During prenatal brain development, a wide variety of specialized neurons is generated from multipotent progenitors cells. These processes are finely regulated in a spatiotemporal-specific manner by the generation of signalling molecules, which in turn are released from specific organizing centers (Osumi et al., 1997; Lee et al., 2000; Schuurmans and Guillemot, 2002). Major inducers characterized so far include highly conserved signal molecules such as bone morphogenetic proteins (BMPs), fibroblast growth factor (FGFs), Hedgehogs, and Wnts (Hogan, 1999). During brain development, gradients of these morphogens act on early neuronal progenitor cells by eliciting the activation of specific transcription factors, which in turn allow precursor cells to acquire their differentiative traits (Ben-Arie et al., 1997; Gurdon and Bourillot, 2001).
Among the transcription factors that mediate fate specification in neuronal precursor cells, the LIM-homodomain (LIM-hd) transcription factors (for review, see Hobert and Westphal, 2000) act in a combinatorial manner, which leads the neuronal cell to acquire unique properties (for review, see Shirasaky and Pfaff, 2002). The role played by specific combinations of LIM-hd transcription factors in defining a given neuron subtype identity has so far been elucidated only in a few systems, including the target specificity of individual motor neurons (Jessell, 2000) and the developmental regulation of thalamus (Nakagawa and O'Leary, 2001) and amygdala (Remedios et al., 2004). However, other targets of LIM codes still remain to be characterized. In this context, the pituitary development is of particular interest, owing to its dependence on the LIM gene Lhx3 (Sheng et al., 1996, 1997). Studying this system, it was possible to identify and clone a gene, Thg1-pit, which acts downstream from Lhx3 during early steps of pituitary organogenesis and is transcriptionally activated in concomitance with the onset of pituitary cell differentiation (Fiorenza et al., 2001). In addition to the pituitary, Thg-1pit expression was also detected in several areas of developing brain, suggesting that the product of this gene is relevant to fate shaping in a variety of neuronal cells (Fiorenza et al., 2001).
In this study, we have investigated the expression of Thg-1pit in mouse brain during development and in the adult. We show here that, at early stages of embryonal brain development, this gene is transcriptionally activated in the areas of granule cell precursor specification. Thereafter, Thg-1pit expression landmarks both the differentiative steps and the mature function of granule/interneuron cells in several brain districts, including cerebellum, basal forebrain, olfactory bulb, and hippocampus. In the adult, this gene becomes activated also in mitral cells of olfactory bulb and in Purkinje cells of cerebellum. Such activation occurs in concomitance with full development of the synaptic contacts Purkinje and mitral cells establish with granule cells.
Thg-1pit Is Expressed in Specific Brain Areas in E12.5 Embryos
Previous reverse transcriptase-polymerase chain reaction (RT-PCR) analyses had shown that Thg1-pit is first activated at the embryonic day (E) 9.5 stage of mouse embryo development and that this gene is expressed at E12.5 in several areas of the developing nervous system (Fiorenza et al., 2001). Therefore, we first identified the regions where this gene is specifically expressed at that stage. The E12.5 stage is of particular interest because, at this time, important differentiative events take place at the levels of ventricles, including massive neurons differentiation, expansion of differentiating zone, formation of subventricular zone, and induction of neuron-specific genes (Bayer and Altman, 1995). Whereas at the beginning of Thg1-pit activation, transcripts of this gene are spread all over the entire neural tube (data not shown), whole-mount in situ hybridization of E12.5 embryos (Fig. 1A,B) showed specific presence of Thg1-pit transcripts at the level of all ventricles. The strongest signals were observed at the fourth ventricle, with particular reference to the rhombic lip (Fig. 1C). This term indicates the free hindbrain margin that surrounds the dorsal opening of the fourth ventricle at this developmental stage (Miale and Sidman, 1961; Fujita et al., 1966; Altman et al., 1969). As for the lateral ventricle, hybridization of histological sections showed the presence of Thg1-pit transcripts restricted to subventricular neuroepitelium/mantle zone and lateral ganglionic eminence (Fig. 1D).
Thg-1pit Expression Marks Interneuron Cell Migration/Differentiation During Postnatal Brain Development
In the light of the finding that, in E12.5 embryos, Thg-1pit is expressed in subventricular germinal zones, i.e., where progenitor cells of different neuron types are generated, we next determined whether this gene is maintained active throughout various steps of precursor cell differentiation. As for rostral rhombic lip, it gives rise to cerebellar granule cells by an ordered sequence of proliferation, migration, and synaptogenesis steps occurring during the first 3 weeks after birth. In details, granule cells precursors migrate out over the rhombic lip and establish a secondary proliferative zone in the outermost layer of the cortical region of developing cerebellum, the so-called external granular layer (EGL). Postmitotic granule cells then leave EGL and migrate inward, forming the internal granular layer (IGL) where granule neurons differentiate terminally (Rakic, 1971; Hatten et al., 1997).
To determine whether Thg1-pit is expressed during those steps, we analyzed sections of postnatal day (PN) 6 brains. At the level of cerebellum, sharp hybridization signals were detected in the entire EGL, including both the most superficial layer (containing rapidly proliferating granule cell precursors) and the deeper one (containing postmitotic, differentiating granule cells; Fig. 2A,B).
In addition to the cerebellum, Thg-1pit expression was also prominent in subventricular zones of anterior lateral ventricles (Fig. 2C). Departing from these areas, strong hybridization signals were also prominent in the entire rostral migratory stream (Fig. 2C) and the olfactory bulb (Fig. 2D), where positive cells included differentiated granule cells of the granular layer and the periglomerular cells (Fig. 2E). In addition to the rostral migratory stream, Thg-1pit expression was also apparent in the so-called ventral migratory mass (VMM; Fig. 2C). VMM has been described recently as a cellular stream that, departing from lateral ventricles toward the basal forebrain, represents the source of Island of Calleja and basal forebrain interneurons (De Marchis et al., 2004). Interestingly, Thg-1pit was also expressed in these interneurons (Fig. 2F). Other brain areas displaying sharp Thg-1pit positivity were the dentate gyrus of hippocampus (Fig. 2G) and scattered cells located in pons/medulla areas of hindbrain (data not shown).
Thg-1pit Is Also Expressed in the Adult Brain
The in situ hybridization analyses performed at PN6 showed that presence of Thg-1pit transcripts is also a property of postmitotic granule cells after they reach their final location. Therefore, it was of interest to analyze in detail whether Thg-1pit was expressed in differentiated cells of adult brain. This issue was first addressed by semiquantitative RT-PCR amplification of messages extracted from different brain areas (Fig. 3). Thg-1pit expression was detected through the entire brain, the highest level being found in cerebellum, accordingly to results obtained at PN6. Other Thg-1pit positive areas were posterior cortex, striatum, pons/medulla oblongata, and olfactory bulb.
As large brain regions such as striatum, pons/medulla, and mesencephalon include several different structures, Thg-1pit expression was more finely analyzed by in situ hybridization on adult brain sections. As for cerebellum, the strong positivity observed in previous RT-PCR assays was fully justified by marked labeling of granule cells. These cells per se approximate the 90% of total cerebellar cells (Fig. 4A,B). Of interest, Thg-1pit transcripts were also consistently found in most Purkinje cells (Fig. 4C). It is interesting to note that Thg-1pit expression was undetectable in these cells at PN6 (Fig. 4D), whereas it was barely detected in a few of these cells at an intermediate developmental time (PN21; data not shown). In the light of the very asynchronous maturation of Purkinje cells (Armengol and Sotelo, 1991), positive cells at this stage are likely to correspond to the early maturing ones.
The olfactory bulb was another area of prominent Thg-1pit expression, with particular reference to both granule cells and periglomerular cells (Fig. 5A,B). Thg-1pit expression was also apparent in mitral cells (Fig. 5C), a cell type that, at PN6, was also positive (Fig. 5D), likely because of its maturation timing earlier than that of Purkinje cells (Switzer et al., 1985; Sotelo, 2004). In addition to the olfactory bulb, the entire rostral migratory stream was also labeled, including the subventricular zone of anterior lateral ventricle (Fig. 5E), whereas no ventral migratory mass was apparent, in agreement with disappearence of this structure in the mature brain (De Marchis et al., 2004). However, cells derived from ventral migratory mass in basal forebrain were positively labeled (Fig. 5F). Thg-1pit expression was also apparent in granular layer of dentate gyrus (Fig. 5G) and in scattered cells of pons–medulla (data not shown).
In this study, we have investigated the expression of the Thg-1pit gene during brain development in the mouse. Thg-1pit belongs to TSC-22/DIP/bun genes, a family coding for transcriptional regulators expressed in a variety of systems (Shibanuma et al., 1992; Treisman et al., 1995; Dobens et al., 2000). Due to the lack of a canonical DNA-binding domain and the presence of a leucine zipper domain, proteins of this family are believed to regulate gene expression by homodimerization and/or heterodimerization with other leucine zipper-containing factors (Kester et al., 1999). This mechanism allows the formation of a variety of homo-/heteromultimers, having unique and specific regulatory properties that differ from those of individual multimer components. To this respect, a transcriptional repressor activity was described both for the TSC-22 homodimer and the TSC-22/THG-1 heterodimer (Kester et al., 1999). TSC-22 was also characterized functionally and shown to induce growth inhibition and differentiation in a variety of cell types (Gupta et al., 2003; Kawamata et al., 2004). In addition, TSC-22 expression was studied on mouse development and found to be scattered in the neural tube at E8.5 and then, since E10.5, to become restricted to the ventricles' mantle layer and to discrete sites of the hindbrain, suggesting a role(s) in specification/differentiation of neural progenitors cells (Kester et al., 2000).
In the present study, a similar restriction to specific areas during prenatal/postnatal brain development has also been found for the expression of the Thg-1pit gene. During development, in fact, this gene is expressed in several brain regions anatomically and functionally different as cerebellum, hippocampus, olfactory bulb, and basal forebrain. At the E12.5 stage, Thg-1pit expression becomes restricted and prominent in brain areas that represent the primary sources of granule cells precursors, including the anterior-most region of the rhombic lip and lateral ganglionic eminences. These regions give rise to several neural cell types, including the granule cell progenitors of cerebellum and olfactory bulb, respectively (Wingate and Hatten, 1999; Lin et al., 2001; Wichterle et al., 2001). In addition, Thg-1pit expression is maintained during postnatal (PN6) granule cells migration/differentiation, as it was particularly evident at the level of cerebellum in the proliferating granule cells of EGL and the postmitotic granules populating the IGL. In addition to the cerebellum, a similar developmental pattern was also observed in the specification/differentiation of granule cells of the olfactory bulb, dentate gyrus of hippocampus, and basal forebrain. In these cases, in fact, differentiating granule cells displayed a continuous expression of the Thg-1pit gene during proliferation/migration and terminal differentiation, even though with different developmental schedules, depending on the specific cell lineage considered. As for the olfactory bulb and dentate gyrus, granule cells are continuously produced both in the neonatal and the adult mouse, owing to the long-term persistence of a germinal layer in the subventricular zone that surrounds the ependymal layer of dorsolateral wall of the lateral ventricle (Luskin, 1993; Lois and Alvarez-Buylla, 1994). Cells arisen from this area migrate tangentially to the olfactory bulb through the rostral migratory stream (Luskin, 1993; Lois and Alvarez-Buylla, 1994) and then, once arrived in the olfactory bulb core, they disperse radially over both the granule and glomerular layers, by eventually differentiating and integrating into olfactory circuits (for review, see Carleton et al., 2003).
Our study has also shown that, in addition to rostral migratory stream, Thg-1pit expression is also a feature of granule cell/interneuron progenitors that migrate ventrally through the VMM (De Marchis et al., 2004). VMM represents an additional route of migration that a subpopulation of subventricular zone–derived cells follows in the neonatal, but not adult, mouse, by migrating ventrally across the nucleus accumbens (De Marchis et al., 2004). These VMM-derived cells eventually give rise to populations of subcortical forebrain interneurons that form the Islands of Calleja and the olfactory tubercle (Meyer et al., 1989). We, therefore, conclude that the expression of Thg-1pit is a constant marker of, and likely relevant to, the specification, proliferation/migration and acquisition of fully differentiative traits of mouse granule/interneuron cells, irrespectively of brain areas and developmental times where/when these processes take place. A similar conclusion was drawn previously for the zinc-finger transcription factors zipro 1 (formerly RU49), which specifically marks granule cell specification and differentiation, in olfactory bulb, dentate gyrus, and cerebellum (Yang et al., 1996, 1999). In addition, it is also interesting to note that Thg-1pit expression is not limited to the developmental periods indicated above, but it also represents a typical feature of fully differentiated granule/interneuron cells of the adult brain, suggesting that this factor is also very relevant to the specific functions of the mature cells.
As for the factors that putatively regulate Thg-1pit expression during granule cell development, major candidates are molecules of the TGF-β family, with particular reference to TGF-β1 and BMPs. In fact, at the level of rhombic lip, the specification of granule cell precursors depends on BMP signals released from hindbrain roof plate (Lee et al., 2000). Moreover, BMPs initiate the program of granule cell specification in neural plate-derived cells (Alder et al., 1999). In agreement with these findings, preliminary experiments performed in our laboratory have shown that a short-term treatment of in vitro cultured cerebellar granule cells elicits a significant and transient increase in Thg-1pit expression (data not shown).
Another interesting finding of this study was that Thg-1pit expression is also evident in the mature Purkinje cells of cerebellum and mitral cells of olfactory bulb, despite the Thg-1pit transcriptional inactivation these cells display during their postnatal development. The earlier developmental time at which Thg-1pit expression was detected in mitral cells (PN6) with respect to that of Purkinje cells (adult) is likely related to the different developmental times at which functional interactions between those cells and granule cells are fully stabilized. Purkinje cell maturation is a long-lasting process, which occurs postnatally and eventually results in the typical espalier arrangement of dendritic trees. An essential role in this process is played by either synaptic transmission or the release of neurotrophins by the parallel fibers of granule cells (Sotelo, 1978; Baptista et al., 1994; Morrison and Mason, 1998; Yuste and Bonhoeffer, 2004). Different features by which Thg-pit expression is elicited in developing granule cells and mature Purkinje/mitral cells indicate that Thg-pit activation is subject to distinct regulatory mechanisms during development. In light of the present findings, it is tempting to hypothesize that early Thg-pit activation occurring during granule/interneuron cell development is mostly controlled by cell–environment interaction (such as gradients of secreted signalling molecules), whereas Thg-pit expression in mature granule cells and Purkinje/mitral cells is mostly dependent on synaptic activity. Further study is required to address these questions in detail.
C57BL/6J mice were obtained from Charles River (Calco, Italy). Pregnant females were killed by cervical dislocation, and embryos were removed rapidly from the uterus, dissected free from their membranes, and processed as needed. Embryos were staged taking the midday of the day of vaginal plug as E0.5. Before total-body perfusion with fixative, animals were deeply anesthetized by an intraperitoneal injection of 150 mg/kg ketamine and 10 mg/kg xylazine. Experimental protocols and procedures were approved by the Ministry of Public Health, Rome, Italy. Animals were treated in accordance with La Sapienza University Guidelines for the Care and Use of Laboratory Animals.
RT-PCR Amplification of Thg-1pit mRNA
A total of 5 μg of DNAse-treated total RNA from discrete brain regions of adult mice (ChemiRNA TOTAL RNA panel mouse brain; Chemicon International, Temecula, CA) were first converted into cDNA using oligo-dT and MMLV reverse transcriptase (Superscript TM II, Life Technologies, GIBCO, Milano, Italy) and then amplified by PCR using Advantage cDNA Polymerase mix (Clontech, Becton Dickinson Italia, Milano, Italy), according to manufacturer's instructions. Oligonucleotide primer pairs used were as follows: Thg-1pit amplification fragment (302-bp), 5′-GAGAGCCTTATCGTCGAGGT-3′ (sense, nucleotides [nt] 1263–1282), 5′-TCTCCACCTTAGCCTTGCC-3′ (antisense, nt 1547–1565); S16 ribosomal protein amplification fragment (103-bp), 5′-AGGAGCGATTTGCTGGTGTGG-3′ (sense, nt 1451–1471), 5′-GCTACCAGGGCCTTTGAGATGGA-3′ (antisense, nt 1621–1641). PCR amplification conditions were as follows: 95°C for 1 min and 35 cycles of the following steps: denaturation at 95°C for 30 sec, annealing at 64°C for 30 sec, and elongation at 72°C for 50 sec in the presence of 32P-αdCTP as tracer. Under our RT-PCR amplification conditions, S16 and Thg-1pit cDNAs were amplified linearly for at least 40 PCR cycles. Amplification products were fractionated by 5% polyacrylamide gel electrophoresis and visualized by autoradiography. 32P incorporation into amplification bands was measured with an InstantImagerMT (Packard Instruments Co., Milano, Italy).
Preparation of Digoxigenin-Labeled Riboprobes for In Situ Hybridization
Digoxigenin (DIG)-labeled probes were synthesized by transcribing 1 μg of template DNA with T7 and SP6 RNA polymerase using DIG RNA labeling kit protocol (Boehringer Mannheim, Milano, Italy), according to manufacturer's instructions. Riboprobes corresponded to the 5′ untranslated region (nt 159–863; Fiorenza et al., 2001). Riboprobes including almost the entire coding sequence (nt 1263–1929) gave results comparable to those obtained with the shorter probes indicated above.
In Situ Hybridization
Whole-mount in situ hybridization was performed on E12.5 embryos according to published procedures (Belo et al., 1997) with minor modifications. Embryos were rapidly dissected in phosphate-buffered saline (PBS), fixed overnight with 4% paraformaldehyde (PFA) in PBS, rinsed in PBS containing 0.1% Tween-20 (PBSw), dehydrated, and eventually stored at −20°C. As needed, embryos were transferred to 0–4°C, rehydrated with PBSw, bleached with 6% H2O2 in PBSw for 10 min, transferred again to PBSw and eventually treated with 10 μg/ml proteinase K (Life Technologies, GIBCO) in PBSw for 13 min at room temperature. After rinsing with PBSw for 5 min, embryos were post-fixed with 4 % PFA and 0.2% glutaraldehyde for 20 min at 0–4°C, then transferred to hybridization mix (50% formamide, 50 μg/ml yeast RNA, 2% Boehringer Block, 5× standard saline citrate [SSC], 100 mg/ml heparin, 0.5% CHAPS, and 0.5 μM ethylenediaminetetraacetic acid [EDTA]) at 65°C for 3 hr, and eventually incubated overnight in the presence of 10 μg/ml DIG-labeled probe in hybridization mix at 70°C. At the end of incubation, embryos were sequentially rinsed at room temperature with the following solutions: hybridization mix, 2× SSC pH 4.5, 2× SSC pH 7 containing 0.1% CHAPS, PBS, and PBSw; and eventually transferred to blocking buffer (10% heat-inactivated goat serum, 1% Boehringer Block, 0,1% Tween-20) at 4°C. After a 2-hr incubation in this buffer, embryos were further incubated overnight in the presence of alkaline phosphatase-conjugated anti-DIG antibody (1/2,000, Roche Diagnostic, Milano, Italy) in blocking buffer at 4°C and then sequentially rinsed with 0.1% bovine serum albumin (BSA) in PBSw, PBSw, and AP1 buffer (0.1 M NaCl, 0.1 M Tris pH 9.5, 50 mM MgCl2). The staining was eventually developed using BM Purple AP Substrate Precipitating (Roche Diagnostic).
In situ hybridization on paraffin sections was performed as previously described (Sassoon and Rosenthal, 1993). Brains of PN6 male mice were rapidly dissected after cervical dislocation and then immediately fixed by immersion into 4% PFA. Adult brains were obtained from male mice deeply anesthetized and then perfused with 4% PFA. Dissected brains were post-fixed by immersion into 4% PFA overnight, then dehydrated, embedded in Paraplast Plus Tissue Embedding Medium (Società Italiana Chimici, Roma, Italy), and serially sectioned with a slice thickness of 9 μm. Sections were placed on silanized slides, dewaxed, rehydrated, digested with 20 μg/ml proteinase K in TE buffer (100 mM Tris, 50 mM EDTA, pH 8.0) at 37°C for 20 min, rinsed twice with PBS, and then subjected to deproteination with 0.2 M HCl for 10 min at room temperature. Slices were then postfixed with 4% PFA, rinsed again in PBS, dehydrated, and air-dried. Hybridization was carried out overnight at 70°C in the presence of 1 μg/ml DIG-labeled probe, 50% formamide, 10 mM Tris pH 7.5, 1 mM EDTA, 600 mM NaCl, 1 mg/ml yeast tRNA, 10% dextran sulfate, and 1× Denhardt's solution (Sigma Aldrich, Milano, Italy). Posthybridization rinses included a 20 min rinse in 2× SSPE (180 mM NaCl, 10 mM NaH2PO4 and 1 mM EDTA, [pH 7.4]), followed by a 30-min incubation in a solution made of 50% hybridization mix and 50% 2× SSPE and supplemented with 0.3% CHAPS. After extensive rinsing with 2× SSPE and PBSw, nonspecific probe-binding sites were blocked by a 30-min incubation with blocking buffer. Sections were then incubated overnight at 4°C in the presence of alkaline phosphatase-conjugated anti-DIG goat Fab fragment (1:1,500, Roche Diagnostic) and then sequentially rinsed with 0.1% BSA in PBSw and AP1 buffer. The staining was eventually developed using BM Purple AP Substrate, Precipitating (Roche Diagnostic).
Unless otherwise indicated, all chemicals were from Sigma Aldrich (Milano, Italy).
We thank Dr. Heiner Westphal for advice and critical reading of the manuscript and Dr. Domenico Grillo for figure editing.