Report

Change in Chemoattractant Responsiveness of Developing Axons at an Intermediate Target

See allHide authors and affiliations

Science  02 Jan 1998:
Vol. 279, Issue 5347, pp. 105-107
DOI: 10.1126/science.279.5347.105

Abstract

Developing axons reach their final targets as a result of a series of axonal projections to successive intermediate targets. Long-range chemoattraction by intermediate targets plays a key role in this process. Growing axons, however, do not stall at the intermediate targets, where the chemoattractant concentration is expected to be maximal. Commissural axons in the metencephalon, initially attracted by a chemoattractant released from the floor plate, were shown to lose responsiveness to the chemoattractant when they crossed the floor plate in vitro. Such changes in axon responsiveness to chemoattractants may enable developing axons to continue to navigate toward their final destinations.

In the developing nervous system, axons navigate considerable distances toward their final targets in a highly stereotyped and directed manner. This process is achieved by a series of axonal projections to successive intermediate targets under the influence of local guidance cues (1). Accumulating evidence has indicated the importance of long-range chemoattraction in guiding developing axons not only to final (2) but also to intermediate targets (3-10). Commissural axons originating from the alar plate of the vertebrate central nervous system, for example, initially grow ventrally, attracted by a diffusible chemotropic molecule secreted from the ventral midline floor plate (3-9), an intermediate target of these axons (6, 11). These axons, however, grow past the floor plate to extend contralaterally. Since the first demonstration of the existence of the floor plate–derived chemoattractant (3), an intriguing question has been why growing axons do not stall at their intermediate targets, where the chemoattractant concentration is expected to be maximal. Here, we provide evidence for a change in the chemoattractant responsiveness of growing axons during their growth across an intermediate target.

Metencephalon commissural axons, which originate from the cerebellar plate (CP) of the rat embryo, initially grow circumferentially toward floor plate cells at the ventral midline of the metencephalon (6). In vitro studies have suggested that these commissural axons (referred to hereafter as CP axons) are guided toward the midline by a diffusible chemoattractant released from floor plate cells (6, 7, 9). Here, we used an in vitro preparation that reproduces the crossing of the midline floor plate by CP axons to examine possible changes in the chemoattractant responsiveness of CP axons when they cross the floor plate (12). When a strip of the rostral metencephalon that included the entire circumferential trajectory of CP axons (Fig.1A) was cultured alone in collagen gel, axons originating from the CP grew across the midline floor plate to extend contralaterally (n = 16) (Fig. 1B) (13). We next tested whether CP axons, after they have crossed the floor plate, are attracted by an ectopic floor plate explant. We juxtaposed a floor plate explant to the metencephalic strip on one side (Fig. 1A) and examined the behavior of CP axons by implanting the fluorescent tracer 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI) into the CP contralateral to the explant. Under such conditions, CP axons that had crossed the midline floor plate did not show directed growth toward the ectopic floor plate explant (Fig. 1C). To compare directly the behavior of CP axons extending from both sides, we implanted DiI crystals into the contralateral CP and 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO) crystals into the ipsilateral CP of the same strip preparations (Fig. 1D). Although DiO-labeled CP axons that had not crossed the midline floor plate showed directed growth toward the ectopic explant, DiI-labeled axons that had crossed showed no sign of directed growth (Fig. 1, D and E).

Figure 1

Loss of CP axon responsiveness to a floor plate (FP)–derived chemoattractant when the CP axons cross the floor plate. (A) (Left) A diagrammatical view of the coronal plane of the neural tube with the trajectory of CP axons. (Center) Flat, whole-mount preparation that was obtained by cutting the dorsal midline of the neural tube. (Right) A metencephalic strip preparation of E13 rat brain including the circumferential trajectory of CP axons was further dissected from the whole-mount preparation as represented in the center diagram. An ectopic floor plate (eFP) explant was placed on one side of the preparation. (B) Fluorescent view of DiI-labeled CP axons in a metencephalic strip preparation that had been cultured for 2 days. (C) Fluorescent micrograph of DiI-labeled CP axons in a strip preparation with an ectopic floor plate explant on one side. (D and E) Different behaviors of CP axons in the same preparation with an ectopic floor plate explant on one side (lower right). DiO crystals were implanted into one side (right) and DiI crystals into the other side (left). (E) is a higher magnification view of (D). (F) Immunofluorescence micrograph of a strip preparation after staining with antibody to TAG-1. (G) In the presence of an ectopic floor plate explant, ipsilateral TAG-1+ CP axons are deflected toward the explant. Asterisks in (B) and (C) indicate DiI injection sites, and arrows indicate the ventral midline floor plate (B to D, F, and G). Scale bar, 250 μm (B to D, F, and G) and 100 μm (E).

Because implantation of DiI or DiO into the CP might also label axonal populations other than those from commissural neurons, such as longitudinally growing axons (Fig. 1C) (6), we next assessed chemoattraction of CP axons by labeling them with a molecular marker for commissural axons (14). Commissural axons at all axial levels from the spinal cord to the mesencephalon express TAG-1, an axonal surface glycoprotein, during their circumferential growth until they reach the floor plate in vivo (9, 15). Moreover, TAG-1–positive (TAG-1+) commissural axons including CP axons are attracted by the floor plate in vitro (3, 9). We found that most TAG-1+ CP axons in control strip preparations grew straight along the circumferential axis (n = 15) (Fig. 1F), although some TAG-1+ CP axons near the cut edges of the preparations spontaneously exited the preparations, deflected from their circumferential trajectory. However, TAG-1+ CP axons at a distance of >0.16 mm away from the cut edges were never observed to exit from control strip preparations. Thus, we judged CP axons to be attracted by an ectopic floor plate explant if CP axons at a distance of >0.16 mm away from the cut edge facing the explant showed turning toward and penetration into the explant. We found that, although TAG-1+ CP axons even at a distance of 0.56 ± 0.02 mm (mean ± SEM, n = 17) away from the explant showed turning toward the ipsilaterally placed explant (Fig. 1G and Table 1), DiI-labeled CP axons, once they had crossed the floor plate, were not attracted by the explant (Table 1) (16). Thus, CP axons, which are attracted by the chemoattractant until they reach the floor plate, lack responsiveness to it after they cross the floor plate.

Table 1

Chemoattraction of CP axons in metencephalic strip preparations. Culture preparations were scored as positive if they contained any CP axons that were qualified as “attraction” by the criteria described in the text. FP, floor plate.

View this table:

At least two possible mechanisms can be considered regarding this change in growth cone responsiveness: one is a cell-autonomous loss of responsiveness at the floor plate, and the other is a change caused by an interaction of commissural axons with floor plate cells. If the latter is the case, one would predict that the growth cones extending into the contralateral side without encountering the midline floor plate should retain responsiveness to the chemoattractant. We found that when a ventral region including the floor plate was surgically removed from the strip preparation (Fig.2A), CP axons growing into the contralateral part showed directed growth toward the ectopic floor plate explant (Fig. 2B and Table 1). In contrast, when a ventral region that did not contain the floor plate was removed (Fig. 2C), CP axons, once they had crossed the midline floor plate, grew circumferentially without being attracted by the ectopic floor plate explant (Fig. 2D and Table 1). These results support the view that the growth cone encounter with floor plate cells abolishes its responsiveness to the floor plate–derived chemoattractant.

Figure 2

Abolishment of chemoattractant responsiveness of CP axons by the midline floor plate. (A and C) Diagrams showing experimental manipulations. The hatched regions were removed from the strip preparation, and the remaining explants were put together to fuse the cut aspects of the preparation. (B) Fluorescent micrograph of a culture preparation after the experimental manipulation shown in (A), demonstrating that DiI-labeled CP axons growing into the contralateral part are attracted by an ectopic floor plate explant. Asterisks indicate DiI injection sites. (D) When a ventral region excluding the midline floor plate is removed, DiI-labeled CP axons, once they cross the midline floor plate, grow circumferentially on the contralateral side, showing no sign of being attracted by an ectopic floor plate explant. Scale bar, 250 μm.

We next tested whether CP axons, after crossing the floor plate, continue to have responsiveness to netrin-1, a laminin-related diffusible protein secreted by floor plate cells (5, 17). Netrin-1 can attract commissural axons at all axial levels from the spinal cord to the mesencephalon in vitro (5, 6, 9). We found that CP axons, which are attracted by aggregates of heterologous cells secreting netrin-1 (Fig. 3A and Table 1) (18), were not attracted by the aggregates after they had crossed the floor plate (Fig. 3B and Table 1) (19), indicating that CP axons lose responsiveness to netrin-1 during their growth across the floor plate (20).

Figure 3

Loss of CP axon responsiveness to netrin-1 when the CP axons cross the floor plate. A netrin-1–secreting cell aggregate was placed on one side of the strip preparation. (A) Immunofluorescence micrograph showing that TAG-1+ CP axons are attracted by ipsilaterally placed aggregates of netrin-1–secreting cells (Net). (B) DiI-labeled CP axons are not attracted by netrin-1–secreting cell aggregates after they cross the floor plate. Asterisks indicate DiI injection sites. Scale bar, 250 μm.

Our results show that floor plate chemoattraction of commissural axons is effective only until they arrive at the floor plate. Although it remains unknown how modification of growth cone responsiveness to the chemoattractant is achieved, one possible mechanism is that the interaction of commissural axons with floor plate cells directly changes the chemoattractant responsiveness of the axons by modifying chemoattractant-receptor mechanisms locally in the growth cones. Alternatively, the modification might be mediated by changes in transcription of the receptor or one or more receptor-related molecules. In any case, receptor mechanisms mediated by netrin-1 may in some way be involved. Deleted in Colorectal Cancer (DCC) is a receptor or a component of a receptor that mediates the effect of netrin-1 on spinal commissural axons (21,22) and is also expressed on CP axons during their growth toward the floor plate in vivo (23). Thus, down-regulation of DCC expression at the floor plate may underlie the loss of the chemoattractant responsiveness. However, DCC appears to be expressed on commissural axons both as they project toward the floor plate and after they cross the floor plate in the developing rat spinal cord (21), suggesting that another as yet unidentified component of the netrin-1 receptor may be involved (9, 21,24). It is also possible that the transfer of proteins from midline cells to crossing axons at the midline (25) is in some way involved in the regulation of the chemoattractant responsiveness of the axons.

After they cross the floor plate, commissural axons at all axial levels from the spinal cord to the mesencephalon abruptly change their growth direction to extend longitudinally (6,11, 26). Although the molecular mechanisms subserving such a change in growth direction remain unknown, changes in growth cone responsiveness to guidance cues as demonstrated here might explain why commissural axons can grow along the longitudinal axis only on the contralateral side, despite the existence of identical ipsilateral cues.

In conclusion, loss of chemoattractant responsiveness of developing axons at intermediate targets as described here may contribute to their leaving the intermediate targets. In addition, encounters with their intermediate targets might also cause sensitization of growing axons to subsequently encountered cues that guide them to their final destinations.

  • * To whom correspondence should be addressed. E-mail: sirasaki{at}bpe.es.osaka-u.ac.jp

REFERENCES AND NOTES

View Abstract

Navigate This Article