GFAP Immunoreactivity in Octopus Nervous System Before and After Injury Abstract Astrogliosis following injury to the mammalian central nervous system is accompanied by increased synthesis of glial fibrillary acidic protein (GFAP) and the lack of regeneration. In order to extend studies of GFAP expression in relation to injury in taxa of higher regenerative capacity, we chose Octopus rubescens as a model system. We used a polyclonal anti-human GFAP in conjunction with immunoblotting and immunohistochemistry techniques applied to optic lobes of the brain and to peripheral axial nerve cords in the arms. Animals were also studied before and after lesion to the axial nerves that contain ganglionic regions of neuronal and glial somata. In uninjured Octopus, Western blots with boh tissue sources revealed a single protein band (140KD). Pre-absorption of the antibody with purified human GFAP eliminated this band, which is larger than the 48-51 KD protein found in vertebrates. Immunostaining of fixed and sectioned material from axial nerve cords showed that anti-GFAP preferentially stained cellular regions. This pattern is distinct from that produced by anti¬ neurofilament, which stained primarily neuropil and axons. Studies were also carried out following injury (amputation) to the axial nerve cords. Nine days after injury, immunoblotting revealed an increased level of the 140 KD band. Experiments on other cephalopods were also carried out. In uninjured Sepia officianalis, Sepioteuthis lessoniana, and Loligo opalescens, the 140 KD band was not detected by Western blots with optic lobe proteins. This product was obtained, however, with axial nerve material from Sepia. These results (i) indicate the presence of a GFAP-like protein in the nervous system of cephalopods, (ii) show that increased protein levels accompany injury in the peripheral nerve cords in Octopus, and (iii) suggest that the protein may be associated with glial cells. Introduction In mammalian systems, there is experimental evidence that glial scarring presents an impediment to regeneration in the central nervous system (Reier et al., 1983; Windle et al., 1952). Such glial scarring is characterized by swelling of the astrocyte cell bodies and extension of the processes. This hypertrophy of astrocytes following injury, termed astrogliosis, is accompanied by increased synthesis of glial fibrillary acidic protein (GFAP) (Eng. 1988). In most cases, regenerating axons are unable to penetrate the scar tissue. Although the accumulation of GFAP in these reactive glial cells appear to inhibit regeneration in mammalian CNS, they do not seem to present an obstacle to regenerating axons in lower vertebrates such as the teleosts (Stafford et al., 1990: Anderson et al., 1984). In invertebrates, the ability to repair damage to the nervous system surpasses that of the vertebrates (Fredman and Nutz, 1988). However, the roles of glial cells in response to injury in these model systems have not been examined extensively. The lack of glial markers for immunochemistry studies and the limited knowledge of glial cells in these organisms beyond morphology make these studies difficult. In the present study, we investigate a glial filament protein in cephalopods that appears to be related to GFAP in vertebrates. In a previous report, it was found that a monoclonal antibody to GFAP cross reacted with three bands (56, 49, 37 KD) in protein blots of Octopus vulgaris optic lobe homogenate. Using immunoelectron microscopy, it was found that the antibody labelled filaments that may be glial in nature (Cardone and Roots, 1990). In this study, we use a polyclonal antibody to study GFAP immunoreactivity in normal and injured axial nerve cords from Octopus rubescens, which are able to regenerate these nerve cords. These tissues basically consist of a series of ganglia arranged along the arms and embedded within the musculature of the arms. In each ganglion, there is an outer cellular layer that surrounds the neuropil (Young, 1971). Immunohistochemistry techniques were utilized to examine if GFAP immunoreactivity was restricted to certain regions of the tissue. We also carried out comparative studies of GFAP immunoreactivity in other uninjured cephalopods, Loligo opalescens, Sepioteuthis lessoniana, and Sepia officianalis. Materials and Methods ANIMAL PREPARATION: Octopi (Octopus rubescens) about 8 inches in arm span were collected in Monterey Bay, California. They were kept in holding tanks with circulating sea water and fed small live crabs. For injury experiments, the animals were anesthesized in 23 ethanol until arm movements stopped (about 5 to 10 minutes), and four of the arms were amputated about 3 cm from the tip. The animals were placed back in the holding tanks. Nine days after amputation, axial nerve cords were removed from the proximal stumps and control arms for immunoblotting. For comparative studies of GFAP immunoreactivity in other cepahalopods, Loligo opalescens, Sepioteuthis lessoniana, and Sepia officianalis were used, L. opalescens as collected in Monterey Bay, Ca.; S. lessoniana as provided by Dr. Roger Hanlon at the University of Texas, Galveston, Medical Branch; and S. officianalis as donated by the Monterey Bay Aquarium. All three were kept in tanks with circulating sea water. The nervous system tissues used for immunoblotting and immunohistochemistry were removed from these animals following decapitation without anesthesia. IMMUNOBLOTTING: Nervous system tissues were removed from the animals and placed directly in phosphate buffer (pH 8.0) containing 13 SDS, 2 mM EDTA, and 3 mM EGTA. The samples were homogenized in a motor driven homogenizer, centrifuged at 12,000 rpm (Eppendorf microfuge) for 30 minutes, and the supernatant collected. Protein concentrations were determined by the bicinchoninic acid (BCA) assay (Brown et al., 1989). The proteins were separated using 103 polyacrylamide gel electrophoresis according to the method of Laemmli (1970) and transferred onto nitrocellulose membrane (Towbin et al., 1979). Antibody detections were done on the blots using a previously described polyclonal rabbit antiserum to human GFAP (Eng and DeArmond, 1983), monoclonal mouse antiserum to neurofilament (Sternberger and Sternberger, 1983), and the peroxidase anti¬ peroxidase method (PAP) (Sternberger et al., 1970). To test specificity of staining, the same rabbit anti-GFAP pre-absorbed with purified GFAP was obtained from Dr. Lawrence Eng, Department of Pathology, Stanford University Medical School. Nonspecific binding was minimized by preincubation with normal swine serum for polyclonal or normal goat serum for monoclonal; and all incubations were done in 18 milk in buffer. The blots were incubated in primary antibodies overnight at a dilution of 1:500. After incubation in the secondary antibody and the PAP complex, cross-reacting proteins were visualized by development in 3,3'- diaminobenzidine. IMMUNOHISTOCHEMISTRY: Tissues were removed from animals and fixed in 43 paraformaldehyde at 4'C for 48 hours. Paraffin sections of these tissues were prepared by standard techniques (Galigher and Kozloff, 1964). Briefly, tissues were dehydrated through graded alcohol, alcohol replaced with xylene, and tissues gradually infiltrated with melted paraffin. The infiltrated tissues were then embedded in paraffin and sectioned at 10 microns with a steel knife microtome and mounted on albumin coated slides. Prior to immuostaining, the slides were deparaffinized in xylene, rehydrated through graded alcohol, and treated with acetylated Trypsin (Sigma Chemicals) and hydrogen peroxide. Immunostaining was done using the same antibodies as described above and the PAP method. The sections were then counterstained and mounted for viewing. Results IMMUNOBLOTTING: The polyclonal anti-GFAP cross reacted with one band (140 KD) in transblots of optic lobe and axial nerve cords from uninjured Octopus (fig. la). This protein is of higher molecular weight than mammalian GFAP (48-51 KD). However, using the same antibody pre¬ absorbed with the purified mammalian antigen, the band was eliminated (data not shown). To further demonstrate that the antibody was not directed to neurofilament protein (NF), the same transblots were immunostained with a monoclonal anti-NF. NF staining revealed bands clustered around 200 KD and one strong band at 110 KD, none of which corresponded with the GFAP band (fig. 1b). INJURY EXPERIMENTS: Nine days after amputation of the arms, there were no visible signs of regeneration. However, the stumps were still active, responsive to touch, and showed no signs of degeneration. The axial nerve cords from injured and control arms removed at this point showed differing levels of GFAP immunoreactivity. From the transblot stained with anti-GFAP (fig. 2a), it appears that the 140 kD band is more intense in the injured tissue than in the control. This difference is evident despite the greater amount of protein loaded for the control tissue (fig. 2b). This rise in the level of GFAP immunoreactivity is accompanied by the appearance of another cross reactive band at 80 KD. However, the significance of this band in relation to injury is questionable since the band sometimes appears in uninjured tissues as well (data not shown). IMMUNOHISTOCHEMISTRY: Immunostaining of paraffin sections of Octopus axial nerve cords and Sepioteuthis stellate ganglia with anti-GFAP and anti-NF showed different patterns of staining. With anti-GFAP, staining was more diffuse. However, it appears that staining was primarily in layers rich in cell bodies, in particular the interganglionic regions (fig. 3a). In Sepioteuthis stellate ganglion, the antibody stained material which seemed to surround and adhere tightly to the large neuronal cell bodies (fig. 4a). In contrast, anti-NF stained the cytoplasm of these cell bodies (fig. 4b). In sections of Octopus axial nerve cords, anti-NF stained mostly the neuropil and nerves (fig 3b). GFAP IMMUNOREACTIVITY IN CEPHALOPODS: Immunoblotting was also carried out using tissue homogenates from other uninjured cephalopods. We were able to detect the 140 KD protein in the axial nerve cords of Sepia. Results with Loligo stellate ganglion were not conclusive, probably because of an insufficient amount of tissue to work with. The results of experiments on various tissue sources are summarized in figure 5. We were not able to detect this protein in the central nervous systems of the cuttlefish or squid; this is in contrast to the case in Octopus. Discussion The level of differentiation and specialization of glial cells tends to be a function of the complexity of the nervous system of an organism (Cardone and Roots, 1990). At the level of cephalopods, the development of the gliovascular system (Young, 1971) is an indication of this trend of specialization and differentiation. In vertebrates, the glial cells have become highly specialized cells in function and structure. Although little is known about the properties of glia in cephalopods, the fibers that support the processes of the glial cells as described by Young (1970) may be related to glial filament proteins in mammals. GFAP has been shown to be the major structural protein in differentiated astrocytes (Eng, 1985). Using a polyclonal rabbit antiserum to human GFAP and immunochemistry techniques, we have shown cross reactivity of the antibody to a specific protein in Octopus central and peripheral nervous systems, as well as parts of the peripheral nervous system of other cephalopods. Furthermore, using immunohistochemistry, we were able to show that this protein does not co-localize with neurofilament protein, an indication that the protein may be associated with glial cells. The product we obtained was a 140 kD protein. This is unusual as variations in the molecular weight of mammalian GFAP (48-51 KD) are small (Dahl and Bignami, 1973). The size of the Octopus protein we obtained conflicts with the findings of Cardone and Roots, who detected three protein bands in Western blots with a monoclonal anti-GFAP (56, 49, and 37 KD). It is interesting that the molecular weights of the three protein bands sum up to approximately 140 KD. Although the reducing conditions of the gel sample buffer were similar in both studies, it is still possible that (i) the high molecular weight we obtained may be the result of aggregation of the three lower molecular weight bands, or (ii) the three smaller bands may be degradation products of the larger protein we found. To draw further connections between this 140 KD protein and vertebrate GFAP, there appears to be an increase in expression following injury. The role of GFAP in glial scarring is well established in vertebrates (Eng, 1988; Reier et al., 1983). This increase in the level of GFAP immunoreactivity may constitute a similar event in the Octopus. Although in the present study we did not wait long enough for the arms to regenerate, observations have confirmed that the arms are able to do so (Cousteau and Diole, 1973). Thus it seems that the increase in this GFAP-like protein expression does not present an obstacle to neural regeneration. It is possible that the high regenerative capacity of the nervous system of the Octopus may be related to the high molecular weight GFAP. In central nervous system tissue from cephalopods of presumably lower regenerative capacity, Sepia, Loligo, and Sepioteuthis, we could not detect the 140 KD protein by immunoblotting. In the peripheral arm nerves of the Sepia and the peripheral stellate ganglion of the Sepioteuthis, however, we were able to see cross reactivity with the antibody. Thus, there may be a link between this GFAP-like protein and vertebrate GFAP beyond shared antigenic properties. Structural and biochemical similarities may be reflected in functional similarities as well. Acknowledgements This work was made possible with valuable contributions from Dr. W. F. Gilly. We thank the laboratories of Dr. D. Epel and Dr. D. Powers for the use of equipment, Y. L. Lee for valuable advice, Bruce Hopkins for assistance in collecting animals, and Chris Patton for expertise in photography of the blots. The antibodies were generous gifts from Dr. Lawrence F. Eng. Literature Cited Anderson, M.J., Swanson, K.A., Waxman, S.G., and Eng, L.F. (1984) Glial fibrillary acidic protein in regenerating teleost spinal cord. J. Histochem. and Cytochem. 31(10): 1099-1106. (1989) Protein Brown, R.E., Jarris, K.L., and Hyland, K.J. measurement using bicinchoninic acid: elimination of interfering substances. Analytical Biochemistry. 180: 136-139. Cardone, B. and Roots, B. (1990) Comparative immunohistochemical study of glial filament proteins (glial fibrillary acidic protein and vimentin) in goldfish, octopus, and snail. Glia. 3: 180-192. Cousteau, J.Y. and Diole, P. (1973) Octopus and Squid, the Soft Intelligence. Doubleday and Co., Inc., Garden City, pp. 11-184. and (1973) Immunochemical and Dahl, Bignami immunofluorescence studies of the glial fibrillary acidic protein in vertebrates. Brain Research. 61: 279-293. Eng, L.F. and DeArmond S.J. (1983) Immunochemistry of the glial fibrillary acidic protein. In Progress in Neuropathology, vol 5, edited by H.M. Zimmerman. Raven Press, New York, pp. 19-39. Eng, L.F. (1985) Glial fibrillary acidic protein (GFAP): the major protein of glial intermediate filaments in diffrentiated astrocytes. J. Neuroimmunology. 8: 203-214. Eng, L.F. (1988) Astrocytic response to injury. In Current Issues in Neuronal Regeneration Research. Edited by P.J. Reier, R.P. Bunge, and F. Seil. Alan R. Liss, Inc., New York, pp. 247-255. Fredman, S.M. and Nutz, P.G. (1988) Regeneration of identified neurons and their synaptic connections in the central nervous system of aplysia. Amer. Zoology. 28: 1099-1108. Galigher, and E.N. (1964) Essentials of A.E. Kozloff, Microtechnique. Lea and Febiger, Philadelphia, pp. 212-256. Laemmli, U.K. (1970) Cleavage of structure proteins during the assembly of the head of the bacteriophage T4. Nature. 227: 680-685. Reier, P.J., Stensaas L.J., Guth, L. (1983) The astrocytic scar as an impediment to regeneration in the central nervous system. In Spinal Cord Regeneration. Edited by C.C. Kao, R.P. Bunge, and P.J. Reier. Raven Press, New York, pp. 163-195. Stafford, C.A., Shehab, S.A.S., Nona, S.N., and Cronly-Dillon, J.R. (1990) Expression of glial fibrillary acidic protein (GFAP) in goldfish optic nerve following injury. Glia. 3: 33-42. Sternberger, L.A., Hardy, P.H. Jr., and Cuculis, J.J. (1970) The unlabeled antibody enzyme method of immunohistochemistry preparation and properties of stable antigen-antibody complex (horseradish peroxidase-antiperoxidase) and its use in the identification of spirochetes. J. Histochem. Cytochem. 18: 315-333. Sternberger, L.A. and Sternberger, N.H. Monoclonal (1983) antibodies distinguish phosphorylated and nonphosphorylated forms of neurofilaments in situ. Proc. Natl. Acad. Sci. USA. 80: 6126- 6130. Towbin, H., Staehelin, T., and Gordon, J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedures and some applications. Proc. Natl. Acad. Sci. USA. 76: 4350-4354. Windle, W.F., Clemente, C.D., and Chambers, w.W. (1952) Inhibition of formation of glial barrier as a means of permitting a peripheral nerve to grow into the brain. J. Comp. Neurology. 96: 359. Young, J.Z. (1971) The Anatomy of the Nervous System of Octopus Vulgaris. Clarendon Press, Oxford, pp. 45-59, 531-594. B. GFAP NE A. 200- 97.4- 69- 46- 30- 2 3 45 Figure 1. (a) GFAP staining of protein blot of Octopus left optic lobe (lane 1); right optic lobe (lane 2); axial nerve cord (lane 3); rock fish (Sebastes mystinus spinal cord (lane 4); mouse brain (lane 5). (b) GFAP and NF staining of protein from axial nerve cord of Octopus. The GFAP band does not correspond with any of the bands stained by anti-NF. A. Anti-GFAP B. Total Protein 200- 97.4- 69- 46- 30 2 Figure 2. (a) GFAP staining of proteins from Octopus axial nerve cords after injury (lane 1) and before injury (lane (b) Coomassie stained total protein from (a) show lower amounts of protein loaded for the injured tissue and the bands corresponding to the GFAP bands. Figure 3. (a) GFAP staining of longitudinal section of Octopus axial nerve cord showing strong staining in the interganglionic region (arrows) and the cellular layer. This section was not counterstained. (b) NF staining of the same tissue with staining strongest in the neuropil. Figure 4. (a) GFAP staining of Sepioteuthis stellate ganglion showing staining of possible glial cells that extend processes to wrap around the large neuronal cell bodies (arrows). (b) NF staining of same section showing staining of the cytoplasm of the cell bodies. GFAP IMMUNOREACTIVITY IN VARIOUS CEPHALOPOD NERVOUS SYSTEM TISSUE SOURCES Brain Optic Stellate Axial Lobe Ganglion Nerve Cord Octopus rubescens Sepia officianalis Loligo opalescens + ? Sepioteuthis lessoniana Figure 5.