls, processes of a subset of astrocytes are clearly labeled by the specific anti-SMV a actin antibodies.
MATERIALS AND METHODS Cell Cultures The D19 cell line can be regarded as the in vitro counterpart of the velate protoplasmic astrocytes. Its culture conditions have been previously described (Alliot and Pessac, 1984; Alliot et a]., 1988). Primary cultures of astrocytes from ED 15 to PN2 forebrain, cerebellum, and brain stem were prepared according to the following procedure. After the meninges were peeled, brain cells were dissociated by trypsinization and gentle pipetting and plated at a density of 150,000-200,000 cells/cm’ in Eagle’s basal medium supplemented with 10%fetal bovine serum (FBS). These cultures were used 5-30 days later. Immunolabeling showed that such cultures are composed of large flat
Received November 20, 1990; revised January 30, 1991; accepted February 6 , 1991. Address reprint requests to Dr. Bernard PESSAC, C.B.C.-C.N.R.S. UPK 3101. 67, rue Maurice Gunsbourg. 94205 lvry sur Seine. Cedex. France.
602
Lecain et al.
intensely glial fibrillary acidic protein (GFAP) immunoreactive astrocytes.
Brain Sections Adult C57BI mice were killed by neck elongation. Their brains were rapidly dissected, snap frozen in isopentane cooled at -40°C in Dry Ice for 1 min, and immediately sectioned in a cryostat. 10-12 pm brain parasagittal and transversal sections were made. Frozen sections were kept at -20°C.
antiserum diluted 1/100followed by an incubation with a 1/20 solution of a FITC goat anti-rabbit IgG.
RESULTS Astroglial cultures The D19 astroglial cell line. Figure 1 shows a typical actin network in cultures of the D19 astroglial clone after labeling with a monospecific anti-SMV a actin antiserum. A similar result was obtained with a monoclonal antibody specific for SMV a actin (not Immunocytochemistry shown). These cultures were made of confluent nonCultures and sections were fixed in paraformalde- synchronous cells in which many, but not all, cells were hyde 4% in phosphate buffer saline (PBS) for 20 min, immunoreactive to either SMV a actin probe. However, then permeabilized with Triton X 100 1% for 5 min. All in double labeling experiments with phalloidin, which incubations were carried out at room temperature for 1 binds to all actin isoforms, all cells in these confluent hr . cultures were strongly labeled. Moreover, when D19 asSingle immunolabeling. The cultures or sections torocytes were seeded at low density and as long as they were first incubated with a 1/100 dilution of the anti- remained sparse, all cells were labeled by the SMV a SMV a actin antiserum (Hanin et al., 1989) or 1/50 actin antibodies (now shown). dilution of the monoclonal anti-SMV a actin antibody Brain primary cultures. Primary cultures from (PROGEN GmbH ASM-I) (Skalli et al., 1986), or a various brain regions at different developmental stages 1/100dilution of the monoclonal anti-GFAP (GA5, ICN, were double-immunolabeled with a GFAP specific Cat, no. 694-1 102) antibody; after extensive washings, monoclonal antibody (GA5) and the anti-SMV a actin they were incubated with a 1/20 FITC goat anti-rabbit Ig antiserum. Figure 2 shows a primary culture of 2-day (Southern Biotechnology Associates, Cat. no. 4010-02) postnatal forebrain after 14 days in vitro. Comparison of or a 1/20 FITC goat anti-mouse IgG2a (Southern Bio- Figure 2a and 2b shows that most, if not all, GAS-postechnology Associates, Cat. no. 1080-02) or a 1/20 FITC itive astrocytes (b) are immunoreactive to the anti-SMV goat anti-mouse (Southern Biotechnology Associates, a actin (a), although to different extents. In addition, Cat. no. 1070-02) solution, respectively. No immunor- there are cells which show a typical actin network but are eactivity was detected when the first antibody was omit- not labeled by GA5 (arrows). It is possible that these cells are not astrocytes or that they are immature astroted. As a control, the monospecific antiserum was ad- cytes that do not yet express GFAP. Figure 2c shows that sorbed overnight at 4°C with smooth muscle vascular a when the anti-SMV a actin antiserum was adsorbed by actin extracted from aorta (a gift from Dr. J . C. Cava- SMV a actin, no actin network was seen. Taken todore) under shaking. After centrifugation at 4,000g for gether, these results show unambiguously that in vitro 30 min at 4"C, the supernatant that contained the ad- astrocytes produce SMV a actin into a typical network. sorbed SMV a actin antibody was used. Double labeling. The anti-SMV a actin poly- Brain Sections This unexpected finding suggested that in vivo asclonal antiserum was mixed with the anti-GFAP monoclonal (GA5). After washings, a 1/20 dilution of a RITC trocytes, or a subset of astrocytes, might also be immugoat anti-mouse IgG 1 (Southern Biotechnology Associ- noreactive to monospecific SMV (Y actin antibodies. To ates, Cat. no. 1070-03) was mixed with the FITC goat investigate whether astrocytes express SMV a actin in anti-rabbit Ig antiserum. When the monoclonal anti- vivo, we then carried out immunocytochemistry experiSMV a actin was used. slides were incubated succes- ments in parasagittal and frontal sections throughout the sively with the anti-SMV 01 actin monoclonal, the FITC adult mouse brain. goat anti-mouse IgG2a, the monoclonal GA5 and the A frontal section through cerebellum and brainstem RITC goat anti-mouse IgG 1 antibodies. No immunore- is shown in Figure 3. Figure 3a shows that phalloidin activity to SMV a actin or GFAP was detected when the binds to actin in all cell types in this region of the CNS. corresponding antibody was omitted. By contrast, Figure 3b shows that the monospecific SMV For double-labeling with phalloidin and the anti- a actin only labels vessel walls. However, other sections clearly show immunoreSMV a actin antiserum, slides were first incubated for 5 min at 37°C in a 1/1000 dilution of a rhodamine-conju- activity of astrocytes to SMV a actin antibodies. Figure gated phalloidin solution, then with the anti-SMV a actin 4a,b shows a parasagittal section, through the third ven-
Smooth Muscle Vascular (Y Actin in Astrocytes
tricle and corpus callosum, labeled with GA5 antibody. Figure 4b is a field of Figure 4a at higher magnification. In Figure 4a,b astrocytes, very abundant in this region, are strongly immunoreactive to the anti-GFAP antibody. Figure 4b,c represents the same field labeled by antiGFAP (b) and anti-SMV a actin (c) antibodies, respectively. The striking result is that most, if not all astrocytes, in these parasagittal sections, are immunoreactive to the SMV a actin antiserum. Figure 5a-c shows frontal sections of forebrain through corpus callosum at the limit of the two brain hemispheres. Figure 5a,b is immunolabeled with the GA5 antibody; Figure 5b is a magnification of Figure 5a. Figure 5c shows a section labeled by the anti-SMV a actin antiserum, adjacent to Figure 5b. In this frontal section, few corpus callosum astrocytes are immunoreactive to SMV a actin although processes of other astrocytes are strongly immunoreactive to this antibody (arrows). Figure 6 represents another frontal section of corpus callosum which also shows that in that plane, few astrocyte processes are immunoreactive to SMV a actin. In an other parasagittal section through the fornix, astrocytes are also strongly immunoreactive to GA5 (not shown) and labeled by the monoclonal antibody antiSMV a actin (Fig. 7).
DISCUSSION The fact that D19 astrocytes display an SMV a actin network was not unexpected, since they express the corresponding mRNA. However, D19 cells are an immortal astroglial cell line and their expression of SMV a actin might have been related to the immortalization process or to their astroglial nature. This led us to investigate whether primary astrocytes in short term cultures also contain SMV a actin. Our data clearly show that, indeed, this is the case. The next step was to investigate whether astrocytes or a subset of astrocytes also express SMV a actin in vivo. Indeed, the most striking result of the present study is the demonstration of an immunoreactivity to SMV a actin antibodies in astrocytes of distinct brain regions. These data rely on the specificity of the anti-SMV a actin probes used in this study. First, the actin network labeled by the anti-SMV a actin antiserum was no longer visible when this antiserum was preadsorbed with SMV a actin. Second, experiments carried out with a monoclonal anti-SMV a actin antibody gave results identical to those obtained with the SMV a actin antiserum. Third, in many brain sections, only vessel walls which contain smooth muscle cells, are labeled by the antiSMV a actin antiserum. These data clearly show that the
603
monospecific antiserum and the monoclonal antibody used in this study only label the SMV a actin isoform. Although these results do not formally eliminate the possibility that the antibodies might recognize an other antigen sharing the amino-terminal of SMV a actin, two types of data indicate that this is very improbably. In Western blots of brain extracts, the anti-SMV a actin monoclonal antibody recognized a 43 kD band, which corresponds to the molecular weight of actin (Omri B., data not shown); furthermore, a search in NBRF protein data base shows that no protein has been reported to have the same amino-terminal as SMV a actin. The fact that processes of corpus callosum astrocytes are strongly immunoreactive to anti-SMV a actin antibodies only in sagittal sections came as a surprise. This suggests the existence of structural differences between distinct astrocyte processes. Indeed, it has been reported that different actin isoforms are not equivalent (Bcrnstein, 1990). It is possible, therefore, that SMV a actin, present in the astrocyte processes that follow the sagittal planes of the corpus callosum and fornix, play a specific physiological role. For instance, SMV a actin in these processes might act as a scaffold for long myelinated fibers. In this context, it will be interesting to investigate if the SMV a actin begins to be expressed when long myelinated fibers appear. Another possibility is that SMV a actin is expressed throughout the brain during development and becomes restricted to astrocytes in some regions of the adult CNS, as already described for skeletal muscles (Sawtell and Lessard, 1989). In conclusion, our results clearly show that distinct populations of astrocytes can be distinguished by a biochemical marker, SMV a actin. It will be of interest to determine whether SMV a actin is also present in human astrocytes and in tumors derived from astrocytes.
ACKNOWLEDGMENTS We thank J.C. Cavadore for the monospecific smooth muscle vascular (Y actin antiserum, for helpful suggestions and critical reading of the manuscript and B. Omri for the Western blot experiments. We are grateful to Michel Louette for the photographic work and Patricia Gaudoin and Isabelle Angelchic for typing the manuscript. This work was supported by the Centre National de la Recherche Scientifique, Ministere de la Recherche et de la Technologie (M.R.T., grant 87C0527) and Association de la Recherche contre le Cancer (A.R.C., grant 6777) and Fondation pour la Recherche Medicale Fransaise. Eric Lecain was supported by Ministere de la Recherche et de la Technologie and Association Fransaise contre les Myopathies (A.F.M.).
Fig. 1. Actin network visualized with a monospecific antiSMV a actin antiserum in the astroglial D19 cell line. x 250.
double labe!ed. b. With GAS antibody. x 350. c. With antiSMV a actin antiserum. x 250.
Fig. 2 . Primary cultures of a 2-day postnatal forebrain after 14 days in vitro. a,b. Represent the same field. a. Immunolabeled by GAS antibody. X 250. b. Labeled by anti-SMV a actin antiserum. X 250. c. Another field immunolabeled by antiSMV a actin antiserum adsorbed by smooth muscle a actin from aorta.
Fig. 5 . Frontal brain sections through corpus calloaum. a,b. Same field immunolabeled with the GAS antibody. a. X 100. b. x 250. c. An adjacent section immunolabeled with antiSMV a actin antiserum. x 250.
Fig.3 Frontal brain section through cerebellum and brain stem double-labeled. a. With phalloidin rhodamine. b. With antiSMV a actin antiserum. x SO. Fig. 4. Sagittal brain section through the third ventricle and corpus callosum. a. Immunolabeled with the GA5 antibody. X 100. b,c. Higher magnification of a field from Figure 4a
Fig. 6. Frontal sections show another field of the corpus callosum of the frontal sections shown in Figure 5.a. Immunolabeled with GA.5 antibody. X 250. b. Immunolabeled with antiSMV a actin antiserum. X 2.50. Fig. 7. Parasagittal brain section through the fornix immunolabeled with the monoclonal anti-SMV a actin antibody. x 250.
Figs. 4-7.
606
Lecain et al.
REFERENCES Alliot F. Pessac B (1984): Astrocytic cell line clones derived from established cultures of 8 day postnatal mouse cerebella. Brain Res 306:283-29 1. Alliot F, Delhaye-Bouchaud N. Geffard M, Pessac B (1988): Role of astroglial cell clones in the survival and differentiation of cerebellar embryonic neurons. Dev Brain Res 44247-257. Bernstein PA (1990): The functional importance of multiple actin isoforms. Bio Essays 12:309-315. Fatigati V, Murphy RA (1984): Actin and tropomyosin variants in smooth muscles. Dependence on tissue type. J Biol Chem 259: 14383-14388. Carrels JI. Gibson W (1976): Identification and characterization of multiple fornis of actin. Cell 9:793-805. Hanin V, Naharisoa H. Sarrade V. Calas B (1989): Production of oligoclonal antibodies directed to the N-terminus of smooth
muscle a-actin using peptidyl polyacrylic resin as direct immunogen. Peptide Res. 2367-372. Rhyner TA. Lecain E, Mallet J. Pessac B (1990): Isolation of cDNAs from a mouse astroglial cell line by a subtracted cDNA library. J Neurosci Res 27:145-152. Sawtell NM, Lessard J (1989):Cellular distribution of smooth muscle actin during mammalian embryogenesis: Expression of the avascular but not y-entcric isoform in differentiating striated myocyte. J Cell Biol 109:2929-2937. Skalli 0, Ropraz P, Traeciak A , Benzonana G , Gillessen D. Gabbiani G (1986): A monoclonal antibody against alpha-smooth muscle actin: A new probe for smooth muscle differentiation. J Cell Biol 1032787-2796. Vandekerckhove J, Weber K (1978): At least six different actins are expressed in higher mammal: An analysis based on the aminoacid sequence of the amino-terminal peptide. J Mol Biol 126:783-802.
