In order to determine the role of the extracellular matrix in regulating the directed growth of embryonic neurites, antisera to retina (a-RBL I and II), to pigment epithelium (a-PBL) and to glomerular (a-GBL) basal lamina were probed for an effect on the ordered extension of neurites. In the assays, retina explants from chick and quail were cultured on basal lamina from embryonic chick retina and pigment epithelium either in the presence of anti-basal lamina antisera or in the presence of the corresponding preimmune sera. In the presence of all anti-basal lamina antisera, normal extension of axons was greatly inhibited both on retina and on pigment epithelium basal lamina. The antisera affected the growth pattern and the morphology of the individual axons in two ways: in the presence of a-RBL I the short axons were less directed, developed more and longer side branches, and the lamellipodia of the growth cones were reduced in size compared to axons from control cultures. In the presence of a-RBL II and a-GBL, axons grew slowly out from the explants as very thick bundles, strikingly different from axons in control cultures. The antiserum to pigment epithelium basal lamina induced both strong fasciculation and disorganization of the linear fiber extension, being intermediate between the two types of effects observed after antiserum addition. The results suggest that adhesive matrix molecules in basal laminae have important functions in elongation, fasciculation and in the morphology of growing axons.

A variety of investigations has shown that the initial outgrowth of axons is not random, but rather that growth cones follow defined directions (Roberts and Taylor, 1982; Bentley and Caudy, 1983; Berlot and Goodman, 1984; Halfter et al. 1985; Jacobson and Huang, 1985; Eisen et al. 1986; Harris, 1986; Schubiger and Palka, 1986; Dale et al. 1987). In the avian embryonic retina, for example, a statistical analysis of the initial growth orientation of a large number of newly formed optic axons has revealed that the frequencies of directional errors of the axons follows a Gaussian distribution, with the majority of axons being polarized towards the optic nerve head (Halfter et al. 1985). A similar polarity of axons is found in the mammalian retina (Maffei and Perry, 1988). The mechanisms underlying the directed axonal growth are largely unknown; one hypothesis, however, proposes that nerve fibers are guided by gradients of axon-specific markers in the environment (Sperry, 1963; Bonhoeffer and Gierer, 1984). Although gradients of molecules have been detected in the retina of the chick (Trisler et al. 1981) and in the mouse (Constantin-Patton et al. 1986), a contribution of these gradients in axonal guidance has not been assessed as yet. Another possibility to explain the directional outgrowth of axons assumes the existence of pathways that promote nonspecifically neurite outgrowth (Lewis, 1978; Katz et al. 1980) and, independent from the growth-promoting components in the pathways, additional information that causes the orientation of growth. In the early avian embryonic retina, axons are found at the basal side of the neuroepithelium within an environment that consists of the endfeet of the ventricular cells and a basal lamina that delineates the vitreal side of the neuroepithelium. If one deflects the axons from this basal position, aberrant fiber tracts result (Goldberg, 1977) suggesting that information for directed fiber growth is located at the vitreal border of the tissue (i.e. where the endfeet of the ventricular cells and the basal lamina is found). The basal surface of the retina has recently been isolated (Halfter et al. 1987) and shown to consist of the vitreal basal lamina (inner limiting membrane) and a dense layer of vesicles that are derived from the endfeet of the ventricular cells, the natural environment of growing optic nerve fibers in vivo. The endfeet can be removed by detergent treatment leaving only the basal lamina. Experiments using these retina basal lamina preparations (both the endfeet monolayer as well as the denuded basal lamina) as a substratum for neurites from dorsal root ganglia and retina explants have shown that these matrices have far better neurite outgrowth-promoting properties than any other known matrix component (Halfter et al. 1987).Another basal lamina preparation was obtained from the embryonic pigment epithelium (Halfter,1988). This basal lamina, which is normally not in contact with nerve fibers in vivo, also promotes neurite outgrowth. Despite the growth-promoting properties of the basal laminae, guiding cues were not detectable in these preparations (Halfter et al. 1987), suggesting that the factors responsible for neurite growth and neurite orientation are independent from each other.

Does this mean that the composition of the substratum is only necessary for neurite elongation? In order to obtain a better understanding of the role of the extracellular matrix components involved in axonal outgrowth, antisera were raised against retina and pigment epithelium basal lamina preparations. Another antiserum was raised against glomerular basal lamina from the rat. In a number of functional assays, the effects of these anti-basal lamina antisera on the outgrowth of axons on basal lamina and commercially available extracellular matrix molecules was studied.

Basal lamina preparation

(1) Retina basal lamina preparation

For the axonal growth assays, basal laminae from embryonic day 7 (E7) chick and E6 quail embryos were prepared on polylysine-coated petriperm dishes (Haereus, Zuerich, Switzerland) as previously described (Halfter et al. 1987, 1988). Explants were cultured either on the endfeet of the ventricular cells that cover the basal lamina sheet, or on the denuded basal laminae that were obtained by treatment of the basal lamina preparations with detergent to remove the endfeet (Halfter et al. 1987). Alternating stripes of plain basal lamina and basal lamina covered with endfeet were made by scraping off the endfeet from the underlying basal lamina with a spatula. This scraping removed the endfeet but left the basal lamina intact as seen by immunocytochemistry using antilaminin antibodies. For the isolation of basal laminae as immunogen, the basal lamina preparation technique was slightly modified. Flat-mounted E7 chick retinae were placed with the vitreous surface on nitrocellulose filters (Sartorius, Goettingen, FRG). After 10 to 15 min attachment, the retina flat mounts were removed, leaving the vitreal surface of the retina attached to the nitrocellulose filters. The filters with the basal lamina preparations were fixed in 4% paraformaldehyde, dehydrated in a graded series of isopropanol and finally placed in 100% acetone. Acetone dissolved the nitrocellulose filters leaving the basal laminae preparations intact. The laminae were collected in acetone, washed several times for a total of 2 h and rehydrated in Ca2+-/Mg2+-free Hank’s solution (CMF).

(2) Pigment epithelium basal lamina preparation

For neurite outgrowth assays, the basal lamina of the pigment epithelium of the eye was isolated from E7 chick or quail embryos as described (Halfter, 1988). For immunization and for gel electrophoresis, basal laminae from pigment epithelium were obtained in the following way: dissected pigment epithelia were incubated in 2 % Triton X-100 for 2 min and then drawn up and down repetitively in a Pasteur pipette. This removed the epithelial cells, and the cell-free matrix could then be isolated. The preparations were washed in Triton solution for a total of 1 h, treated with SO/igml-1 DNase in CMF for 1 h and finally washed in CMF.

(3) Rat glomerular basal lamina preparation

The basal laminae were isolated from rat kidney as described (Bhan et al. 1978).

Antibody production

60–80 chick retina or pigment epithelium basal laminae were homogenized in CMF, mixed with an equal volume of complete Freunds adjuvant and injected subcutaneously into rabbits. The animals were boosted 4 weeks after the first injection and again 2 weeks after the second injection with the same amount of basal lamina homogenate with incomplete Freunds adjuvant. Serum was taken 10 days after the final boost. Two rabbits were injected with retina basal lamina plus ventricular cell endfeet (a-RBL I and II) and one was injected with pigment epithelium basal lamina (a-PBL). A purified IgG fraction (10 mg ml-1) of an antiserum to rat glomerular basal lamina raised in sheep was kindly provided by Dr D. Adu, Renal Research Laboratory, Queen Elisabeth Medical Center, Birmingham, UK). Antibodies to human plasma fibronectin and EHS-mouse tumor laminin were obtained from BRL (Basel, Switzerland).

Explant cultures

E6 chick and E5 quail retina strips were explanted either on basal laminae from chick retina and pigment epithelium, on laminin- or fibronectin-coated plastic (Halfter et al. 1983). The culture dishes (Falcon) were coated with 20 μg ml-1 human plasma fibronectin, or 20 μgml-1 chick embryo fibronectin (isolated from chick embryo fibroblast-conditioned medium by gelatin-sepharose affinity chromatography; kindly provided by Dr R. Chiquet-Ehrisman, FMI, Basel) or 20 μgml-1 EHS-mouse tumor laminin (Gibco/BRL, Basel, Switzerland) for 1 h. Mouse or rat retina explants were taken from E15 embryos, and dorsal root ganglia were isolated from E7 chick or quail embryos. Some explants from retina were labeled in vivo with the vital dye Dil (Honig and Hume, 1986) by placing the flat mounted retina before explantation onto an evenly spread layer of the dye. The explants were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal calf serum (serum-containing medium) or in DMEM with Imgml-1 bovine serum albumin (serum-free condition). For the culture of the dorsal root ganglia, 100 μg ml-1 7S NGF (Calbiochem, Lucerne, Switzerland) was added to the culture medium. The explants were fixed after 30h of incubation by adding 4% paraformaldehyde to the cultures.

Functional assays and histology

Antibody inhibition assays were performed by incubating the explants in the presence of various dilutions of the antisera. Control cultures were incubated with the corresponding preimmune sera and with anti-EHS-mouse tumor laminin antiserum (BRL). Outgrowth of axons from the explants was evaluated by measuring the average length of the axons (i.e. the general front of the axons measured at three different positions of the explants) from camera-lucida drawings. The means of at least 3 different explants from 3 different experiments were calculated. The percentage inhibition of axonal growth is blotted as mean axon length (in %) with the preimmune sera as 100%. Another control experiment was done using preabsorbed anti-basal lamina antiserum: 200//I culture medium with 1:10 diluted a-RBL I was incubated with seven E7 basal laminae, or 100 gg EHS-mouse tumor laminin at 37°C for Ih. The inhibitory effect of the preabsorbed antiserum were tested in a subsequent outgrowth assay.

Other ncuritc outgrowth assays were performed after preincubation of the basal laminae with the antisera; after a 1h preincubation with 50 μl of the undiluted antisera, the dishes were washed several times with culture medium for a total of 1 h, and were then used for the assay. In these cultures, no antisera were added during the neurite outgrowth assay. Binding of antibodies to the basal lamina preparations was confirmed by staining of the cultures with FITC-labeled goat anti-rabbit antiserum (Dianova, Hamburg, FRG). The explants were photographed by dark-field, phase-contrast or bright-field microscopy after silver staining (Halfter et al. 1986). For scanning electron microscopy, the explants were fixed in 2% glutaraldehyde in CMF overnight. The preparations were dehydrated, critical-point dried, and sputter coated according to standard procedures.

Immunoblotting

Homogenates of E7 chick pigment epithelium basal laminae (10 –15 basal laminae per slot) were suspended in sample buffer (Laemmli, 1970), boiled for 10 min, and separated by a 3·6–14% gradient SDS-PAGE. The proteins were electro-phoretically transferred to nitrocellulose filters (Towbin et al. 1979) and visualized with Ponceau red staining. Strips of the blot were blocked with 1 % gelatin, 10% FCS in PBS for 4h at 37°C and incubated overnight with the anti-basal lamina antisera (a-GBL; a-RBL I and II: 1:30; a-PBL: 1:1000) or anti-EHS laminin (1:200) in TENB-N (50mM-Tris, pH7·4, 5mM-EDTA, 150mM-NaCl, 0·25% BSA, 0-5% Nonidet p-40) and 5 % FCS at 20°C. The nitrocellulose filters were washed in TENB-N and incubated with 1:1000 diluted alka-line-phosphatase-conjugated swine anti-rabbit IgG (Dako-patts, Denmark) or rabbit-anti-sheep (Dianova, FRG) for 2 h in TENB-N. The filters were rinsed four times for 20 min each in TENB-N, and the alkaline phosphatase was visualized in 100mM-Tris (pH 8·8), 100mM-NaCl, 5mM-MgC12, 0·004% nitrobluetetrazolium (Sigma, St Louis), 0·015% 4-bromo-5-chloroindolylphosphate (Boehringer, Mannheim).

Anti-basal lamina antibodies

Antisera to embryonic chick or pigment epithelium basal laminae were obtained by injecting formaldehyde-fixed basal lamina preparations into rabbits. Two rabbits were inoculated with retina basal lamina plus the adherent ventricular endfeet (a-RBL I and II) and one with basal lamina from pigment epithelium (a-PBL). All three rabbits developed immunglobulins to the antigens as assayed by immunocytochemistry: basal laminae from the embryonic retina and pigment epithelium of both chick and quail retina were recognized by the antisera, but were not reactive with the pre-immune sera (Fig. 1A, C). The antisera stained the preparations at a dilution of 1:30–1:50 (a-RBL I and II) and at a dilution of up to 1:500 with the a-PBL. When the basal laminae were incubated with the antisera for one hour, washed several times, and then cultured for another 24 or 30 h in the absence of antisera, a strong staining was still obtained for all three antisera (Fig. 1B). Another antiserum to rat glomerular basal lamina (a-GBL) was raised in sheep and used as purified IgG fraction. It stained the chick retina and pigment epithelium basal laminae at a dilution of 1:10 to 1:30 (i.e. 1–0·3mgml-1); the staining was also detectable when the basal lamina preparations were washed after antibody incubations and cultured for another 24 h in the absence of antiserum. Axons from embryonic avian retina or from dorsal root ganglia explants were faintly stained (a-RBL I and II) or not stained (a-PBL; a-GBL; results not shown).

Fig. 1.

Fluorescence micrographs showing the staining of chick retina basal lamina preparations (A,B,C) and quail retinal tissue (D-G) with anti-basal lamina antisera. The nonfixed basal lamina preparations were incubated with the a-RBL I and the bound antiserum was visualized with a FITC-labeled secondary antibody (A). In B, the basal lamina was incubated with the antiserum, washed three times and incubated for another 24h in the absence of the antiserum. Subsequent staining of the preparation with a FITC-labeled secondary antibody revealed a strong positive signal. Preimmune serum did not label the basal lamina (C). To label the retinal tissue in situ, a-RBL I (E), a-GBL (F) and a-PBL (G) was injected into the eye of a living quail embryo, and after 12h the embryos were fixed. Sections of the entire head were incubated with a FITC-labeled secondary antibody. The antisera stained the retina in different patterns depending on the antiserum, but always most prominently on the vitreal side. Staining was also seen at the basal lamina of the lens and at the cornea (E-G). Injection of the preimmune serum did not label the retina (D). Bar: A-G: 100 μm.

Fig. 1.

Fluorescence micrographs showing the staining of chick retina basal lamina preparations (A,B,C) and quail retinal tissue (D-G) with anti-basal lamina antisera. The nonfixed basal lamina preparations were incubated with the a-RBL I and the bound antiserum was visualized with a FITC-labeled secondary antibody (A). In B, the basal lamina was incubated with the antiserum, washed three times and incubated for another 24h in the absence of the antiserum. Subsequent staining of the preparation with a FITC-labeled secondary antibody revealed a strong positive signal. Preimmune serum did not label the basal lamina (C). To label the retinal tissue in situ, a-RBL I (E), a-GBL (F) and a-PBL (G) was injected into the eye of a living quail embryo, and after 12h the embryos were fixed. Sections of the entire head were incubated with a FITC-labeled secondary antibody. The antisera stained the retina in different patterns depending on the antiserum, but always most prominently on the vitreal side. Staining was also seen at the basal lamina of the lens and at the cornea (E-G). Injection of the preimmune serum did not label the retina (D). Bar: A-G: 100 μm.

There was no labeling of fibronectin- or laminin-coated plastic with a-RBL II, and only a very faint staining of laminin-coated plastic by a-RBL I (results not shown). The a-PBL stained fibronectin-but not laminin-coated plastic, whereas the a-GBL stained the laminin-but not fibronectin-coated plastic.

Labeling of extracellular antigens in embryonic retina tissue in situ was investigated by in vivo injection of the antisera into the eyes of E5 quail embryos (Fig. 1D-G). The in vivo labeling was used because only extracellular matrix and cell surface-bound, but not intracellular, antigens are recognized. Binding of the injected antibodies to the eye tissue was demonstrated by incubating the sections with a labeled secondary antibody (Halfter and Chen, 1987). With the a-RBL I and II, strong staining of the entire retina was obtained with a more prominent labeling of the vitreal surface (Fig. 1E). Little cell surface staining but a strong labeling of the vitreal basal lamina were obtained using the a-GBL (Fig. 1F). A similar labeling of the vitreal basal lamina and a strong reaction with the vitreous body were observed using the a-PBL antibody (Fig. 1G). Injection of the preimmune sera did not label the tissue (Fig. 1D). Except for the a-PBL (see below), staining of fixed chick tissue with the anti-basal lamina antisera was not useful because the antibodies had to be applied at a high concentration (1:30–1:50) that always caused considerable background. a-PBL could be used to stain tissue sections at dilutions of 1:500 to 1:1000. In the chick and quail central nervous system, the antiserum selectively labeled basal laminae, including those from blood vessels, whereas in the embryonic mesenchyme the extracellular matrix was stained throughout.

For biochemical characterization of the antisera, Western blot analysis was performed (Figs 2, 3, 4). Since the very time-consuming preparation of retina basal lamina also uses organic solvents to remove the nitrocellulose filter supports and, since the sera showed identical effects on axonal outgrowth both using retina or pigment epithelium basal lamina (see below), the samples for electrophoresis were made from pigment epithelium basal lamina. The antisera stained a variety of proteins (Figs 2, 3, 4). Only the a-GBL showed a restricted labeling of mainly the area around Mr 180–200×103 (Fig. 3). The protein staining patterns were reproduced in at least 3 blots per antibody.

Fig. 2.

Western blot analysis of pigment epithelium basal lamina proteins (lane 1, 3, 5; 12–15 basal lamina preparations per slot), or EHS-mouse tumor laminin (LN; lane 2, 4; 5 μg) using a-RBL I (lane 1, 2), a-RBL II (lane 3), or anti-EHS mouse tumor laminin antiserum (lane 4, 5). The position of relative molecular mass markers are indicated by arrows. With a-RBL I, a cross-reactivity with the 400K laminin chain but not with the 200K band (lane 2), was observed. Anti-laminin antibodies stained a triplet around 190K in basal lamina samples (lane 5). The same antibody recognizes both the 400 and the 200K band in samples from EHS-mouse laminin (lane 4; note that the laminin concentration in lane 2 is 10 times higher than in lane 4. The stars label the nonspecific staining of the immunoblots due to high antibody concentration.

Fig. 2.

Western blot analysis of pigment epithelium basal lamina proteins (lane 1, 3, 5; 12–15 basal lamina preparations per slot), or EHS-mouse tumor laminin (LN; lane 2, 4; 5 μg) using a-RBL I (lane 1, 2), a-RBL II (lane 3), or anti-EHS mouse tumor laminin antiserum (lane 4, 5). The position of relative molecular mass markers are indicated by arrows. With a-RBL I, a cross-reactivity with the 400K laminin chain but not with the 200K band (lane 2), was observed. Anti-laminin antibodies stained a triplet around 190K in basal lamina samples (lane 5). The same antibody recognizes both the 400 and the 200K band in samples from EHS-mouse laminin (lane 4; note that the laminin concentration in lane 2 is 10 times higher than in lane 4. The stars label the nonspecific staining of the immunoblots due to high antibody concentration.

Fig. 3.

Western blot analysis using a-GBL antibodies. Proteins from pigment epithelium basal lamina proteins (lane 1 and 3) and EHS-mouse tumor laminin (lane 2 and 4) were stained by an a-GBL antiserum (lane 1 and 2) or by anti-mouse laminin antiserum. Note that the anti-basal lamina antiserum mainly recognizes proteins at 190K that migrate at the same height as the laminin immunoreactivity in the basal lamina preparation.

Fig. 3.

Western blot analysis using a-GBL antibodies. Proteins from pigment epithelium basal lamina proteins (lane 1 and 3) and EHS-mouse tumor laminin (lane 2 and 4) were stained by an a-GBL antiserum (lane 1 and 2) or by anti-mouse laminin antiserum. Note that the anti-basal lamina antiserum mainly recognizes proteins at 190K that migrate at the same height as the laminin immunoreactivity in the basal lamina preparation.

Fig. 4.

Western blot analysis using the a-PBL antiserum. Pigment basal lamina preparations (lane 1 and 4), EHS-mouse tumor laminin (lane 2) and fibronectin from chicken fibroblasts (lane 3) were separated by SDS-PAGE. The Western blots were stained with either a-PBL antiserum (lane 1, 2, 3) or anti-laminin (lane 4). The a-PBL labeled many proteins, but showed no cross-reactivity with mouse tumor laminin. Fibronectin is clearly recognized. Another strip of the same gel is stained with anti-laminin antiserum to show the localization of this protein.

Fig. 4.

Western blot analysis using the a-PBL antiserum. Pigment basal lamina preparations (lane 1 and 4), EHS-mouse tumor laminin (lane 2) and fibronectin from chicken fibroblasts (lane 3) were separated by SDS-PAGE. The Western blots were stained with either a-PBL antiserum (lane 1, 2, 3) or anti-laminin (lane 4). The a-PBL labeled many proteins, but showed no cross-reactivity with mouse tumor laminin. Fibronectin is clearly recognized. Another strip of the same gel is stained with anti-laminin antiserum to show the localization of this protein.

Cross-reactivity with laminin and fibronectin was checked for all antisera. Using mouse tumor laminin as a sample, the a-RBL I reacted faintly with the 400K subunit of laminin, but the a-RBL II showed no staining of the laminin despite the fact that high amounts of laminin (3 μg per slot) were loaded on the gels. There was no cross-reactivity of the a-RBL I and II with fibronectin or proteins from fetal calf serum, even at very high concentrations of the proteins (3μg, 5 μ1 per slot). The a-PBL recognized fibronectin even at dilutions up to 1:1000 (Fig. 4) and the a-GBL stained the 200K band of mouse tumor laminin (Fig. 3).

The immunoblots were also incubated with antisera to EHS-mouse tumor laminin in order to identify laminin-like components in the samples (Figs 2, 3, 4). In pigment epithelium basal lamina, as in chick retina basal lamina (Halfter et al. 1987), two major and one minor laminin-like proteins, slightly below 200K, were detectable, but no staining was observed at 400K. Under nonreducing conditions, laminin bands were not detectable in the immunoblots. The mercaptoethanol-dependent extraction of laminin from embryonic basal lamina was confirmed by an immunofluorescence study (not shown). In corresponding controls using EHS-mouse tumor laminin, the same antisera recognized in the Western blots both the 400 and the 200K bands (Figs 2, 3), and, under nonreducing conditions, a single band of 1000K (not shown). In all of the blots, including the blots stained with the preimmune sera, there was a nonspecific staining at 50K when high concentrations of the primary antibodies were used.

Anti-basal lamina antisera disturb normal axonal extension on basal lamina

For functional tests, stripes of embryonic chick and quail retina, and dorsal root ganglia were cultured on chick retina or pigment epithelium basal lamina in the presence of various dilutions of the antisera. The controls were incubated in the presence of the corresponding preimmune sera at the same dilutions. The effects of the antisera to growing neurites could be classified into two types: one class of antisera disturbed the linear extension of axons and induced a disorganized fiber network close to the expiant (a-RBL I and partially a-PBL). The other type of effect was an inhibition of the fiber outgrowth combined with a strong fasciculation of the neurites (a-RBL II and a-GBL and partially a-PBL). The antiserum to pigment epithelium basal lamina induced both strong fasciculation and disorganized the linear fiber extension, being intermediate between both groups.

(1) Inhibition of linear axonal growth by a-RBL I

The antibody a-RBL I inhibited normal axonal extension on chick (Figs 5, 6) and quail retina and pigment epithelium basal lamina and gave the following morphological results in the inhibition tests: at a dilution of 1:10, retinal axons grew out from the explants for 50-–100 μm within 30 h of incubation compared to over 1 mm in control cultures, and formed a dense, disorganized network around the explant (Figs 5, 6). It was not possible to quantify neurite outgrowth, because of the high density of the fiber network around the explant; several layers of axons were growing on top of each other (Fig. 6). However, the high number of axons emerging from the explant indicates that the antisera did not inhibit axogenesis as such (compare Fig. 6B and C, 6D and E), but instead made the substratum inappropriate for neurite extension. Inhibition of axonal growth was demonstrated both on plain basal laminae and on preparations covered with the ventricular endfeet. Likewise, axonal growth inhibition was equal on chick and on quail basal lamina preparations, and it was independent of the presence or absence of fetal calf serum in the culture medium (Fig. 5A-D). The outgrowth inhibition was seen on several types of explants since the extension of axons from rat retina (Fig. 5E, F) or dorsal root ganglia was inhibited as well.

Fig. 5.

a-RBL I inhibit fiber outgrowth from chick retina explants growing on chick retina basal lamina. Explants were cultured in the absence (A,C) or in the presence of FCS (B,D). At an antiserum dilution of 1:10, fiber outgrowth was greatly inhibited (C, D), whereas incubation with the preimmune serum had no effect (A, B). Axonal growth was also inhibited when rat retina explants were used in the assay (compare E and F). The inhibition of fiber outgrowth was dependent on the concentration of the anti-basal lamina antiserum in the culture medium (G-I; the antiserum dilutions are indicated; C: control). Quantification of the inhibitory effect (see graph) showed a linear relationship between antiserum dilution and axonal growth inhibition (the inhibition is related to the 100% fiber outgrowth (=0% inhibition) from control cultures; mean length of axons). A-F, dark-field micrographs; G-l, silver-stained preparations. The explants were cultured for 30h. The enboxed area in C demonstrates the position of the SEM micrograph in Fig. 6A. Bar: A-F, 1 mm; G-l, 250 μm.

Fig. 5.

a-RBL I inhibit fiber outgrowth from chick retina explants growing on chick retina basal lamina. Explants were cultured in the absence (A,C) or in the presence of FCS (B,D). At an antiserum dilution of 1:10, fiber outgrowth was greatly inhibited (C, D), whereas incubation with the preimmune serum had no effect (A, B). Axonal growth was also inhibited when rat retina explants were used in the assay (compare E and F). The inhibition of fiber outgrowth was dependent on the concentration of the anti-basal lamina antiserum in the culture medium (G-I; the antiserum dilutions are indicated; C: control). Quantification of the inhibitory effect (see graph) showed a linear relationship between antiserum dilution and axonal growth inhibition (the inhibition is related to the 100% fiber outgrowth (=0% inhibition) from control cultures; mean length of axons). A-F, dark-field micrographs; G-l, silver-stained preparations. The explants were cultured for 30h. The enboxed area in C demonstrates the position of the SEM micrograph in Fig. 6A. Bar: A-F, 1 mm; G-l, 250 μm.

Fig. 6.

SEM micrographs showing corresponding areas of axon outgrowth from chick retina explants on chick pigment epithelial (A-C) or retinal basal lamina (D, E) in the presence of a-RBL I (A, B, D), and in the presence of the preimmune serum (C, E; serum dilution 1:10). An overview of the fiber outgrowth from the explant in the presence of the antiserum is shown in A (for localization of the micrograph see boxed area in Fig. 5C). In cultures with antiserum (A, B, D), the fiber density in the vicinity of the explant was very high, and the retinal axons formed a dense meshwork next to the explant. In control cultures (i.e. in the presence of preimmune serum; C, E), fibers were more parallel and less abundant. The explants were cultured for 30 h. Bar: A, 50 μm; B-E, 10 μm.

Fig. 6.

SEM micrographs showing corresponding areas of axon outgrowth from chick retina explants on chick pigment epithelial (A-C) or retinal basal lamina (D, E) in the presence of a-RBL I (A, B, D), and in the presence of the preimmune serum (C, E; serum dilution 1:10). An overview of the fiber outgrowth from the explant in the presence of the antiserum is shown in A (for localization of the micrograph see boxed area in Fig. 5C). In cultures with antiserum (A, B, D), the fiber density in the vicinity of the explant was very high, and the retinal axons formed a dense meshwork next to the explant. In control cultures (i.e. in the presence of preimmune serum; C, E), fibers were more parallel and less abundant. The explants were cultured for 30 h. Bar: A, 50 μm; B-E, 10 μm.

On basal lamina, the inhibition of axonal growth was always dependent on the dilution of the antiserum. For example, a maximum inhibition (94 %) by a-RBL I was obtained at a dilution of 1:10, corresponding to a final IgG concentration in the culture medium of approximately 0·6–l·2mgml-1. Axonal growth inhibition decreased linearly with higher dilutions of the antiserum whereby growth inhibition was still detectable at a dilution of 1:60 (Fig. 5G-I and graph in Fig. 5). Inhibition of axonal growth was observed at the same concentration ranges when the explants were cultured on basal lamina from the chick (Fig. 6A, B) and quail pigment epithelium. For example, at a dilution of 1:20, axon growth was inhibited by 84 % and at a dilution of 1:10 by 94 %. The inhibitory activity of a-RBL I could also be assayed by a preincubation procedure: when basal laminae from retina or pigment epithelium were preincubated with the anti-basal lamina antiserum, washed extensively, and then used for explant cultures, axonal growth was also inhibited: on both retina and pigment epithelium basal lamina, fiber outgrowth was inhibited by 75–80%. Preincubation with the pre-immune sera had no effect. The specificity of the antisera to basal lamina was further demonstrated by using fibronectin and laminin as growth substrates for axons. The extent of axonal growth on EHS-mouse tumor laminin was reduced by 30% (serum dilution 1:10; Fig. 7). Since retinal axons do not grow on fibronectin-coated plastic, dorsal root ganglia explants (that grow many nerve fibers on fibronectin) were used instead and incubated in the presence of the basal lamina antisera. a-RBL I inhibited axonal growth of the dorsal root neurites by 45 %, but did not interfere with axonal growth on fibronectin.

Fig. 7.

Quantification of the inhibitory activity of a-RBL I on the extension of chick retinal or dorsal root ganglion axons plated on chick retina basal lamina or on fibronectin- or laminin-coated plastic. With preimmune serum (C), 100% fiber outgrowth was obtained (vertical axis gives percent of the average neurite length, n = 6 explants, with the control as 100%). The labels on top of the column indicate the substratum used in the experiment; BL, chick retina basal lamina; LN, EHS-mouse tumor laminin; FN, fibronectin from chicken-fibroblast-conditioned medium). The antiserum was used at a dilution of 1:10. In the presence of a-RBL 1, axonal growth is greatly inhibited (aBL). Preincubation of the antiserum with seven retina basal laminae (aBL+BL) absorbed the inhibitory activity and restored fiber outgrowth to over 90 % of controls, whereas incubation with 100 μg laminin had a much smaller effect (aBL+LN). Likewise, outgrowth of retinal axons on laminin was inhibited by 30% in the presence of the antiserum. Similar data were obtained using dorsal root ganglia explants: fiber outgrowth was not inhibited when the axons were grown on fibronectin in the presence of the anti-basal lamina antiserum, but was strongly affected using basal lamina as substratum. The explants were cultured for 30 h.

Fig. 7.

Quantification of the inhibitory activity of a-RBL I on the extension of chick retinal or dorsal root ganglion axons plated on chick retina basal lamina or on fibronectin- or laminin-coated plastic. With preimmune serum (C), 100% fiber outgrowth was obtained (vertical axis gives percent of the average neurite length, n = 6 explants, with the control as 100%). The labels on top of the column indicate the substratum used in the experiment; BL, chick retina basal lamina; LN, EHS-mouse tumor laminin; FN, fibronectin from chicken-fibroblast-conditioned medium). The antiserum was used at a dilution of 1:10. In the presence of a-RBL 1, axonal growth is greatly inhibited (aBL). Preincubation of the antiserum with seven retina basal laminae (aBL+BL) absorbed the inhibitory activity and restored fiber outgrowth to over 90 % of controls, whereas incubation with 100 μg laminin had a much smaller effect (aBL+LN). Likewise, outgrowth of retinal axons on laminin was inhibited by 30% in the presence of the antiserum. Similar data were obtained using dorsal root ganglia explants: fiber outgrowth was not inhibited when the axons were grown on fibronectin in the presence of the anti-basal lamina antiserum, but was strongly affected using basal lamina as substratum. The explants were cultured for 30 h.

The specificity of the antiserum to only the underlying basal lamina substratum and not to growing nerve fibers was further demonstrated by preincubating the antiserum with several retinal basal lamina preparations prior to the axonal growth assay. By this procedure, the inhibitory activity of the antiserum was greatly diminished, and in the subsequent growth assay, axonal growth was inhibited by only 20 % compared to 90 % inhibition in cultures where the antiserum was not preabsorbed. However, when the antiserum was preincubated with 100 fig ml-1 EHS-mouse tumor laminin, axonal growth was not restored but was still inhibited by 70%. The results from the outgrowth inhibition tests for a-RBL I are summarized in Fig. 7.

Morphology of growing fibers in the presence of a-RBL

When retina or dorsal root explants were cultured on retina or pigment epithelium basal lamina in the presence of a-RBL I at submaximal concentrations (e.g. 50–80% growth inhibition), the orientation and the morphological appearance of the growing axons was impaired. Light microscopy showed that axons in the presence of the antiserum formed a dense fiber network around the explants, whereas, in the control cultures, fibers run straight and parallel from the explant tissue (see Figs 5, 6). In order to see single fibers within the huge fiber population, axons were labeled with the fluorescent vital dye Dil. This dye stains only few axons out of a large fiber population, enabling the tracing of single axons in the dense fiber net around the explants (Fig. 8A,B). From these investigations, it appeared that single axons in the presence of a-RBL I were less straight and less parallel than single axons from control cultures. Also, unusually many side branches were seen in cultures with the anti-basal lamina antiserum (compare Fig. 8C and D). These findings were confirmed by SEM studies and quantitative analysis of the frequency of side branches (number of side branches per 100 um axon length; Fig. 9; Table 1): axons that grew out in the presence of the antiserum had three times more side branches than axons in the control cultures, and the side branches were twice as long as the side branches of the controls; likewise, the surface areas of the growth cones in cultures with the antiserum were more than 50% smaller compared to the controls (Fig. 9; Table 1), but the growth cones had very long filopodia.

Table 1.

Altered morphology of axons and growth cones on basal lamina in response to a-RBL I. The measurements were made from SEM micrographs of 4 chick retina explants from two experiments

Altered morphology of axons and growth cones on basal lamina in response to a-RBL I. The measurements were made from SEM micrographs of 4 chick retina explants from two experiments
Altered morphology of axons and growth cones on basal lamina in response to a-RBL I. The measurements were made from SEM micrographs of 4 chick retina explants from two experiments
Fig. 8.

a-RBL I disturbed the orderly growth of chick retinal axons on chick retina basal lamina. The explant fibers had been labeled with the vital fluorescent dye Dil in order to trace single fibers within a dense population of axons (compare A and B, showing the same field with fluorescent microscopy (A) and phase contrast (B)). A stained fiber is indicated in both micrographs by arrowheads. The axons growing in the presence of anti-basal lamina antiserum (A,B,C) were less straight and had more side branches than the fibers in control cultures (D). The explants were cultured for 30h. Bar: 100 μm.

Fig. 8.

a-RBL I disturbed the orderly growth of chick retinal axons on chick retina basal lamina. The explant fibers had been labeled with the vital fluorescent dye Dil in order to trace single fibers within a dense population of axons (compare A and B, showing the same field with fluorescent microscopy (A) and phase contrast (B)). A stained fiber is indicated in both micrographs by arrowheads. The axons growing in the presence of anti-basal lamina antiserum (A,B,C) were less straight and had more side branches than the fibers in control cultures (D). The explants were cultured for 30h. Bar: 100 μm.

Fig. 9.

Alterations in the morphology of chick retinal axons growing on chick retina basal lamina in the presence of a-RBL I as shown by SEM. In the presence of the antiserum, growth cones had small lamellipodia (A), but long filopodia, whereas in control culture (B) the growth cones had large lamellipodia. Along the axon shaft, many side branches were seen when fibers were grown in the presence of the antiserum (C), whereas in the presence of the preimmune serum few side branches were found (D). Bar: 5 μm.

Fig. 9.

Alterations in the morphology of chick retinal axons growing on chick retina basal lamina in the presence of a-RBL I as shown by SEM. In the presence of the antiserum, growth cones had small lamellipodia (A), but long filopodia, whereas in control culture (B) the growth cones had large lamellipodia. Along the axon shaft, many side branches were seen when fibers were grown in the presence of the antiserum (C), whereas in the presence of the preimmune serum few side branches were found (D). Bar: 5 μm.

Despite the dramatic effects of the antisera on the linear elongation and the morphology of the nerve fibers, the principal direction of fiber outgrowth from the retina explants was not affected. As in the controls, most fibers grew out from that side of the explant strips that had been closer to the optic nerve head or fissure in the parent retina (see Halfter et al. 1983).

The morphological alterations in the axon morphology were not only observed with axons from the retina, but also with nerve fibers from the dorsal root ganglia (Fig. 10). In the control cultures, axons from dorsal root ganglia had 1·3 side branches per 100 μm axon length, whereas axons grown in the presence of a-RBL I had 11 side branches per 100 μm axon length, which corresponds to an 8·5-fold increase over the controls. In addition, the surface area of the growth cones of the sensory neurons was smaller in cultures with the antisera than in cultures with the preimmune serum (Fig. 10). The alterations in the growth pattern or the morphology of axons were also observed when the basal laminae were preincubated with the antisera and the subsequent assays run in the absence of the antibodies; however, alterations were not detectable when axons were grown on laminin or fibronectin (results not shown).

Fig. 10.

The morphology of chick dorsal root ganglia explant fibers grown on chick retina basal lamina in the presence (B, C) or absence (A) of a-RBL I (serum dilution 1:20). The preparations were silver-stained. Note that in control cultures (A) the growth cones are much larger than in cultures incubated in the presence of the antiserum (B, C). Also, the presence of the antiserum induces the outgrowth of many side branches that are absent in the controls. The explants were cultured for 30 h. Bar: 100 μm.

Fig. 10.

The morphology of chick dorsal root ganglia explant fibers grown on chick retina basal lamina in the presence (B, C) or absence (A) of a-RBL I (serum dilution 1:20). The preparations were silver-stained. Note that in control cultures (A) the growth cones are much larger than in cultures incubated in the presence of the antiserum (B, C). Also, the presence of the antiserum induces the outgrowth of many side branches that are absent in the controls. The explants were cultured for 30 h. Bar: 100 μm.

(2) Inhibition of normal axonal growth by a-RBL II and a-GBL

Upon exposure to a-RBL II and a-GBL, chick and quail retinal axons grew out from the explants on retina and pigment epithelium basal lamina as very thick and short bundles (Figs 11, 12). Quantitative measurements showed that a-RBL II at a dilution of 1:10 inhibited the growth of retinal axons both on chick and quail retina and pigment epithelium basal lamina by 80%. No inhibition was found with antiserum dilution of more than 1:50. a-GBL inhibited the extension of retinal fibers on the basal lamina preparations by 70 %, and no inhibition was detectable with a serum dilution of more than 1:40. As with the other antisera, a quantification of the amount of fiber growth in the presence of a-RBL II and a-GBL was not possible because of the different degree of axon fasciculation between test samples and controls.

Fig. 11.

a-RBL II inhibits the extension of chick retinal axons on embryonic chick retina basal lamina (compare A and B; C and D) but does not inhibit retinal fiber growth on laminin-coated tissue culture plastic (compare E and F). At a dilution of 1:15, the antiserum not only inhibited fiber elongation on basal lamina, but also caused axons to form thick bundles as shown at low (A) and high magnification (C). The control cultures were incubated with preimmune serum (B and D). The cultures were incubated for 24h. Axons growing on laminin were not disturbed in their growth by the a-RBL II (dilution 1:15; E) and no difference in the fiber outgrowth was obvious compared to cultures incubated with preimmune serum (F). On laminin, axons grew slightly slower than on basal lamina, therefore the cultures were incubated for 40 h. A, B, E, F are dark-field micrographs, C and D phase-contrast pictures. Bar: A, B, E, F: 500 μm; C, D: 100 μm.

Fig. 11.

a-RBL II inhibits the extension of chick retinal axons on embryonic chick retina basal lamina (compare A and B; C and D) but does not inhibit retinal fiber growth on laminin-coated tissue culture plastic (compare E and F). At a dilution of 1:15, the antiserum not only inhibited fiber elongation on basal lamina, but also caused axons to form thick bundles as shown at low (A) and high magnification (C). The control cultures were incubated with preimmune serum (B and D). The cultures were incubated for 24h. Axons growing on laminin were not disturbed in their growth by the a-RBL II (dilution 1:15; E) and no difference in the fiber outgrowth was obvious compared to cultures incubated with preimmune serum (F). On laminin, axons grew slightly slower than on basal lamina, therefore the cultures were incubated for 40 h. A, B, E, F are dark-field micrographs, C and D phase-contrast pictures. Bar: A, B, E, F: 500 μm; C, D: 100 μm.

Fig. 12.

SEM micrographs showing the morphological appearance of chick retinal axons on chick retinal basal lamina in the presence of a-RBL II (A, D), a-GBL (B, E) and in the presence of preimmune serum (C, F). The explant shown in B was cultured on basal lamina with ventricular endfeet, all others were cultured on denuded basal laminae. Both a-RBL II and a-GBL inhibit the extension of neurites and induce strong fasciculation compared to controls (A, B, C: low magnification; D, E, F: high magnification). The fasciculation is seen both on plain basal lamina as well as on preparations covered by the ventricular endfeet (B). Note that in B axon bundles begin to lift off from the substratum, indicating a weak adhesion of the bundles to the basal lamina preparation. The explants were cultured for 30h. Bar: A-C: 50 μm; D-F: 5 μm.

Fig. 12.

SEM micrographs showing the morphological appearance of chick retinal axons on chick retinal basal lamina in the presence of a-RBL II (A, D), a-GBL (B, E) and in the presence of preimmune serum (C, F). The explant shown in B was cultured on basal lamina with ventricular endfeet, all others were cultured on denuded basal laminae. Both a-RBL II and a-GBL inhibit the extension of neurites and induce strong fasciculation compared to controls (A, B, C: low magnification; D, E, F: high magnification). The fasciculation is seen both on plain basal lamina as well as on preparations covered by the ventricular endfeet (B). Note that in B axon bundles begin to lift off from the substratum, indicating a weak adhesion of the bundles to the basal lamina preparation. The explants were cultured for 30h. Bar: A-C: 50 μm; D-F: 5 μm.

The extension of retinal fibers was inhibited by a-RBL II and a-GBL both on denuded basal lamina as well as on the endfeet monolayer. This was confirmed in experiments using alternating stripes of plain basal lamina and basal lamina with endfeet (Fig. 13). In the control as well as in the test samples, axons given a choice between these alternating stripes did not show a preference for either, and in the presence of the antibasal lamina antisera, axonal growth was equally inhibited on the endfeet as well as on the denuded basal lamina stripes. a-RBL II did not inhibit the extension of retinal fibers on laminin- or fibronectin-coated plastic, whereas a-GBL did inhibit fiber extension on both EHS-mouse tumor laminin and fibronectin.

Fig. 13.

Chick retinal axons do not distinguish in distance and density of outgrowth between plain retinal basal lamina and the surface of ventricular endfeet. Alternating stripes were made by scraping the endfeet from the basal lamina (A; dark-field micrograph). Axons on both areas had the same length and the same density. The area enboxed in A is shown at higher magnification (B; phase-contrast micrograph). The arrowhead indicates the border between basal lamina without and with ventricular endfeet. Another preparation was processed for SEM (C, D). The border between areas with endfeet and without endfeet is shown in C. a-GBL induces strong fasciculation on both basal lamina with and without ventricular endfeet (D). Bar: A: 1mm; B: 100 μm; C and D: 20 μm.

Fig. 13.

Chick retinal axons do not distinguish in distance and density of outgrowth between plain retinal basal lamina and the surface of ventricular endfeet. Alternating stripes were made by scraping the endfeet from the basal lamina (A; dark-field micrograph). Axons on both areas had the same length and the same density. The area enboxed in A is shown at higher magnification (B; phase-contrast micrograph). The arrowhead indicates the border between basal lamina without and with ventricular endfeet. Another preparation was processed for SEM (C, D). The border between areas with endfeet and without endfeet is shown in C. a-GBL induces strong fasciculation on both basal lamina with and without ventricular endfeet (D). Bar: A: 1mm; B: 100 μm; C and D: 20 μm.

(3) Inhibition of axonal extension by a-PBL

The antiserum to embryonic pigment epithelium basal lamina inhibited growth of retinal axons both on retina and on pigment epithelium basal lamina slightly less effectively than a-RBL I: at a dilution of 1:10, axonal extension was inhibited by 72 % and at a dilution of 1:50 by 20 % ; no inhibitory activity was detectable at a dilution of more than 1:60. On retina and pigment epithelium basal lamina, a preincubation with a-PBL inhibited retinal fiber growth by 60%. On the basal laminae, a-PBL induced both a disorganization of the linear fiber extension as well as strong neurite fasciculation. On retina basal lamina, on average the antiserum induced strong fasciculation in 50 % of the retinal explants and, on pigment epithelium basal lamina, it affected 70 % of the explants. Both effects could also be detected sometimes on the same explant where, in the center of the fiber layer, a dense fiber network was seen whereas at the periphery of the explant strong fasciculation was prominent. On pigment epithelium basal lamina, the fascicles often had a very weak attachment to the substratum, and the fibers very often retracted upon movement of the culture dish.

There was no inhibition of fiber extension when the explants were cultured on laminin but there was a complete outgrowth inhibition when the explants were grown on fibronectin.

In the early embryo, growing axons are frequently found at the basal side of the neuroepithelium (Rager, 1980; Easter et al. 1984; Silver and Rutishauser, 1986; Scott and Bunt, 1986; Halfter and Chen, 1987; Anderson and Tucker, 1988) suggesting that important information for both extension and pathfinding of axons is localized at this site. The present investigation shows that antisera directed against the retinal glia endfeet, its apposed retinal basal lamina and basal laminae from other tissues cause dramatic changes in the elongation and appearance of axons in an in vitro test system. Three basal lamina preparations were used as immunogens for antiserum production: (1) the stroma-free vitreal surface of the retina, which consists of a laminin-rich basal lamina and a monolayer of endfeet of the ventricular cells that represents the natural substratum of growing nerve fibers in vivo, (2) the stroma-containing pigment epithelium basal lamina preparation that consists of a basal lamina sitting on a layer of mesenchymal extracellular matrix, and (3) the glomerular basal lamina that represents a stroma-free basal lamina from the kidney glomeruli. The preparations were immunogenic, and rabbits and sheep injected with this immunogen developed antisera that stained basal laminae in situ and could be used to perturb normal axonal growth on embryonic basal laminae from retina and pigment epithelium.

The perturbing activity of the antisera primarily to the substrata and not to the axons was confirmed since none of the preimmune sera had any effect on axonal growth. Also, the axonal growth was inhibited even in the absence of antiserum when the basal laminae were preincubated with the antiserum, and lastly, the inhibition effects were also seen with antisera that did not label axons at all (a-PBL; a-GBL). What are the molecular components that are recognized by the basal lamina antisera? Western blot analysis of pigment basal lamina homogenate with a-RBL and a-PBL showed a variety of labeled protein bands. The reactivity of the antisera with many proteins was due to the fact that the antibodies also recognize tissue and matrix components that are not from basal lamina and the surface of glia endfeet. For example, the pigment epithelium basal lamina preparation contains a relatively thick stroma layer that is immunogenic and probably responsible for the strong cross-reactivity of the antiserum with fibronectin. Further, the retina basal lamina preparation often has cellular contaminations at the periphery of the samples and part of the antiserum is also directed against intracellular components of the ventricular endfeet. The fact that many protein bands are recognized by the antisera may also explain why, in functional tests, a relatively high concentration of the antisera had to be used. It is, thus, very likely that only a minority of the IgG molecules are directed against the components critical for neurite outgrowth, and of these antibodies only a fraction have a high binding affinity to these molecules. Moreover, the high concentrations of antiserum may also indicate that not only one but several matrix components have to be blocked at the same time to obtain a major effect on neurite outgrowth. The experiments, on the other hand, show that the titer of the antiserum to stain tissue sections or label proteins in Western blots did not parallel its ability to inhibit axonal extension in the functional assays. For example, a-PBL stained sections and blots up to dilutions of 1:1000, but was less effective in the neurite outgrowth assays than a-RBL I which had to be used at dilutions of 1:30 for sections and of 1:70 for Western blots for successful staining.

Fewer protein bands of the immunoblots are labeled by a-GBL. The major reactivity was found around Mr 180−200×103, at the same location as laminin is detected by using anti-laminin antibodies. A variety of observations indicate that one of these laminin immunoreactivities is identical or related to the recently identified s-laminin isolated from rat glomerular basal lamina, s-laminin has a relative molecular mass of 190×103, has sequence homologies with the laminin B-l chain and is only extractable using S-S-bridge-reducing agents (Hunter et al. 1989). Laminin immunoreactivity from retina and pigment epithelium basal lamina has the same molecular weight and is also only extractable in the presence of reducing agents. It is therefore possible that a protein similar to s-laminin is one of the important factors for neurite outgrowth in the avian embryonic retina. The restricted homology of laminin from retina basal lamina with laminin from mouse tumor may also explain why EHS-tumor mouse laminin was not or only very faintly recognized by the anti-basal lamina antisera: a-RBL I labeled only the 400K band of the mouse laminin in immunoblots, and neither the a-RBL nor the a-PBL recognized laminin bands at all.

All anti-basal lamina antisera showed the same inhibitory effects when fibers were growing on the surface of the ventricular endfeet or on plain basal laminae without the ventricular endfeet indicating that the neurite growth-promoting molecules may occur both on the endfeet as well as in the basal lamina. Since the basal lamina is formed by secretion of matrix components from the basal processes of the neuroepithelial cells, it would be logical to assume that during early development on the surface of the endfeet of the ventricular cells and the retinal basal lamina a similar composition of matrix molecules is found. This is confirmed by previous immunocytochemical studies showing that a variety of matrix components are present on the endfeet and on the basal lamina (Halfter and Chen, 1987; Cohen et al. 1987; McLoon et al. 1988).

The antisera directed against pigment and glomerular basal lamina cross-reacted with constituents of embryonic retina basal lamina and vice versa and disturbed neurite extension indicating that basal laminae, even from tissues that are normally not innervated (e.g. pigment epithelium; glomeruli), may also have neurite outgrowth-promoting components that are identical or similar to those found in the embryonic retina basal lamina.

It is worth mentioning that, in the functional assays, the antisera were consistently less effective in neurite outgrowth inhibition when dorsal root ganglia instead of retinal explants were employed. Fibers from dorsal root ganglia are obviously less sensitive to alterations in the matrix by having more or different types of extracellular matrix receptors than fibers from retinal tissue.

When using antisera that are directed to an entire tissue element with many different constituents, the identification of single axonal outgrowth-promoting molecules is not possible. Rather, the major goal of this study was to see what kind of effects on neurite outgrowth can be observed when a matrix that normally promotes neurite outgrowth is covered with antibodies and made nonadhesive and whether this blockade has an effect on the directed outgrowth of nerve fibers.

Antisera to basal lamina affect the growth pattern and the morphology of growing axons

A variety of antisera are known to inhibit axonal growth in vitro and in vivo. Most of these antibodies, however, react with the surface of the axon and block receptor molecules for laminin or fibronectin (for review see Jessel, 1988; Rathjen, 1988). When assayed in vitro, these antibodies stop further axonal growth, or they induce defasciculation (Rutishauser et al. 1978; Rathjen et al. 1987; Neugebauer et al. 1988). One monoclonal antibody is known that can block axonal growth by interfering with the environment of axons. This antibody is directed to a laminin/heparan sulfate proteoglycan complex (Sandrock and Matthews, 1987a, b), and inhibits the growth of peripheral neurites in vitro and in vivo by about 20 –50 % ; however, morphological alterations of axons growing in the presence of this antibody have not been described.

The anti-basal lamina antisera described here not only inhibited axonal advance, but also disturbed the organization of growing neurites by inducing strong fasciculation, and altering the morphology of the individual fibers. The presence of antisera in the culture medium did not decrease the number of axons growing out from the explants (see Fig. 4A), but, since axons were unable to attach efficiently to the substrate, they meandered more frequently, they extended many side branches as focal attachment points, or they converged into thick fascicles using each other’s surface as an alternative substratum. Previous studies have shown that nonadhesive substrata induce the formation of long filopodia and decrease the size of the lamellipodia (Luduena, 1973; Letourneau, 1975; 1979; Roberts and Taylor, 1983). The reduced size of the growth cones and the many and long filopodia clearly show that, after treatment with the antisera, the basal laminae are less adhesive for growing neurites. All deviations from the normal appearance of the fibers can be explained as a reaction of the fibers when exposed to a nonadhesive substratum. The morphological changes that are observed suggest that the axons try to increase the attachment to the underlying substratum by putting out additional filopodia or to find conditions conducive to further growth (fasciculation; using other neurons as growth substratum).

The experiments also show that adhesion components of the extracellular matrix may have a regulatory function in the fasciculation of axons. A strong fasciculation results from either a strong adhesivity between neurites or a weak interaction of the neurites with the substratum, thus favoring axon-axon adhesion. Defasciculation of neurites predominates when fibers have a strong preference for the substratum, or when only a weak interaction between neurites exists. The degree of neurite fasciculation then would result from the ratio of axon-axon interaction, on the one hand, and from axon-matrix interaction, on the other hand.

Two types of alterations of neurite outgrowth were consistently observed. Whether either of these effects reflects the blockade of two or more separate basal lamina constituents or the blockade of different epitopes of one component is at the moment not clear.

The fact that the linear axonal extension is impaired when antibodies bind to the basal lamina indicates that, for normal axonal pathfinding, the substratum plays an important role. It has been proposed that growing axons are guided by gradients of chemical markers in the environment (Sperry, 1963; Bonhoeffer and Gierer, 1984). The present results indicate that the normal constitutive adhesion proteins in the basal lamina and on the surface of the ventricular cells possess some sort of organizing function, in that they allow the linear extension of the nerve fibers. Our antibody experiments, on the other hand, also show that the oriented outgrowth of the nerve fibers from only one side of the retinal explants is not disturbed and, therefore, the mechanism underlying the initial outgrowth of axons from the ganglion cells, as it is seen in vivo, is not dependent on the molecular composition of the substratum.

A variety of experiments have shown that axons can grow along the axonal fiber tracts aberrantly in reverse direction, both in the optic nerve (Bohn and Stelzner, 1981; Bunt and Lund, 1981; Halfter, 1987) and in the retina fiber layer (Halfter and Deiss, 1984) indicating that directional cues in the nerve fiber pathways are not present or not very efficient. Adhesion proteins in the axonal environment are therefore responsible for favorable conditions for rapid and straight neurite extension and for regulating fascicle formation, but not for the initial fiber orientation. A so far unknown mechanism causes the initial directed outgrowth of the nerve fiber from the ganglion cell. The substratum at the basal side of the neuroepithelium then allows the axons to pursue this initial orientation and fiber-fiber interactions of neighboring axons may help to correct minor orientation errors of the individual nerve fibers.

I would like to thank Ch. Bruecher (Ciba-Geigy, Basel) for technical assistance with the SEM, Dr D. Adu (Queen Elisabeth Medical Center, Birmingham, UK) for providing me with the a-GBL antibodies, and Drs S. Lambert, R. Tucker, D. Monard and Ch. Viereck (FMI, Basel) for critical review of the manuscript.

Anderson
,
H.
&
Tucker
,
R. P.
(
1988
).
Pioneer neurons use basal lamina as a substratum for outgrowth in the embryonic grasshopper limb
.
Development
104
,
601
608
.
Bentley
,
D.
&
Caudy
,
M.
(
1983
).
Navigational substrates for peripheral pioneer growth cones: Limb-axis polarity cues, limbsegment boundaries, and guidepost neurons
.
Cold Spring Harbor Symp. quant. Biol
.
48
,
573
585
.
Berlot
,
J.
&
Goodman
,
C. S.
(
1984
).
Guidance of peripheral pioneer neurons in the grasshopper: Adhesive hierarchy of epithelial and neuronal surfaces
.
Science
223
,
493
496
.
Bhan
,
A. A.
,
Schneeberger
,
E. E.
,
Collins
,
A. B.
&
McClusky
,
R. T.
(
1978
).
Evidence for a pathogenic role of cell-mediated immune mechanism in experimental glomerulonephritis
.
J. exp. Med
.
148
,
246
260
.
Bohn
,
R. C.
&
Stelzner
,
D. J.
(
1981
).
The aberrant retino-retinal projection during optic nerve regeneration in the frog. 1. Time course of formation and cells of origin
.
J. comp. Neurol
.
196
,
605
620
.
Bonhoeffer
,
F.
&
Gierer
,
A.
(
1984
).
How do axons find their target on the tectum?
TINS
7
,
378
381
.
Bunt
,
S. M.
&
Lund
,
R. D.
(
1981
).
Development of a transient retino-retinal pathway in the hooded and albino rats
.
Brain Res
.
211
,
399
404
.
Cohen
,
J.
,
Burne
,
J. F.
,
McKinlay
,
C.
&
Winter
,
J.
(
1987
).
The role of laminin and the laminin/fibronectin receptor complex in the outgrowth of retinal ganglion cell axons
.
Devi Biol
.
122
,
407
418
.
Constantin-Paton
,
M.
,
Blum
,
M.
,
Mendez-Otero
,
R.
&
Barnstable
,
C.
(
1986
).
A cell surface molecule distributed in a dorso-ventral gradient in the perinatal rat retina
.
Nature, Land
.
324
,
459
462
.
Dale
,
N.
,
Roberts
,
A.
,
Ottersen
,
O. P.
&
Storm-Mathisen
,
X. X.
(
1987
).
The development of a population of spinal cord neurons and their axonal projection revealed by GABA immunocytochemistry in frog embryos
.
Proc. R. Soc. Land. B
232
,
205
215
.
Easter
,
S. S.
,
Bratton
,
B.
&
Scherer
,
S.
(
1984
).
Growth-related order of the retinal fiber layer in goldfish
.
J. Neurosci
.
4
,
2173
2190
.
Eisen
,
J. S.
,
Myers
,
R. Z.
&
Westerfield
,
M.
(
1986
).
Pathway selection by growth cones of identified motoneurons in live fish embryos
.
Nature, Lond
.
320
,
269
271
.
Goldberg
,
S.
(
1977
).
Unidirectional, bidirectional and random growth of embryonic optic axons
.
Expl Eye Res
.
25
,
399
404
.
Halfter
,
W.
,
Newgreen
D. F.
,
Sauter
,
J.
&
Schwarz
,
U.
(
1983
).
Oriented axons outgrowth from avian embryonic retinae in culture
.
Devi Biol
.
95
,
56
64
.
Halfter
,
W.
,
Deius
,
S.
&
Schwarz
,
U.
(
1985
).
The formation of the axonal pattern in the embryonic avian retina
.
J. comp. Neurol
.
232
,
466
480
.
Halfter
,
W.
(
1987
).
Anterograde tracing of retinal axons in the avian embryo by low-molecular weight derivatives of biotin
.
Devi Biol
.
119
,
322
335
.
Halfter
,
W.
(
1988
).
Aberrant optic axons in the retinal pigment epithelium during chick and quail visual pathway development
.
J. comp. Neurol
.
268
,
161
170
.
Halfter
,
W.
&
Chen
,
S. F.
(
1987
).
Immunohistochemical localization of laminin, N-CAM, collagen type IV and T61 antigen in the Japanese quail retina by in vivo injection of antibodies
.
Cell and Tissue Res
.
249
,
487
496
.
Halfter
,
W.
&
Deiss
,
S.
(
1984
).
Axon growth in embryonic chick and quail retina whole mounts in vitro
.
Devi Biol
.
102
,
344
355
.
Haleter
,
W.
,
Diamantis
,
I.
&
Monard
,
D.
(
1988
).
Migratory behavior of cells on embryonic retina basal lamina
.
Devi Biol
.
130
,
259
275
.
Halfter
,
W.
,
Reckhaus
,
W.
&
Kroeger
,
S.
(
1987
).
Nondirected axonal growth on basal lamina from the embryonic avian retina
.
J. Neurosci
.
7
,
3712
3722
.
Harris
,
W. A.
(
1986
).
Homing behavior of axons in the embryonic vertebrate brain
.
Nature, Lond
.
320
,
266
269
.
Honig
,
M. G.
&
Hume
,
R. I.
(
1986
).
Fluorecent carbocyanine dyes allow living neurons of identified origin to be studied in longterm cultures
.
J. Cell Biol
.
103
,
171
187
.
Hunter
,
D. D.
,
Shah
,
V.
,
Merlie
,
J. P.
&
Sanes
,
J. R.
(
1989
).
A laminin-like adhesive protein concentrated in the synaptic cleft of the neuromuscular junction
.
Nature, Lond
.
338
,
229
234
.
Jacobson
,
M.
&
Huang
,
S.
(
1985
).
Neurite outgrowth traced by means of horseradish peroxidase inherited from neuronal ancestral cells in frog embryos
.
Devi Biol
.
110
,
102
113
.
Jessel
,
T. M.
(
1988
).
Adhesion molecules and the hierarchy of neural development
.
Neuron
1
,
3
13
.
Katz
,
M. J.
,
Lasek
,
R. J.
&
Nauta
,
H. J. W.
(
1980
).
Ontogeny of substrate pathways and the origin of the neural circuit pattern
.
Neurosci
.
5
,
821
833
.
Laemmli
,
U. K.
(
1970
).
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
.
Nature, Lond
.
227
,
680
685
.
Letourneau
,
P.
(
1975
).
Possible role of cell: substrate adhesion in neuronal morphogenesis
.
Devi Biol
.
44
,
77
91
.
Letourneau
,
P.
(
1979
).
Cell-substratum adhesion of neurite growth cones, and its role in enurite elongation
.
Expl Cell Res
.
124
,
127
138
.
Lewis
,
J.
(
1978
).
Pathways of axons in the developing wing: evidence against chemo-specific guidance
.
Zoon Suppl
.
6
,
175
179
.
Luduena
,
M. A.
(
1973
).
The growth of spinal ganglia neurons in serum-free medium
.
Devi Biol
.
33
,
470
476
.
Maffei
,
L.
&
Perry
,
V. H.
(
1988
).
The axon initial segment as a possible determinant of retinal ganglion cell dendrite geometry
.
Dev. Brain Res
.
41
,
185
194
.
McLoon
,
S. C.
,
McLoon
,
L. K.
,
Palm
,
S. L.
&
Furcht
,
L. T.
(
1988
).
Transient expression of laminin in the optic nerve of the developing rat
.
J. Neuroscience
8
,
1981
1990
.
Neugebauer
,
K. M.
,
Tomaselli
,
K. J.
,
Lilien
,
J.
&
Reichhardt
,
L. F.
(
1988
).
N-cadherin, N-CAM, and integrins promote neurite outgrowth on astrocytes in vitro
.
J. Cell Biol
.
107
,
1177
1187
.
Rager
,
G.
(
1980
).
Development of the retinotectal projection in the chicken
.
Adv. Anat. Embryol. Cell Biol
.
63
,
1
92
.
Rathjen
,
F. G.
(
1988
).
A neurite outgrowth-promoting molecule in developing fiber tracts
.
TINS
11
,
183
184
.
Rathjen
,
F. G.
,
Wolff
,
J. M.
,
Frank
,
R.
,
Bonhoeffer
,
F.
&
Rutishauser
,
U.
(
1987
).
Membrane glycoproteins involved in neurite fasciculation
.
J. Cell Biol
.
103
,
343
353
.
Roberts
,
A.
&
Taylor
,
J. S. H.
(
1982
).
A scanning electron microscope study of the development of a peripheral sensory neurite network
.
J. Embryol. exp. Morph
.
69
,
237
250
.
Roberts
,
A.
&
Taylor
,
J. S. H.
(
1983
).
A study of the growth cones of developing embryonic sensory neurites
.
J. Embryol. exp. Morph
.
73
,
31
47
.
Rutishauser
,
U.
,
Gall
,
W. E.
&
Edelman
,
G. M.
(
1978
).
Adhesion among neural cells of the chick embryo. IV. Role of the cell surface molecule CAM in the formation of neurite bundles in cultures of spinal ganglia
.
J. Cell Biol
.
79
,
382
393
.
Sandrock
,
A. W.
&
Matthews
,
W. D.
(
1987a
).
An in vito neurite-promoting antigen functions in axonal regeneration in vivo
.
Science
237
,
1605
1608
.
Sandrock
,
A. W.
&
Mattews
,
W. D.
(
1987b
).
Identification of a peripheral nerve neurite growth-promoting activity by development and use of an in vitro bioassay
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
6934
6938
.
Schubiger
,
M.
&
Palka
,
J.
(
1986
).
Axon polarity in Drosophila wings with mutant cuticular polarity patterns
.
Devi Biol
.
113
,
461
466
.
Scott
,
T. M.
&
Bunt
,
S. M.
(
1986
).
An examination of the evidence for the existence of preformed pathways in the neural tube of Xenopus laevis
.
J. Embryol. exp. Morph
.
91
,
181
195
.
Silver
,
J.
&
Rutishauser
,
U.
(
1984
).
Guidance of optic axons in vivo by a preformed adhesive pathway on neuroepithelia endfeet
.
Devi Biol
.
106
,
485
499
.
Sperry
,
R.
(
1963
).
Chemoaffinity in the orderly growth of nerve fibers and connections
.
Proc. natn. Acad. Sci. U.S.A
.
50
,
703
710
.
Towbin
,
H.
,
Staehlin
,
T.
&
Gordon
,
J.
(
1979
).
Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedures and some applications
.
Proc. natn. Acad. Set. U.S.A
.
76
,
4350
4354
.
Trisler
,
G. D.
,
Schneider
,
M. D.
,
Moskal
,
J. R.
&
Nierenberg
,
M.
(
1981
).
A topographic gradient of molecules in the retina can be used to identify neuron position
.
Proc. natn. Acad. Sci. U.S.A
.
78
,
2145
2149
.