Cell recognition and pattern formation in the developing nervous system

Development and function of the nervous system involves the formation of a complex network of cellular organization. Highly stereotyped patterns of synaptic connections are formed between diverse classes of neurons in various and distant locations in the nervous system. In the retinotectal system of lower vertebrates, for example, the topographic order of retinal cells is conserved when retinal ganglion cells project to the optic tectum and other regions of the brain. Attardi and Sperry (1963), using the goldfish, demonstrated that after optic nerve transection regenerating axons from dorsal retina project to ventral tectum and those from ventral retina project to dorsal tectum (Fig. 1A). Posterior retina grows to anterior tectum and anterior retina to posterior tectum (Fig. IB). Similar topography is present in the avian visual system (DeLong and Coulombre, 1965). Resolution of the topographic map in greater detail, determined by measuring electrophysiological responses of tectal cells to stimulation of discrete retinal fields, revealed a continuous point-to-point correspondence between cells of the retina and cells of the tectum (Fig. 1C; for a review, see Fraser and Hunt, 1980). The intricate specificity of synapses in the continuum map develops in a stereotyped pattern. However, the map is flexible to a degree. The retinal map will compress or expand to fill the available synaptic field of tectum after ablation of part of the retina or tectum, yet maintain the topographic order of the map (Gaze et al. 1963; Gaze and Sharma, 1970; Yoon, 1971; Chung and Cooke, 1975; Schmidt et al. 1978). The orderly retinotectal projection apparently relies, in part, on the ability of retinal axons to read tectal cues. In the developing chick retina, misrouted ganglion cell axons, disrupted in the retina and optic nerve, can reorient on the tectal surface and project to the correct target site (Fig. ID; Thanos et al. 1984). Some misrouted growth cones migrate to the appropriate anterior-posterior tectal longitude, make a 90° turn and project to the correct position along the dorsoventral axis of the tectum.

Developmental mechanisms involved in axonal guidance that result in synaptic specificity are not fully understood. Molecular gradients have been proposed as a mechanism for encoding cell positional information and establishing pattern in the embryo (Boveri, 1901; Dalcq and Pasteels, 1937; Child, 1941; Sperry, 1963; Grierer and Meinhardt, 1972; Fraser and Hunt, 1980; Whitelaw and Cowan, 1982; Trisler, 1982). Topographically distributed molecules have been reported in retina and optic tectum (Trisler et al. 1981; Constantine-Paton et al. 1986; Rabacchi and Drager, 1987; Trisler and Collins, 1987; Müller et al. 1990).

Topographically graded molecules in retina were detected with monoclonal antibodies generated by the fusion of spleen cells from mice immunized with a small portion of dorsal or posterior retina to P3X63Ag8 mouse myeloma cells (Fig. 2). One antibody to a cell surface molecule bound more abundantly to cells from dorsal retina than to cells from the remainder of the retina. A second antibody bound preferentially to cells of posterior retina. These molecules, termed TOP for toponymic (i.e. a marker of position), are present throughout the retina but are distributed in topographic gradients (Fig. 3). TOP_DV is graded dorsoventrally and TOP_AP is graded anteroposteriorly. Bilaterally symmetrical gradients of TOP molecules are present in the retinas of both right and left eyes.

A 35-fold gradient of TOP_DV was found extending from the dorsal to the ventral. margins of the retina aligned parallel to the long axis of the choroid fissure (Fig. 4), and a 16-fold gradient of TOP_AP is present from the anterior retinal margin to the posterior margin perpendicular to TOP_DV. The concentration of TOP molecules detected varied continuously and logarithmically with the logarithm of distance along the circumference of retina.

Indirect immunofluorescence of anti-TOP_DV antibody binding to cells in tissue sections taken along the dorsoventral axis of retina of E5 embryos revealed TOP_DV was most abundant in dorsal retina, intermediate in middle, and least abundant in ventral retina (Fig. 5). No obvious heterogeneity was observed in the cell population from each location. Most or all cells across the thickness of retina from the vitreal surface to the pigmented epithelium stained. Visual scanning of fluorescently stained transverse sections of whole retina showed that TOP_DV was continuously graded. This indicates that the dorsoventral gradient is due to differences in the amount of TOP_DV per cell rather than to differences in the number of cells expressing TOP_DV. The ring fluorescence pattern around each cell is consistent with the fact that TOP_DV is a cell surface molecule. In older embryos and in adult retina, TOP_DV is most abundant in the synaptic layers and in the ganglion cell axon layer (Fig. 5G).

Similarly, TOP_AP is distributed in a graded pattern along the anteroposterior axis (Fig. 5). The ring fluorescence of cells from anterior retina was weakest, that of middle cells was intermediate and that of cells from posterior retina exhibited the greatest binding of anti-TOP_AP in E5 embryos. Autoradiography of anti-TOP_AP binding along the anterior-posterior equatorial meridian of the entire E8 retina revealed the TOP_AP gradient. Since TOP_DV and TOP_AP molecules are graded on the basis of number of molecules per cell, TOP_DV can be used to identify cell position along the dorsoventral axis, while TOP_AP identified cell position along the anteroposterior axis of retina.

Gradients of TOP molecules were present in retina at all developmental ages tested from E4 to E18 and in hatchling chicks (Fig. 6). The orientation of the gradients remained constant throughout development; however, the magnitude of the TOP_DV gradient increased 10 times from a threefold gradient at E4 to a 35-fold gradient at E12. The TOP_AP gradient increased fivefold from E5 to E18. Thus, the TOP gradients are present early during blast cell proliferation and persist throughout development when both neurogenesis and gliogenesis occur and tissue organization is established.

Both TOP_DV and TOP_AP are cell surface proteins. TOP_Dv has an apparent relative molecular mass of 47×10³ (Moskal et al. 1986) and TOP_AP of 40×10³. Topographic differences in TOP_AP distribution were found in immunoblots of retina cell proteins resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Approximately 16-fold more TOP_AP was detected in cells from the posterior margin of retina than from the anterior margin (Fig. 7). Analogously, more TOP_AP was present in the posterior half-retina than in the anterior half-retina. These results suggest that the graded nature of TOP_AP demonstrated by radioimmuno-binding assay and by indirect immunofluorescence reflects the distribution of the antigen itself and not that of a masking molecule or of antigen accessibility. However, a gradient of post-translational modification of evenly distributed TOP_AP molecules has not been ruled out.

TOP_DV and TOP_AP molecules were also detected in the optic nerve (Fig. 8). Their distribution matched the organization of ganglion cell axons in the nerve. Axons from peripheral dorsal retina cells of E12 embryos were shown by fluorescent dye-tracing to be present in dorsal nerve. Indirect immunofluorescence revealed a greater abundance of TOP_DV in dorsal nerve than in ventral nerve. Ganglion cell axons from posterior retina were in anterior nerve, as were higher levels of TOP_AP. Although the distribution of TOP molecules reiterates the organization of ganglion cell axons in the nerve, the question of whether TOP molecules in the nerve are associated with axons, with glia or with both is not resolved. In E12 embryos, at the age when retinal axons have covered the tectal surface, TOP_Ap was most abundant on the anterior medial portion of tectum (Fig. 8) where posterior retinal axons project (DeLong and Coulombre, 1965).

TOP molecules are present on optic tectum cells as well as in retina (Trisler and Collins, 1987). They are distributed in gradients in E5 embryos 1 day before ganglion cell axons arrive at the tectum (Fig. 9). TOP_DV and TOP_AP gradients in tectum are inverted with respect to the retinal gradients. TOP_DV is 10-fold higher in ventral than in dorsal tectum and TOP_AP is eightfold higher in anterior than in posterior tectum. The quantities of TOP_DV and TOP_AP detected per cell by indirect immunofluorescence varied continuously along the axes of the tectal gradients (Fig. 5). Little or no heterogeneity of TOP_DV or TOP_AP expression was found among cells from ventricular to pial surfaces at a given position along the respective gradients. Thus, as in retina, TOP molecules can be used to mark cell position in tectum.

The orthogonal monotonic gradients of TOP_DV and TOP_AP topographic molecules constitute a possible Cartesian coordinate system that can be used to designate cell position at all points in the plane of the retina and optic tectum (Fig. 10). The presence of corresponding TOP gradients in retina and tectum suggests a possible role for the molecules in orienting the dorsoventral and anteroposterior axes of the retinal projection onto the tectum.

The retina grows by accretion of rings of cells in the proliferative zone at the peripheral margin (Coulombre, 1955; Kahn, 1974). Expression of the TOP gradient during retinal growth was examined by comparison of the amount of TOP_DV in the proliferative zone at the poles of the gradient in E4-E10 retinas with the amount of TOP_DV at the corresponding distances along the gradient axis in E12 retina (Fig. 11). The cells in the proliferative zone at the dorsal margin of retina expressed progressively more TOP_DV with retinal growth, while those in the proliferative zone at the ventral margin expressed progressively less TOP_DV. thereby increasing the magnitude of the gradient (Trisler, 1987). There was close agreement in the amount of TOP_DV detected on cells at a given distance along the gradient axis from the fundus, the oldest portion of the retina, throughout the developmental period tested. Thus, each position along the axis has a constant. TOP_DV value through these developmental ages.

The level of TOP_DV expressed in the progeny of cells dividing at various angles to the axis of the antigen gradient was determined by comparing TOP_DV expression in cells of the proliferative zone of E4 and E12 retinas at different positions around the peripheral margin (Fig. 12). Both the magnitude and the sign of change in TOP_DV expression during development varied depending on the position of the parental cells. Progeny cells in dorsal retina expressed more TOP_DV than parental cells while those in ventral retina expressed less. The greatest magnitude of change in expression was along the axis of the gradient (0°). Little or no change in TOP_DV expression occurred in cells with retinal growth along the perpendicular axis (90°). Intermediate rates of change in TOP_DV expression were found during cell division at 45° to the gradient axis.

Ablation of parental cells in the proliferative zone at the dorsal and ventral poles of the gradient in embryos 60 h after fertilization altered the level of TOP_DV expression in dorsal and ventral retina later in development (Fig. 13). Progeny cells that replaced the ablated dorsal region expressed 60% less TOP_DV than normal in E16 retina. After ventral ablation, cells in ventral retina expressed 300 % more TOP_DV.These values are similar to that expected for the progeny of parental cells from regions neighboring the ablated portion. This suggests that by 60 h of development the level of TOP_DV expressed by the retina cells and their progeny is already determined.

The role of molecular markers of cell position in the development of the nervous system was examined using a monoclonal antibody to TOP_DV (Trisler, 1983; Trisler et al. 1986). Antibodies provide a means of blocking molecular function (Levi-Montalcini and Booker, 1960; Crawford et al. 1982; Schwartz and Spirma, 1982; Warner et al. 1984). Several groups have demonstrated effects of antibodies against nervous system molecules on growth cone behavior, neurite outgrowth and tissue organization in vivo and in vitro. Antibodies against chicken cognin and N-CAM inhibit cell-cell adhesion (Hausman and Moscona, 1979; Thiery et al. 1977, respectively) and anti-N-CAM disrupts axonal fasciculation (Thanos et al. 1984; Fraser et al. 1984). Antibody T61/3/12 blocks neurite outgrowth of chick retina cells in vitro (Henke-Fahle and Bonhoeffer, 1983), whereas antibody to Thy-1 stimulates neurite outgrowth of rat retinal ganglion cells in vitro (Leifer et al. 1984). Antibody LI inhibits granular cell migration in rat cerebellar explants (Lindner et al. 1983). Our objectives were to determine the accessibility of TOP_DV to antibody in the in vivo retina, to determine the persistence of antibody in retina after injection into the embryo, and to identify changes in the development of retina continuously exposed to anti-TOP_DV antibody.

Antibody to TOP_DV injected into the amniotic cavity of in ovo chick embryos 2–4 days after fertilization was detected on retina cells 1 day after injection (Fig. 14). The concentration of [Anti-TOP_DV TOP_DV] complexes (Ab-TOP_DV) detected was higher in dorsal retina than in ventral retina. Anti-TOP_Dv antibody was injected intraocularly into the vitreal space of E7-E19 eyes. A dorsal to ventral gradient of Ab-TOP_DV complexes of the same magnitude and orientation detected by in vitro binding studies was present 24 h after intraocular injection of antibody.

A gradient of Ab-TOP_DV complexes in retina persisted for 4 days after intraocular injection of mouse ascites fluid containing anti-TOP_DV antibody (Fig. 15). The Ab-TOP_DV gradient was maintained for 10 days when hybridoma cells, that synthesize anti-TOP_DV antibody, were injected. The hybridoma cells provide a continuous source of antibody for long-term maintenance of Ab-TOP_DV complexes in the retina. The decrease in Ab-TOP_DV complexes in retina after E18 represents a loss in accessibility of TOP_DV to intraocular antibody, perhaps due to developmental changes in permeability of the inner limiting membrane of the retina.

Two aspects of the process of synapse formation -the disappearance of growth Cones and the appearance of synapses -were measured in retina from embryos exposed to anti-TOP_DV antibody. Antibody was injected into eyes of E11 embryos at the time neuron process layers are forming and 2 days before the first structurally identifiable synapses appear (Sheffield and Fishman, 1970; Hughes and LaVelle, 1974; Daniels and Vogel, 1980). For the electron microscopic analysis of injected retinas, growth cones were defined as bodies containing large, irregular membrane cisternae or vesicles (Del Cerro and Snider, 1968; Kawana et al. 1971). These bodies were larger in section than most neurites and were sometimes seen in continuity with neurites and filopodia. Anti-TOP_DV antibody from the injected hybridoma source reached maximal binding in the retina 4 days after injection (E15 embryos). At this time, retinal development appeared normal. However, with continued exposure to anti-TOP_DV antibody retinal development was altered. The normal progressive loss of growth cones during development was delayed in retinas containing Ab-TOP_DV complexes and synapse formation was inhibited (Fig. 16). The inhibition of synapse formation was not restricted to any particular cell type. Both conventional synapses and bipolar cell ribbon synapses were present in reduced numbers from 5 to 7 days after injection of anti-TOP antibody. Synapse formation appeared to be interrupted across the entire inner synaptic layer in both dorsal and ventral retina. Our working hypothesis is that TOP_DV molecules mark cell position in the retina and that, in the presence of Ab-TOP_DV complexes, growing neurites fail to detect the TOP_DV gradient. This could result in the persistence of growth cones and the delay in synaptogenesis. Synapses eventually form after exposure to anti-TOP antibody. It is not known whether these synapses are positionally correct.

The topographic map of cell position in retina is conserved and inverted when retinal axons project to and synapse with neurons of the optic tectum. Developmental mechanisms based on molecular gradients that specify positional information and pattern formation have been postulated in the establishment of the topographic map of cells in avian retina and optic tectum. Orthogonal gradients of toponymic (TOP) cell membrane molecules are present in the retina and optic tectum. TOP_DV molecules are distributed dorsoventrally and TOP_AP molecules are graded anteroposteriorly. The polarities of the gradients are inverted in tectum with respect to retina. TOP_DV is most abundant in dorsal retina and ventral tectum and TOP_AP is most abundant in posterior retina and anterior tectum. The quantities of TOP_DV and TOP_AP detected per cell vary continuously along the axis of the respective gradient. Thus, TOP molecules can be used to identify cell position along the dorsoventral and anteroposterior axes of the developing retina and optic tectum.

The orthogonal monotonic gradients of TOP_DV and TOP_AP molecules constitute a possible Cartesian coordinate system that can be used to identify cell position in the topographic map of retina and optic tectum. Synapse formation in retina was inhibited in the presence of anti-TOP_DV antibody. This suggests that TOP_DV may be involved in the recognition of cell position that is required for normal synapse formation. The inverted configuration of the gradients in retina and tectum corresponds to the inverted topographic map of cell position in the retinotectal projection. Thus, TOP molecules may be involved in orienting the retinotectal map.