We first examined which FGF receptors (FGFRs) are expressed in the developing cerebral cortex of ferrets at postnatal day 0 (P0), when cortical folds are about to be formed. Reverse transcription-PCR (RT-PCR) demonstrated that FGFR1, FGFR2 and FGFR3 were expressed in the ferret cerebral cortex (Figure 1A). To examine the role of FGF signaling in cortical folding, we inhibited FGF signaling in the developing ferret cortex. Because multiple FGFRs are expressed in the cortex, we introduced a soluble extracellular domain of FGFR3 (sFGFR3), which acts in a dominant-negative manner, into the ferret cerebral cortex using our IUE technique to suppress all FGFR activity (Figure 1B). It should be noted that sFGFR3, which consists of the extracellular domain of FGFR3 without the transmembrane domain, is released from transfected cells and binds to endogenous FGFs (Fukuchi-Shimogori and Grove, 2001). It was therefore supposed that sFGFR3 suppresses FGF signaling not only in transfected cells but also in neighboring non-transfected cells non-cell-autonomously, presumably resulting in stronger phenotypes. To test this, we introduced sFGFR3 into the ferret cerebral cortex using IUE and performed in situ hybridization for Sprouty2, whose expression is known to be up-regulated by FGFR activation (Tsang and Dawid, 2004). Sprouty2 signals were abundantly distributed in the OSVZ and the VZ of the control ferret cortex (Figure 1—figure supplement 1). In contrast, Sprouty2 signals in non-transfected cells, which did not show GFP signals in their somata, were markedly suppressed by sFGFR3 electroporation in the OSVZ (Figure 1—figure supplement 1). These results suggest that sFGFR3 indeed suppresses FGF signaling non-cell-autonomously.

Figure 1 with 4 supplements see all

Download asset Open asset

We introduced sFGFR3 into the ferret cortex using IUE at embryonic day 33 (E33) and examined cortical folds at P16. Interestingly, we found that cortical folding was impaired in sFGFR3-transfected areas (Figure 1B). Consistently, coronal sections showed that cortical folding was inhibited in the sFGFR3-transfected side of the cortex compared with the contralateral side of the cortex and the EGFP-transfected control cortex (Figure 1C,D). To quantify the effects of sFGFR3 on cortical folding, we utilized the local gyrus size (GS) ratio (Figure 1—figure supplement 2), the local sulcus depth (SD) ratio (Figure 1—figure supplement 3) and the local gyrification index (GI) ratio (Figure 1—figure supplement 4). Consistent with our observation, the local GS ratio, the local SD ratio and the local GI ratio were significantly reduced by sFGFR3 (local GS ratio: control, 0.75 ± 0.05; sFGFR3, 0.25 ± 0.05; p=0.003; Student's t-test) (local SD ratio: control, 0.88 ± 0.03; sFGFR3, 0.47 ± 0.10; p=0.019; Student's t-test) (local GI ratio: control, 0.95 ± 0.04; sFGFR3, 0.72 ± 0.05; p=0.025; Student's t-test) (Figure 1E–G, Supplementary file 1). These results clearly indicate that FGF signaling is crucial for cortical folding. It should be noted that cortical folding in peripheral regions of the GFP-positive area did not seem to be affected in the sFGFR3-transfected cortex. It seemed likely that the expression level of sFGFR3 was not sufficient to inhibit FGF signaling strongly in peripheral regions of the GFP-positive area. We also noticed that the local GS ratio and the local SD ratio were slightly smaller than 1 in the GFP-transfected control cortex (local GS ratio, p=0.009; local SD ratio, p=0.02; local GI ratio, p=0.19; one sample t-test), raising the possibility that cortical folding can be affected by electroporation itself. It would be important to be careful when interpreting experimental results using IUE. In addition, it is worth noting that because GFP-positive areas were wide enough to include both the gyrus and sulcus regions, it seemed possible that sFGFR3 affected both the formation of the gyrus and that of the sulcus.

We examined the detailed expression patterns of FGFR1, FGFR2 and FGFR3 in coronal sections of the ferret cerebral cortex during development. In situ hybridization showed that FGFR2 and FGFR3 mRNAs were predominantly expressed in the germinal zones such as the VZ at E32, and the VZ and the SVZ at E36 and E40 (Figure 2—figure supplement 1A–C). FGFR1 was also expressed in the VZ, but its signals were weaker than those of FGFR2 and FGFR3 (Figure 2—figure supplement 1A–C). Interestingly, the expression of FGFR2 and that of FGFR3 in the OSVZ were not uniform throughout the cortex at E36 and E40. Some areas contained abundant FGFR signals in the OSVZ, whereas other areas had fewer FGFR signals (Figure 2—figure supplement 1B,C). These results led to the hypothesis that the regional differences in FGFR expression in the OSVZ mediate cortical folding. We quantified FGFR signal intensities along the tangential axis in the OSVZ at E40 (Figure 2—figure supplement 1D–G). FGFR signals indeed showed regional differences in the OSVZ (Figure 2—figure supplement 1D–G).

Previously, lateral to medial gradients in cell proliferation were reported in the rodent cerebral cortex (Bayer and Altman, 1991). Similarly, it was also reported that Fgfr expression showed medio-lateral gradients in the VZ of the rodent cerebral cortex during early corticogenesis, although it became uniform afterward (Hasegawa et al., 2004; Hébert et al., 2003). It therefore seemed possible that the regional differences in FGFR signals found in the ferret cortex reflected the medio-lateral gradients found in mice. Another possibility was that the regional differences in FGFR expression in ferrets corresponded to the positions of gyri and sulci. To address these possibilities, we next performed in situ hybridization for FGFR2 and FGFR3 at P6, when the gyrus was beginning to be formed. We found that FGFR2 and FGFR3 were abundantly expressed in the region where the gyrus would be formed, but less abundant where the sulcus would be formed (Figure 2—figure supplement 2). These results suggest that the regional difference in FGFR expression in the OSVZ corresponds to the positions of the gyrus and the sulcus. Our findings were consistent with a previous work showing uneven distribution patterns of FGFR2 and FGFR3 in the ferret cortex (de Juan Romero et al., 2015). In contrast, there were no similar regional differences in Fgfr expression in the lissencephalic cerebral cortex of developing mice at E15, which corresponds to E40 in ferrets (Figure 2—figure supplement 3) (Bansal et al., 2003). The regional differences in FGFR expression in the OSVZ during development therefore seemed to be an important feature of the gyrencephalic cortex of mammals.

Previous studies uncovered regional differences in the distribution of SVZ progenitors in the developing ferret cerebral cortex (Toda et al., 2016; de Juan Romero et al., 2015) and reported that the regional differences of SVZ progenitors were correlated with the positions of gyri (Toda et al., 2016; Reillo et al., 2011; Nowakowski et al., 2016; Smart et al., 2002). It seemed possible that the regional differences of FGFRs correspond to those of SVZ progenitors. We therefore examined the expression pattern of Pax6, which is expressed in RG cells and oRG cells, and found that FGFRs and Pax6 showed similar regional differences in the OSVZ (Figure 2—figure supplement 1B,C). These findings are consistent with the idea that FGF signaling in the OSVZ underlies cortical folding of the ferret cerebral cortex.

To determine the cellular identities of FGFR-positive cells in the OSVZ at a single-cell level, we performed triple staining with in situ hybridization for FGFRs and immunostaining for Pax6 and Tbr2. We found that FGFR2-positive cells and FGFR3-positive cells in the OSVZ were mostly positive for Pax6 (Figure 2A, arrows) but negative for Tbr2 (Figure 2B, arrowheads and 2C) (FGFR2: Pax6+/Tbr2- cells, 91.4 ± 6.2%; Pax6+/Tbr2+ cells, 3.8 ± 3.1%; Pax6-/Tbr2+ cells, 0.0%; Pax6-/Tbr2- cells, 4.9 ± 4.4%) (FGFR3: Pax6+/Tbr2- cells, 76.7 ± 5.7%; Pax6+/Tbr2+ cells, 12.4 ± 5.7%; Pax6-/Tbr2+ cells, 0.0%; Pax6-/Tbr2- cells, 10.8 ± 10.5%). Similar results were obtained for FGFR1 (Pax6+/Tbr2- cells, 78.7 ± 10.0%; Pax6+/Tbr2+ cells, 5.8 ± 5.4%; Pax6-/Tbr2+ cells, 0.0%; Pax6-/Tbr2- cells, 15.4 ± 6.5%) (Figure 2A, arrows; 2B, arrowheads and 2C), although FGFR1-positive cells were fewer. Because it was previously reported that astrocytes also expressed Pax6 in the mouse cerebral cortex (Sakurai and Osumi, 2008), we next examined whether Pax6-positive cells in the OSVZ included astrocytes. We found that Pax6-positive cells in the OSVZ did not express GFAP (Figure 2—figure supplement 4). These results suggest that FGFR-positive cells in the OSVZ are mostly Pax6-positive/Tbr2-negative oRG cells rather than Tbr2-positive IP cells or astrocytes.

Figure 2 with 4 supplements see all

Download asset Open asset

Because FGFRs are expressed in oRG cells, we next examined the role of FGF signaling in the production of oRG cells by expressing sFGFR3. We found that sFGFR3 markedly reduced Pax6-positive cells in the OSVZ underneath EGFP-positive transfected areas at P6 (Figure 3B–D). Our quantification showed that Pax6-positive cells in the OSVZ were significantly decreased by sFGFR3 (OSVZ: Control, 9.40 ± 2.25; sFGFR3, 2.54 ± 0.77; p=0.008; Student's t-test), while those in the VZ and the ISVZ were not (VZ: Control, 8.70 ± 2.30; sFGFR3, 11.12 ± 1.71; p=0.15; Student's t-test) (ISVZ: Control, 17.9 ± 7.1; sFGFR3, 13.4 ± 4.7; p=0.25; Student's t-test) (Figure 3E, Supplementary file 1). Interestingly, Tbr2-positive cells were also reduced by sFGFR3 in the OSVZ selectively (VZ: Control, 1.96 ± 0.80; sFGFR3, 2.02 ± 0.29; p=0.46; Student's t-test) (ISVZ: Control, 8.79 ± 3.54; sFGFR3, 6.74 ± 2.41; p=0.27; Student's t-test) (OSVZ: Control, 6.37 ± 2.08; sFGFR3, 2.26 ± 0.32; p=0.025; Student's t-test) (Figure 3F,G, Supplementary file 1), although FGFR was rarely expressed in Tbr2-positive cells (Figure 2). We further examined the detailed identities of the affected cells in the OSVZ using double immunostaining for Pax6 and Tbr2. We found that both Pax6-positve/Tbr2-negative oRG cells and Pax6-positive/Tbr2-positive IP cells were significantly decreased by sFGFR3 in the OSVZ (Pax6-positive/Tbr2-negative cells: Control, 2.38 ± 0.41 per 1000 μm²; sFGFR3, 1.61 ± 0.26 per 1000 μm²; p=0.044; Student's t-test) (Pax6-positive/Tbr2-positive cells: Control, 1.49 ± 0.69 per 1000 μm²; sFGFR3, 0.35 ± 0.13 per 1000 μm²; p=0.041; Student's t-test) (Figure 3—figure supplement 1, Supplementary file 1). These results suggest that FGF signaling is essential for producing OSVZ progenitors. Because a previous study showed that oRG cells in OSVZ have the ability to produce IP cells (Hansen et al., 2010), it seemed plausible that reduction of Pax6-positive/Tbr2-negative oRG cells resulted in that of Tbr2-positive cells.

Figure 3 with 2 supplements see all

Download asset Open asset

We next tested whether cell proliferation in the OSVZ was affected by sFGFR3. As expected, we found that Ki-67-positive cells in the OSVZ, but not in the VZ and the ISVZ, were significantly reduced by sFGFR3 (VZ: Control, 3.22 ± 0.38 per 1000 μm²; sFGFR3, 3.07 ± 0.38 per 1000 μm²; p=0.36; Student's t-test) (ISVZ: Control, 2.15 ± 0.55 per 1000 μm²; sFGFR3, 2.19 ± 1.07 per 1000 μm²; p=0.48; Student's t-test) (OSVZ: Control, 1.75 ± 0.08 per 1000 μm²; sFGFR3, 0.76 ± 0.28 per 1000 μm²; p=0.004; Student's t-test) (Figure 3—figure supplement 2A,B, Supplementary file 1). Consistent results were obtained using anti-phospho-histone H3 (pHH3) (VZ: Control, 0.28 ± 0.02 per 1000 μm²; sFGFR3, 0.29 ± 0.07 per 1000 μm²; p=0.39; Student's t-test) (ISVZ: Control, 0.28 ± 0.08 per 1000 μm²; sFGFR3, 0.23 ± 0.06 per 1000 μm²; p=0.26; Student's t-test) (OSVZ: Control, 0.091 ± 0.041 per 1000 μm²; sFGFR3, 0.024 ± 0.008 per 1000 μm²; p=0.043; Student's t-test) and anti-phosphorylated vimentin (pVim) (VZ: Control, 0.57 ± 0.33 per 1000 μm²; sFGFR3, 0.69 ± 0.07 per 1000 μm²; p=0.33; Student's t-test) (ISVZ: Control, 0.20 ± 0.10 per 1000 μm²; sFGFR3, 0.34 ± 0.15 per 1000 μm²; p=0.16; Student's t-test) (OSVZ: Control, 0.096 ± 0.032 per 1000 μm²; sFGFR3, 0.038 ± 0.020 per 1000 μm²; p=0.049; Student's t-test) antibodies (Figure 3—figure supplement 2C–F, Supplementary file 1). These results indicate that the number of proliferating cells in the OSVZ was reduced by sFGFR3.

We then examined which cell types had their proliferation affected by sFGFR3 in the OSVZ. We performed triple immunostaining for Pax6, Tbr2 and pHH3, and found that the percentage of Pax6-positive/Tbr2-negative cells co-expressing pHH3 in the OSVZ was significantly reduced by sFGFR3 (Control, 2.57 ± 0.91%; sFGFR3, 0.67% ± 0.34%; p=0.025; Student's t-test), while that of Pax6/Tbr2-double positive cells co-expressing pHH3 was not (Control, 1.28 ± 0.60%; sFGFR3, 0.89% ± 1.25%; p=0.36; Student's t-test) (Figure 4, Supplementary file 1). Thus, our findings indicate that proliferation of Pax6-positive/Tbr2-negative oRG cells in the OSVZ is regulated by FGF signaling. This is consistent with our data that FGFRs are expressed in Pax6-positive/Tbr2-negative cells, but not in Tbr2-positive cells in the OSVZ (Figure 2). In addition, we also examined apoptosis using anti-cleaved caspase-3 antibody, but did not find significant effects of sFGFR3 on the number of cleaved caspase-3-positive cells (VZ: Control, 0.27 ± 0.07 per 1000 μm²; sFGFR3, 0.42 ± 0.23 per 1000 μm²; p=0.23; Student's t-test) (ISVZ: Control, 0.22 ± 0.06 per 1000 μm²; sFGFR3, 0.30 ± 0.06 per 1000 μm²; p=0.13; Student's t-test) (OSVZ: Control, 0.019 ± 0.005 per 1000 μm²; sFGFR3, 0.013 ± 0.010 per 1000 μm²; p=0.25; Student's t-test) (Figure 3—figure supplement 2G,H, Supplementary file 1), suggesting that FGF signaling is irrelevant to apoptosis.

Figure 4

Download asset Open asset

We examined whether the layer structure of the cortex was affected by sFGFR3 expression. Immunohistochemistry for FoxP2 and Ctip2, which are expressed in layers 5 and 6 (Figure 5A,B), demonstrated normal layer structure in the cerebral cortex electroporated with sFGFR3. Consistently, in situ hybridization for Rorb and Cux1, which are expressed in layer 4 and layers 2–4, respectively, showed the layer structure of the cerebral cortex was not affected by expressing sFGFR3 (Figure 5C,D). The locations of GFP-positive cells were not affected by sFGFR3; GFP-positive cells were distributed in layers 4 and 5 in both control and sFGFR3-transfected cortex (Figure 5—figure supplement 1). In addition, the locations of GFP-positive migrating cells and layer formation of the cerebral cortex during development seemed to be unaffected by sFGFR3 (Figure 5—figure supplement 2).

Figure 5 with 2 supplements see all

Download asset Open asset

Finally, we quantified the thickness of cortical layers. Because it has been hypothesized that evolutional expansion of upper layers in gyrencephalic mammals resulted from an increase in the number of oRG cells (Borrell and Götz, 2014; Lui et al., 2011; Lukaszewicz et al., 2005; Nowakowski et al., 2016), it seemed plausible that the reduction of Pax6-positive/Tbr2-negative oRG cells by sFGFR3 led to preferential reduction of upper layers. Consistently, we found that the thickness of layer 2/3 was significantly reduced by sFGFR3 (Control, 0.97 ± 0.06; sFGFR3, 0.77 ± 0.06; p=0.017; Student's t-test) (Figure 5E, Supplementary file 1). In contrast, the thicknesses of layer 4, layer 5 and layer 6 were not affected by sFGFR3 (Layer 4: Control, 0.99 ± 0.05; sFGFR3, 0.96 ± 0.06; p=0.34; Student's t-test) (Layer 5: Control, 0.95 ± 0.03; sFGFR3, 1.12 ± 0.13; p=0.078; Student's t-test) (Layer 6: Control, 1.08 ± 0.13; sFGFR3, 1.00 ± 0.08; p=0.26; Student's t-test) (Figure 5F–H, Supplementary file 1). It is worth noting that GFP-positive transfected neurons were distributed in layers 4 and 5, but there were almost no GFP-positive cells in layer 2/3 (Figure 5—figure supplement 1), supporting our results that sFGFR3 exhibits a non-cell-autonomous effect. These findings suggest that FGF signaling affects the ratio between upper and lower layers of the cerebral cortex, and that the increased upper/lower ratio results in cortical folding. The most conceivable scenario would be that FGF increases OSVZ progenitors and preferentially increases upper layer neurons, generating the protrusion of cortical folds (Figure 5I).

Although the mechanisms underlying the formation of cortical folds have been of great interest, they had remained largely unknown. Here we have shown that regional differences in FGFR expression exist in the ferret cerebral cortex even before cortical folds are formed. Using our IUE technique, we uncovered that FGF signaling is required for cortical folding and the proliferation of OSVZ progenitors. Furthermore, FGF signaling preferentially affects the thickness of upper layers. We previously demonstrated that FGF overexpression was sufficient to increase cortical folds and OSVZ progenitors (Masuda et al., 2015). Furthermore, FGF overexpression preferentially increased the thickness of layer 2/3 (Masuda et al., 2015). Taken together, our findings indicate that FGF signaling is crucial for cortical folding in gyrencephalic mammals and is a pivotal upstream regulator of the production of OSVZ progenitors. FGF signaling is the first signaling pathway found to regulate cortical folding.

Our results showed that sFGFR3 inhibited both cortical folding and the production of SVZ progenitors. Our previous study showed that activation of FGF signaling resulted in both additional cortical folds and increased SVZ progenitors (Masuda et al., 2015). These correlations between cortical folding and SVZ progenitors raised the possibility that SVZ progenitors are responsible for generating cortical folds. Consistently, previous pioneering studies using monkeys reported that even before cortical folds are formed during development, SVZ progenitors are abundant in the cortical regions where the gyri will be formed (Smart et al., 2002). Recently, we directly tested this possibility and uncovered that the formation of cortical folds was inhibited when SVZ progenitors were reduced by expressing dominant-negative Tbr2 transcription factor (DN-Tbr2) (Toda et al., 2016). Thus, these results suggest that SVZ progenitors are indeed crucial for cortical folding during development.

It has been hypothesized that the ratio between upper layers and lower layers of the cerebral cortex is crucial for cortical folding (Richman et al., 1975; Kriegstein et al., 2006). A preferential increase in upper layers relative to lower layers is supposed to produce the protrusion (i.e. the gyrus) of the cortex. Consistent with this hypothesis, current and previous our findings demonstrated that sFGFR3 and DN-Tbr2 preferentially reduced upper layers where cortical folding was impaired (Toda et al., 2016). Furthermore, when FGF signaling was enhanced, upper layers were predominantly increased and additional gyri were formed (Masuda et al., 2015).

Besides the ratio between upper layers and lower layers mentioned above, several hypotheses have been proposed to explain the mechanisms of cortical folding (Lewitus et al., 2013; Sun and Hevner, 2014). It was previously proposed that radial fibers provided by oRG cells are involved in gyrification (Reillo et al., 2011). In prospective gyral regions, oRG cells provide oblique radial fibers for conical neuronal migration, which could contribute to the expansion of the cortical surface. Because our results showed that the numbers of oRG cells were reduced by sFGFR3, reduction of oRG fibers could also be involved in the effect of sFGFR3. Another hypothesis is that the tension of nerve fibers plays a key role in gyrification (Van Essen, 1997). This also seemed possible because axonal trajectories could also be affected by sFGFR3. Future investigations would be necessary in order to gain a complete understanding of the mechanisms underlying cortical folding.

Our findings demonstrated that the protrusion of the gyrus was reduced by sFGFR3, whereas the pattern of the gyrus was preserved. These results suggest that FGF signaling is involved in making the protrusion of the gyri rather than regulating the positions of the gyri. Consistently, the overall gyral patterns seemed to be unaffected by FGF8, although small additional gyri appeared (Masuda et al., 2015). On the other hand, the role of FGF signaling in determining the number of the gyri is unclear. sFGFR3 did not seem to affect the number of the gyri, though FGF8 increased it, leading to polymicrogyria (Masuda et al., 2015). Because a previous study reported that an increase in the number of neurons resulted in additional folding (Nonaka-Kinoshita et al., 2013), the increase in neurons induced by FGF8 seemed to be sufficient for producing additional gyri.

Our findings indicate that FGF signaling regulates cortical folding and the proliferation of OSVZ progenitors. In contrast to OSVZ progenitors, the proliferation of RG cells in the VZ was not affected by inhibition of FGF signaling, even though FGFR is expressed in the VZ. Previous studies have demonstrated that several ligands (e.g. growth factors and morphogens) other than FGF, such as BMP7, promoted the proliferation of RG cells (Segklia et al., 2012). Therefore, it seemed plausible that FGFs and other growth factors act redundantly to promote the proliferation of RG cells in the VZ. It should be noted that although our findings indicate that FGF signaling regulates the proliferation of OSVZ progenitors, it remains possible that FGF also regulates the differentiation of RG cells into SVZ progenitors because FGFRs are also expressed in RG cells. Future investigations would be necessary for addressing these points.

Although our findings clearly uncovered a crucial role of FGF signaling in cortical folding, currently it is unknown which types of FGFR and FGF are responsible for cortical folding. Our in situ hybridization showed that FGFR1, 2 and 3 were expressed in the developing ferret cortex. Because similar downstream signaling pathways are activated by FGFR1, 2 and 3 (Iwata and Hevner, 2009), it seems likely that they work redundantly. Similarly, because several FGFs are expressed in the developing mouse cortex (Allen Brain Atlas), they may work cooperatively.

Although several hypotheses about the mechanisms of cortical folding have been proposed (Kelava et al., 2013; Fernández et al., 2016), it had been difficult to test these hypotheses directly through experiments. This was mainly because efficient genetic manipulation techniques that can be used for gyrencephalic mammals were missing. Taking advantage of our IUE technique for ferrets, here we uncovered the crucial role of FGF signaling in cortical folding. Using the same technique, we previously reported that SVZ progenitors were indispensable for cortical folding (Toda et al., 2016), and produced polymicrogyria model ferrets (Masuda et al., 2015). Ferrets are clearly an important option for investigating the mechanisms of cortical folding.

Ferrets have several important advantages. First, anatomical and electrophysiological information about the ferret brain is available because the ferret brain has been widely used for anatomical and electrophysiological research. The structures of cortical gyri and sulci are well described (Smart and McSherry, 1986b). Second, the processes of the formation of cortical gyri during development have been reported (Smart and McSherry, 1986b; Smart and McSherry, 1986a; Neal et al., 2007). For example, when ferret babies are born, their brains are lissencephalic, and cortical folds are formed after birth. Therefore, the mechanisms of cortical folding can be analyzed using neonatal ferrets rather than embryos. Third, usually more than six ferret babies are born from one pregnant mother. This large number of babies per pregnant mother facilitates examination under various experimental conditions and the ability to obtain a sufficient number of experimental samples. Finally, because our IUE technique is now available, genetically manipulated ferret brains can be obtained within a few weeks. Uncovering the molecular mechanisms underlying the formation of gyrencephalic brains using ferrets would help lead to our ultimate goal of understanding the human brain and its diseases.

Normally pigmented, sable ferrets (Mustela putorius furo) were purchased from Marshall Farms (North Rose, NY). Ferrets were maintained as described previously (Kawasaki et al., 2004; Iwai and Kawasaki, 2009; Iwai et al., 2013). ICR mice were purchased from SLC (Hamamatsu, Japan) and were reared on a normal 12 hr light/dark schedule. The day of conception and that of birth were counted as embryonic day 0 (E0) and postnatal day 0 (P0), respectively. All procedures were approved by the Animal Care Committee of Kanazawa University. Experiments were repeated at least three times and gave consistent results.

The IUE procedure to express transgenes in the ferret brain was described previously (Kawasaki et al., 2012, 2013). Briefly, pregnant ferrets at E33 were anesthetized, and their body temperature was monitored and maintained using a heating pad. The uterine horns were exposed and kept wet by adding drops of PBS intermittently. The location of embryos was visualized with transmitted light delivered through an optical fiber cable. The pigmented iris was visible, and this enabled us to assume the location of the lateral ventricle. Approximately 2–5 µl of DNA solution was injected into the lateral ventricle at the indicated ages using a pulled glass micropipette. Each embryo within the uterus was placed between tweezer-type electrodes with a diameter of 5 mm (CUY650-P5; NEPA Gene, Japan). Square electric pulses (50–100 V, 50 ms) were passed 5 times at 1 s intervals using an electroporator (ECM830, BTX, Holliston, MA). The wall and skin of the abdominal cavity were sutured, and the embryos were allowed to develop normally.

pCAG-EGFP was described previously (Sehara et al., 2010). pCAG-sFGFR3 was made in accordance with a previous report (Fukuchi-Shimogori and Grove, 2001). Briefly, the extracellular domain of human FGFR3c was amplified using the following primers and was subcloned into pCAG plasmid vector, yielding pCAG-sFGFR3. Forward, CCATGGGCGCCCCTGCC; reverse, GGAGCCCAGGCCTTTCTT. Plasmids were purified using the Endofree Plasmid Maxi Kit (Qiagen, Valencia, CA). For co-transfection, a mixture of pCAG-EGFP plus either pCAG-sFGFR3 or pCAG control plasmid was used. Prior to IUE procedures, plasmid DNA was diluted to 2.5–6.0 mg/ml in PBS, and Fast Green solution was added to a final concentration of 0.5% to monitor the injection.

Preparation of sections was performed as described previously with slight modifications (Toda et al., 2013; Kawasaki et al., 2000). Briefly, ferrets or mice were deeply anesthetized with pentobarbital and transcardially perfused with 4% paraformaldehyde (PFA), and the brains were dissected. Then, the brains were cryoprotected by three-days immersion in 30% sucrose and embedded in OCT compound. Sections of 18 or 50 µm thickness were prepared using a cryostat.

Epifluorescence microscopy and confocal microscopy were performed with a BIOREVO BZ-9000 (Keyence, Japan) and a FLUOVIEW FV10i (Olympus, Japan), respectively.

Immunohistochemistry was performed as described previously with slight modifications (Toda et al., 2013; Kawasaki et al., 2000). Coronal sections were permeabilized with 0.3% Triton X-100/PBS and incubated overnight with primary antibodies, which included anti-Tbr2 (Abcam, UK, RRID: AB_778267), anti-Pax6 (Covance, Princeton, NJ, RRID: AB_291612), anti-Ki-67 (Leica, Germany, RRID: AB_442102), anti-phospho-histone H3 (Millipore, Billerica, MA, RRID: AB_310016), anti-phosphorylated vimentin (Medical and Biological Laboratories, Japan, RRID: AB_592963), anti-cleaved caspase 3 (BD Pharmingen, San Diego, CA, RRID: AB_397274), anti-Ctip2 (Abcam, UK, RRID: AB_2064130), anti-FOXP2 (Atlas antibodies, Sweden, RRID: AB_1078908), anti-GFAP (Sigma-Aldrich, St. Louis, MO, RRID: AB_477010) and anti-GFP antibodies (Nacalai tesque, Japan, RRID: AB_2313652; Medical and Biological Laboratories, Japan, RRID: AB_591819). After incubation with secondary antibodies and Hoechst 33342, the sections were washed and mounted.

For triple immunostaining, after double immunostaining was performed as described above, the sections were incubated with biotin-conjugated anti-phospho-histone H3 antibody (Millipore, Billerica, MA, RRID: AB_310794) and subsequently with fluorescent-dye conjugated streptavidin.

In situ hybridization was performed as described previously (Matsumoto et al., 2017). Sections prepared from fresh-frozen or fixed tissues were treated with 4% PFA for 10 min, 1 µg/ml proteinase K for 10 min and 0.25% acetic anhydride for 10 min. After prehybridization, the sections were incubated overnight at 58 ˚C with digoxigenin-labeled RNA probes diluted in hybridization buffer (50% formamide, 5x SSC, 5x Denhardt's solution, 0.3 mg/ml yeast RNA, 0.1 mg/ml herring sperm DNA, and 1 mM DTT). The sections were then incubated with alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche, Indianapolis, IN, RRID: AB_514497) and Hoechst 33342, and were visualized using NBT/BCIP as substrates.

For combination of in situ hybridization and immunostaining, after hybridization was performed, the sections were incubated with anti-Pax6 (Abcam, UK, RRID: AB_1566562), anti-Tbr2, anti-GFP (Medical and Biological Laboratories, Japan, RRID: AB_591819) and alkaline phosphatase-conjugated anti-digoxigenin antibodies. After being incubated with secondary antibodies, in situ signals were visualized with NBT/BCIP. Probes used here were as follows. Ferret FGFR2 and FGFR3 probes were described previously (Masuda et al., 2015). Two ferret FGFR1 cDNA fragments were amplified by RT-PCR and inserted into the pCRII vector. The sequences of primers used to amplify ferret FGFR1 cDNA fragments were as follows: forward1, GGAGCTGGAAGTGCCTCCTCTTCTG; reverse1, TGATGCGGGTACGGTTGCTT; forward2, CAGGGGAGGAGGTGGAGGTG; reverse2 CGGCACCGCATGCAATTTCTTTTCCATC. A mixture of two probes made from these two FGFR1 cDNA fragments were used as the FGFR1 probe. Ferret Sprouty2, mouse Fgfr1, mouse Fgfr2 and mouse Fgfr3 cDNA fragments were amplified using RT-PCR and inserted into the pCRII vector. The sequences of primers used to amplify ferret Sprouty2, mouse Fgfr1, mouse Fgfr2 and mouse Fgfr3 cDNA fragments were as follows: ferret Sprouty2, forward ATCGCAGGAAGACGAGAATCCAAGG, reverse CTGGGTGGGACAGTGGGAACTTTGC; mouse Fgfr1, forward CTGCATGGTTGACCGTTCTGGAAGC, reverse TGTAGATCCGGTCAAACAACGCCTC; mouse Fgfr2, forward GGAAGGAGTTTAAGCAGGAGCATCG, reverse CGATTCCCACTGCTTCAGCCATGAC; mouse Fgfr3, forward GAAAGTGTGGTACCCTCCGATCGTG, reverse GTCCAAAGCAGCCTTCTCCAAGAGG. The ferret Cux1 and Rorb probes were described previously (Matsumoto et al., 2017; Rowell et al., 2010; Shinmyo et al., 2017).

RT-PCR was performed as described previously with modification (Kawasaki et al., 2002). Total RNA was isolated from the P0 ferret cerebral cortex using the RNeasy Mini Kit (Qiagen, Valencia, CA). Reverse transcription was performed using oligo(dT)12–18 (Thermo Fisher Scientific, Waltham, MA) and Superscript III (Thermo Fisher Scientific, Waltham, MA). Samples without Superscript III were also made as negative controls. PCR reaction was performed using the following primers: FGFR1-forward, GGAGCTGGAAGTGCCTCCTCTTCTG; FGFR1-reverse, CGGCACCGCATGCAATTTCTTTTCCATC; FGFR2-forward, AGAGATAAGCTGACGCTGGGCAAACC; FGFR2-reverse, GAGGAAGGCAGGGTTCGTAAGGC; FGFR3-forward, GAGGCTAAATTACGGGTACCTGA; FGFR3-reverse, GAGAACAAAGACCACCCTGAAC. PCR products made using these primers were confirmed by DNA sequencing.

Coronal sections were stained with anti-Pax6, anti-Tbr2, anti-Ki-67, anti-pHH3, anti-pVim and anti-cleaved caspase 3 antibodies and Hoechst 33342. After background signals were removed using the ‘threshold’ tool of ImageJ software (Default or MaxEntropy option), the numbers of immunopositive cells in GFP-positive areas were manually counted using the ‘cell counter’ tool and were divided by the total area of the ROI to calculate cell densities. To calculate the numbers of Pax6-positive and Tbr2-positive cells, coronal sections containing the striatum were used. The densities of Pax6-positive and Tbr2-positive cells in each germinal layer in dorsolateral regions of the cerebral cortex were multiplied by its thickness. One region per each animal was used. The thickness of each germinal layer was measured as follows using three sections for each animal: the area of each germinal layer was divided by its tangential length. The cell-dense layer next to the VZ was identified as the ISVZ and the cell-sparse layer between the ISVZ and the IZ was identified as the OSVZ. In addition, the border between the ISVZ and the OSVZ was also identified by the presence of GFP-positive fibers (Kawasaki et al., 2013).

Serial coronal sections containing the suprasylvian gyrus (SSG) were prepared from electroporated brains. One section per every 100 µm was selected. The selected three serial sections containing the SSG for each brain were stained with Hoechst 33342, and images of whole sections were acquired using a BZ-9000 microscope (Keyence, Japan). The averages of the local GS, local SD and local GI ratios were calculated using the three sections for each animal.

To calculate the local GS ratio, the area surrounded by the brain surface (Figure 1—figure supplement 2, green line) and the red line connecting the bottom of the suprasylvian sulcus (SSS) and that of the lateral sulcus (LS) was measured (local GS value) (Figure 1—figure supplement 2). To minimize the variation of the local GS values depending on the positions of coronal sections in the brain, the local GS value on the electroporated side and that on the contralateral non-electroporated side of the cerebral cortex in the same brain sections were measured, and the former was divided by the latter (local GS ratio). The local GS ratio would be 1 if the size of SSG was the same between the electroporated side and the other side (i.e. non-electroporated side), and would be smaller than 1 if cortical folding was suppressed by genetic manipulation.

To calculate the local SD ratio, a line connecting the top of the SSG and the top of the middle ectosylvian gyrus (MEG) was drawn (Figure 1—figure supplement 3, red line). Then a green line connecting the bottom of the suprasylvian sulcus (SSS) and the red line was drawn, so that the green line was perpendicular to the red line (Figure 1—figure supplement 3). The length of the green line was used as the local SD value. To minimize the variation of the local SD values depending on the positions of coronal sections in the brain, the local SD value on the electroporated side and that on the contralateral non-electroporated side of the cerebral cortex in the same brain sections were measured, and the former was divided by the latter (local SD ratio). The local SD ratio would be 1 if the depth of the SSS was the same between the electroporated side and the other side (i.e. non-electroporated side), and would be smaller than 1 if the depth of the SSS was reduced by genetic manipulation.

To calculate the local GI ratio, a line connecting the top of the SSG and that of the MEG was drawn (Figure 1—figure supplement 4, red line). The length of the complete contour between the top of the SSG and that of the MEG (Figure 1—figure supplement 4, green line) was then divided by that of the red line (local GI value). To minimize the variation of the local GI values depending on the positions of coronal sections in the brain, the local GI value on the electroporated side and that on the contralateral non-electroporated side of the cerebral cortex in the same brain sections were measured, and the former was divided by the latter (local GI ratio). The local GI ratio would be 1 if cortical folding was the same between the electroporated side and the other side (i.e. non-electroporated side), and would be smaller than 1 if the cortical folding was suppressed by genetic manipulation.

Coronal sections containing the SSG were used for quantification. The sections were subjected to Hoechst 33342 staining plus either in situ hybridization for Rorb or immunohistochemistry for Ctip2, and images of whole sections were acquired using a BZ-9000 microscope. To calculate the thickness of each cortical layer in the SSG, the thicknesses of layer 2/3 and layer 4 in the SSG, which were identified using Hoechst and Rorb signals, and those of layer 5 and layer 6 in the SSG, which were identified using Hoechst and Ctip2 signals, were measured, and were subsequently divided by the tangential lengths of these areas. To minimize the variation of the thickness value depending on the positions of coronal sections in the brain, the thickness value on the electroporated side and that on the contralateral non-electroporated side of the cerebral cortex in the same brain sections were measured, and the former was divided by the latter (thickness ratio). The thickness ratio would be 1 if the thickness was the same between the electroporated side and the other side (i.e. non-electroporated side), and would be smaller than 1 if the thickness was reduced by genetic manipulation.

Coronal sections prepared at E40 were subjected to in situ hybridization for FGFR2 and FGFR3 and Hoechst 33342 staining, and images of whole sections were acquired using a BZ-9000 microscope (Keyence, Japan). The borders of the OSVZ were determined using Hoechst images. After background signals were subtracted, total FGFR signal intensities per area were measured. The entire OSVZ regions were selected and straightened using the ‘Straighten’ function of ImageJ. Signal intensities along the tangential axis were then measured using the ‘Plot Profile’ of ImageJ.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We apologize to all of the researchers whose work could not be cited due to space limitation. We are grateful to Drs. Eisuke Nishida (Kyoto University), the late Yoshiki Sasai and Shigetada Nakanishi (Suntory Foundation for Life Science) for their continuous encouragement. We thank Dr. Clifton W Ragsdale (University of Chicago) for plasmids, and Zachary Blalock, Dr. Tomohisa Toda, Dr. Yoshio Hoshiba, Miwako Hirota, Yuki Nishita and Kawasaki lab members for their helpful support. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Life Science Foundation of Japan Senri Life Science Foundation, the Uehara Memorial Foundation and Takeda Science Foundation.

Animal experimentation: All procedures were approved by the Animal Care Committee of Kanazawa University.

© 2017, Matsumoto et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.