Proc. Nat. Acad. Sci. USA Vol. 71, No. 4, pp. 1188-1192, April 1974

Transfer of Newly Synthesized Proteins from Schwann Cells to the Squid Giant Axon (glia/neurons/secretion)

R. J. LASEK*, H. GAINERt, AND R. J. PRZYBYLSKI* Marine Biological Laboratory, Woods Hole, Massachusetts 02543 Communicated by Walle J. H. Nauta, November 28, 1973

ABSTRACT The squid giant axon is presented as a teins synthesized in the Schwann cells surrounding the axon model for the study of macromolecular interaction be- tween cells in the nervous system. When the isolated giant are subsequently transferred into the axoplasm. axon was incubated in sea water containing [3Hjleucine MATERIALS AND METHODS for 0.5-5 hr, newly synthesized proteins appeared in the sheath and axoplasm as demonstrated by: (i) radioautogra- Protein synthesis was studied in squid giant axons obtained phy, (ii) separation of the -sheath and axoplasm by extru- from live squid which were kept in a sea tank and used within sion, and (iii) perfusion of electrically excitable axons. hr of obtained The absence of ribosomal RNA in the axoplasm [Lasek, 48 capture. The giant axons were by decapitat- R. J. et al. (1973) Nature 244, 162-165] coupled with other ing the squid and dissecting the axons under a stream of run- evidence indicates that the labeled proteins that are found ning sea water. The axons, 4-6 cm long, were tied with thread in the axoplasm originate in the Schwann cells surrounding at both ends, removed from the mantle, and cleaned of ad- the axon. Approximately 50%70 of the newly synthesized hering connective tissue in a petri dish filled with sea water Schwann cell proteins are transferred to the giant axon. These transferred proteins are soluble for the most part under observation with a binocular dissecting microscope. and range in molecular size from 12,000 to greater than Axons were incubated for 0-4 hr at 18-20° in 1 ml of Milli- 200,000 daltons. It is suggested that proteins transferred pore-filtered (MPF) sea water containing either 10 or 100 from the Schwann cell to the axon have a regulatory role 1,Ci of L-[4,5-3H]leucine. The incubation was terminated by in neuronal function. rinsing the axon in a large volume of MIPF sea water at 40 for min. The was extruded from The supply of macromolecules from the neuron cell body to 5-10 axoplasm immediately the sheath a method described except that the axon is well established (1). However, some proteins by previously (8) a of was substituted for the rub- may be supplied to the axon by synthetic mechanisms which length polyethylene tubing ber roller. All of the data below were obtained exist at the level of the axon. Two possible sites have been presented by this method. alternate methods of the suggested for the local synthesis of axonal proteins: (i) pro- However, extruding with a roller or with fine teins are synthesized in the axoplasm (2), and (ii)- proteins axoplasm by compression forceps produced comparable results. The sheath and axoplasm were are synthesized in the Schwann cells surrounding the axon homogenized in 50 mM Tris- HCL buffer, pH 7.4. Aliquots and subsequently transferred to the axon (3). Singer has been the strongest proponent of the Schwann-cell-to-axon transfer of the homogenate were taken for the following analysis: hypothesis and has reviewed the literature regarding this total radioactivity; hot (950 for 30 min) trichloroacetic acid question (4). precipitable radioactivity measured on filters; protein mea- sured the method of et al. and in some cases The squid giant axon represents an ideal model for the study by Lowry (9); of axonal macromolecular synthesis because the axoplasm sodium dodecyl sulfate, gel electrophoresis (10) or isoelec- tric on In other experi- can readily be separated from the sheath by simple extrusion focusing (11) polyacrylamide gels. (5). Amino-acid incorporation into protein has been demon- ments the giant axons were prepared for radioautography at the end of the incubation. strated in the isolated giant axon (6, 7). Furthermore, Lasek et al. (8) have recently characterized the RNA which is present RESULTS AND DISCUSSION in the axoplasm of the squid giant axon and have found little Table 1 compares the amount of trichloroacetic acid-precipit- if any ribosomal RNA. Ninety-five percent or more of the found in and sheath after 120 RNA in the axoplasm is 4 S in size, apparently transfer RNA. able radioactivity axoplasm min of incubation. The axoplasm contains 5-24% of the These results appear to rule out protein synthesis in the axo- labeled proteins. The variation in the relative labeling of the plasm by generally accepted mechanisms, and an alternate axoplasm is not completely understood; however, it may be source of the labeled proteins seems probable. The findings a result of variations in the amount of connective tissue presented below are consistent with the hypothesis that pro- which remained attached to the sheath after dissection. Abbreviations: MPF, Millipore-filtered; AXM, acetoxycyclo- Puromycin and acetoxycycloheximide (AXM) inhibited heximide. the incorporation into both axoplasm and sheath by 80% * Present address: Department of Anatomy, Case Western Re- or more, whereas chloramphenicol had little effect on in- serve University, Cleveland, Ohio 44106. corporation. The inhibition produced by puromycin and AXM t Present address: Behavioral Biology Branch, National Institutes did not result from reduced precursor entry into the axon of Health, N.I.C.H.D., Bethesda, Md. 20014. because these drugs had no demonstrable effect upon the solu- 1188 Downloaded by guest on October 1, 2021 Proc. Nat. Acad. Sci. USA 71 (1974) Transfer of Proteins from Glia to Neurons 1189 TABLE 1. Incorporation of [3H]leucine into proteins in the axoplasm (A) and sheath (S) of the squid giant axon after 120 min of incubation cpm/jig of protein Chloram- Control AXM t Puromycin phenicol A S A S A S A S 122 (5.3 %)* 4490 2 123 2 201 158 2013 42 (6.6%) 1853 6 241 8 69 48 1340 130 (12%) 2170 24 1148 138 (24%) 1409 20(7.9%) 503 26 (4.7%) 1036 Mean 79 (10%) 1910 4 182 5 135 76 1500

* In the control column, the figures in parentheses represent the amount of total hot trichloroacetic acid-precipitable radioactivity found in the axoplasm and are expressed as a percentage of the total found in the sheath. The protein content of the axoplasm (A) and sheath (S) obtained from each giant axon is on the order FIG. 1. Radioautographs of squid giant axons incubated in I of 100 jig and 40 jig, respectively. ml of MPF sea water containing [Htlleucine (100 jiCi/ml). r'he t AXM (Acetoxycycloheximide) = 50 jg/ml; puromycin = 50 emulsion is saturated in some regions of the sheath because of the jig/ml; and chloramphenicol = 100 or 200 jig/ml. exposure time of 3 weeks. Marker bar equals 50 jim. (A) Portion of a control giant axon with smaller axons left attached (top of ble radioactivity in either the sheath or axon. These results figure). The axon was incubated for 120 min in [3Hlleucine plus suggest that the incorporation of radioactivity into proteins 50 jg/ml of acetoxycycloheximide and 50 jig/ml of puromycin to in the axon and sheath represents translation on cellular ribo- inhibit protein synthesis. (B) Giant axon treated-the same as in somes and that relatively little incorporation occurred in A except that AXM and puromycin were omitted. Note the mitochondria or bacteria. presence of silver grains over the axoplasm. (C) Giant axon with Radioautographs verify that the labeled proteins extruded smaller axons attached incubated for 180 min. Note that the with the axoplasm represent labeled proteins present in the smaller axons also have grains over the axoplasm. (D) Higher magnification of the axon shown in B. Note the concentration of TABLE 2. Comparison of radioautographic grain counts in the grains in the region of the Schwann cell layer (arrows) and in axoplasm and sheath of squid giant axions scattered cells in the outer layer of the sheath.

Minutes of incubation axoplasm, rather than a contaminant from the sheath (Fig. Section 60 120 180 1). Grain counts of labeled axons incubated for 60, 120, and 180 min indicate that the axoplasm contains 12-28% of the A Axoplasm 406 (28%) 398 (28%) 760 (18%) of Sheath 1434 1398 4290 counts in the sheath (Table 2). The percentage radioactivity B Axoplasm 202 (26%) 556 (12%) 2722 (23%) in the axoplasm was higher in the radioautographic analysis Sheath 768 4507 11621 (Table 2) than in the analysis of extruded axoplasm (Table 1). This difference probably results from the inclusion of A and B represent two different- sections from the same axon. connective tissue elements and some small axons which adhere Giant axons were isolated and incubated as described in the text. to the sheath in the extrusion experiments; in the radio- The incubation was terminated by immersion in 5% trichloro- autographic estimates these elements were systematically acetic acid containing excess unlabeled leucine (4°). The solu- disregarded when the grains were counted. Axons incubated tion was changed five times over a period of 6 hr, the axon ex- for 120 min with 50 /Ag/ml of AXM and 50 jg/ml of puro- tracted with two changes of ethanol at room temperature over- no grains above the background level night and imbedded in Epon by standard methods. All of the mycin contained grains in the axoplasm were counted at X 1000 magnification. (Fig. 1A). Grains in the sheath were estimated by counting the grains in a Local differences in incorporation were found when two 25-jim2 ocular micrometer which was moved around the sheath levels of the same axon were compared (compare A and B twice so that every seventh 25-jim2 field was counted. In the first in Table 2). This regional variation in the incorporation of traverse around the sheath the inner Schwann cell layer was proteins into the sheath was also found in single cross sections counted. In the second traverse the outer connective tissue layer of axons (Fig. 1B). It appears significant that a positive cor- was counted. The number of grains per jIm2 was calculated and relation was found between the number of grains in the axo- multiplied by the area of the sheath which was determined by plasm and the amount of label present in the adjacent region planimetry. Background was estimated in an equivalent area of the sheath. The lack of homogeneity in the labeling of of the section adjacent to the tissue. All values are corrected for in the background (which never exceeded 10% of the counts shown and the axon might result either from variation penetra- was much less for the higher values). The radioautographs were tion of the precursor into the sheath cells or from regional all coated with emulsion at the same time and exposed for one differences in the rates of synthesis in the sheath cells. A week. The emulsion was not saturated in any region of these gradient of radioactivity was found within the axoplasm preparations. which decreased with increasing distance from the sheath. Downloaded by guest on October 1, 2021 1190 Cell Biology: Lasek et al. Proc. Nat. Acad. Sci. USA 71 (1974)

MOLECULAR WEIGHT/1000 MOLECULAR WEIGHT/1000 68 43 26 12 1.5 212 130 68 43 26 12 1.5

1600

1400

1200

1000 I 600 . 800

400 600 O 200 400 2800- K0 200

212 130 68 43 26 12 1.5 0 0.25 0.5 0.75 1.0 1.15 212 130 68 43 26 12 '.5

2400

2000

o 600

200

800

400

,, 0 025 0.5 075 1.0 125 0 0.25 0.5 0.75 1.0 1.25 RELATIVE MOBLITY RELATIVE MOBLrrY FIG. 2. Gel electrophoretograms of radioactive proteins in the squid giant axon on 7.5% polyacrylamide gels. These axons were in- cubated in 1.0 ml of MPF sea water containing 100 lcCi of [3H]leucine for 180 min and the axoplasm and sheath separated as indicated in the text. The homogenates were separated into soluble and particulate fractions by centrifugation at 27,000 X g for 1 hr prior to solubilization in 1% sodium dodecyl sulfate. The relative mobility of molecular weight standards run in a parallel gel are indicated at the top of each figure for reference. Arrows on the sheath patterns indicate the location of major peaks in the profiles of the labeled axoplasmic proteins.

That is, the outer regions of the axoplasm near the sheath number of grains in the Schwann cell layer represents at cells exhibited more radioactivity than the more centrally least 1/2 of that in the entire sheath after 60 and 120 min of located regions. incubation. Therefore, the ratio of label in the axoplasm The radioautographs also indicated that the sheath is not and the Schwann cell layer (which probably represents the a homogeneous structure. Fig. iD illustrates that the in- cells interacting with the axon) is approximately 1:2. If the corporation of [3H]leucine into the sheath was most intense Schwann cells are responsible for the synthesis of the axonal in the thin stratum which directly surrounds the axon, and proteins, approximately 50% of the newly synthesized proteins in scattered connective tissue cells in the outer layers of the would thus appear to be destined for transfer to the axon. sheath. This distribution of the radioactivity is consistent A bundle of smaller axons was left attached to the giant with the ultrastructural architecture of the sheath (12). axon for the purpose of comparison in the radioautographs. The sheath consists of at least three primary strata: (i) an It is worth noting that grains were found in the axoplasm inner layer of Schwann cells adjacent to the axon, (ii) an of these smaller axons (Fig. 1C). Therefore, the appearance of intermediate basement membrane, and (iii) an outer con- labeled proteins in the axon is not unique to the giant axon nective tissue layer. The Schwann cell layer adjacent to the in the squid, but appears to be a more general phenomenon in axon contained the highest concentration of radioactivity this mollusc and other species (3, 13). of any region of the giant fiber. The structure defined as the When the homogenates of sheath and axoplasm were cen- sheath in the extrusion experiments contains both the inner trifuged at 27,000 X g for 60 min, approximately 80% of and outer layer. However, in the radioautographs it was pos- the axoplasmic proteins were contained in the supernatant sible to estimate the amount of radioactivity present in the while only about 30% of the sheath proteins remained in the Schwann cell region (layer adjacent to the axon) and the por- supernatant. This differential solubility of axoplasmic and tion of the sheath outside the basement membrane. The sheath proteins, reported previously (7), indicates that labeled Downloaded by guest on October 1, 2021 Proc. Nat. Acad. Sci. USA 71 (1974) Transfer of Proteins from Glia to Neurons 1191 axoplasmic proteins are present either in the soluble fraction with other amino acids such as arginine (16), leucine is a of the axoplasm or in very small vesicles. rather poor substrate for NH2-terminal addition in eukaryotic The soluble and insoluble proteins of the axoplasm and cells. (iii) NHrterminal addition has never been demonstrated sheath were analyzed by sodium dodecyl sulfate-polyacryl- in whole cells. (iv) The correspondence in the electrophoretic amide disc gel electrophoresis (Fig. 2). The profiles of the profiles from axoplasm and sheath, respectively, indicates labeled proteins were complex, exhibiting a wide spectrum that these proteins were labeled by a similar mechanism. of labeled peaks in both axoplasm and sheath. Although the (v) When isolated axoplasm was incubated in a small drop of profiles were not identical, certain similarities are apparent MPF sea water, no [3H]leucine was incorporated into protein when the axoplasm and sheath are compared. The major (personal, unpublished observation). peaks found in the soluble axoplasmic proteins appear to If proteins synthesized in sheath cells are indeed transferred correspond to minor peaks in the soluble sheath proteins. into the axoplasm, then it might be expected that labeled The profile of the soluble sheath proteins was dominated by proteins would appear in the sheath prior to their appearance a peak with a molecular weight of 50,000. This peak was not in the axoplasm. We undertook a preliminary kinetic analysis pronounced in the soluble proteins from the axoplasm, pos- of labeled proteins in sheath and axoplasm of giant axons sibly indicating selectivity in the proteins which migrate from incubated for 15, 30, 45, and 120 min. These preliminary ex- the sheath cells to the axon. In the case of the proteins as- periments, however, failed to resolve any difference in the sociated with the pellet, the sheath and axoplasm both con- rate at which proteins appeared in the sheath and axoplasm. tained a high-molecular-weight peak (approximately 160,000). Labeled proteins appeared in the axoplasm and sheath within The two major peaks in the sheath pellet did not corre- the first 30 min of incubation. A critical study of the period spond exactly to the two peaks found in the axoplasm. between 15 and 45 min of incubation by pulse-chase analysis The significance of comparisons between the sheath and axo- will be needed to determine whether the kinetics of incorpora- plasm pellets is difficult to determine, because the axoplasm tion are consistent with the notion of protein-transfer from pellet contained a relatively small amount of radioactivity, Schwann cells to the axon. while the sheath pellet contained the major fraction of the Another line of evidence supporting the transfer hypothesis labeled sheath proteins. comes from experiments employing the perfusion method of The soluble proteins of sheath and axoplasin were also Lerman et al. (27). In these studies, perfusion of the giant compared by isoelectric focusing on polyacrylamide gels. axon with artificial axoplasm while the axon was immersed The profiles obtained for axoplasm and sheath were similar. in sea water containing 50 ,Ci/ml of [3H]leucine caused hot- A major peak of labeled proteins with an isoelectric point trichloroacetic-acid-precipitable radioactivity to appear in the of 7.5 is present in both the sheath and axoplasm (Lasek, perfusate (Gainer, H., Tasaki, I., Barker, J., and Lasek, R. R. J., Gainer, H., and Barker, J., unpublished). J., unpublished results). Throughout the duration of the per- Giuditta et al. (14) have demonstrated that [1311]albumin fusion the axons were able to conduct action potentials. placed in the incubation medium is actively taken up into These results rule out the possibility that the labeled proteins the isolated squid giant axon. It might be argued that the enter the axon through damaged spots or "holes" in the axon's occurrence of labeled proteins in the axoplasm of the squid plasma membrane. Addition of 50,4g/ml of pancreatic ribo- giant axon incubated in [3H]leucine represents a nonselective nuclease to the perfusion fluid had no effect on the appearance uptake of labeled proteins which are released by the sheath of labeled proteins in the perfusate, an observation suggest- cells into the incubation medium. Therefore, we undertook ing that RNA-directed protein synthesis in the axoplasm an analysis of the appearance of radioactive proteins in the cannot account for the appearance of labeled proteins in the medium after 120 min of incubation. Although trichloroacetic axoplasm. acid-precipitable proteins appeared in the medium, they repre- Of the possible mechanisms which might explain the ap- sented less than 50% of the radioactivity found in the axo- pearance of labeled proteins in isolated giant axons, a transfer plasm. The medium was dialyzed against distilled H20 at of proteins from the Schwann cells to the axon would seem 40, lyophilized, and analyzed for proteins by sodium dodecyl most consistent with the data at hand. The following evi- sulfate-disc gel electrophoresis. Two major peaks of radio- dence from the present experiments supports this hypothesis: activity were found, one which had a molecular weight of (i) Axoplasm contains little if any ribosomal RNA, but labeled 150,000 and another which migrated with a molecular weight proteins were found within the axoplasm by three separate of 12,Q00. No significant peaks were found between 12,000 methods (extrusion, radioautography, and perfusion). The and 150,000 daltons. This experiment suggests that the labeled appearance of these proteins in the axoplasm was inhibited proteins found in the axoplasm do not represent uptake of by puromycin and AXM. (ii) Perfusion of RNase through the labeled proteins present in the medium. axoplasm had no effect on the appearance of labeled proteins The large amount of 4S RNA present in the axoplasm (6) in the axoplasm. (iii) A direct quantitative correlation was raises the possibility that amino-terminal addition occurs observed between the regional distribution of labeled pro- actively in the axoplasm. A soluble enzymatic system which teins in, respectively, the sheath and the axoplasm. (iv) requires tRNA but not ribosomes, and adds certain amino The amount of labeled protein in the axoplasm was greater acids to proteins has been found in prokaryotes (15) and near the surface of the axon than in the center of the axon. eukaryotes (16). This mechanism does not offer a likely ex- (v) The labeled proteins found in the axoplasm corresponded planation for the radioactivity incorporated into axoplasmic to labeled proteins in the sheath with respect to molecular proteins of the squid giant axon, for the following reasons. weight and isoelectric point, and no labeled proteins were (i) While in the present experiments the incorporation of found in the axoplasm which were not represented in the leucine into axoplasmic proteins was inhibited by both puro- sheath. (vi) The labeled proteins found in the axoplasm are mycin and AXM, the soluble NH2-terminal addition system is primarily soluble. much less sensitive to these drugs (15, 16). (ii) In comparison Further, incidental, support of the transfer mechanism Downloaded by guest on October 1, 2021 1192 Cell Biology: Lasek et al. Proc. Nat. Acad. Sci. USA 71 (1974)

here postulated comes from the following observations: (i) transfer of "informational macromolecules" from one cell The squid giant axon actively takes up proteins from the sur- to another; however, such a mechanism has not been critically rounding medium, apparently by pinocytosis at a high rate demonstrated. The evidence presented here indicates that a (14). (ii) Electrophysiological studies on the squid giant axon substantial fraction of newly synthesized proteins can be indicate that the Schwann cells and the axon are electro- transferred from one cell to another in a mature nervous tonically coupled (17). (iii) Marker proteins such as peroxidase system. In future work it should be possible to employ the enters the axon from the extracellular space (18). (iv) Mor- squid giant axon to critically probe the mechanisms under- phological analysis of another giant axon (i.e., that of Myx- lying cell-to-cell transfer between glial cells and neurons. icola infundibulum) has demonstrated cytoplasmic continuity However, the major problem remains: whether or not pro- across the plasma membranes of the Schwann cell and axo- teins which are transferred from the Schwann cell to the axon plasm (Krishnan, N. and Lasek, R. J., unpublished results). convey biologically significant information. Incorporation of labeled amino acids into axonal proteins The authors wish to thank C. Dabrowski and C. Boehm for of axons severed from their cell bodies has been demonstrated their valuable technical asistance. This research was supported in various species, indicating that the transfer of proteins by grants from the National Institutes of Health, U.S.A., and from Schwann cell to axon may be a general phenomenon Muscular Dystrophy Association. R.J.L. is the recipient of a (2,3,13). NIH Research Career Development Award and R.J.P. is an The concept that macromolecules are exchanged between Established Investigator of the American Heart Association. adjacent cells finds support from several lines of evidence in 1. Lasek, R. J. (1970) Int. Rev. Neurobiol. 13, 289-324. other systems (19-22, 26). A substantial literature indicates 2. Koenig, E. (1965) J. Neurochem. 12, 343-355. 3. Singer, M. & Salpeter, M. M. (1966) J. Morphol. 120, 281- that pinocytosis may be an important physiological mecha- 316. nism for the uptake of exogenous proteins into cells (28), 4. Singer, M. (1968) in Ciba Foundation Symposium on Growth and this mechanism has been suggested as the means by which of the Nervous System, ed. Wolstenholme, G. E. W. & protein hormones effect their action on receptive cells (28). O'Connor, M. (J. & A. Churchill Ltd., London), pp. 200- This line of reasoning could be extended to the squid giant 215. 5. Baker, P. F., Hodgkin, A. L. & Shaw, T. I. (1962) J. axon: it is possible that the transfer of proteins from Schwann Physiol. (London) 164, 330-354. cell to axon may take place by some form of secretion (pos- 6. Fischer, S. & Litvak, S. (1967) J. Cell Physiol. 70, 69-74. sibly exocytosis) coupled with pinocytosis into the axon. 7. Giuditta, A., Dettbarn, W.-D. & Brzin, M. (1968) Proc. The large number of vesicles which is present in the axon Nat. Acad. Sci. USA 59, 1284-1287. 8. Lasek, R. J., Dabrowski, C. & Nordlander, R. (1973) near the axolemma is consistent with this possibility (12). Nature 224, 162-165. The possible functional role of the transfer of proteins from 9. Lowry, 0. H., Rosenbrough, N. J., Farr, A. L. & Randall, Schwann cell to axon can only be speculated upon. It is note- R. J. (1951) J. Biol. Chem. 193, 265-275. worthy that the motor axon of at least one invertebrate, the 10. Neville, D. M. (1971) J. Biol. Chem. 246, 6328-6334. can remain excitable for as 11. Gainer, H. (1973) Anal. Biochem. 51, 646-650. crayfish, electrically periods long 12. Villegas, G. M. (1969) Ultrastruc. Res. 26, 501-514. as months after being disconnected from its neuron cell 13. Edstr6m, A. & Sj6strand, J. (1969) J. Neurochem. 16, body (23), even though the axon shrinks significantly during 67-87. this period (24). Support of the crayfish axon might be pro- 14. Giuditta, A., Udine, B. D. & Pepe, M. (1971) Nature New vided in part by the transfer of proteins from the Schwann Biol. 29-30. 15. Kaji, A., Kaji, H. & Novelli, G. D. (1965) J. Biol. Chem. cell to the axon. On the other hand, axons of cephalopods in- 240, 1185-1191. cluding the squid are reported to undergo degenerative 16. Soffer, R. L. (1968) Biochim. Biophys. Acta 155, 228-240. changes after being severed from the perikaryon (25). Thus, 17. Villegas, J. (1972) J. Physiol. (London) 225, 275-296. if the Schwann cells provide the squid giant axon with physio- 18. Krishnan, N. & Singer, M. (1973) Amer. J. Anat. 136, this would seem insuf- 1-14. logically important molecules, supply 19. Loewenstein, W. R. (1973) Fed. Proc. 32, 60-64. ficient to maintain the axon. The interaction between the 20. Bennett, M. (1973) Fed. Proc. 32, 65-75. Schwann cell and the axon may represent a local regulatory 21. Bier, K. (1963) J. Cell Biol. 16, 436-440. mechanism rather than some "nutritive" process. This view 22. Kolodny, G. M. (1972) J. Cell Physiol. 79, 147-150. seems most consistent with the widely held "neuron doctrine" 23. Hoy, R. R., Bittner, G. D. & Kennedy, D. (1967) Science 156, 251-252. and the substantial evidence which indicates that the neu- 24. Norlander, R. H. & Singer, M. (1972) Z. Zellforsch. 126, ron's cell body is the primary source of axonal macromolecules 157-181. (1). 25. Sereni, E. & Young, J. Z. (1932) Pubbl. Staz. Zool. Napoli A large body of evidence indicates that during develop- 12, 173-208. ment interactions between cells are involved 26. Grafstein, B. & Laureno, R. (1973) Exp. Neurol. 39, 44-57. specific adjacent 27. Lerman, L., Watanabe, A. & Tasaki, I. (1969) Neurosci. directly in the regulation of morphogenesis, and that, sub- Res. 2, 71-96. sequently, these interactions are essential for the maintenance 28. Ryser, H. J.-P. (1970) Proceedings of the Fourth International of the differentiated state in mature organs. It has been sug- Congress on Pharmacology, (ed. Eigenmann, R. (Schwabe & gested that these interactions between cells involve the Co., Basel, Switzerland), Vol. 3. Downloaded by guest on October 1, 2021