Molecules that become redistributed during regeneration of the leech central nervous system

A key question in regeneration of the nervous system concerns the identity of factors required for growth and synapse formation to occur. To what extent does the target, by virtue of secreting diffusible factors, play a role and to what extent do environmental factors encountered by the growing axons direct regeneration? It is now clear that cell-adhesion molecules as well as extracellular matrix molecules can appear at the site of a lesion coincident with regrowth and synapse formation (Martini and Schachner, 1988; Gatchalian et al. 1989; Paschke et al. 1992).

One such molecule, which has been studied extensively in vertebrates as well as in the leech, is laminin. It is a glycoprotein of the extracellular matrix and is an effective substratum for a wide range of neurones in culture (Manthorpe et al. 1983). The domains responsible for its effects on neurite outgrowth and neuronal survival have been identified (Edgar et al. 1984). In the central nervous system (CNS) of the adult leech, laminin has a distribution characteristic of a basement-membrane protein. In the nerve connectives, which contain the axons running from one ganglion to the next, laminin is restricted to the capsule tissue surrounding the nerve fibres. Thus, neurones are not in contact with laminin in the CNS of the adult leech. However, when connectives are severed, resulting in massive axonal damage, glial cell injury and the development of a neuropile (Fernández and Fernandez, 1974), it has been shown that leech laminin becomes closely associated with areas of neurite growth at the site of lesion (Masuda-Nakagawa et al. 1990, 1993).

An interesting question concerns the involvement of proteins other than laminin in axonal regrowth. Initiation of growth, axon elongation, termination of growth and synapse formation might be directed by a variety of molecules. Accordingly, a major goal of this study was to look for the presence of molecules other than laminin at the site of lesion during axonal regeneration. Two monoclonal antibodies, raised against proteins of the leech CNS, have been identified that recognize antigens appearing at different times during regeneration.

Leeches, Hirudo medicinalis (L.), were purchased from Ricarimpex (France) or from Biopharm (England) and maintained at 16 °C in artificial pond water.

Ganglionic chains were dissected and put into 10 mmol l^-1 Tris–HCl buffer, pH 7.4, containing 2 % Triton X-100 (E. Merck GmbH, Darmstadt, Germany) and protease inhibitors (2 mmol l^-1 phenylmethylsulphonylfluoride, 5 mmol l^-1N-ethylmaleamide, 1 µg ml^-1 antipain, 0.5 µg ml^-1 leupeptin, 1 µg ml^-1 pepstatin, 5 µg ml^-1 aprotinin; Boehringer Mannheim). The ganglia were homogenized in a 2 ml Dounce homogenizer using a mechanical rotor and rotated at 4 °C overnight. The following day, the insoluble material was centrifuged at 15 000 g for 3 min and washed three times with 50 mmol l^-1 Tris–HCl. The pellet was extracted in 20 mmol l^-1 Caps [3-(cyclohexyl-amino)-1-propanesulphonic acid], pH 11, containing 10 mmol l^-1 EDTA, 150 mmol l^-1 NaCl and protease inhibitors, for 18–24 h at 4 °C. 100 µl of Caps buffer was used per chain of ganglia. The extract was dialysed against Tris-buffered saline (TBS) (50 mmol l^-1 Tris–HCl, pH 7.4, 150 mmol l^-1 NaCl) before further use.

For fractionation on a glycerol gradient, 5 ml of extract corresponding to 50 ganglia chains was concentrated to a final volume of 0.75 ml using Sephadex G100. The extract was applied to a 14 ml 15 %–40 % continuous glycerol gradient in 0.2 mol l⁻¹ NH₄HCO₃, pH 8.0. The proteins were separated by ultracentrifugation at 302 500 g for 18 h at 4 °C in a Kontron TST 41.14 rotor and a Kontron ultracentrifuge. 0.5 ml fractions were collected from the bottom of the tube.

To 1 ml of dialysed Caps extract, CaCl₂, MgCl₂ and MnCl₂ were added to a final concentration of 1 mmol l^-1. This was added to 500 µl of Concanavalin A–sepharose (Pharmacia) equilibrated in the same buffer and incubated at room temperature (24 °C) for 1 h. Extensive washing preceded elution with 0.5 ml of 500 mmol l^-1 methylglucopyranoside.

Balb/c mice were purchased from Madörin AG Kleintierfarm, Füllinsdorf, Switzerland. 140 µl of fraction 14 of the glycerol gradient was suspended in the same volume of complete Freund’s adjuvant (Gibco Laboratories, Grand Island, NY) and injected intraperitoneally into a mouse. Four weeks later, the mouse was boosted by intraperitoneal injection of 150 µl of fraction 14 in incomplete Freund’s adjuvant. The mouse was killed 8 days later and the spleen removed. Spleen lymphocytes were fused with 5×10³ myeloma cells (known as ‘PAI’) and the hybridoma cells were grown as previously described (Chiquet et al. 1988). Instead of using macrophage feeder cells, interleukin-6 (Institute of Immunology, Basel, Switzerland) was added for initial growth. Culture supernatants were screened by an ELISA on microtitre wells (no. 3911; Falcon Labware, Oxnard, CA) coated with Caps extract diluted 100-fold. Seventy hybridoma lines were selected for further screening on cryosections of leech CNS and by Western blotting. Selected hybridomas were cloned by serial dilution and, after rescreening, the clones (one confluent dish per mouse) were injected into pristane-primed (Aldrich Chemical Co., Milwaukee, WI) Balb/c mice to generate ascites tumours. Mice were killed approximately 10 days later and ascites fluid was collected. Immunoglobulins were precipitated with 50 % saturated NH₄SO₄ and dissolved in the original volume of buffer for subsequent coupling to CNBr-activated Sepharose. The identity of the immunoglobulin type was determined with an Ouchterlony test, using antibodies purchased from Nordic Immunology, Tilburg, Netherlands. Both mAb CT16 and mAb NP17 were identified as IgMs.

The crush operations were kindly performed by John G. Nicholls. Leeches were anaesthetized in 8 % ethanol for 20 min. Using forceps, the connectives were crushed on both sides of one ganglion per leech and allowed to regenerate for 1–34 days (Masuda-Nakagawa et al. 1990). The ganglia and adjacent crushed connectives were dissected out of the leech, frozen in Tissue Tek, sectioned and prepared for immunocytochemistry.

Ganglia and connectives of the leech CNS or leech body wall were embedded in Tissue Tek (Miles Laboratories, Inc., Lenaxa, KS) and frozen on dry ice. Sections 10 µm thick were cut using a cryostat (Slee Mainz type HS) and mounted onto gelatin-coated slides. Sections were preincubated with phosphate-buffered saline (PBS) containing 1 % bovine serum albumin (BSA) and then incubated for 3 h with mAb CT16 or mAb NP17 diluted 200-fold in PBS/1 % BSA. After washing the sections with PBS, FITC-labelled secondary antibody (Cappel) was added at a dilution of 1:100 in PBS/1 % BSA for 1 h. When sections were double-labelled, Hoechst 33258 dye was subsequently added for 10 min at a concentration of 5 µg ml^-1 PBS before washing with PBS and mounting in PBS containing 50 % glycerol and 0.5 % N-propylgalate. Samples were viewed using an Olympus microscope equipped with 40×1.3 NA and 10×0.4 NA objectives.

To examine whether mAb CT16 binds to collagen or to a collagen-associated protein, its staining pattern was analysed on cryosections treated with collagenase. Slides were preheated to 100 °C for 2 min, before adding bacterial collagenase (Clostridiopeptidase type VII, Sigma, Giessenhofen, Germany) at 40 units ml^-1 (units defined by Sigma). The tissue was treated with the enzyme in the presence of 5 mmol l^-1 CaCl₂ and then washed several times in PBS. mAb CT16 or control mAb CS39, a monoclonal antibody against an intermediate filament protein, was added and immunostaining of the slides was performed as described above.

SDS–PAGE was performed according to Laemmli (1970) with reagents from the Bio-Rad Laboratories (Richmond, CA). Samples of protein extracted from whole CNS and fractionated on a glycerol gradient were run with or without reduction by dithiothreitol (Sigma Chemical Co., St Louis, MO) on 3 %–15 % gradient acrylamide gels. Gels were stained with 0.2 % Coomassie Brilliant Blue. Relative molecular mass standards for the gels were obtained from Sigma Chemical Co. Rainbow markers were used for Western blots and purchased from Amersham International. Western blots were performed as described by Towbin et al. (1979). Proteins were transferred and the nitrocellulose (BA-85; Schleicher and Schuell, Inc., Dassel, Germany) was preincubated with PBS containing 5 % milk powder (SanoLait, Coop, Switzerland) and 0.05 % Tween-20 and then incubated for 3 h at room temperature (24 °C) in mAb CT16 or mAb NP17 (1:200). After washing the filter in PBS, 5 % milk powder and 0.05 % Tween-20, it was incubated with the secondary antibody goat anti-mouse IgG (1:100; Cappel Laboratories, Cochranville, PA) for 1 h at room temperature (24 °C). This was followed by a 1 h incubation in mouse peroxidase anti-peroxidase (PAP) (1:3000; Jackson Immunoresearch Laboratories, Avondale, PA) before the blots were developed with 4-chloro-1-naphthol (E. Merck GmbH) as described by Imhof et al. (1983).

Proteins extracted from leech CNS were separated on a 15 %–40 % glycerol gradient by ultracentrifugation and a fraction, number 14, running with a sedimentation coefficient of approximately 10S (10⁻¹² s) was selected for production of monoclonal antibodies. Two other proteins of the leech CNS, laminin and tenascin, were not present in this fraction, since they occur in fractions with lower sedimentation coefficients (Masuda-Nakagawa and Nicholls, 1991; Masuda-Nakagawa and Wiedemann, 1992). After fusion, the hybridomas were screened for differences in staining patterns on cryosections of intact and regenerating nerve fibres of leech CNS. Two antibodies with distinctive properties, which selectively stained neuropile (mAb NP17) or connective tissue (mAb CT16), were selected.

To assess the properties of the NP17 antigen, a CNS extract was prepared for Western blotting. As seen in Fig. 1, mAb NP17 recognized two main bands with relative molecular masses of approximately 80×10³ and 60×10³. Running the antigen in the presence of reducing agent (dithiothreitol) did not result in a change in its electrophoretic mobility, indicating that the antigen was not linked by disulphide bonds. The antibody was tested on the various fractions of CNS extract separated by ultracentrifugation. As shown in Fig. 1A, the bands appeared in fractions 13 and 14, confirming the specificity of this antigen to the fraction used for immunization.

To test whether the NP17 antigen was glycosylated with mannosyl or glycosyl oligosaccharides, CNS extract was applied to a Concanavalin A–sepharose column, and the bound proteins were eluted with 0.5 mol l^-1 methylglucopyranoside. The CNS extract, the flow-through proteins and the proteins eluted with 0.5 mol l^-1 methylglucopyranoside were separated on a gel and transferred to nitrocellulose. Fig. 1B shows that the 80×10³ and 60×10³M_r bands are seen only in lanes a and b, and did not bind to Concanavalin A. These results indicate that mAb NP17 antigen may not be glycosylated and may thus represent an intracellular protein.

mAb NP17 recognized an antigen situated predominantly in the neuropile, the structure where synaptic connections are made. Slight staining was also observed in neuronal perikarya and within the fibre tracts of the connectives (Figs 2 and 3).

To determine whether the NP17 antigen was localized intracellularly or extracellularly, neurones were isolated from the CNS and plated on coverslips coated with Concanavalin A. The cells were then stained with mAb NP17 in the presence and in the absence of 1 % Triton X-100. There was a slight increase in staining after permeabilization with 1 % Triton X-100 (not shown) with immunoreactivity visible in neuronal cell bodies as well as in the neurites, suggesting that, under these conditions, the NP17 antigen has an intracellular location.

When regenerating fibres were analysed for NP17 immunoreactivity in crushed connectives, punctate staining was first seen at the border of the lesion after 2 days (Fig. 3). At 4 days, when regrowing fibres had crossed the lesion and when synapses had begun to be formed, increased staining was detected at the crush site. To indicate the site of lesion, sections were simultaneously stained with the Hoechst dye 33258, which labels the microglial cells that accumulate there (Morgese et al. 1983). These remain at the crush site for several weeks, but in time also spread along the connectives. In a 16-day regenerating connective, spread of immunoreactivity was observed through a wide area containing large numbers of microglial cells. A few days later, however, the staining in crushed connectives became indistinguishable from that of control connectives (not shown). Thus, the NP17 antigen, which is localized predominantly in the neuropile of the intact CNS, was localized within the connectives at the crush site.

Fig. 4 shows the staining pattern of mAb CT16 in a ganglion of the leech CNS. Immunoreactivity was present in connective tissue of the outer and inner neural capsules, as well as in the partitions that subdivide each ganglion into packets (Fig. 4A). At higher magnification, some label was also seen within the neuropile, especially in areas close to the inner capsule (Fig. 4B). This staining probably represents extracellular material that can be found surrounding the neural arborizations (Gray and Guillery, 1963; Coggeshall and Fawcett, 1964). Some openings were seen within the inner capsule (arrow Fig. 4B). These may represent the areas through which the apical processes of the neuronal perikarya enter the neuropile (Fernández, 1978). In Fig. 4C, finger-like structures are visible at the ganglion–connective junction zone. These are the connective tissue partitions, derived from the adjacent inner capsule of the packets, that funnel axons into distinct fibre pathways as they leave the ganglion.

mAb CT16 distribution was examined in other tissues of the leech. The body wall of leech is largely made up of three layers of muscles running circumferentially, obliquely and longitudinally. Staining this area revealed extracellular CT16 immunoreactivity surrounding single muscle fibres (not shown).

To determine whether the antigen might have an intracellular location in neurones, cells were stained with mAb CT16 in the presence and in the absence of 1 % Triton X-100. No staining was apparent in either permeabilized or non-permeabilized cells (not shown). Staining of a section through the outer capsule of a leech ganglion showed a mesh-like pattern, reminiscent of a collagen or collagen-associated antigen (Fig. 5A).

CT16 immunoreactivity disappeared in cryosections digested with collagenase (Fig. 5C), whereas a control mAb recognizing a cytoskeletal antigen of the leech CNS (T. E. Lüthi, D. L. Brodbeck and P. Jenö, in preparation) retained its staining pattern after similar treatment (Fig. 5D). These results suggest that mAb CT16 recognizes a collagen or collagen-associated protein of the extracellular matrix.

Since molecules of the extracellular matrix have been implicated in neuronal outgrowth and regeneration, crush experiments were performed to examine the staining pattern of mAb CT16 during regeneration of the leech CNS. Fig. 6 shows connectives stained with mAb CT16 1 day and 9 days after injury. No changes in CT16 immunoreactivity were observed compared with control connectives. Label remained confined to the connective tissue capsule surrounding the fibres. The crush site could be unambiguously recognized by the presence of microglial cells. After 10 days, punctate staining was first detected at the site of crush on sections incubated with mAb CT16. Thereafter, fibre-like immunoreactivity spread throughout the connective, as shown in a section of a connective 16 days after injury. Interestingly, in two instances the staining was observed to extend away from the crush site towards the nearest ganglion (Fig. 7). Smith et al. (1991) have proposed that non-neuronal elements of the nervous system provide the critical microenvironment for neuronal repair in insects. They observed migration of glial cells from the ganglion to the lesion site. Thirty-four days after the crush, axon tracts within the connectives were still stained with a fibre-like pattern. No changes in the staining pattern were observed within the ganglia.

Regeneration of the leech central nervous system involves a characteristic sequence of events. Microglial cells move to the site of injury immediately after damage and remain in contact with regrowing axons as the glial cells of the connective eventually degenerate (McGlade-McCulloh et al. 1989). Severed axons begin to sprout fine processes 2 days after injury, a time when the neurite-promoting molecule laminin has been shown to appear in regions where nerve outgrowth occurs (Wallace et al. 1977; Masuda-Nakagawa et al. 1990). At 4 days, fibres of S-cells meet across the lesion and form synapses on their severed segments (Carbonetto and Muller, 1977). Restoration of connections is achieved by 1–2 months after injury.

Little is known about the molecular events accompanying regeneration, and the aim of the present experiments was to analyse the molecular processes involved in nerve fibre regrowth. It seems likely that damage to the CNS induces local signals between cells, initiating a series of events that lead to functional recovery. The study presented here demonstrates that the appearance of molecules other than laminin is part of the response to nerve injury.

One molecule, the NP17 antigen, appears in the connectives at the site of lesion at increased levels after approximately 3 days and the immunoreactivity remains elevated for several weeks. In the intact animal, the NP17 antigen is restricted to the neuropile and only a very little immunoreactivity is observed within the fibre tracts of the connectives. The association of the NP17 antigen with the site of synaptic junctions and the appearance of increased immunoreactivity within the fibre tracts after 4 days is interesting. Fernández and Fernandez (1974) have shown that crushing connectives results in the induction of a neuropile at the site of injury. Whether this antigen plays a role in synapse stability or function remains to be determined. Two lines of evidence point to an intracellular location of the NP17 antigen. First, permeabilization of leech neurones with Triton X-100 resulted in increased staining, indicating that neurones may synthesize this molecule. Second, Concanavalin A did not bind to the antigen. This suggests that, as one would expect for an intracellular molecule, it may not be glycosylated. It will be of interest to examine the localization of this antigen at a subcellular level.

Another molecule, the CT16 antigen, is, like laminin, located in the extracellular matrix of the leech CNS. However, there are differences in the staining pattern shown by this antibody and by mAbs 203 and 206, which bind to leech laminin (Chiquet et al. 1988). The CT16 antigen is not restricted to the basement membrane; instead, it labels connective tissue in a more general way and staining is not found around the nerve cell bodies. The mesh-like staining of the outer connective capsule and the lack of staining on cryosections treated with collagenase suggest an association of this antigen with collagen. It cannot, however, be ruled out that this antibody is recognizing collagen itself. It is distinct from the laminin antigen in its temporal distribution during regeneration. Whereas laminin staining appears at the crush site after 2 days, when nerve fibres initiate regrowth, CT16 immunoreactivity remains unaltered for 10 days after lesion. Only then does staining occur at the sites of fibre regrowth. The reason for this late expression has yet to be determined. The cells that are activated to produce the newly appearing molecules have not been identified. Molecules of the extracellular matrix have been implicated in neuronal outgrowth and regeneration (Sanes, 1989). However, because of the nature of the immunoglobulin, it has not been possible to purify sufficient amounts of antigen to test for neurite-promoting properties.

A number of molecules have been implicated in nerve regeneration in other systems. Changes in the expression of tenascin during nerve repair in adult mouse sciatic nerve and in the chicken neuromuscular junction have been demonstrated (Daniloff et al. 1989; Martini et al. 1990). The application of antibodies directed against tenascin to regenerating frog neuromuscular junctions resulted in delayed or inhibited reinnervation (Mège et al. 1992). Levels of N-CAM and Ng-CAM increase in the neurones and Schwann cells surrounding the lesion site following damage to mouse sciatic nerve (Martini and Schachner, 1988). Another candidate cell adhesion molecule, neurolin, has been shown to be re-expressed in the fish retinal tectal system after optic nerve transection (Paschke et al. 1992) and significant levels of ciliary neurotrophic factor (CNTF) protein have been detected at sites distal to nerve lesions (Sendtner et al. 1992).

The NP17 and CT16 antigens described in this study add possible candidates to the list of molecules that appear in a specific spatio-temporal distribution during regeneration. Analysis of the distinct steps involved in regeneration and the characterisation of the various molecules involved may allow a more detailed understanding of this phenomenon.

I wish to thank Drs John Nicholls and Diana Neely for helpful suggestions and critical reading of the manuscript. This work was supported by a stipend from the Boehringer Ingelheim Fonds and by a grant from the Swiss Nationalfonds to J.G.N. (no. 312781489).