Nuclear derivatives and axonal projections originating from rhombomere 4 in the mouse hindbrain
- 1k Downloads
The r4-derived territory is located in the pontine region of the brainstem, forming a wedge-shaped slice that broadens from the choroidal roof to the ventral midline. R4-derived neuronal populations migrate radially inside and tangentially outside this rhombomere, forming nuclei of the sensorimotor auditory, vestibular, trigeminal and reticular systems. R4-derived fibre tracts contribute to the lateral lemniscus, the trigeminothalamic tracts, the medial tegmental tract and the medial forebrain bundle, which variously project to the midbrain, thalamus, hypothalamus and telencephalon. Other tracts such as the trigeminocerebellar and vestibulocerebellar tracts reach the cerebellum, while the medial and lateral vestibulospinal tracts, and the reticulospinal and trigeminal oro-spinal tracts extend into the spinal cord. Many r4-derived fibres are crossed; they decussate to the contralateral side traversing the midline through the cerebellar, collicular and intercollicular commissures, as well as the supraoptic decussation. Moreover, some fibres enter into the posterior and anterior commissures and some terminals reach the septum. Overall, this study provides an overview of all r4 neuronal populations and axonal tracts from their embryonic origin to the adult final location and target.
KeywordsFate map Rhombomere 4 b1r4-Cre line Sensorimotor systems
Facial nerve or root
Anteroventral cochlear nucleus
Anterior pretectal nucleus, dorsal part
Anterior pretectal nucleus, ventral part
Caudal periolivary nucleus
Dorsal cochlear nucleus
Dorsomedial tegmental nucleus
Dorsal periolivary area
Dorsal lateral geniculate nucleus
Dorsal nucleus of the lateral lemniscus
Facial branchiomotor neuron
Inner efferent neuron
Intermediate nucleus of the lateral lemniscus
Intermediate reticular nucleus
Inferior vestibular nucleus
Lateral olivocochlear neurons
Lateral nucleus of the trapezoid body
Lateral superior olive
Lateral vestibular nucleus
Lateroventral periolivary nucleus
Medial geniculate nucleus
Medial geniculate nucleus, dorsal part
Medial geniculate nucleus, medial part
Medial geniculate nucleus, ventral part
Motor trigeminal nucleus
Medial olivocochlear neurons
Medial nucleus of trapezoid body
Medial superior olive
Posterior thalamic nucleus
Pontine grey nucleus
Pontine reticular nucleus, caudal part
Pontine reticular nucleus, oral part
Posterodorsal tegmental area
Principal trigeminal nucleus
Posteroventral cochlear nucleus
Ventral cochlear nucleus
Rostral periolivary region
Reticulotegmental nucleus, central part
Reticulotegmental nucleus, pericentral part
Spinal trigeminal nucleus, interpolar part
Spinal trigeminal nucleus, oral part
Spinal trigeminal nucleus, dorsomedial oral part
Spinal trigeminal nucleus, ventrolateral oral part
Superior olivary complex
Superior vestibular nucleus
Subcoeruleus nucleus, dorsal part
Subcoeruleus nucleus, ventral part
Vestibular efferent neurons
Ventral nucleus of the lateral lemniscus
Ventral nucleus of the trapezoid body
Ventral posteromedial nucleus
Zona incerta, dorsal part
Zona incerta, ventral part
Hindbrain neurons process and relay sensory information, control vital functions and contribute to motor coordination. The complex circuitry formed by sensory, reticular/interneuronal and motor neurons is established during development and depends on a spatially and temporally ordered sequence of neuronal specification, cell migration and axonal pathfinding.
The immature rostral hindbrain of vertebrates is overtly segmented into molecularly and morphologically distinct developmental (histogenetic) compartments, called rhombomeres (r), forming the series r1–r6 along the anteroposterior (AP) axis (Fraser et al. 1990; Graham et al. 2014; Kiecker and Lumsden 2005; Lumsden 1990; Puelles et al. 2013; Vaage 1969). Each of these rhombomeres, except r1, expresses a particular combination of Hox and other genes, which specify their respective molecular identity and developmental fate (Krumlauf et al. 1993; Tomas-Roca et al. 2016; Tumpel et al. 2009). The paralogue Hox groups 1–3 operate instructively across this domain in an anteroposterior series that follows the principle of 3′–5′ collinearity (Krumlauf et al. 1993; Parrish et al. 2009). The caudal medullary hindbrain has instead a hidden rhombomeric organization, which is molecularly detectable using Hox gene markers, but is not distinct morphologically as visible transverse bulges. This domain was divided into five cryptorhombomeres (r7–r11) (Cambronero and Puelles 2000; Puelles et al. 2013), which also display typical anteroposterior (3′–5′) spatial colinearity of Hox gene expression; the rostral expression limits of the paralogue Hox groups 4–8 sequentially coincide with the experimentally fate-mapped limits of r7–r11 (Marin et al. 2008; Tomas-Roca et al. 2016). A similar case is posed by the isthmic cryptorhombomere (r0), which only can be delimited molecularly from r1 (Aroca and Puelles 2005; Puelles et al. 2013). Embryonic hindbrain neuromeres or segments give rise to the prepontine region (isthmus plus r1 and r2; includes the cerebellum), the pons (r3, r4), the retropontine region (r5, r6) and the medulla oblongata (r7–r11) in the adult brain (Puelles 2013; Puelles et al. 2013). The cryptically delimited isthmus (r0) and r1 are patterned under the influence of the isthmic organizer, which also induces across their dorsal aspect the cerebellum.
Long-term fate-mapping studies of rhombomeres in chicken (Cambronero and Puelles 2000; Cramer et al. 2000; Diaz et al. 1998; Marin et al. 2008; Marin and Puelles 1995; Puelles et al. 2013; Tan and Le Douarin 1991) and mouse (Di Bonito et al. 2013a, 2015; Di Meglio et al. 2013; Farago et al. 2006; Gray 2013; Oury et al. 2006; Pasqualetti et al. 2007; Tomas-Roca et al. 2016) have revealed that various rhombomeres contribute to neurons of the auditory, vestibular, trigeminal, somatosensory, reticular and precerebellar systems.
Within each rhombomere, inductive signals and specific transcriptional pathways confer a positional pattern on the neural progenitors, linking their position along the AP and DV axes to differential specification of characteristic neuron subtypes, specific migratory behaviour of given cell populations and axonal projections to characteristic targets, contributing to the complexity of hindbrain sensorimotor circuits (reviews by Di Bonito et al. 2013b; Philippidou and Dasen 2013). Neuromeric classification of classic anatomic entities in the hindbrain often entails the recognition of their bi- or plurineuromeric composition (e.g. postmigratory basilar pons placed across r3 and r4, or trigeminal motor nucleus originated across r2 and r3). Moreover, the interaction between the unique molecular identity of each rhombomere and a shared mechanism of dorsoventral patterning (dorsalization versus ventralization) leads to functionally characteristic alar and basal derivatives of each rhombomere and cryptorhombomere, which participate in the plurineuromeric modular neuronal arrangements obtained within hindbrain columns (e.g. the cochlear, vestibular and trigeminal sensory columns).
In the mouse, the organization of the facial somatosensory map is related to the r2 and r3 components of the principal trigeminal sensory nucleus (Oury et al. 2006; Pouchelon et al. 2012). Vestibular nuclei originate from multiple rhombomeres ranging from r2 to r9, and the corresponding segmental modules display specific axonal projections to distinct targets (Chen et al. 2012; Di Bonito et al. 2015; Pasqualetti et al. 2007). Some neuromere-selective projection patterns have been described. For instance, the chicken r4 vestibular module does not produce vestibulo-ocular projection neurons and only few of them arise in mice r4, while r4 contributes mainly contralaterally and ipsilaterally projecting vestibulospinal neurons (Di Bonito et al. 2015; Diaz et al. 1998, 2003).
In the auditory system, the cochlear nuclear complex represents an important plurineuromeric relay station of the auditory pathway, which derives from the dorsal rim of the r2–r5 units, in particular from Atoh1- and Ptf1-expressing progenitor domains (Farago et al. 2006; Fujiyama et al. 2009; Maricich et al. 2009; Rose et al. 2009). R5 contributes massively to the superior olivary complex (SOC) (Maricich et al. 2009), whereas r4 contributes only partially to the LSO (Di Bonito et al. 2013a; Marin and Puelles 1995). Along the DV axis, the SOC derives from both rhombic lip and non-rhombic lip progenitor domains (Altieri et al. 2015; Machold and Fishell 2005; Marrs et al. 2013), namely from Atoh1- (Maricich et al. 2009), Wnt1- (Fu et al. 2011; Marrs et al. 2013) and Wnt3a-positive (Louvi et al. 2007) rhombic lip domains dorsally and from an En1-positive lineage (Altieri et al. 2015; Marrs et al. 2013) ventrally, respectively. The ventral nucleus of the lateral lemniscus (VLL) derives from r4 (Di Bonito et al. 2013a), while the dorsal nucleus of the lateral lemniscal (DLL) derives from r1 (as is suggested by its Pax7 expression, which only occurs in the r1 mantle; Lorente-Canovas et al. 2012; Moreno-Bravo et al. 2014 misidentified it as ILL; Allen Developing Mouse Brain Atlas).
Different rhombomeres thus contribute to different functional modular subsystems of the auditory system. R4 contributes mainly to sound perception; moreover, r4-derived cochlear sensory neurons contribute jointly with basal plate-derived motor structures to the two distinct auditory sensorimotor feedback subcircuits that protect the cochlea from acoustic overstimulation; r2, r3 and r5 instead contribute to sound localization circuitry (Di Bonito et al. 2013a, b).
Precerebellar nuclei, essential for the relay of peripheral proprioceptive signals and cortical collateral copy input to the cerebellum, are generated from the rhombic lip in the caudal hindbrain (precerebellar lip), which extends from r6 to r11 (Cambronero and Puelles 2000; Hidalgo-Sanchez et al. 2012; Landsberg et al. 2005; Machold and Fishell 2005; Marin and Puelles 1995; Puelles et al. 2013; Rodriguez and Dymecki 2000; Wang et al. 2005; Wingate 2001). They originate along the dorsoventral axis from discrete molecularly defined pools of progenitor cells expressing Atoh1, Ngn1, Mash1 and Ptf1a (reviewed by Ray and Dymecki 2009). In particular, the basilar pontine nuclei derive from the lower rhombic lip over r6–r8. They follow a subpial tangential migration path into the basilar region of r3 and r4, during which neuronal subsets maintain their relative positions (Di Meglio et al. 2013).
Finally, r4 is known to produce in its basal plate the facial branchiomotor neurons (FBM) as well as inner ear efferent neurons (IEE) whose axons course through the facial (7n) and vestibulocochlear (8n) cranial nerves (Auclair et al. 1996; Bruce et al. 1997; Simon and Lumsden 1993).
As is well known, the Hoxb1 gene specifies singularly the identity of the dorsoventral sets of progenitor fields found within r4 (Di Bonito et al. 2013a, 2015; Studer et al. 1996). A late r4-Hoxb1 enhancer specifically maintains Hoxb1 transcription at high levels in the whole r4 (Studer et al. 1998, 1994); later in development Hoxb1 expression becomes restricted to distinct dorsoventral neuronal subdomains corresponding to specific sensory and motor columns (Gaufo et al. 2000; Gavalas et al. 2003). Hoxb1-deficient mouse mutants and human patients carrying mutations in the HOXB1 gene have severe impairments in the auditory, trigeminal and vestibular systems (Chen et al. 2012; Di Bonito et al. 2013a, 2015; Studer et al. 1996; Webb et al. 2012). Hoxb1 activates an r4-specific developmental programme and represses locally the emergence of r3-like molecular and cellular features by inhibiting Hoxa2, Atoh1 and Ascl1 expression in the trigeminal, auditory and vestibular columns via different molecular pathways (Chen et al. 2012; Di Bonito et al. 2013a, b, 2015; Gaufo et al. 2000). As a consequence, lack of function of Hoxb1, the major r4 selector gene, entails that r4-derived motor and sensory neurons of the auditory, vestibular and trigeminal systems acquire molecular profiles and neuronal behaviours (migration and axonal pathfinding) characteristic of their r3 analogs. Thus, repatterning related to Hoxb1 loss-of-function changes the molecular identity of r4 progenitors into an r3-like identity and promotes a complete reorganization of several crucial axonal pathways and neuronal circuits in the mature brainstem.
In mouse, the use of the Cre–lox system and of various Cre-recombinase lines, either specific to distinct rhombomeres (Altieri et al. 2015; Di Bonito et al. 2013a, 2015; Di Meglio et al. 2013; Farago et al. 2006; Jensen et al. 2008; Maricich et al. 2009; Oury et al. 2006; Pasqualetti et al. 2007), or to specific DV subdomains (Altieri et al. 2016; Fujiyama et al. 2009; Kim et al. 2008; Marrs et al. 2013; Rose et al. 2009; Storm et al. 2009; Wang et al. 2005), have been instrumental in mapping the origin of some hindbrain neuronal subtypes, eventually following their migratory displacements during development and their projections. However, only few rhombomere-specific lines including Rse-Cre (r2) (Awatramani et al. 2003), Krox20 Cre or Egr2 Cre (r3/r5) (Voiculescu et al. 2000), Hoxb1 Cre (r4 to posterior) (Arenkiel et al. 2003) and Hoxa3-Cre (r5/r6) (Di Meglio et al. 2013) have been used to dissect the assembly of sensorimotor systems.
In order to study the neuronal subtypes originating from r4, we generated an r4-restricted Cre driver in the mouse (named b1r4-Cre) (Di Bonito et al. 2013a). The b1r4-Cre is a transgenic mouse line that has been obtained using a well-characterized r4 enhancer from the Hoxb1 locus (Studer et al. 1994) to express the Cre-recombinase gene exclusively in r4. This contrasts with the previous lines Hoxb1 Cre (Altieri et al. 2016; Arenkiel et al. 2003; Farago et al. 2006; Maricich et al. 2009; Marrs et al. 2013) and r4-Cre (Geisen et al. 2008; Oury et al. 2006) which were described as r4-specific lines, though they invariably also display additional Cre expression caudal to r4. In the CNS, no ectopic Cre expression has been observed in the transgenic b1r4-Cre so far (Di Bonito et al. 2013a; their Figure S1D, and data not shown), but the theoretic possibility of some ectopic expression cannot be fully discounted. Crossing of the b1r4-Cre line with the ROSA26-YFP reporter line has allowed us to fate map selectively all r4 derivatives from E9.0 until adulthood. Previously, we described summarily the auditory, trigeminal and vestibular r4 derivatives (Di Bonito et al. 2013a, b, 2015). In this paper, we provide a full overview at different stages of all r4 neuronal populations and axonal tracts, followed from their early embryonic origins to the final adult locations and targets.
Materials and methods
The b1r4-Cre mice (Di Bonito et al. 2013a) were crossed with ROSA26-YFP reporter mouse (Srinivas et al. 2001) to obtain double heterozygous b1r4-Cre/YFP progeny, in which r4 and r4 derivatives are permanently labelled. All mouse experiments were performed according to the protocols approved by the Institutional Animal Care and Use Committee of the University of Nice Sophia Antipolis, Nice, France.
Adult and P8 mice were perfused with cold 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) pH 7.4. Embryos and perfused brains were fixed overnight in buffered 4% PFA. Tissues were cryoprotected with 10, 20 and 30% sucrose in PBS and frozen in OCT embedding matrix (Kaltek), to be cryostat-sectioned at 16 µm thickness (E10.5, E11.5 and E12.5 embryos, sagittal plane; E14.5 embryos, sagittal and r4-adapted horizontal and coronal section planes; E16.5 brains, sagittally and horizontally, and E18.5 brains, sagittally, as well as in r4-adapted horizontal and coronal section planes). P8 and adult brains were cut 20 µm thick sagittally or coronally, respectively.
Immunohistochemistry and Nissl staining
Immunohistochemistry was performed as previously described (Di Bonito et al. 2013a), using rabbit polyclonal anti-GFP antibody (1:500, Molecular Probes) as a primary antibody, which cross-reacts with the YFP protein. Nissl staining of cryostat sections was performed using standard procedures and the slides were mounted with EUKITT mounting medium.
Sagittal and coronal brain cryosections were incubated for 1 h at room temperature with blocking solution (10% goat serum; 0.3% Triton X-100 in PBS) and then overnight at 4 °C with primary antibodies diluted in the hybridization buffer (3% goat serum; 0.3% Triton X-100 in PBS): rabbit anti-GFP (1:500, Molecular Probes) and mouse anti-Pax7 (1:100, Developmental Studies Hybridoma Bank). Tissue was washed in PBS and incubated for 1 h at room temperature with secondary antibodies: Alexa Fluor 488 (green) goat anti-rabbit and Alexa Fluor 555 (red) goat anti-mouse (1:400, Molecular Probes). The slides were mounted with 2% N-propyl gallate in 90% glycerol in PBS (Sigma P3130).
In situ hybridization
In situ hybridization for Atoh7, vGluT2, Gad67 and Gata3 was performed as previously described (Di Bonito et al. 2013a).
Digital bright-field microphotographs were taken on a Leica DM 6000B microscope equipped with a Leica DFC310 FX colour camera, and processed with Adobe Photoshop CS5 software using the Photomerge function to obtain a panorama of each brain section. Digital microphotographs of immunofluorescence images were acquired by laser scanning confocal microscopy using a Zeiss LSM 710 confocal microscope and converted by ZEN software. The figures were mounted and labelled using Adobe Photoshop CS5.
The r4-derived domain forms an intensely labelled wedge in the pontine region that expands from the choroidal roof down to the midline raphe. As the development proceeds, the domain becomes more and more compressed anteroposteriorly, particularly at the choroidal roof, rhombic lip area and neighbouring dorsalmost alar plate (cochlear complex).
Our description below of the diverse derivatives of r4 will be divided into those which migrate out from this rhombomere into the neighbouring ones (tangentially migrated derivatives) and those which remain inside the r4 domain (radially migrated derivatives). In addition, we will consider the labelled fibre tracts in the last section.
Radial migration refers to ventriculo-pial displacement of newborn postmitotic cells within a subregion of the neural tube wall. Radially migrated cells are stabilized permanently in the mantle close to their ventricular matrix site, building the local radial histogenetic domain. In contrast, tangentially migrating postmitotic neurons may secondarily move away from their source, invading other parts of the neural wall, either of the same or a different neuromere, at close or longer range. It is widely thought that causal mechanisms, molecular guidance and even cytoskeletal translocation mechanisms for these two types of cell migration are substantially different. Accurate description of the histogenetic processes occurring within a developmental unit such as a neuromere requires tracking any cases of tangential migration, either entering or sorting out of the unit.
Tangentially migrated r4 derivatives
Early on, two major streams of migratory cells leave r4 and invade adjacent regions.
Facial motor nucleus
Caudally, the well-known caudalward medial migratory stream of the facial motor nucleus starts to exit r4 at E10 (Fig. 1a, b/bʹ) and enters r6 at E11.5 (Fig. 1g, h/hʹ). At the latter stage, the massively labelled stream extends caudalwards in a slightly diverging course just outside of the r5 paramedian basal ventricular zone, adjacent to the floor plate and then turns laterally as it reaches the r6 basal mantle (Fig. 1g, h/hʹ–j/jʹ). At E12.5, the stream still appears connected rostrally to r4 (Fig. 2a), but already proceeds lateralwards (ventrodorsally) within r6, reaching a locus judged to correspond to the ventral rim of the r6 alar plate (this is consistent with the chick results reported by Ju et al. 2004, as well as with unpublished observations of LP in mouse). The facial nucleus primordium next spreads out progressively as the cells stream ventrolaterally, approaching the pial surface (Fig. 2b–f). In this migratory process, the motoneurons initially arch medially around the abducens motor nucleus in r5, forming the prospective genu of the facial nerve (Fig. 3k). A much smaller but distinct group of labelled neurons appears associated to the periventricular facial genu; these may represent the efferent vestibular neurons described in the literature. At higher magnification, the facial nerve axons are distinctly labelled and can be followed from the migrated facial nucleus (r6) into the genu (r5), and from there across r4 to their exit into the facial nerve root.
Lateral lemniscus complex
Another sizeable, though less compact, exiting stream of labelled r4 cells emerges rostrally from the basal longitudinal zone after E10 (Fig. 2d–f). These cells are preceded by a growing ascending tract that was identified at older stages as the lateral lemniscus (Fig. 1b/bʹ–d/dʹ, h/hʹ–k/kʹ, 2b–g; this tract carries crossed fibres from the cochlear nuclei, which comprise an r4 module, plus uncrossed fibres of the largely r5-derived superior olivary complex and of the interstitial cell populations known as bed nuclei of the lateral lemniscus; Malmierca and Merchán 2004). The migrating cells move along or inside this superficial tract, which first courses obliquely rostralwards through the basal plate of r3 and then passes into the alar plate to ascend through the prepontine alar hindbrain (r2, r1, isthmus) into the inferior colliculus of the midbrain. By E12.5, the cells have advanced about 140 micrometres beyond the rostral boundary of r4, apparently lying still mainly within r3 (Fig. 2e, f). By E14.5, the pioneering elements of this cell population extend farther along the growing lateral lemniscus, stopping roughly at the estimated boundary between r2 and r1 (Fig. 3b–g). At subsequent stages (we examined E16.5, Figs. 4, 5, 6; E18.5, Figs. 7, 8, 9, 10; P8, Figs. 10, 11, 12, 13; adult Fig. 14), the migrated labelled cell population lying interstitial to the lateral lemniscus remains stretched between r2 and r4 and clearly correlates topographically with the ventral nucleus of the lateral lemniscus (VLL; Di Bonito et al. 2013a; Ito et al. 2011; Malmierca and Merchán 2004; Figs. 4b–f, 7f–k, 11h–p, 14o–s). No labelled cells were found at the more rostral r1 locus of the Pax7- and Gad67-positive dorsal nucleus of the lateral lemniscus (Figs. 15b–i, k–p, 16; Moreno-Bravo et al. 2014; Allen Developing Mouse Brain Atlas). The green-labelled cells Gad67- and Gata3-positive clearly represent the majority of neurons in the VLL nucleus, with few green neurons spread in the space between DLL and the massive VLL occupied by the VGluT2-positive intermediate lemniscal nucleus (ILL) (Figs. 10a, b, d, e, 14o–s, 15, 16).
At dorsoventral levels corresponding to the vestibular column, a number of labelled neurons exit r4 rostrally and caudally; they come to a stable new position in r3 and r5, respectively, always within the vestibular complex. The r4 vestibular elements invading r5 (i.e. the inferior vestibular nucleus) first appear at E10.5 (Fig. 1a/aʹ, e/eʹ), and their number increases at E11.5 (Fig. 1g/gʹ, k/kʹ, l/lʹ) and E12.5 (Fig. 2g–i). The migrated cell group is constituted by distinct large stellate neurons which remain visible at E14.5 (Fig. 3d, e, k), E16.5 (Figs. 4c, 5n–t), E18.5 (Figs. 7h–k, 9g, h, 10 f), P8 (Fig. 11l–o) and the adult (Fig. 14a). We found no labelled migrated cells within the inferior vestibular nucleus at r6 levels, or more caudally. On the other hand, the r4 vestibular elements invading the r3 vestibular column module (i.e., the superior vestibular nucleus) start to sort out of r4 between E10.5 and E12.5 (Fig. 1a, e/eʹ, l/lʹ, 2h, i). They incipiently aggregate within r3 at E14.5 (Fig. 3a, b) and later form a dense, well-circumscribed and large-celled aggregate at E16.5 (Figs. 4a, 5n–t, 6l), E18.5 (Figs. 7f, 9f–h, 10f), P8 (Fig. 11i, j) and the adult (Fig. 14b, c). We observed some sparser labelled cells more rostrally in the vestibular column (superior vestibular nucleus), which possibly lie within r2 (Fig. 5m).
At E18.5 and postnatally, we observed some dispersed labelled neurons in the ventral and dorsal cochlear nuclei outside of the central labelled wedge representing the r4 cochlear column module (Figs. 9g, h, 10g, 11a–h, 12a–e). The apparently tangentially migrated population is more important in the anteroventral cochlear nucleus (rostral to r4). Notably, r4 does not contribute to the granule cells of the microneuronal shell of the cochlear complex (Fig. 12a–e; see also Di Bonito et al. 2013a; their Fig. 1J).
After E10.5, there appear some dispersed labelled reticular neurons in r5, usually close to r4 (Fig. 1i/iʹ–k/kʹ; 2b–h; 3c–i; 4d–i; 5u–y; 7l–s; 11l–q). Dispersed labelled reticular elements were also observed after E14.5 rostral to r4, mainly medial to the trigeminal sensory column in r3 and r2, or in a neighbouring paramedian tegmental position (Figs. 3d–i, 4d–i, 5u–w, 7l–t, 11l–q).
Superior olivary complex
The lateral olivocochlear (LOC) and medial olivocochlear (MOC) efferent neurons (Brown and Levine 2008; Simmons 2002) are part of the inner ear efferent system (IEE); they are born in ventral (basal) r4 and are identified as migrated into the r5 territory by their characteristic ChAT expression (Di Bonito et al. 2013a). They lie in or near the lateral superior olive (LSO; Fig. 14g–j and inset), or in the rostral and ventral periolivary region (Fig. 14l, m and inset).
The cerebellum is a derivative of the isthmus (vermis) and r1 (hemisphere). We observed a number of small labelled cells in the cerebellar white matter, which we identified as oligodendrocytes (Buffo and Rossi 2013) migrated out of r4 at E18.5 (Fig. 7a–e), P8 (Figs. 11a–j, 12a–e; see also the merged YFP/Nissl images) and in the adult (Fig. 14a–k).
R4 apparently contributes via rostralward tangential migration to the dorsomedial tegmental nucleus (DMTg), which is found postnatally within caudal r1 or r2 (Fig. 14g, h). Labelled neurons also characterize the nucleus subcoeruleus, found caudally to the locus coeruleus, which is restricted to r1 (SubC; Fig. 14h, i).
Tangentially migrated medullary derivatives invading r4 mix with local oligodendroglia
The medullo-pontine migration of r6–r8 rhombic lip precerebellar cells that eventually builds up the basilar pontine nuclei and the reticulotegmental nucleus in r3 and r4 is very well known and will thus not be described in detail here. The relationship of r4 with the prospective pons only begins to be appreciated after E16.5, when the unlabelled pontine migration stream reaches its target locus (Figs. 4c–f, 5v–y). Although the migrated neuronal population of the basilar pons primordium is largely unlabelled, it contains internally after E16.5 and postnatally a thinly dispersed labelled population as well as an outer crust of labelled cells, which may be glial in nature (Figs. 4d–i, 5v–y, 7l–s, 11p, q). This result was expected, since the basilar pontine neurons migrate in from more caudal parts of the rhombic lip (Altman and Bayer 1978) and thus should be unlabelled in our material, and a large production of oligodendrocytes occurs in the pontine r4 territory (Miguez et al. 2012). The scattered small (presumably glial) labelled pontine cells are present exclusively inside the r4 pontine area, being absent in the pontine sector belonging to r3 (compare Fig. 10a, b). The relatively small r3 pontine sector can be best identified at E18.5 and P8 (Figs. 7l–s, 10a, b, d, e, 11p, q). In the adult, YFP-labelled oligodendrocytes strongly populate the pontine grey (Pn) (Fig. 14n–u), as well as the central and pericentral parts of the reticulotegmental nucleus, which derives likewise from the bulbopontine migration (RtTg, RtTgP; Fig. 14j–s).
Radially migrated r4 derivatives
At early stages, the identification of specific derivatives is handicapped both by their immaturity and the dense labelling obtained after the GFP immunoreaction. The facial and vestibulocochlear roots are useful r4 landmarks, since they are enclosed by the transverse boundaries of this segment. The r3/r4 boundary passes internally rostral to the facial knee and just caudal to the trigeminal motor nucleus (the trigeminal motor nucleus is placed in r2 and r3, as we corroborated in parallel ChAT-immunoreacted and Nissl-stained sectioned material; not shown). The r4/r5 limit passes rostral to the abducens nucleus lying in r5 (as we verified with ChAT-immunoreacted and Nissl-stained parallel series; not shown).
Our interpretation of subsequent stages, in which the mantle layer starts to reveal structures more or less distinguishable one from another, relies on the assumption of a well-known columnar structure within alar r4, which is shared with the neighbouring neuromeric parts of the hindbrain; that is, we assumed that structural sectors labelled within r4 that are serially continuous with characteristic sensory columns identified outside r4 represent r4 components or modules of these columns. We accordingly found, as expected, a selectively labelled r4 sector intercalated in the cochlear, vestibular and trigeminal columns, as well as in the reticular formation.
The pattern found at the cochlear column was clearest in lateral sagittal sections at P8 (Figs. 10g–i, 11a–h, 12), in which a wedge-shaped labelled sector intersects both the ventral and dorsal nuclei of the cochlear complex, consistently with adult coronal sections (Fig. 14a–j). Most of the posteroventral cochlear nucleus appears labelled (which agrees with the classic observation that the cochlear nerve root enters through it; Lorente de Nó 1981; Malmierca and Merchán 2004), whereas the anteroventral cochlear nucleus, which is differentially positive for Atoh7, is largely unlabelled, excepting the dispersed labelled cells mentioned above (Fig. 12). A rostroventral part of the dorsal cochlear nucleus also becomes densely labelled, while the caudodorsal part and the extreme rostroventral part remain unlabelled, except for few dispersed elements (Figs. 10g, 11a–h, 12). More medial sagittal sections and horizontal sections show a progressive lateromedial thinning of the labelled wedge intercalated within the cochlear complex as the latter extends medialwards periventricularly (Figs. 5m–o, 7n, 9g, h). Lateral sections show that the labelled wedge extends uninterruptedly dorsalwards into the rhombic lip and the attached labelled portion of the choroidal tela of the IV ventricle (Fig. 4a).
The findings for the vestibular column are already implicit in our description of the tangentially migrated components of this column. Rhombomere 4 apparently generates a good number of large multipolar vestibular neurons, which form locally the lateral vestibular nucleus, whereas the labelled multipolar vestibular cells that invade r3 and r5 are integrated within the superior vestibular and inferior vestibular parts of the same column (Figs. 10f, 11i–o, 14a–e).
The vestibular efferent neurons (VEN) born in basal r4 move into a position dorsal to the facial nerve genu within r4-derived territory (Martinez-de-la-Torre et al. 2017; Simmons 2002; see also Di Bonito et al. 2015; their Fig. 5a–c), and their axons enter the vestibulocochlear nerve together with those of the lateral and medial olivocochlear neurons.
The alar r4 domain labelled ventral to the vestibular column corresponds to the trigeminal column (Fig. 5p–w). This sector represents a segmental module of the oral subnucleus of the descending trigeminal nucleus (Sp5O) (Fig. 14a–e), which is divided into dorsomedial and ventrolateral parts (Sp5ODM, Sp5OVL). This interpretation rests on the assumption that the main trigeminal sensory nucleus (Pr5) is restricted to r2 and r3 correlatively with the ascending branch of the trigeminal nerve root (Oury et al. 2006). At r4 levels, no part of the viscerosensory column is present yet (it starts at r7 level; Martinez-de-la-Torre et al. 2017).
Since the branchiomotor elements of r4 migrate tangentially away into r6 (the facial nucleus), while the parasympathetic preganglionic superior salivatory nucleus supposedly originates and remains in r5, it can be assumed that the rest of labelled r4 tegmentum found ventral to the trigeminal column and dorsal to the basilar pons primordium should largely contain lateral, intermediate and medial populations of the r4 reticular formation. The rostral part of the intermediate reticular nucleus (IRt) lies in the YFP-positive region in adult coronal sections (Fig. 14e–g). R4 possibly also contributes to the caudal and oral pontine reticular nuclei of the medial reticular formation (PnC, PnO; Fig. 14i–k). We already commented above that at this level the boundaries of r4 are made fuzzy by the dispersion of some reticular neurons out of r4 into neighbouring areas of r3 and r5. Rostral to the facial nucleus, in a subpial position, the most lateral part of the lateral superior olive (LSO) (Fig. 14f–j) and the dorsal periolivary area (DPO) (Fig. 14f–j) are labelled, implying that they contain r4-derived cells; it is unclear whether these cells are located within r4 territory (being radially migrated), or have migrated tangentially into r5 (the SOC largely lies in r5). In a more rostral position, the rostral periolivary region (RPO) is strongly and entirely YFP positive, and therefore we ascribe it to r4 proper (Fig. 14k–p), jointly with the r4-located caudal portion of the VLL nucleus. Labelled LOC and MOC efferent neurons seem to be located both in r5- and r4-derived territory (Fig. 14g–j, l, m and insets; see also Di Bonito et al. 2013a; their Fig. 1f and S7).
The r4-derived radial histogenetic domain is delimited by the choroidal roof dorsally and the midline raphe ventrally.
The flattened neuroepithelial cells of the r4 choroidal roof also appear intensely labelled; they form a distinct transverse labelled band across the hindbrain roof (Figs. 1f, j–l, 2c–i, 3a–e, h, p, 4a–j, 7i–t, 9e, f, 11, 12).
The hindbrain floor plate apparently lacks neurogenetic activity (no known neuronal derivatives) and its neuroepithelial cells directly differentiate into radial glial cells. Their cell bodies lie in the local median ventricular zone, and they have basal (radial) cytoplasmic processes that reach the medioventral pial surface. Collectively, these cells build a median palisade that is conventionally known as “the hindbrain median raphe”. Labelled r4 glial cells of the median raphe which show the characteristic radial ventriculo-pial morphology are observed in our sagittally sectioned material.
Midsagittal sections show that the intensely labelled median glial raphe cells belonging to r4 encompass the major part of the pontine bulge in a fan-like median sagittal expansion (Figs. 3j, 4j, 5w–y, 7t). A smaller rostral sector of the pontine bulge whose raphe remains largely unlabelled probably corresponds to r3 (e.g. Fig. 7s, t). Interestingly, there usually appear some labelled median radial glial cells outside of r4, either in r3 or r5. This was not observed in other parts of the neural wall and may be due to the reported absence of interrhombomeric clonal restriction boundaries across the floor plate (Fraser et al. 1990); unfortunately, this cannot be verified in our material, since our labelling is polyclonal.
Labelled fibre tracts
A surprising number of tracts appear labelled in our material (Fig. 17). Some of them can be identified confidently, though they present novel aspects, whereas others must be given tentative identities. Clearly, we can provide here only a preliminary descriptive analysis and this subject will need additional experimental work.
As mentioned above in connection with the ventral nucleus of the lateral lemniscus, the tangentially migrating lemniscal neurons are accompanied by labelled growing fibres within the tract, which presumably come from projection neurons in the r4 module of the contralateral cochlear nuclei as well as from the migrating VLL neurons themselves, plus the small r4-derived component of ipsilateral SOC cells (Malmierca and Merchán 2004). These fibres reach the inferior colliculus (deep to its marginal stratum) between E14.5 and E16.5 (Figs. 3u, 4a–c, 5c–g), but do not penetrate radially into its deeper parts yet; this happens between E16.5 and E18.5 (Figs. 7c–h, 9c–f). Surprisingly, already at E14.5 we observed that many of these labelled lemniscal fibres extend through the brachium of the inferior colliculus into the thalamus (Fig. 3l–o) and beyond. In the thalamus, the labelled lateral lemniscal fibres course longitudinally roughly through the area occupied by the medial geniculate nucleus (Figs. 8a–c, 11d–j, 13a–e); the packet of fibres then progressively approaches the brain surface more rostrally (Figs. 4a, b, 5m–o, 6a–d). Rostral to the medial geniculate body (prethalamus), the lemniscal fibres adopt a subpial position ventral to the optic tract, and at E16.5 they extend across the hypothalamus all the way to the supraoptic decussation (Figs. 5p–y, 6g–i). At E18.5 and P8, the lateral lemniscus shows labelled profuse terminal innervation of the inferior colliculus, as well as labelled fibres in the intercollicular commissure; moreover, collaterals from the ascending lateral lemniscus fibres in the brachium of the inferior colliculus enter the deep stratum of the superior colliculus (superficial to the periaqueductal grey), and some of them extend into the tectal commissure (Figs. 7b–t, 8, 9a–f, 11e–q).
Medial tegmental tract and medial forebrain bundle
Another massive ascending longitudinal tract coming out of r4 forms a thick ascending packet in the medial tegmentum of the upper hindbrain; afterwards, it crosses the midbrain and diencephalic tegmentum in a similar position and apparently contributes within the lateral hypothalamus to the medial forebrain bundle; the tract eventually reaches the telencephalon via the peduncular hypothalamus (Puelles et al. 2012, 2013; Puelles and Rubenstein 2015). We assume that this packet of fibres is a segmental r4 component of a larger tract with other segmental contributors, probably the central tegmental tract (see “Discussion”), though it might include some other components. The earliest appearance of the r4-labelled medial tegmental tract is at E11.5. Its fibres barely reach the isthmus at this stage (Fig. 1h/hʹ, i/iʹ). At E12.5, they extend across the midbrain and diencephalic tegmentum, following the curve of the cephalic flexure (Fig. 2b–d), and they enter next the peduncular domain of the basal hypothalamus at E14.5 (Fig. 3e–h, p–x). At E16.5, some labelled fibres of this tract course longitudinally through the whole basal hypothalamus, reaching the supraoptic decussation, whereas others turn sharply into the medial forebrain bundle as it ascends through the alar peduncular hypothalamus (Figs. 4b–i, 5d–y, 6a–i). These accordingly incorporate into the telencephalic peduncle and penetrate the telencephalon. As the tract passes through the subparaventricular area of the peduncular hypothalamus, some fibres diverge radially, approaching a specific unidentified area of the hypothalamic periventricular stratum possibly coinciding with the preincertal area; collaterals are given also ventrally to the mamillary/retromamillary periventricular area (not shown). Detailed examination at high magnification indicated that most labelled fibres reaching the telencephalon through this pathway disperse and end within the subpallium (mainly substantia innominata and pallidum). We found only one or two fibres that crossed the striatum and passed beyond the palliosubpallial boundary into the pallium (not shown). Some terminal fibres alternatively extend medially into the anterior commissure and the septum (not shown). At E18.5, the medial tegmental tract displays collaterals that grow into the posterior commissure and nearby areas of the pretectum, and it may contribute as well to the superior colliculus and the tectal commissure (Figs. 7j–t, 8g–i). The origin of the labelled fibres of the medial tegmental tract is uncertain, but the best candidate seems to be the r4 reticular formation (see “Discussion”).
Superficial ascending tract
At E16.5, there appears subpially a small labelled ascending axon bundle lying superficial to the medial tegmental bundle, which might arise from a superficial group of dispersed labelled neurons found just rostral to the pons; this tract could be followed into the midbrain paramedian tegmentum, where it either ends or is incorporated by the medial tegmental tract (Figs. 5l–t, 6k, l).
We observed at E16.5 a small packet of labelled fibres that occupy a position intermediate between the lateral lemniscus and the medial tegmental tract in the upper brainstem (Fig. 5g–k). These fibres can be followed at higher magnification into the posterior thalamic nucleus (Po) and the medial part of the ventrobasal complex (VPM) (Figs. 7i–k, 8d–f, 11h–m, 13d–f). We therefore assume that they represent labelled r4 components of one of the trigeminothalamic projection tracts, possibly the ventral trigeminothalamic tract, given the presumed origin of such projections in the oral subnucleus of the trigeminal descending column (De Chazeron et al. 2004; Guy et al. 2005; Veinante et al. 2000; Fig. 14a–e). These fibres probably cross the r4 floor plate (in contrast, the decussation of trigeminothalamic fibres originating in the principal sensory nucleus—the trigeminal lemniscus—apparently occurs across the r2 floor plate, just rostral to the pons, where we do not observe labelled decussated fibres).
Cerebellopetal trigeminal and vestibular tracts
We found two separate sets of labelled fibres that enter the cerebellum: one composed of thick, coarse fibres originated in the r4 part of the oral sensory trigeminal subnucleus and the other formed by thin fibres related to the vestibular system. The thicker fibres were observed already at E12.5 (Fig. 2g–i). At E14.5 and E16.5, they occupy an intermediate radial position in the r1 alar plate and seem to stream out of the trigeminal sensory column (Figs. 3u, 4a–c, 5k–w); they approach in an arch the superior cerebellar peduncle at the roof of the cerebellar plate, behind the isthmus, passing through a nuclear aggregate (possibly a cerebellar nucleus, or the parabrachial complex) and penetrating finally the incipient cerebellar commissure (Figs. 3b–h, j, p, 4a–i, 5f–j); this cerebellopetal trajectory recalls that of the indirect spino-cerebellar tract (which carries spinal collateral motor copy signals to the cerebellum), but has a trigeminal source in r4, thus possibly mediating analogous cerebellar input from the oral region.
The thinner fibres were only observed after E16.5, but then they already cross the midline of the cerebellar nodule (Fig. 5f), so their invasion of the cerebellar primordium via the inferior cerebellar peduncle possibly occurs at E15. These fibres can be traced back along a superficial pathway into the vestibular complex and may thus be vestibulocerebellar (or reticulo-cerebellar) in nature (Fig. 5f–w).
Medial reticulospinal and vestibulospinal tracts
Fibres descending from r4 into the spinal cord already exit from the medial r4 tegmentum at E11.5 (Figs. 1i–k, 2a–e). At E14.5, they approach the upper limit of the spinal cord (Fig. 3h, i) and at E16.5 they build a massive bundle that descends medially through the medulla ventral to the medial longitudinal tract (Figs. 4c–i, 5q–y, 7f–t, 9e–h) and then adopts a superficial position in the ventral column of the spinal cord (Fig. 5v–y). These descending fibres may originate in reticular and vestibular neurons from r4.
Lateral trigeminal oro-spinal tract
We observed this tract from E16.5 onwards. It consists of thin fibres that course in an intermediate (lateral basal) tegmental position, intercalated between the medial reticulospinal and vestibulospinal tracts and the lateral vestibulospinal tract (Figs. 5r–w, 9d–h). These fibres finally arch into the lateral column of the spinal cord (not shown). They may correspond to the descending projections of the oral trigeminal subnucleus to the cervical spinal cord (Devoize et al. 2010).
Lateral vestibulospinal tract
These fibres were distinguished at E16.5. They seem to originate from the labelled neurons of the lateral vestibular nucleus in r4 (and from those migrated into r3 and r5) (Di Bonito et al. 2015) and have a separate far lateral descending course into the lateral column of the spinal cord (Figs. 4c, d, 5l–u, 7h–l, 9a, b, d–h).
In this work, we used a rhombomere-specific Cre-recombinase mouse line crossed with a floxed YFP reporter line to genetically label rhombomere 4 and its derivatives. Going beyond the preliminary data recorded by Di Bonito et al. (2015) and Di Bonito et al. (2013a, b), we mapped in detail the anatomical fate of tangentially and radially migrated r4-derived neuronal populations from embryogenesis to adult, analysing also the r4-originated fibre tracts.
Quail-chick grafting experiments (Cambronero and Puelles 2000; Marin and Puelles 1995; Wingate and Lumsden 1996) and mouse rhombomere-specific transgenic fate-mapping data (Di Bonito et al. 2013a, 2015; Di Meglio et al. 2013; Farago et al. 2006; Oury et al. 2006; Pasqualetti et al. 2007; present work) have shown the persistence of both overt and cryptic rhombomere-derived territories in the brain wall until adult stages, irrespective of the fact that advancing maturation of the hindbrain eventually hinders their non-experimental visualization.
Such data have corroborated that the hindbrain is organized in longitudinal columns of sensory and motor nuclei, which are subdivided into discrete segmental or neuromeric units. The molecular boundaries of overt and cryptic rhombomeres correlate topographically with the transverse limits of nuclei, or of distinct columnar modular subdivisions, as has been visualized according to co-linear differential expression of Hox genes (Cambronero and Puelles 2000; Marin et al. 2008; Marin and Puelles 1995; Tomas-Roca et al. 2016). The classic columns thus have a plurisegmental origin; it is believed that the subtle molecular differences that distinguish segmental modules one from another (a result of anteroposterior patterning) causes the columns to be structurally and functionally heterogeneous lengthwise (cell properties, local circuitry, long-range afferents and projections). The original Hox gene code of each neuromeric columnar subdomain, jointly with other AP molecular determinants and differential DV expression of transcription factors, appears to determine the ulterior development of specific neuronal identities or cell arrangements inside the intracolumnar modules. This pattern underlies on the whole the observed heterogeneity of neuronal populations within the DV sensorimotor columns along the AP axis (Di Bonito et al. 2013a, b, 2015; Di Meglio et al. 2013; Marin et al. 2008; Oury et al. 2006; Philippidou and Dasen 2013; Puelles et al. 2013; Tomas-Roca et al. 2016).
In particular, we show here that r4 contributes from dorsal to ventral to a particular neuromeric module of the auditory, vestibular and trigeminal sensory columns, as well as of the intermediate and lateral reticular formation, and it produces likewise specific tangentially migrated cell populations.
R4-derived auditory system
The fact that r4 contributes both to the posteroventral cochlear nucleus and the overlying rostroventral part of the dorsal cochlear nucleus, while only few labelled cells appear in the anteroventral cochlear nucleus and rostralmost dorsal nucleus, or at the caudal end of the cochlear column, implies that the cochlear column is double, that is, it is truly composed of dorsal and ventral subcolumns, rather than representing unitary cochlear nuclei aligned in a deformed topologically longitudinal series, as has been hitherto assumed (Malmierca and Merchán 2004). Each subcolumn displays its own set of neuromeric modules. This conclusion seems to be of considerable interest for comparative considerations, since homologs of the mammalian dorsal and ventral cochlear nuclei in non-mammals were generally expected to be arranged singly along the rostrocaudal axis, and no satisfactory solution has emerged so far. A small rostral part of the lateral superior olive likewise shows r4-derived neurons, as described in the chick superior olive (Marin and Puelles 1995), though it is not clear that these formations are really homologous. Other r4-derived audition-related neurons migrate rostralwards out of r4, apparently guided by the incipient lateral lemniscus tract, eventually forming the interstitial ventral nucleus of the lateral lemniscus that stretches after its tangential migration across r4, r3 and r2. Its dorsoventral origin within r4 is suggested by the facts that VLL cells share a GABAergic/glycinergic phenotype (e.g. expression of GlyT2, GAD67, VIAAT) with neurons in the lateral, ventral and medial nuclei of the trapezoid body, and both VLL and the trapezoid complex derive from an En1Cre genetic lineage (Altieri et al. 2015, 2016). The latter is topographically restricted to the basal plate, which indicates that the VLL migration probably has a basal origin. Curiously, there are three conventionally described lemniscal nuclei, namely ventral, intermediate and dorsal ones (VLL, ILL, DLL), but all three seem to have different origins and are molecularly diverse, and their lineal order along the lateral lemniscus is actually caudo-rostral, rather than ventro-dorsal (there is a local bending of the axial dimension at the lower limb of the cephalic flexure).
We previously reported that VLL coming from r4 reaches r1 levels (Di Bonito et al. 2013a), but we now hold that the rostral alar r1 locus within the lateral lemniscus is occupied instead by the Pax7-positive dorsal lemniscal nucleus (DLL), thought to originate from alar r1 (Pax7 is not expressed in the hindbrain basal plate—Allen Developing Mouse Brain Atlas; but is expressed ubiquitously in the ventricular zone of the hindbrain alar plate; nevertheless, Pax7-positive neurons entering the mantle layer are generated exclusively in r1; see Ju et al. 2004; Lorente-Canovas et al. 2012; Moreno-Bravo et al. 2014). Both DLL and VLL are GABAergic/glycinergic populations, but alar-derived DLL is unlabelled in our material, while basal-derived VLL is YFP positive. The small intermediate lemniscal nucleus or ILL also remains unlabelled in our material; however, it is Pax7 negative (i.e. is not r1 derived) and expresses glutamatergic markers. Accordingly, it represents a separate lemniscal population whose so far unknown AP origin must be different from those of DLL and VLL, possibly somewhere in r2 or r3; as regards its DV origin, it derives from an Atoh1/Wnt1-positive alar plate domain (Machold and Fishell 2005; Rose et al. 2009).
The lateral lemniscus fibres reach the deep stratum (central nucleus) of the inferior colliculus and also target the superior colliculus: some fibres decussate through the intercollicular and tectal commissures. It was a surprise to find that many lateral lemniscus fibres extend through the brachium of the inferior colliculus into the primordium of the medial geniculate nucleus in the thalamus and even farther ahead. Remarkably, it turns out that both the lateral lemniscal connection with the superior colliculus and the thalamic connection were illustrated by Ramón y Cajal on the basis of Golgi observations (Ramon Cajal 1911; Fig. 17b). Our results indicate that this ascending branch of the lateral lemniscus extends beyond the medial geniculate body, adopting thereafter a subpial position ventral to the optic tract, and reaches the supraoptic decussation, presumably targeting thereafter the contralateral medial geniculate nucleus.
In the superior olivary complex, r4 forms a small part of LSO and contributes to the dorsal (DPO) and rostral (RPO) periolivary regions, as well as to the lateral and medial olivocochlear efferent neurons (LOC, MOC). The cholinergic LOC and MOC populations originate jointly with facial branchiomotor motoneurons and vestibular efferent neurons in the basal plate of r4. The axons of the vestibular efferent neurons enter the vestibulocochlear nerve together with the olivocochlear fibres.
R4-derived vestibular system
The r4 module of the vestibular column produces large stellate neurons that in part remain in the lateral vestibular nucleus within r4 and in part invade r3 rostrally and r5 caudally (neighbouring neuromeric portions of the superior and inferior vestibular nuclei, respectively). These large vestibular cells represent an anatomically heterogeneous group of descending vestibular projection neurons. The r4-derived vestibular neurons largely project ipsilaterally to the spinal cord forming the lateral vestibulospinal tract; other neurons project contralaterally via the medial vestibulospinal tract (Di Bonito et al. 2015; Diaz et al. 2003, 1998). Thinner fibres originating likewise in the r4 module of the vestibular complex enter the cerebellum via the vestibulocerebellar tract, which crosses the midline of the cerebellar nodule.
R4-derived trigeminal system
In the trigeminal column, r4 forms the rostralmost segmental module of the oral subnucleus of the descending trigeminal column; this unit contributes to the lateral trigeminothalamic tract projecting to the posterior and ventral posteromedial thalamic nuclei (De Chazeron et al. 2004; Guy et al. 2005; Veinante et al. 2000). In contrast, the principal sensory nucleus maps to r2 and r3 (Oury et al. 2006) and projects to the thalamus via the ventral (crossed) and dorsal (ipsilateral) trigeminothalamic tracts. Thus, Hox collinearity differentiating the rhombomeric origins of Pr5 in r2 (Hoxa2 low )-r3 (Hoxa2 high ) and Sp5O in r4 (Hoxb1) is consistent with the differential hodologic properties of their distinct trigeminothalamic pathways (Di Bonito et al. 2013b; Pouchelon et al. 2012). In addition, thick labelled fibres stream out of the r4 module of the trigeminal sensory column as part of the trigeminocerebellar tract, which decussates within the intracerebellar commissure. Watson and Switzer (1978) and Fu et al. (2011) described rhombic lip-derived precerebellar neurons in the Pr5, but deduced an origin caudal to r4, which excludes a correlation with our data; however, it may be speculated that the coarse labelled trigeminocerebellar projections we see arising from the oral trigeminal subnucleus may come from similar precerebellar rhombic lip cells originated within r4 that migrate into Sp5O. The lateral trigeminal oro-spinal tract described by Devoize et al. (2010) also contains fibres originated in r4.
R4-derived reticular system
The hindbrain reticular formation generally consists of medial, intermediate and lateral domains (Jones 1995); the medial large-cell component occupies the basal plate, whereas the intermediate medium-size component occupies the ventral rim of the alar plate where tangentially migrated preganglionic parasympathetic, branchiomotor and noradrenergic neurons tend to aggregate (Blessing 1997; see his Figs. 8.3, 8.6; see also Ju et al. 2004). The lateral parvocellular component lies deep to the sensory columns, farther dorsally in the alar plate. The small lateral reticular neurons serve as interneurons for reflex sensorimotor circuitry. The r4 module contributes as expected to the whole local reticular formation, roughly at the level of the caudal pontine reticular nucleus and rostral to the nucleus raphe magnus (ascribed to r5–r6; Alonso et al. 2013); this sector seems to contribute likewise to the shell of the posterodorsal tegmental area and the nucleus subcoeruleus; these entities lie rostrally in r2 (caudal to locus coeruleus in r1; see shell of PDTg in Fig. 14d). Reticular fibres exit caudalwards from the r4 tegmentum forming the medial reticulospinal descending tract, which adopts a superficial position in the ventral spinal cord column. Some r4-derived reticular neurons projecting to the spinal cord translocate into r5 (Di Bonito et al. 2015).
A remarkable finding of the present study was the thick longitudinal tract that ascends through the medial (basal) tegmentum of the upper hindbrain, midbrain and diencephalon. This tract reaches the supraoptic decussation and, separately, the telencephalon via the medial forebrain bundle, coursing ventrodorsally through the peduncular lateral hypothalamus (hypothalamic prosomere 1 of Puelles 2013; Puelles et al. 2012, 2013; Puelles and Rubenstein 2015). As this tract transits through the midbrain and caudal diencephalon, it emits branches into the superior colliculus and tectal commissure, as well as into the posterior commissure and nearby areas of the pretectum, but was not seen to penetrate the thalamus. The r4-derived fibres that ascend through the medial forebrain bundle seem to end mainly at the basal telencephalon (subpallium).
The observed tract probably represents just the r4 segmental (caudal pontine) component of a larger entity to which other rhombomeres may also contribute (Jones 1995). Its remarkable volume and sparse en route terminations is somewhat surprising. The precise neuronal origin of these ascending r4 fibres is not distinguishable in our material. The most distinct possibility is that the ascending r4 medial tegmental tract originates from the local medial large-celled component of the reticular formation, known to send fibres through the diencephalic tegmentum into the medial forebrain bundle, connecting there with hypothalamo-cortical neurons (Jones and Yang 1985; Saper 1985; Vertes and Martin 1988; Vertes et al. 1986). As a potential alerting system, this pathway is unusual in seemingly bypassing the intralaminar thalamus (though more detailed observations might detect such connections).
Alternatively, or in addition, the observed terminal projections into superior colliculus, pretectum and mamillary/retromamillary areas might be consistent with an r4-derived vestibular input to circuitry generating head-direction coding properties in neurons at these target sites, extending as well into the basal telencephalon. It is generally assumed that this functional specialization requires visual, head proprioception and vestibular input, which might come in part from r4. This conjecture unfortunately lacks presently a strong basis, since relevant connections into the head-direction system have been associated so far to the mutually connected lateral mamillary nucleus and dorsal tegmental nucleus (in r1) and to the afferents of the latter from the nucleus prepositus hypoglossi and the medial vestibular nucleus, which convey horizontal canal input (e.g. Bassett and Taube 2005; Hopkins 2005). However, the nucleus prepositus hypoglossi, jointly with the medial vestibular nucleus, lies intercalated between the abducens nucleus (r5) and the rostral end of the hypoglossal nucleus (r8), and therefore would not be expected to be labelled in our material. It remains nevertheless possible that other than horizontal head velocity vestibular input—e.g. head pitch, or eye position—is conveyed to the head-direction system by other parts of the vestibular column, part of which may lie in r4.
R4-derived functional subcircuits in sensorimotor systems
In the auditory system, the r4-derived part of the CN and VLL contribute to the main sound transmission pathway (Di Bonito et al. 2013a). The parallel pathway for sound localization in space instead involves the r2/r3-derived parts of the cochlear nucleus and the superior olivary complex (SOC), which is mostly derived from r5 (Di Bonito et al. 2013a; Farago et al. 2006; Maricich et al. 2009; Marin and Puelles 1995). Moreover, centrifugal acoustic cell populations originated in the r4 basal plate (FBM and MOC) represent two distinct auditory sensorimotor feedback subcircuits (MEM and MOC reflexes) essential to protect the cochlea from acoustic overstimulation. Innervation of auditory receptor cells by MOC is also probably important for their survival, as well as for their function in the cochlear amplification process (Di Bonito et al. 2013a).
In the vestibular system, r4 derivatives contribute to the lateral and medial vestibulospinal pathways that regulate trunk and limb musculature that counteracts perturbations of body position (Di Bonito et al. 2015), as well as to the vestibular efferent neurons. Apart from postural, vestibulo-ocular and optokinetic reflexes, segmental components of the vestibular column might also be involved in the generation of head-direction properties.
Facial somatosensory inputs are variously transmitted through the heterogeneous trigeminal column to the brainstem and thalamus (Pouchelon et al. 2012). Along the AP axis, four distinct trigeminal pathways have been reported to convey input from the face to somatosensory cortex, associated with different parts of the principal (Pr5) and spinal (Sp5) nuclei. Crossed and ipsilateral trigeminal lemniscal pathways are associated to r2- and r3-derived parts of Pr5 conveying proprioceptive and tactile facial stimuli; other crossed trigeminothalamic pathways arise from the Sp5O and the Sp5I subnuclei (Oury et al. 2006; Pouchelon et al. 2012). We found that some Sp5O neurons projecting to the VPM and Po derive from r4, but we ignore whether additional such elements arise likewise more caudally in r5 (Di Bonito et al. 2013b). The Sp5O is involved in the processing of somatosensory input and nociceptive stimuli from the orofacial region (Dallel et al. 1999, 1990; Raboisson et al. 1995) and contains interneurons taking part in various reflex activities (Abrahams and Richmond 1977; Olsson et al. 1986). The Sp5O has also been suggested to be involved in the coordination of head–neck movements and related modulation of incoming sensory information from the cervical spinal cord (Devoize et al. 2010). Each trigeminal subnucleus apparently has distinct cytoarchitecture, connectivity and function, a feature that seems correlated to their differential rhombomeric origins and the underlying molecular identities.
In conclusion, different rhombomeres contribute via multiple distinct modular units to different subcircuits with specific functions within distinct sensorimotor systems. Indeed, r4-derived nuclei and fibre tracts contribute to rhombomere-specific functional subcircuits in the auditory, trigeminal and vestibular sensorimotor systems. The r4-derived descending reticular tracts attend postural reflexes, while the corresponding ascending reticular circuitry needs to be better characterized, since it seems relevant at least for the ascending reticulo-hypothalamic alerting system, but possibly also for the head-direction system, jointly with other vestibular and reticular modular elements.
We thank S. Srinivas for the ROSA26-YFP reporter line. This work was supported by the Spanish MICINN grant BFU2014-56517P (with European FEDER support) and Séneca Foundation 19904/GERM/15 contract to LP, and by the French “Agence National de la Recherche” (ANR) “2009 Chaire d’Excellence” #R09125AA and “DEAF” #ANR-15-CE15-0016-01 Programs to MS.
- Abrahams VC, Richmond FJ (1977) Motor role of the spinal projections of the trigeminal system. In: Anderson DJ, Matthews B (eds) Pain in the Trigeminal Region. Elsevier, Amsterdam, pp 405–411Google Scholar
- Alonso A, Merchán P, Sandoval JE, Sánchez-Arrones L, Garcia-Cazorla A, Artuch R, Ferrán JL, Martínez-de-la-Torre M, Puelles L (2013) Development of the serotonergic cells in murine raphe nuclei and their relations with rhombomeric domains. Brain Struct Funct 218:1229–1277 doi: 10.1007/s00429-012-0456-8 CrossRefPubMedGoogle Scholar
- Auclair F, Valdes N, Marchand R (1996) Rhombomere-specific origin of branchial and visceral motoneurons of the facial nerve in the rat embryo. J Comp Neurol 369:451–461. doi: 10.1002/(SICI)1096-9861(19960603)369:3<451::AID-CNE9>3.0.CO;2-4 CrossRefPubMedGoogle Scholar
- Bassett JP, Taube JS (2005) Head direction signal generation: ascending and descending information streams. In: Wiener SI, Taube JS (eds) Head Direction Cells and the Neural Mechanisms of Spatial Orientation. MIT Press, Cambridge, pp 83–109Google Scholar
- Blessing WW (1997) The lower brainstem and bodily homeostasis. Oxford Univ. Press, New YorkGoogle Scholar
- Devoize L, Domejean S, Melin C, Raboisson P, Artola A, Dallel R (2010) Organization of projections from the spinal trigeminal subnucleus oralis to the spinal cord in the rat: a neuroanatomical substrate for reciprocal orofacial-cervical interactions. Brain Res 1343:75–82. doi: 10.1016/j.brainres.2010.04.076 CrossRefPubMedGoogle Scholar
- Di Meglio T, Kratochwil CF, Vilain N, Loche A, Vitobello A, Yonehara K, Hrycaj SM, Roska B, Peters AH, Eichmann A, Wellik D, Ducret S, Rijli FM (2013) Ezh2 orchestrates topographic migration and connectivity of mouse precerebellar neurons. Science 339:204–207. doi: 10.1126/science.1229326 CrossRefPubMedPubMedCentralGoogle Scholar
- Fujiyama T, Yamada M, Terao M, Terashima T, Hioki H, Inoue YU, Inoue T, Masuyama N, Obata K, Yanagawa Y, Kawaguchi Y, Nabeshima Y, Hoshino M (2009) Inhibitory and excitatory subtypes of cochlear nucleus neurons are defined by distinct bHLH transcription factors, Ptf1a and Atoh1. Development 136:2049–2058 doi: 10.1242/dev.033480 CrossRefPubMedGoogle Scholar
- Gray PA (2013) Transcription factors define the neuroanatomical organization of the medullary reticular formation. Frontiers in Neuroanatomy 7:7 doi: 10.3389/fnana.2013.00007
- Hopkins DA (2005) Neuroanatomy of head direction cell circuits. In: S.I. W J.S. T (ed) Head Direction Cells and the Neural Mechanisms of Spatial Orientation. MIT Press, Cambridge, pp 83–109Google Scholar
- Jones BE (1995) Reticular formation: cytoarchitecture, transmitters, and projections. In: Paxinos G (ed) The rat nervous system. 2nd edn. Academic Press, San Diego, pp 155–171Google Scholar
- Lorente de Nó R (1981) The primary acoustic nuclei. Raven Press, New YorkGoogle Scholar
- Maricich SM, Xia A, Mathes EL, Wang VY, Oghalai JS, Fritzsch B, Zoghbi HY (2009) Atoh1-lineal neurons are required for hearing and for the survival of neurons in the spiral ganglion and brainstem accessory auditory nuclei. J Neurosci 29:11123–11133. doi: 10.1523/JNEUROSCI.2232-09.2009 CrossRefPubMedPubMedCentralGoogle Scholar
- Martinez-de-la-Torre M, Lambertos A, Peñafiel R, Puelles L (2017) An exercise in brain genoarchitectonics: analysis of Azin2-LacZ expressing neuronal populations in the mouse hindbrain. J Neurosci Res (in press)Google Scholar
- Miguez A, Ducret S, Di Meglio T, Parras C, Hmidan H, Haton C, Sekizar S, Mannioui A, Vidal M, Kerever A, Nyabi O, Haigh J, Zalc B, Rijli FM, Thomas JL (2012) Opposing roles for Hoxa2 and Hoxb2 in hindbrain oligodendrocyte patterning. J Neurosci 32:17172–17185. doi: 10.1523/JNEUROSCI.0885-12.2012 CrossRefPubMedGoogle Scholar
- Parrish M, Nolte C, Krumlauf H (2009) Hox gene expression. In: Lemke G (ed) Developmental neurobiology. Academic Press, New York, pp 61–71Google Scholar
- Pasqualetti M, Diaz C, Renaud JS, Rijli FM, Glover JC (2007) Fate-mapping the mammalian hindbrain: segmental origins of vestibular projection neurons assessed using rhombomere-specific Hoxa2 enhancer elements in the mouse embryo. J Neurosci 27:9670–9681. doi: 10.1523/JNEUROSCI.2189-07.2007 CrossRefPubMedGoogle Scholar
- Paxinos G, Franklin KBJ (2001) The mouse brain in stereotaxic coordinates. 2nd edn. Academic press, San DiegoGoogle Scholar
- Puelles L (2013) Plan of the developing vertebrate nervous system relating embryology to the adult nervous system (prosomere model, overview of brain organization). In: Rubenstein JLR, Rakic P (eds) Comprehensive developmental neuroscience: patterning and cell type specification in the developing CNS and PNS. Academic Press, Amsterdam, pp 187–209CrossRefGoogle Scholar
- Puelles L, Rubenstein JL (2015) A new scenario of hypothalamic organization: rationale of new hypotheses introduced in the updated prosomeric model. In: Alvarez-Bolado G, Grinevich V, Puelles L (eds) Development of the Hypothalamus. Front. Neuroanat, Lausanne, pp 9–27. doi: 10.3389/fnana.2015.00027.
- Ramon y Cajal S (1911) Texture du Systeme Nerveux de l’Homme et des Vertebrás. vol 2. Paris: Maloine, re-edit. 1954 Madrid: CSICGoogle Scholar
- Storm R, Cholewa-Waclaw J, Reuter K, Bröhl D, Sieber M, Treier M, Müller T, Birchmeier C (2009) The bHLH transcription factor Olig3 marks the dorsal neuroepithelium of the hindbrain and is essential for the development of brainstem nuclei. Development 136:295–305 doi: 10.1242/dev.027193 CrossRefPubMedGoogle Scholar
- Vaage S (1969) The segmentation of the primitive neural tube in chick embryos (Gallus domesticus). A morphological, histochemical and autoradiographical investigation. Ergebnisse der Anatomie und Entwicklungsgeschichte 41:3–87Google Scholar
- Webb BD, Shaaban S, Gaspar H, Cunha LF, Schubert CR, Hao K, Robson CD, Chan WM, Andrews C, MacKinnon S, Oystreck DT, Hunter DG, Iacovelli AJ, Ye X, Camminady A, Engle EC, Jabs EW (2012) HOXB1 founder mutation in humans recapitulates the phenotype of Hoxb1-/- mice. Am J Hum Genet 91:171–179. doi: 10.1016/j.ajhg.2012.05.018 CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.