Quail-chick grafting experiments corroborate that Tbr1-positive eminential prethalamic neurons migrate along three streams into hypothalamus, subpallium and septocommissural areas

Our first example of short-survival chimeras is case M6, fixed at stage HH25/26 (4.5–5 d.i.o.), cut coronally (levels of Fig. 2c–h schematically indicated in Fig. 2b), and processed for Tbr1 ISH and QCPN immunoreaction (IMR). The estimated graft extent relative to the fate map appears in Fig. 2a. At HH25/26, the grafted neuroepithelium area appears as a dense brown domain (enclosed between two blue arrowheads; Fig. 2c–h). Quail-labelled graft derivatives include the di-telencephalic chorioidal tela (Fig. 2c–f), the dorsal part of the PThE (Fig. 2f–h; note partially unlabelled ventricular zone of PThE in Fig. 2c–e), and parts of the habenula and thalamus (Hb, Th; Fig. 2c–h). Tbr1 ISH signal appears throughout the eminential mantle layer, as well as at a subpial aggregation of Tbr1-positive cells on top of the cerebral peduncle, suggestive of an incipient peripeduncular migratory stream (red arrowheads; Fig. 2c, c’, d; the peduncle [pe] appears as a superficial packet of unstained fibers). Sections through the peduncular hypothalamus show a continuity of the Tbr1-positive PThE mantle deep to the peduncle with blue cells within the lateral hypothalamus, indicating presence of an advanced juxtapeduncular migratory stream at this stage, which occupies the lateral hypothalamus mantle superficial to the paraventricular hypothalamic nucleus (black arrowheads; Fig. 2e–g). However, QCPN-labelling shows no quail migrated cells at the lateral hypothalamus, and only a few labelled cells were observed at the peripeduncular stream (small black arrows in Fig. 2d). It may be deduced from this discrepancy at the lateral hypothalamus that the earliest cells that exit from the PThE largely must originate from its unlabelled ventral part. The dorsal PThE part labelled selectively in this experiment contributes few if any cells to juxtapeduncular (hypothalamic) migrated PThE cells at this stage but enters incipiently the peripeduncular stream. This result suggests that the juxtapeduncular stream slightly precedes the peripeduncular stream at this stage.

A second example of our short survival eminential chimeras is case Q292, sacrificed at stage HH28 (5.5–6 d.i.o.; graft map in Fig. 3a; section orientation in Fig. 3b, b’; grafted derivatives in Fig. 3c–l, with schematic maps of grafted cells in Fig. 3c’–l’, and high magnification microphotographic details in Fig. 4a–g). This brain, doubly reacted for Tbr1 ISH and QCPN IMR, was accidentally sectioned in an oblique sagittal plane (see Fig. 3b, b’). We were guided in our interpretation of these sections by the known Tbr1 pattern (Alonso et al. 2020). The series shown starts close to the midline at the control side (Fig. 3c), where we see continuity of eminential Tbr1 signal with that of the telencephalic caudal ‘ventricular ridge’ pallium (including periventricular parts of pallial amygdala; PallA); separate cortical pallial parts (Pall/MPall) are also labelled with Tbr1, as well as subpial migrated pallial cells in the subpallial olfactory tuberculum (SPall; ot). The diminishing contralateral PThE is seen again in Fig. 3d, e, after which we have two other sections illustrating the (QCPN-labelled) median chorioidal roofplate (ech; Fig. 3f, g; compare Fig. 3b’). The experimental PThE is then sectioned first at its ventral end (Fig. 3h; compare Fig. 3b’), and the eminential section series then proceeds dorsalward across PThE, showing its relationships with the cerebral peduncle and associated labelled migration streams (PThE, pe; black and red arrowheads; Fig. 3h–l). At levels through Fig. 3k, l, the experimental PThE reaches its dorsal contact with the chorioidal roof (ech; Fig. 3l; compare Fig. 3b’). The non-eminential prethalamus appears laterally to the peduncle (PTh; Fig. 3k, l). Due to their obliquity, these last sections also provide an unusual insight on the neighborhood relationships of PThE with the pallial amygdala, as well as with the neighboring caudolateral end of the hippocampus, at the caudal tip of the hemisphere (PallA, Hi; Fig. 3k, l).

We mapped schematically the location of labelled quail cell nuclei at ventricular and mantle zone levels (Fig. 3e’–l’) and we show in Fig. 4 correlative microphotographic details of framed areas (Fig. 4a–g; note sometimes the QCPN-labelled brown cell nuclei are partly hidden by the blue Tbr1 ISH signal where it is massive, as occurs inside the PThE mantle; nevertheless, the brown label usually can be detected at high magnification). As regards the grafted domain, a large part of the diencephalic (eminential, ech; thalamic, thch) and telencephalic/fimbrial (fich) chorioidal tela of the transplanted hemisphere was graft-derived (ech, thch, fich; Fig. 3e–g, i–l), as well as most of the right habenula (Hb; Fig. 3j–l), and the whole dorsal part of the PThE (compare Fig. 3a; QCPN labelling of the ventricular zone is seen at levels of Figs. 3j, j’, k, k’ and 4e, f). Ventral section levels through the PThE did not show clearcut ventricular zone labelling (Fig. 3h, h’, i, i’), though migrated labelled cells were observed there in the mantle. The non-eminential prethalamus was not included in the graft (PTh; Fig.3a; k, l). There was also no QCPN-labelled ventricular zone at the pallial amygdala or hippocampus, or any other part of the telencephalic pallium.

According to these data, the Q292 graft included massively the dorsal part of the PThE, in contrast to the partial dorsal labelling observed in the M6 case, which also had a shorter survival. By stage HH28 many Tbr1- and QCPN-positive eminential quail cells have moved outside the PThE, following either the juxta- or peripeduncular streams described previously (Alonso et al. 2020). The double-labelled juxtapeduncular stream is best developed and is found typically just medial (deep) to the cerebral peduncle, the site of the lateral hypothalamus; some of its cells reach the lateral preoptic area (black arrowheads; POA, LPO, LH; Fig. 3g, h). The peripeduncular stream is less advanced; most of the corresponding double-labelled cells appear aggregated between the PThE mantle and the back of the peduncle, where this stream begins (pe; red arrowheads in Fig. 3i, j). However, more lateral sections tangent to the outside of the peduncle show labelled cells of this stream squeezing subpially around the peduncle and advancing into the basal telencephalon. Some labelled cells even seem to have reached the extended amygdala (red arrowheads, pe, EA; Fig. 3k, l). However, more medial sections in Fig. 3i/i’, j/j’ and correlative details in Fig. 4d, e show that practically no Tbr1- or QCPN-labelled cells can be seen subpially ventrally to the peduncle, indicating that migrating peripeduncular eminential cells do not yet reach the preoptic and diagonal areas at stage HH28, as they will do at later stages (Alonso et al. 2020). There is practically no indication of the third, eminentio-septal migratory stream at this stage, with the possible exception of a few double-labelled cells observed outside of the PThE, next to the paraventricular hypothalamic ventricular zone (Figs. 3g, g’ and 4b). These may represent the earliest sign of the eminentio-septal stream.

The coronally sectioned chimeric brains Q24 and Q303 (fixed at stage HH33/34-8 d.i.o.), illustrate the behaviour of migrating PThE cells at these stages, particularly with regard to their relationship with the peduncle. The HH33 Q24 chimera (map in Fig. 5a; ventricular zone marked between blue arrowheads in Fig. 5c; section plane for 5c, d in Fig. 5b) was one of our earliest experiments, performed before we had the possibility to perform in situ hybridization in our laboratory, and was therefore only labelled with QCPN IMR. This labelling revealed as grafted quail derivatives a small dorsocaudal portion of the PThE (Fig. 5a, c), jointly with a part of the adjacent di-telencephalic chorioidal tela (not shown), and parts of the rostral habenula (Hb) and thalamus (Th) (Fig. 5a). Quail-derived eminential cells fill up the dorsal deep and superficial mantle of the PThE (strata separated by the stria medullaris tract; PThE, sm; Fig. 5c, c’). At the caudal level shown in Fig. 5c, either the mass of peripeduncular cells or a part of non-eminential prethalamus is visible superficially to the peduncle, but this locus contains practically no quail-derived cells (detail in Fig. 5c’). At this level, numerous QCPN-labelled quail cells found within the lateral hypothalamus characterize the juxtapeduncular migratory stream, settling down in the lateral hypothalamus, medially to the peduncle (black arrowheads, LH, pe; Fig.5c, c’). A more rostral section level nevertheless shows abundant quail-derived cells in peripeduncular position; these are also found partly superficially to the peduncle; some juxtapeduncular migrated cells are also seen nearby (red and black arrowheads; pe; Fig.5d, d’).

The likewise coronally sectioned Q303 chimera (fixed at stage HH34 -8 d.i.o.) was processed for Tbr1 ISH and QCPN IMR (operation map in Fig. 5e; section levels in 5e’; quail ventricular zone between blue arrowheads in Fig. 5g, g’–i). In this chimeric brain the di-telencephalic chorioidal tela, a dorsal part of thalamus (Th) and habenula (Hb), a small dorsal part of the PThE, and a dorsalmedial portion of the medial pallium (Hi, MPall) derived from the quail graft (Fig. 5e; blue arrowheads in Fig.3g’, h’, i’). Numerous quail cells appear dispersed within the eminential Tbr1-positive mantle, apparently migrated from the small quail ventricular fragment observed dorsocaudally (Fig. 5g, g’, h, h’). Likewise, a number of quail cells have left by this stage the labelled eminential territory, following the three migratory routes described (juxtapeduncular, peripeduncular, and eminentio-septal streams; Alonso et al. 2020). As regards the peripeduncular stream quail Tbr1-positive cells clearly surround the peduncle, reaching rostrally the extended amygdala (EA), which already appears populated by numerous quail cells (red arrowheads; pe, EA; Fig. 5g, g’, h’, i, i’, j–l). In addition, some quail cells detach from the peripeduncular migratory stream and follow a subpial route into the Dg area (red arrowheads; DB; Fig. 5j–l). The juxtapeduncular migratory stream is also populated by numerous quail eminential cells which have followed a caudo-rostral pathway reaching the LH, and the LPO (Fig. 5g, g’, h’, i, i’, i’’, j–l). Some of these quail cells converge as well towards the anterior commissure (ac; Fig. 5k). The eminentio-septal migratory stream is not yet massive at this stage, but we observed some periventricular cells heading periventricularly rostralwards from the PThE, approaching the area of the anterior commissure, through which they will enter the septocommissural region (green arrowheads; ac; Fig.5h’, i, i’, i’’, j–l). Tbr1 expression mixed with QCPN-positive cell nuclei jointly corroborates the rostral and lateral contact of the PThE with the caudal hippocampus (Hi) and the pallial amygdala (PallA) (Fig.5i, i’).

We will examine next the Q294 chimera, which was sacrificed at HH35 (8.5–9 d.i.o.), sectioned sagittally, and doubly processed for Tbr1 ISH and QCPN IMR. This case consists of a very small eminential graft which also affected the chorioidal roof, the Hb and Th (Fig. 6a). The graft derivatives include a dorsocaudal portion of the PThE, a rostromedial part of Hb and Th, as well as part of the diencephalic chorioidal tela (Fig. 6a–l). The two medial sagittal sections shown (Fig. 6c,f and color-coded framed details 6d, e/g–j) illustrate double-labelled migrated cells (marked by brown nuclei and blue Tbr1-expressing cell bodies) along the three eminential streams. The juxtapeduncular migratory stream (black arrowheads) expands massively the PThE ventralwards and occupies the local lateral hypothalamus at hypothalamic paraventricular area levels (LH; Pa; Fig. 6c, e, f, j); some of its pioneering cells reach the superficial lateral preoptic area, wherein the more compact ‘preoptic island’ described in Alonso et al. (2020) is already distinguished (LPO, POI; Fig. 6e, f, i, j). The eminentio-septal migratory stream (green arrowheads; Fig. 6f) appears at this stage as a relatively compact band of deep periventricular double-labelled cells sorting out rostralward from the PThE; it enters the septocommissural domain represented by the anterior commissure (ac), its associated anterior commissure nucleus (AC), and the septal triangular nucleus (TS). The more compact dorsal elements of this stream are interpreted as the primordium of the commissural septal nucleus (CoS; Fig. 6f, g). The peripeduncular migratory stream is shown here only after it finishes its peripeduncular course proper (i.e., after coursing subpially all around the peduncle), and then passes subpially (superficially) from the lateral preoptic area into the diagonal band (red arrowheads, LPO, DB; Fig. 6c, f), on its way to the abundantly double-labelled extended amygdala (EA; the latter is shown in the more lateral sagittal section illustrated in Fig. 6k, l).

Long survival experiments display advanced stages of the studied tangential migrations and allow more precise anatomic identification of the labelled nuclei. Chimera M25 (Fig. 7a) was sacrificed at HH38 (12 d.o.i.), sectioned roughly orthogonal to the peduncle (Fig. 7b) and double-labelled with Tbr1 ISH and QCPN IMR (Fig.7c–q). The graft includes the local chorioidal tela (Fig. 7g, o–q), as well as a rostrodorsal portion of Hb (Fig. 7e, j), and the dorsalmost part of the presumptive PThE territory (Fig. 7a, g, n). It is similar in extent to the Q294 graft. Interestingly, ventral parts of the PThE whose ventricular zone is QCPN-negative (host-derived) show abundant intermixed quail cells in the mantle, suggesting that at least some migrating cells course through the PThE itself before entering one of the external migratory streams (Fig. 7h, j). Derivatives of the eminential juxtapeduncular and peripeduncular streams start to become histologically stabilized and their previous connective bridges with the PThE start to disappear (red and black arrowheads; Fig. 7c–g). In contrast, the more retarded eminentio-septal still maintains a straightforward migratory appearance well connected to the PThE (green arrowheads; Fig. 7e–g). The juxtapeduncular stream populates the LH, LPO, POI, MPO and MnPO target areas (Fig. 7c–h, k, m, n–q). In more rostral sections, scattered cells also appear in the Dg area (Dg; Fig. 7p, q). As regards peripeduncular stream derivatives, few double-labelled cells remain dorsolateral to the peduncle, forming what we called ‘eminential wings’ (red arrowheads, EW; Fig. 7d, e; Alonso et al. 2020), whereas denser labelled elements of the extratelencephalic portion of this stream are seen ventrolaterally to the peduncle, at the site tentatively identified as supraoptic nucleus (red arrowheads, SO; Fig .7c, d, i). The telencephalic portion shows labelled derivatives at the extended amygdala (red arrowheads, EA; Fig. 7g) and may contribute also to the preoptic and diagonal areas (not shown). The nuclear derivatives of the eminentio-septal stream, forming the eminentio-septal area (ESA) rostral to the PThE, are still immature at this stage on their way to the septocommissural area (green arrowheads, ESA; Fig. 7e, f, j, l). Many double-labelled cells start to aggregate at the primordia of the lateral and medial CoS nuclei (green arrowheads, CoSL, CoSM; Fig. 7g, n–p). In contrast, the nucleus of the hippocampal commissure and the triangular septal nucleus only show a few dispersed cells (green arrowheads, HiC, TS; Fig. 7o–q).

To check whether the septocommissural TS and HiC nuclei, the preoptic MnPO and MPO nuclei, and the diagonal band nuclei receive more massively migrated cells when the topologically dorsal and ventral regions of the PThE are jointly grafted, we performed some grafts covering the whole dorsoventral extent of the presumptive eminential territory. These chimeras usually also included small parts of the caudal MPall (Hi) and/or the presumptive PallA, or of the non-eminential central alar prethalamic territory. A representative example of such experiments is case Q325, sacrificed at HH39 (13 d.o.i.; Fig. 8a, b). At this stage, eminential migrations are very advanced, or have ended (Alonso et al. 2020). In the Q325 chimera almost all the presumptive territory of the PThE was grafted, only excepting its caudalmost portion next to the Hb (Fig. 8a). Parts of alar prethalamus, thalamus, and PallA, were labelled as well (Fig. 8a). In this chimeric brain there is a striking rostral dispersion of PThE-originated quail cells in the septocommissural periventricular stratum (small green arrowheads in Fig. 8c–f; the grafted PThE appears in Fig. 8g, h). We interpret this pattern as quail cells migrating through the deep eminentio-septal stream under the interventricular foramen. All nuclei previously described as recipients of eminential migrating cells (Alonso et al. 2020) display Tbr1-labelled quail cells (Fig. 8c–k). However, the MnPO and MPO nuclei are distinct in not showing massive populations of graft-derived quail cells, in comparison to their level of Tbr1 expression. The graft left out a caudal eminential region from which these prospective preoptic cells may have originated (Fig. 8a); however, we cannot rule out the possibility that these preoptic nuclei, similarly as the HiC, Dg and DB nuclei, may receive migrated Tbr1-positive cells from the telencephalic pallial region (Pombero et al. 2011; Alonso et al. 2020). In this chimera the subpallium appears invaded by numerous dispersed quail cells (Fig. 8h–m). These cells do not express Tbr1, a result which suggests an oligodendrocyte or astroglial nature of these dispersed elements (see Cobos et al. 2001a, b; Olivier et al. 2001).

To rule out non-eminential neighboring parts of the diencephalon as the source of cells migrated to the telencephalon, we made chimeras in which rostral parts of the presumptive diencephalon (Hb, Th, eventually including non-eminential PTh) were selectively grafted, ideally without inciding on the PThE (n = 11). We show here representative cases Q320 and Q278 (Fig. 9). Case Q320, sacrificed at HH37 (11 d.i.o.), is representative of controls including a small part of non-eminential prethalamus. Consistently with the fate map (Fig. 9a, b), this chimera only contained quail-derived cells at the Hb (brown labelled cells; Fig. 9c), dorsal part of Th (Fig. 9d–h), and dorsocaudal part of non-eminential PTh (Fig. 9d, e). The sections of this brain also were processed for Tbr1 in situ hybridization. The reaction was somewhat pale, but nevertheless it identifies a blue PThE devoid of quail cells (Fig. 9e–g), and none such cells appear at the recipient nuclei of the eminential migrations (e.g., CoS, HiC, EA, POA, EW; Fig. 9f–i). Case Q278 received a graft only affecting the habenula and neighboring rostral part of thalamus (Fig. 9j–m). The Tbr1 reaction was clearcut in this case, identifying the PThE (Fig. 9n, o) and some of the targets of its migrations (Fig. 9n–r), all devoid of quail cells. These results jointly allow us concluding that the non-eminential PTh, Th and Hb do not participate in the source of eminential cells that migrate into the lateral hypothalamus and telencephalon.