The innervation of a virgin tectum by a double-temporal or a double-nasal eye in Xenopus

In a recent series of papers (Straznicky & Gaze, 1980; Gaze & Straznicky, 1980 a, b;Straznicky, Gaze & Keating, 1981) we have examined the development and regeneration of retinotectal fibre projections from surgically constructed ‘compound eyes’ in Xenopus. The evidence presented shows that each half of such a reconstructed eye retains some attributes, or markers, which identify it as a half-eye. When a compound double-nasal (NN), doubletemporal (TT) or double-ventral (VV) eye is caused, after metamorphosis, to innervate its ipsilateral tectum along with the projection from the normal eye, the compound eye projection restricts itself to the appropriate part of the tectum. NN fibres innervate caudomedial tectum, TT fibres innervate rostro-lateral tectum and VV fibres innervate medial tectum. This shows that the retinal fibres are recognizing something which constrains them to terminate in particular regions of the tectum. These ‘markers’ on the tectum could be produced and carried by the tectal cells themselves, as supposed by Sperry (reviewed 1951, 1963, 1965), or could be related to the fibre projection, either as the fibres themselves or as a trace left on the tectum by the fibres (Schmidt, 1978).

If the projection from the normal eye is removed before the fibres from the compound eye reach this tectum, the projection from the compound eye is no longer restricted but spreads at once to cover the entire tectum (Straznicky & Tay, 1981). This suggests that the fibres from the compound eye are recognizing the fibres from the normal eye, rather than the tectal cells themselves, or that any markers deposited by the nerve fibres are very short lived. Since an optic fibre projection can mark the tectum in a way that is recognizable to another incoming optic fibre projection, the question arises whether the original establishment of the retinotectal projection, in early development, requires the existence of tectal markers. If so, tectal markers must exist on the tectum before the arrival of the optic fibres, as suggested by Sperry. If not, then some mechanism other than target affinity must be responsible for the first establishment of the projection.

In this paper we describe another approach to the question of whether or not tectal positional markers exist before the first arrival of optic fibres. We have arranged for one tectum to remain without optic innervation (‘virgin’) during development, by removal of the contralateral eye in embryonic life, before the fibres grow to the brain. Then, after metamorphosis, we caused the fibres from the other eye (which was compound, TT or NN) to innervate the virgin tectum, and we studied the resulting projections by means of autoradiography with tritiated proline. The consistent result was that TT projections to the virgin tectum were restricted to rostrolateral tectum but NN projections covered the entire tectum. A brief account of this work has appeared elsewhere (Gaze & Straznicky, 1980c).

Xenopus laevis were obtained from laboratory breeding pairs

Under MS 222 anaesthesia (tricaine methane sulphonate, Sandoz 1:3000), the right and left eyes were operated on at stages 32–33 (Nieuwkoop & Faber, 1967). The nasal half of the right eye anlage of the host embryo was removed and replaced by a temporal half of the left eye of a donor embryo to form a double-temporal (TT) eye. Similarly in other embryos double-nasal (NN) eyes were formed. One to two hours after the first surgery the left eye of the operated embryos was removed. The TT and NN eye animals were kept separately under standard laboratory conditions and reared past metamorphosis.

One week after metamorphosis the remaining right optic nerve was exposed through the pharynx, under MS 222 anaesthesia (1:1000), and cut close to the optic chiasma to facilitate bilateral tectal regeneration. In a few animals with right compound eye the optic nerve was not cut, and these animals served as controls. Four to six weeks after the optic nerve section, or metamorphosis in the case of control animals, the resultant optic fibre projections from the right compound eyes were assessed autoradiographically and, in some cases, electrophysiologically.

Twenty-four hours before sacrifice 10 μCi [³H]proline (³H-P; 21 Ci/mmol Amersham) in 0·25 μl solution was injected into the vitreous. The head of the animal was fixed in Bouin’s solution, the dissected brain was embedded in paraffin and serially sectioned at 10 μm. The sections were mounted on slides and coated with Ilford K2 nuclear emulsion, exposed at 4 °C for 14 days, developed in Kodak Dektol and counterstained with Harris’s haematoxylin.

The extent of the tectum and the tectal projection/s from the right eye were measured on camera-lucida drawings.

In a few animals (control group) the visuotectal projection from the right operated eye was mapped in order to check the success of the embryonic eye operation. Visuotectal recordings were carried out according to standard procedures (Straznicky & Gaze, 1980; Gaze & Straznicky, 1980a, b).

The left (normal) eye was removed at stage 32/33, and the right eye was made NN in two animals and TT in four animals.

One week after metamorphosis, and twenty-four hours before the anijnals were killed, the compound eye was labelled with ³H-P. One animal with an NN eye and two with TT eyes were mapped electrophysiologically. Each animal recorded showed the reduplication of the visuotopic map characteristic of the nature of the compound eye. The NN eye projected to the tectum so that the vertical midline of the field was represented on rostral tectum, while nasal and temporal field extremities were represented caudally (Fig. 1). The maps from the TT eyes showed that the nasal and temporal extremities of the field projected to rostral tectum while the vertical midline of the field projected caudally (Fig. 2).

The distributions of the projections from the control compound eyes, as revealed by autoradiography, are summarized in Fig. 3. It may be seen that the two NN projections gave complete coverage of the tectum contralateral to the compound eye (Fig. 3e,f) and that the tecta ipsilateral to the compound eye are empty of label except for a small amount of faint and atypical labelling rostromedially in Fig. 3f. Three of the four TT projections (Fig. 3 a, b, d) show a deficit in the projection to the caudomedial part of the contralateral tectum, as has been noted previously for TT eyes (Straznicky et al. 1981). One animal (Fig. 3 c) gave a complete coverage of the contralateral tectum. The tecta ipsilateral to the compound eye were devoid of label in two cases (Fig. 3a and c) and show only a minimal amount of atypical labelling rostrally in Fig. 3b and d. In Fig. 3b the fibre projection is unusual in that the optic nerve fibres reach contralateral tectum over two quite separate pathways -one via the chiasma and the other as a separate fibre bundle entering the rostral tectum directly from the cranial cavity.

The nature of the small amount of ipsilateral labelling in Fig. 3 b, d and f is not known ; its relative intensity is shown, for two cases, in Fig. 4. Apart from the aberrant label the ipsilateral tectum appears to be virgin in each case.

The projections from the TT eye to both tecta 28 days after optic nerve section are shown, for all five animals investigated, in Fig. 5 a-e. On both contralateral and ipsilateral (virgin) sides the tectal coverage is incomplete caudomedially, and this deficit is more marked on the virgin tectum. The projections from NN eye to both tecta, 28 days after optic nerve section, are shown, for all five animals investigated, in Fig. 5f–j. With one exception (Fig. 5g) the tectal coverage is virtually complete on both contralateral and ipsilateral (virgin) sides. And with the same exception the label is more dense rostrally on the virgin tectum in all cases.

The projections from TT and NN eyes after 42 days of regeneration are shown in Fig. 6. There is no obvious consistent difference between these projections and those at 28 days regeneration shown in Fig. 5.

In the previous paper of this series we described the development of the retinotectal projection from TT, NN and VV eyes (Straznicky et al. 1981). It was found that TT projections initially restricted themselves to rostrolateral tectum and spread slowly across the tectum, so that tectal coverage was pearly complete shortly after metamorphosis. In a comparable fashion, VV projections were initially restricted to medial tectum and took time to spread across the tectum laterally. NN projections, on the other hand, covered most of the tectum from the start, although these projections were most dense caudally.

In normal Xenopus the retina grows by the addition of rings of cells at the ciliary margin (Straznicky & Gaze, 1971) whereas the tectum grows from rostrolateral (oldest tissue) to caudomedial (youngest tissue; Straznicky & Gaze, 1972). The retina sends a fibre projection to the tectum from around stage 45, when both retina and tectum are small in comparison with the adult; and shortly thereafter the retinotectal map may be shown electrophysiologically to have the normal orientation and order (Gaze, Keating & Chung, 1974). From this time retina, tectum and the retinotectal projection all increase in extent while the ordering of the retinotectal map is maintained (Gaze, Keating, Ostberg & Chung, 1979b). These findings have led to the proposal that the retinotectal connexions shift progressively during normal development (Gaze et al. 1919 b).

The topological problem of fitting the retinal projection on to the tectum, when both retina and tectum are growing differently, is made yet more difficult in the case of compound eyes. These have been shown to grow like normal eyes by cellular addition at the ciliary margin (Feldman & Gaze, 1972); and the pattern of the projection, particularly for TT eyes, is such that the shift of connexions that occurs during growth is much greater than for normal eyes (Straznicky et al. 1981). Thus for normal eyes, and even more for compound eyes, the observed modes of growth of the projections do not readily fit the idea that localized tectal positional markers exist prior to the arrival of the optic nerve fibres. Not, that is, if the tectal markers are stable and the differential affinities between the various retinal markers and the various tectal markers are also stable.

The present paper describes a different approach to the question of whether or not tectal positional markers exist before the arrival of optic fibres. The virgin tectum is one that has had its normal source of optic fibre input removed in embryonic life, before the optic fibres have grown from the eye to the tectum. We have shown that the virgin tecta in these experiments are indeed virgin in that they possess virtually no autoradiographically demonstrable projection from the remaining (compound) eye after metamorphosis. It is conceivable that there was a transient projection to these tecta in early life but this is rendered unlikely by the observations (Steedman, Stirling & Gaze, unpublished; Fawcett, Hirst & Gaze, unpublished) that cobalt or HRP impregnation of fibres from a residual eye shows the ipsilateral tectum to be uninnervated in tadpole life. These virgin tecta, therefore, grow up with no experience of optic fibres, certainly over the greater part of the tectal surface. It is of course obvious that such tecta have extensive connexions with other parts of the nervous system, including, probably, visually related regions such as the nucleus isthmi, some of which send terminals to those tectal layers normally occupied by optic fibres. Even so it remains true that the tecta are devoid of optic fibres and we can then ask how the system behaves when optic fibres are first supplied to the tectum after metamorphosis.

When a compound eye is caused to innervate both tecta after metamorphosis, we find that the TT projections are restricted to rostrolateral tectum to a greater extent on the virgin side than is the case on the contralateral side. NN projections to the virgin tecta, on the other hand, show complete tectal coverage and, in four out of five cases, the autoradiographic density of the projection was greater rostrally than caudally.

The projections from TT eye to virgin tecta are thus compatible with the idea that tectal positional markers already exist on the tectum when the fibres arrive from the compound eye. The projections from NN eyes to virgin tecta, however, do not support this idea. Fibres from NN eyes, which under other circumstances have been shown to restrict themselves to caudomedial tectum (Gaze & Straznicky, 19806), covered the entire surface of the virgin tectum within 28 days of nerve section. There was no rostral deficit such as might have been expected if the nasal fibres restricted themselves initially to caudal tectum.

In a previous paper (Straznicky et al. 1981) we have argued that the arrangement of fibres within the optic tract could play a major role in the establishment of the normal retinotectal projection and might also account for the observed differences in the development of projections from TT, NN and W eyes. Any mapping mechanism must provide for the proper internal ordering of the map, its proper extension across the tectum and its proper orientation. All these characteristics of the normal map could be accounted for if the development mechanism used in the establishment of the map involved selective recognition between individually labelled retinal and tectal units, as previously proposed by Sperry (see 1951). The evidence for positional labelling of retinal fibres is strong (Gaze, Feldman, Cooke & Chung, 1979a; Straznicky et al. 1981) but the evidence for the existence of tectal labels ab initio is weak, as we show here and in the previous paper of this series (Straznicky et al. 1981).

Since we are trying to identify an alternative mapping mechanism, other means for establishing the three prime requisites of a normal map must be found. The internal ordering of the map could probably result from interactions between labelled retinal fibres leading to the preservation of retinal neighbourhood characteristics in the tectal distribution of the optic fibre endings. Evidence for some kind of fibre/fibre interactions is strong although the mechanism is obscure. What controls the overall extent of the retinal fibre projection on the tectum is unknown, but a generalized recognition of tectum (and other optic termination areas elsewhere in the brain) is a strong likelihood. Given the chance, retinal fibres will terminate in the tectum rather than in adjacent foreign tissue.

The provision of a correct orientation for the map, in the absence of a target-affinity mechanism, requires a form of control distinct from that ordering the map internally. Orientation could be provided by a weak system of polarity markers across the tectum, recognizable to incoming optic fibres (Straznicky, 1978; Willshaw & von der Malsberg, 1979). This could represent the tectal part of a general system of rostrocaudal and mediolateral axial cues across the whole brain (Sharma, 1981) or a polarization particular to the tectum, perhaps induced early in embryogenesis by the influence of the diencephalic anlage (Scalia & Fite, 1974; Chung & Cooke, 1978). On the other hand, orientation could be provided by the optic fibres being led on to the tectum from the correct parts of the tectal margin (Straznicky et al. 1981).

In this context we may mention that fibres from different parts of the retina appear to behave differently in the optic tract. Ventral retinal fibres select the medial branch of the optic tract during development, and this positional requirement is absolute rather than relative (Straznicky, Gaze & Horder, 1979; Steedman, 1981). Temporal retinal fibres normally approach the rostral tectal margin in a coherent bundle, and again they seem constrained to do this in an absolute fashion; nasal retinal fibres, on the other hand, are widespread in the optic tract as they approach the tectum (Steedman, 1981).

Since the results of the present experiments do not consistently support the idea that the optic fibre projections to a virgin tectum are established under the influence of pre-existing tectal positional markers, it will be helpful to know how the fibres innervating a virgin tectum are arranged in the tract. Experiments to determine this are in progress.

We thank Mrs June Colville for expert histological assistance.