Human fetal tissue from terminated pregnancies was obtained from the joint MRC/Wellcome Trust-funded Human Developmental Biology Resource (HDBR; Gerrelli et al., 2015).¹ All tissue was collected with appropriate maternal consent and approval from the Newcastle and North Tyneside NHS Health Authority Joint Ethics Committee. Fetal samples from genetically normal cases ranging in age from 7.5 to 21 PCW were used. Ages were estimated from foot and heel to knee length measurements according to Hern (1984).

RNA in situ hybridization experiments were performed using the RNAscope^® technology, which has been previously described (Wang et al., 2012). Paired double-Z oligonucleotide probes were designed and manufactured by ACD Bio Techne against target RNA (HsFEZ1; catalogue number 468471, a 20ZZ probe targeting base pairs 255–1,433 of NM_005103.4). The RNAscope Reagent Kit (ACD Bio Techne, Abingdon, United Kingdom) was used according to the manufacturer’s instructions but with slight modifications. In brief, 8 μm-thick paraffin sections were baked on a heating pad for 10 minutes at 60°C, dewaxed in xylene, and then boiled with target retrieval buffer (ACD) for 20 min at 95°C. Protease digestion was carried out at 40°C for 30 min, followed by probe hybridization for 2 h at 40°C with target probes. The hybridized signals were amplified by a cascade of signal amplification molecules and detected with the RNAscope 2.5 HD detection kit (Fast Red). Slides were counterstained with toluidine blue, positive signals showed as red chromogenic dots in the cytoplasm or nucleus. Each sample was quality controlled for RNA integrity with a probe specific to the housekeeping gene GAPDH. Negative control background staining was evaluated using a probe specific to the bacterial dapB gene.

In situ hybridization was carried out first as described above except that the hybridized signals were amplified and detected using Opal 570 tyramide signal amplification (TSA; Akoya Biosciences, Delaware, United States). Then immunofluorescent staining was carried out according to previously described protocols (Alzu’bi et al., 2022) with various antibodies (Table 1). Sections were boiled in 10 mM citrate buffer pH 6 for 10 min, followed by incubation with primary antibody (diluted in 10% normal blocking serum in Tris buffered saline [TBS] pH 7.6) overnight at 4°C. Sections were then incubated with HRP-conjugated secondary antibody for 30 min (ImmPRESS^™ HRP IgG [Peroxidase] Polymer Detection Kit, Vector Labs). Signals were detected with Opal 520 TSA applied for 10 min (Akoya Biosciences). Sections were mounted using Vectashield antifade mounting medium with DAPI (Vector Labs, Peterborough, United Kingdom). Sections from eight fetal brains ranging in age from 8 to 21 PCW were used.

A number of databases were consulted to provide information on FEZ1 expression, including whole tissue and single cell RNAseq. Using data deposited at www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-4840 (Lindsay et al., 2016). 138 samples of cortical tissue taken at ages ranging from 7.5 to 17 PCW, and from various positions along the anterior posterior axis of the cortex including the temporal lobe were analyzed for tissue RNAseq. As shown in Figure 1A, FEZ1 was highly expressed from the beginning of cortical plate formation (7.5 PCW) until the latest developmental stage studied, with expression increasing significantly with age. For all samples except one expression levels are in the top quartile for protein coding genes (> 40 normalized RPKM) and in 85% of samples expression levels were in the top 5% of protein coding genes (> 160 normalized RPKM; Harkin et al., 2017). This suggests a crucial role for FEZ1 in cortical development, and that it might be predominantly expressed in post-mitotic cells, which increase as a proportion of the cell population in the cortical wall with age. No statistically significant difference in expression between cortical regions was observed (not shown). This suggests that FEZ1 does not play a role in cortical arealisation.

To explore the cell type specificity of FEZ1 expression further, the cortical development expression viewer (CoDEx)² a scRNAseq database, was interrogated. This provides data on expression from human cortical tissue samples at 17/18 PCW (Polioudakis et al., 2019). As shown in Figure 1B, FEZ1 was found to be ubiquitously expressed in a high proportion of cells, including glutamatergic and GABAergic neurons, as well as subtypes of progenitor cell. Expression was consistently highest in more mature glutamatergic neurons (subplate and lower cortical plate) lower in dividing progenitor cells compared to quiescent progenitors, and lowest in cells of non-neuroectodermal origin such as microglia, pericytes and endothelial cells. Expression levels were highly variable in many cell types.

We also examined expression of genes for proteins known to interact with FEZ1 (see Introduction). For tissue RNAseq, we found two neurodevelopmental susceptibility genes that also interact with the SNARE complex and showed the same pattern of high and increasing expression with time (Figure 1A). We found that neurexin 1 (NXRN1) expression, a high probability susceptibility gene for schizophrenia (Hu et al., 2019) was particularly tightly coupled to FEZ1 expression across our samples. STXBP1 expression was also highly correlated with FEZ1, but not as tightly as NRXN1. STXBP1 has been associated with schizophrenia (Behan et al., 2009) although with less evidence than NRXN1 but is more commonly associated with infantile epilepsy (Xian et al., 2022). On the other hand, the synapse associated susceptibility gene DISC1 showed only low levels of expression which did not change over time (not shown). We further examined scRNA seq for the cortex. This confirmed findings of high expression for NRXN1 (Figure 1B) and STXBP1 (not shown) with a strikingly similar cellular distribution to FEZ1. In situ hybridization data deposited at the BrainSpan atlas of the developing human brain (Ding et al., 2022)³ confirmed this expression pattern for NRXN1 at 15 and 21 PCW with particularly high expression in mature glutamatergic neurons.

Retinoic acid receptors interact with FEZ1 at the level of transcriptional control (see introduction) we found there expression to be relatively low and uniform across development by tissue RNAseq (not shown). Cellular distribution of the most highly expressed isoform of retinoic acid receptor A (RARA) showed highest expression in quiescent radial glia and GABAergic interneurons or their immature precursors (Figure 1B) suggesting that FEZ1 may act to regulate transcription in these cell types.

Altered development of thalamocortical circuits are strongly implicated in schizophrenia (Jiang et al., 2021) thus cell specific expression in the thalamus was explored using data from an 18 PCW human thalamus (sample deposited at NeMO Analytics; Orvis et al, 2021). TSNe plots were constructed for FEZ1 and markers of various cell phenotypes (Figure 1C) revealing that FEZ1 is also widely expressed in the thalamus, co-localizing with both glutamatergic (SLC17A6/vGLUT2 expressing) and GABAergic (GAD1/GAD67 expressing) neurons as well as progenitor cells and astrocytes (SOX2 expressing).

Taken together, it appears that FEZ1 is expressed by maturing neurons in certain locations, where it may play a role in axon, dendrite and synapse formation (see introduction) but is also widely expressed in progenitor cells, where it may regulate transcriptional programs but possibly play a role in extension of radial glial fibers.

In the cerebral cortex at 8 post conceptional weeks (PCW) just after the onset of cortical plate formation, expression of both FEZ1 mRNA and FEZ1 protein was highest in the post-mitotic neurons of the subplate (SP) and cortical plate (CP), but expression was detectable in the progenitor cell containing ventricular and subventricular zones VZ and SVZ (Figures 2A–C). Protein expression was higher in neurons (identified by doublecortin immunoreactivity) settled in the cortical plate compared to those migrating through the subplate and intermediate zone suggesting these is no FEZ1 expression in DCX+ migrating neurons and neurites found in these locations (Figure 2C). However, in the proliferative ganglionic eminences (GE) and progenitor zones of the diencephalon (asterisk; Figure 2A) expression is higher than in sub-cortical post-mitotic neurons. This dichotomy is maintained at 10 post-conceptional weeks (Figures 2D–F). Interestingly FEZ1 mRNA expression in dorsal and lateral cortical progenitor zones appeared decreased compared to medial zones, although this was not supported by immunohistochemical detection of FEZ1 protein expression (not shown). FEZ1 mRNA was strongly co-localized with TBR1, a marker for post-mitotic glutamatergic neurons, but also expressed in non-TBR1 positive cells in the VZ and SVZ, presumably progenitor cells (Figure 2G). The transcriptomics studies (Figure 1) suggested actively dividing cortical cells show low levels of FEZ1 expression. Immunohistological evidence suggested that while putative dividing KI67+ progenitors in the SVZ show little evidence of FEZ1 co-expression, ventricular zone radial glial cells with cell bodies located at the apical surface (where they undergo mitosis; Taverna and Huttner, 2010) strongly co-express these proteins (Figure 2H).

FEZ1 mRNA expression was maintained in the forebrain throughout development and increased in the post-mitotic glutamatergic neurons of the thalamus. Different levels of expression were observed in different thalamic nuclei, with the lateral geniculate nucleus (LGN) showing higher levels of expression than the pulvinar, for instance (Figures 3A–C). However, this at least partially reflects a higher cell density in the LGN. Nevertheless, distinct thalamic nuclei are identifiable at this age unlike earlier stages of development shown in Figure 2.

Expression appeared strong in the cortical plate right up until the oldest stage studied (21 PCW; Figures 3A,E) and particularly in the pyramidal layer of the hippocampus (Figure 3D). Expression in the cortical plate was strongest at the boundary between the cortical plate and the marginal zone. This may partially reflect a higher density of cells at this location but also reflects higher levels of expression per individual cell. These neurons are the most recent cells to have stopped migrating to the cortical plate at this stage and thus this may suggest a role for FEZ1 in cell migration. This expression pattern has previously been observed for neurexins (Harkin et al., 2017; Ding et al., 2022) the intracellular transport of which is a mooted role for FEZ1 (see introduction). Intense expression of FEZ1 by glutamatergic pyramidal neurons was a feature of the hippocampus, however, there was low expression in the stratum oriens which contains predominantly GABAergic neurons of various phenotypes (Figure 3A; Freund and Buzsáki, 1996). Similarly, the caudal ganglionic eminence, containing GABAergic progenitor cells, showed relatively high expression compared to the overlying caudate nucleus which is principally composed of post-mitotic GABAergic neurons. Relatively high levels of expression of FEZ1 were still observed in the cortical VZ and SVZ although by this stage neurogenesis is ending and gliogenesis beginning (Cadwell et al., 2019) The VZ shows reduced levels of cell division at this stage (Nowakowski et al., 2016).

We explored the apparent reduced expression of FEZ1 by GABAergic neurons further by examining co-expression of GAD67 (GABA synthesizing enzyme) and FEZ1 by immunohistochemistry at 16 PCW (Figure 4). The reticular nucleus of the thalamus, which in maturity comprises a sheet of GABAergic neurons surrounding the dorsal and lateral portions of the thalamus, showed high expression of GAD67, as expected, but showed little expression of FEZ1 (Figure 4B). Interestingly, there was a gradient of expression of GAD67 from ventral and lateral to dorsal parts of the thalamus (Figures 4A,B). Many FEZ1+ cell bodies were intermingled with GAD67+ cells in more ventral parts of the thalamus, but there was little evidence of co-expression. This suggests that thalamic post-mitotic thalamic glutamatergic neurons express FEZ1 by this stage of development. It also reveals potential pathways of migration of GABAergic neurons into the thalamus from the prethalamus (precursor of the reticular nucleus) and subthalamic midbrain regions at this stage of development. In other areas of high GABAergic cell density such as the nascent basal ganglia, GAD67 immunoreactivity was high but FEZ1 expression relatively low, although FEZ1 expression remained high in adjacent subpallial GABAergic neuroprogenitor zones (Figure 4C).

By 16 PCW GABAergic neurons have increased in the cortex having largely migrated there from the ganglionic eminences (Ma et al., 2013; Alzu’bi and Clowry, 2019). The majority of cortical GAD67+ neurons did not show double labelling for FEZ1 at 16 PCW, although in or near the marginal zone, a major pathway for migrating GABAergic neurons at this stage of development (Zecevic and Rakic, 2001; Meyer and González-Gómez, 2018) some small double-labelled cells were observed predominantly near or close to the MZ (Figure 4D). These may be the small calretinin positive GABAergic interneurons previously described (Meyer and González-Gómez, 2018; Alzu’bi and Clowry, 2020).

FEZ1 protein expression at later developmental stages was not confined to neuronal cell bodies but showed selective enrichment in certain axon pathways. For instance, strong expression was observed in the fornix and certain callosal axons at 19 PCW (Figure 5A) but in the lateral cortical cortex, expression in the cortical intermediate zone, where large numbers of growing axons are found, was weak.

Relatively low FEZ1 expression in the subplate contrasted with stronger expression of both GAP43 and FEZ1 in cells of the cortical plate, and immunoreactivity in the inner fiber layer was less than in the ventricular zone (Figure 5B). Double labelling for the growing axon marker GAP43 and FEZ1 demonstrated a concentration of FEZ1 in selected axon pathways. For, instance, in the internal capsule at 16 PCW, FEZ1 was expressed by what appeared to be thalamocortical axons, whereas adjacent corticofugal axons were GAP43+ only (Figure 5C). We interpret this as showing that FEZ1 is selectively expressed by thalamocortical axons but more strongly in fasciculated axons proximal to the thalamus (Figure 5C) than in the distal terminal branches in the subplate (Figure 5B). FEZ1 expression is stronger in the cell bodies of glutamatergic neurons of the cortical plate than in their growing corticofugal axons in the internal capsule.