The transcription factor 4 (TCF4, Gene ID: 6925) and its paralogues TCF3 and TCF12 form the subgroup of class I basic Helix-Loop-Helix (bHLH) transcription factors (TFs). Class I bHLH TFs are ubiquitously expressed, and their transcriptional output is dependent on their interaction partners. They are obligate partners for proneural class II bHLH TFs, which are unable to exert transcriptional function without dimerization to Class I bHLH. Class I bHLH TF are inhibited by interaction Class V bHLH factors, which do not possess a DNA-binding domain and sequester class I factors [1].

TCF4 has gained major attention following the discoveries that heterozygote loss-of-function mutations in TCF4 cause Pitt-Hopkins syndrome (PTHS) [2, 3]—a neurodevelopmental disorder characterized by early onset developmental delay, moderate to severe intellectual disability, autistic behavior, intermittent breathing abnormalities, seizures, and distinctive facial features—and that single nucleotide polymorphisms (SNPs) in potentially regulatory regions of TCF4 are associated with an increased risk of schizophrenia [4–6].

The extent of cognitive deficits in PTHS suggests that TCF4 dosage is critical for the development of various brain regions, but it remains largely unknown which neural populations develop in a TCF4-dependent fashion. Based on the rationale that TCF4 expression will predict regions that require TCF4 activity for their development, we performed a systematic analysis of the temporo-spatial expression pattern of TCF4 protein in mouse brain development. Moreover, we determined TCF4 expression in the adult mouse brain to identify neural systems, in which TCF4 may serve functions in maintenance and plasticity. Our analyses reveal that TCF4 is broadly expressed in cortical and subcortical structures in the development and adulthood with prominent expression in the cortex, hippocampus, and a subset of hypothalamic and amygdaloid nuclei. Comparison of the developmental and adult murine expression pattern with transcriptomic datasets from rhesus monkey and human brain revealed strong similarities in TCF4 expression between species. Moreover, we found that Tcf4 haploinsuffiency in mice replicated structural brain anomalies that have been observed in individuals with PTHS. Thus, our data suggest that TCF4 regulation and function in brain development may be evolutionary conserved.

We first determined the specificity of the Tcf4-antibody intended for immunohistochemical analysis. To this end, HEK 293T cells were transfected with a C-terminally Flag-tagged TCF4-expression construct or with a control expression construct. The anti-TCF4-antibody as well as the anti-Flag antibody detected a single 100 kDa band in cell lysates from TCF4-transfected cells. The 100 kDa band was absent from cell lysates of control-transfected cells and was reduced in cells, in which a TCF4-specific shRNA construct was co-transfected with the TCF4-expression construct (Fig. 1a). Homozygous Tcf4-knockout (Tcf4 ^lacZ/lacZ ) is associated with high perinatal lethality, and from our breedings, we obtained only one Tcf4 ^lacZ/lacZ mouse that survived into adulthood. We used brain tissue from this mouse to further validate specificity of the TCF4-antibody. Staining with the TCF4 antibody yielded a robust signal in tissue from wildtype mice but no signal in homozygote Tcf4-knockout tissue (Fig. 1b, c). Collectively, these data demonstrate the specificity of the antibody for TCF4.

Next, we systematically analyzed the expression pattern of TCF4 protein in the developing and adult mouse brain. On embryonic day 11.5 (E11.5), TCF4 was broadly expressed in the germinal regions for cortical glutamatergic neurons and GABAergic neurons. Immunoreactive intensity varied between the distinct germinal regions. It was high in the cortical and hippocampal neuroepithelium and remained intense in the expanding neocortex and in the developing hippocampal formation (Fig. 1d–h). The ganglionic eminences are the main germinal zones for GABAergic neurons. The lateral ganglionic eminence (LGE) generates GABAergic interneurons and projection neurons of the olfactory bulb, amygdala, and striatum from E12 until birth [7, 8], while the medial ganglionic eminence (MGE) generates the majority of cortical interneurons between E12 and E16.5. MGE-derived interneuron precursors migrate tangentially to their final neocortical location [9]. Notably, the expression of TCF4 in the MGE correlated with the time of cortical interneuron production [10] (Fig. 1d–f). Moreover, TCF4 expression showed a lateral to medial gradient with moderate to high levels in the most medial part of the MGE and low to absent immunoreactivity in the LGE (Fig. 1d–f).

Regionally distinct TCF4 expression levels were observed in subcortical areas. Developing hypothalamic nuclei showed moderate expression of TCF4 starting from E11.5 onwards (Fig. 1d–h); at E15.5 and postnatal day 7 (P7), TCF4 showed prominent expression in a subset of cells in the paraventricular hypothalamic nucleus and the basal peduncular hypothalamus, respectively (Fig. 2a, d). From E15.5, when the amygdaloid complex can be clearly distinguished, we found continuously high expression of TCF4 in this region (Fig. 2b–e). In the globus pallidus (GP), TCF4 expression was high at E15.5 but decreased at subsequent stages (Fig. 2b, c, e). Scattered cells with low to intermediate TCF4 expression were observed in the caudoputamen (CP) at postnatal stages (Fig. 1h; Fig. 2e, f). A summary of the developmental expression pattern and levels of TCF4 is provided in Fig. 3a.

In young adult mice (P56), the highest TCF4 immunoreactivity was found in the hippocampal formation, the cortex, the purkinje cell layer of the cerebellum, and the amygdala (Fig. 4A–D’). Of the amygdala subnuclei, cells of the lateral (LA) and basolateral (BLA) nucleus highly express TCF4. The central nucleus (CEA) also shows high TCF4 immunoreactivity; in contrast to the LA and BLA nuclei, TCF4 signal in the CEA was not nuclear but was located in fibers, whose distribution strongly resembled the distribution pattern of calcitonin gene-related peptide immunoreactive projections that arise from posterior thalamic nuclei [11]. Low to intermediate TCF4 immunoreactivity was found in structures of the diencephalon, the midbrain, and in scattered cells of the caudoputamen (Fig. 4E). The expression pattern of TCF4 in the adult mouse brain is summarized in Fig. 3d.

Next, we analyzed the expression of TCF4 in glial cells. TCF4 was expressed in the majority of astroglia and some Sox10-positive oligodendroglia (Fig. 4F–H”). Expression levels in both glial populations varied from low to high. Iba1-positive microglial cells did not show TCF4 immunoreactivity (Fig. 4F–I”).

Given the prominent expression of TCF4 in the cortex and hippocampus, we analyzed these areas in more detail. During cortical development, radial glial cells (RGCs) in the ventricular zone (VZ) divide to self-renew and to generate either immature neurons or intermediate progenitor cells (IPCs) [12, 13]. IPCs accumulate in the subventricular zone (SVZ) where they give rise to postmitotic neurons via asymmetric cell division [13]. Neurons migrate radially into the expanding cortical plate (CP), forming the distinct cortical layers in an inside-out fashion with later-born excitatory neurons migrating to more superficial layers of the CP passing over earlier born neurons [14, 15]. At E11.5, TCF4 expression is high in the ventricular zone (VZ), which coincides with the generation of preplate (PP) neurons from neural progenitors in this region (Fig. 5A). Co-staining with the precursor marker Sox2 confirmed that TCF4 was indeed highly expressed by VZ precursors at this time point (Fig. 5G–G”). At later stages (E13.5 and thereafter), TCF4 immunoreactivity was low to moderate in the germinal zones, i.e., the VZ and SVZ, but was high in the PP and the expanding cortical plate (Fig. 5B–D). At E13.5, the highest TCF4 expression levels co-localized with DCX, a marker for immature migrating neurons (Fig. 5H–H”); at E18.5, high TCF4 expression levels frequently co-localized with NeuN, a marker for post-migratory neurons (Fig. 5I–I”). Cells with high TCF4 expression were also observed in the postnatal and adult cortex (Fig. 5E, F, J–L”). These TCF4-expressing cells also expressed the layer V/VI neuronal marker CTIP2 or the layer II-IV neuronal marker Cux1, demonstrating that subsets of cortical neurons expressed high levels of TCF4 (Fig. 5J–K”). In addition, high TCF4 expression was observed in a subset of GAD67-positive cortical interneurons (Fig. 5L–L”).

The hippocampal neuroepithelium is the germinal zone for hippocampal neurons and is subdivided into the Ammon’s horn neuroepithelium (ANE) and the dentate gyrus neuroepithelium (DNE). CA1-CA3 pyramidal neuron precursors are generated in the ANE and migrate radially from the VZ to the prospective CA1-CA3 subfields where they undergo terminal differentiation and maturation. The DNE, which is also called primary matrix, generates dentate granule neuron precursors (secondary matrix) that migrate along a radial glial scaffold to the subpial zone of the hippocampus [16–19]. Reaching the hippocampal fissure, precursors accumulate and form the tertiary matrix [16]. During the first postnatal week, the secondary matrix disappears, while the tertiary matrix remains the proliferative zone generating the inner shell of the granule cell layer [16, 19]. From P14 onwards, the tertiary matrix resolves and the neurogenic niche becomes successively confined to the subgranular zone (SGZ) of the dentate gyrus (DG) [16, 20]. TCF4 is highly expressed in pyramidal and dentate granule precursors and neurons during the course of hippocampal development (Fig. 6a–f). In contrast to the adult cortex, where high TCF4 expression is observed only in a subset of neurons, TCF4 expression in the adult hippocampus remained prominent throughout the CA subfields and the dentate gyrus (Fig. 6g–o). In the dentate gyrus, TCF4 was not only strongly expressed in Calbindin-positive mature dentate granule neurons but also in Parvalbumin-positive interneurons (Fig. 7D–E”). The dentate gyrus is one of two regions of the adult mammalian brain, where stem cells generate neurons throughout life [21]. Nestin-positive radial glia-like stem cells in the subgranular zone displayed moderate TCF4 immunoreactivity (Fig. 7A–A”). Sox2-positive precursors and astroglia in the SGZ showed variable levels of TCF4 expression, whereas NeuroD1-positive immature dentate granule neurons showed strong TCF4 immunoreactivity (Fig. 7B–C”). Collectively, the expression data of TCF4 strongly suggests that TCF4 is intimately linked to developmental and adult hippocampal neurogenesis. The persistence of high TCF4 expression in mature neurons of all hippocampal subfields raises the intriguing possibility that TCF4 is also involved in hippocampal plasticity.

Next, we sought to evaluate whether the TCF4 expression pattern is conserved from mice to humans. The NIH Blueprint Non-Human Primate (NHP) Atlas (2009) (http://www.blueprintnhpatlas.org) provides mRNA expression data of the developing rhesus monkey brain from E40–0 months; the BrainSpan Atlas documents mRNA expression data of human neural development for post conception weeks 15–21 [22]. TCF4 mRNA expression data in the adult human brain was extracted from the Allen Human Brain Atlas (© 2010 Allen Institute for Brain Science. Allen Human Brain Atlas. Available from: human.brain-map.org, [23]). The neurodevelopmental expression pattern of TCF4 was highly similar between mice, rhesus monkeys, and humans: regions with high expression during murine neural development such as the hippocampus, the cortex, the ganglionic eminences, and some nuclei of the amygdaloid complex, consistently show high TCF4 expression in the developing brain of rhesus monkey and humans, whereas low TCF4 expressing regions such as the caudoputamen also show low expression during rhesus monkey and human brain development (Fig. 3a-c).

Striking similarities in the TCF4 expression pattern were also found during adulthood. Regions with high expression in adult mice, i.e., hippocampus, cerebellum, cortex, and nuclei of the amygdaloid complex, were also the highest TCF4-expressing regions in the adult human brain, while, for example, the diencephalon, metencephalon, myelencephalon, and parts of the basal ganglia showed low TCF4 expression levels in humans and mice (Fig. 3d, e). These data indicate that the expression pattern of TCF4 is conserved from mouse to humans.

In humans, TCF4 haploinsufficiency causes Pitt-Hopkins syndrome, a severe neurodevelopmental disorder, associated with psychomotor delay, intellectual disability, and autistic behavior. MRI-based findings of structural anomalies in PTHS are variable, ranging from normal cerebral MRIs (30–50% of cases) to enlarged ventricles, cerebellar atrophy, hippocampal hypoplasia, and hypoplasia of the corpus callosum [24–26]. We analyzed MRIs of two individuals with genetically confirmed PTHS. Consistent with previous studies reporting hypoplasia of the corpus callosum as a frequent structural anomaly in PTHS (24–45% of cases), we found general thinning of the corpus callosum and aplasia of the splenium of the corpus callosum in both individuals (Fig. 8A–D). One of the individuals also displayed a small hippocampal formation (Fig. 8E).

We next investigated whether Tcf4 haploinsufficiency in mice reproduced structural anomalies of PTHS. To determine the influence of Tcf4-haploinsufficiency on cortical architecture, we analyzed the diameter of six different cortical regions along the rostral to caudal axis. Cortical thickness was significantly reduced in all analyzed areas in Tcf4-haploinsufficient (Tcf4ex4 ^lacZ/WT ) mice compared to their wildtype littermates (WT) (Fig. 8I–P). Moreover, the dentate gyrus volume was significantly decreased in Tcf4ex4 ^lacZ/WT mice (Fig. 8F–H). In addition, Tcf4ex4 ^lacZ/WT mice displayed agenesis of the splenium of the corpus callosum (Fig. 8F–G’). Thus, Tcf4 haploinsufficiency in mice results in anatomical deficits that resemble anatomical anomalies observed in PTHS.

Finally, we qualitatively compared TCF4-expression in the cortex and hippocampus at E15.5, E18.5, P7, and P56 between WT and Tcf4-haploinsufficient mice (Fig. 9) The cortex of Tcf4ex4 ^lacZ/WT animals appeared markedly thinner at all time points (Fig. 9A–D’). The density of TCF4-expressing cells in the cortex of Tcf4ex4 ^lacZ/WT mice seemed to be reduced, and TCF4-expressing cells showed a less regular distribution. Similarly, the arrangement of TCF4-expressing cells in the developing hippocampus (E15.5–P7) of Tcf4ex4 ^lacZ/WT mice was less organized (Fig. 9E–H’). Moreover, TCF4-expressing cells in the CA3 subfield of adult Tcf4-haploinsufficient mice appeared to be less densely packed (Fig. 9H–H’).

Here, we used immunohistochemical analysis to systematically investigate the expression of the intellectual disability and schizophrenia-associated Tcf4 gene in the developing and adult murine CNS. Consistent with recent experimental evidence linking TCF4 to precursor proliferation, neuronal positioning and development of excitability in the developing cortex [27–29], and the suggestion that TCF4 regulates neurite branching and synapse formation in prefrontal cortical neurons [30], we found TCF4 to be expressed from the early precursor stage up to the post-migratory neuronal stage in the developing cortex. We also demonstrate that TCF4 is expressed in the major germinal zones for the principal neurons of the hippocampus, for cortical interneurons, as well as in developing neurons of, for example, the cerebellum and some nuclei of the amygdaloid complex. The broad expression of TCF4 strongly suggests that TCF4 regulates early and late development of cortical and subcortical structures.

As a class I bHLH TF, the transcriptional output of TCF4 is dependent on interaction with other bHLH TFs. Through such interactions, TCF4 may act as a network hub coordinating different transcriptional modules and developmental processes. The interaction partners of TCF4 in the developing CNS have not been defined. The bHLH TFs Ascl1, Neurogenin 1 and 2, NeuroD1, and NeuroD6, have a restricted developmental expression pattern and fulfill defined stage-specific functions during hippocampal and cortical development [31–33]. It is tempting to speculate that TCF4 fulfills at least part of its functions in CNS development through interaction with these proneural class II bHLH TFs.

Recent work indicated that TCF4 activity is modulated by post-translational mechanisms such as neuronal activity-dependent phosphorylation [34]. Interestingly, we detected TCF4 immunoreactivity in fibers of the CEA, raising the possibility that TCF4 activity may be modulated by nuclear-cytoplasmatic translocation.

PTHS is associated with deficits in various functional domains including cognition, social interaction, speech, motor skills, and vegetative function, and there is some debate to which extent the PTHS phenotype may be due to post-developmental functions of TCF4 in mature neurons [35]. Indeed, it has been shown that TCF4 directly regulates plasticity of CA1 pyramidal neurons potentially through the control of activity-dependent DNA methylation [35]. Strikingly, we found that TCF4 is highly expressed not only in the hippocampus but also in a number of cortical and subcortical structures during adulthood, raising the possibility that TCF4 function in mature neurons is critical for plasticity in many neural circuits.

TCF4 haploinsufficiency causes PTHS, while SNPs in regulatory regions, which may cause increased TCF4 expression [36], are associated with an elevated risk for schizophrenia. Moreover, behavioral deficits have been reported not only for Tcf4-haploinsufficient mice but also for transgenic mice with mild TCF4 over-expression [37, 38]. These observations have led to the suggestion that TCF4 levels have to be kept in a tight range to allow for neuronal development and function. Intriguingly, we observed cell type- and region-dependent differences in TCF4 expression, raising the questions whether distinct cell populations may require different TCF4 dosages and whether there is cell type-/region-specific vulnerability to pathologically altered TCF4 expression levels.

A recent study reported similarities in the expression of TCF4 mRNA in the developing prefrontal cortex of rats and humans [39]. Here, we show that the expression pattern of TCF4 between mice and humans is highly similar across the developing and adult brain and provide evidence that Tcf4 haploinsufficiency in mice reproduces structural anomalies that are observed in PTHS. These observations strongly suggest that the regulation and function of TCF4 is conserved between species and further validate Tcf4-haploinsufficient mice as a preclinical model for PTHS.

Our preliminary analysis indicated an irregular distribution of TCF4-expressing cells in the cortex and hippocampus of Tcf4-haploinsufficient mice strengthening the notion that TCF4 plays a role in migration and positioning of neurons [40]. A more detailed analysis of Tcf4-haploinsufficient mice will allow to decipher the pathophysiological mechanisms underlying PTHS on the molecular, cellular, and neural circuit level and to explore experimental strategies to ameliorate cognitive and behavioral deficits in this disorder [35, 41, 42].