FGF15 promotes neurogenesis and opposes FGF8 function during neocortical development

Abstract

Background: Growth, differentiation and regional specification of telencephalic domains, such as the cerebral cortex, are regulated by the interplay of secreted proteins produced by patterning centers and signal transduction systems deployed in the surrounding neuroepithelium. Among other signaling molecules, members of the fibroblast growth factor (FGF) family have a prominent role in regulating growth, differentiation and regional specification. In the mouse telencephalon the rostral patterning center expresses members of the Fgf family (Fgf8, Fgf15, Fgf17, Fgf18). FGF8 and FGF17 signaling have major roles in specification and morphogenesis of the rostroventral telencephalon, whereas the functions of FGF15 and FGF18 in the rostral patterning center have not been established.

Results: Using Fgf15-/- mutant mice, we provide evidence that FGF15 suppresses proliferation, and that it promotes differentiation, expression of CoupTF1 and caudoventral fate; thus, reducing Fgf15 and Fgf8 dosage have opposite effects. Furthermore, we show that FGF15 and FGF8 differentially phosphorylate ERK (p42/44), AKT and S6 in cultures of embryonic cortex. Finally, we show that FGF15 inhibits proliferation in cortical cultures.

Conclusion: FGF15 and FGF8 have distinct signaling properties, and opposite effects on neocortical patterning and differentiation; FGF15 promotes CoupTF1 expression, represses proliferation and promotes neural differentiation.

Figure 1

In situRNA hybridization on horizontal sections in wild-type,Fgf8^Null/Neoand Fgf15^-/-E9.5 embryos. (a, b) Fgf15 (a) and Fgf8 (b) expression in adjacent sections of a wild-type embryo. (c-d') Fgf15 expression in wild-type (c) and Fgf8^Null/Neo embryos (d). Higher magnifications of (c, d) are shown in (c', d'). (e-n) Expression analysis in wild-type and Fgf15 mutant embryos showing adjacent sections of CoupTF1 (e, f), Fgf8 (g, h), Spry2 (i, j), Erm (k, l) and cMyc (m, n). Arrowheads in (a, b): rostral midline that is Fgf8⁺ and Fgf15^-; the midline becomes Fgf15⁺ in Fgf8^Null/Neo (d, d'). Arrow in (d): loss of Fgf15 expression in the midbrain of the Fgf8^Null/Neo mutant. Arrows in (e, f): reduced CoupTF1in Fgf15^-/-. Bar in (a) is 200 μm.

Figure 2

Analysis of the cortical patterning defects in the Fgf15 mutants. (a-r) In situ RNA hybridization on coronal sections of Fgf15^+/+ (left column) and Fgf15^-/- (right column) embryos at E12.5 (a-l) and E14.5 (m-r). CoupTF1 (a, b, m, n), Mest (c, d), Sp8 (e, f), Pax6 (g, h), Emx2 (i, j, o, p) and Erm (k, l, q, r). Arrows highlight the changes in the extent and/or intensity of expression. Bar in (a, m) is 200 μm.

Figure 3

Analysis of the cortical neurogenesis defects in the Fgf15 mutants. (a-h) In situ RNA hybridization (a, b, e-h) and immunofluorescence (c, d) on coronal sections of Fgf15^+/+ (left column) and Fgf15^-/- (right column) embryos at E14.5. Ngn2 (a, b), β-III-Tubulin (c, d), Tbr1 (e, f), Mef2C (g, h). Arrows point out the reduced thickness of the cortical plate expression of Tbr1 and Mef2C. Brackets in (a-d) highlight the reduced thickness of the cortical plate. CP, cortical plate; SVZ, subventricular zone; VZ, ventricular zone. Bar in (a) is 200 μm.

Figure 4

Analysis of proliferation and cell cycle in the Fgf15 mutants. (a-i) Comparison of the rate of proliferation of the neuronal progenitors at E14.5 in coronal hemisections from Fgf15^+/+ (left column) and Fgf15^-/- (right column) embryos. The number of neuronal progenitors undergoing mitosis was evaluated by PH3⁺ immunofluorescence (a, b); quantification of the PH3⁺ cells in the VZ and SVZ is shown in (c) (n = 3; p = 0.0019, Student's t-test, indicated by * and **; see Materials and methods). The analysis of the cell cycle length (d, e) (quantification in (f)) and of the number of neuronal progenitors exiting the cell cycle (Q fraction) (f, g) (quantification in (i)) were performed at E14.5 by double labeling with IdU and BrdU (n = 3; p = 0.001, Student's t-test, indicated by *; see Materials and methods). The rectangles in a-b, d-e, g-h indicate the sampled bins. The error bars indicate the standard deviations. MUT, mutant; SVZ, subventricular zone; VZ, ventricular zone; WT, wild type.

Figure 5

Activation of the Wnt/β-catenin and retinoic acid signaling pathways in the Fgf15 mutants.(a-d) Retinoic acid pathway activation was determined at E12.5 (a, b) and E14.5 (c, d) by β-galactosidase staining of coronal sections of Fgf15^-/-; RL⁺ embryos (RL, RARE LacZ, retinoic acid reporter). (e-h) The Wnt/β-catenin pathway activation was revealed by β-galactosidase staining of coronal sections of Fgf15^-/-; BG⁺ (BG, BATgal, Wnt/β-catenin reporter) at E12.5 (e, f) and E14.5 (g, h). Arrows indicate the regions that show the largest changes in the extent and/or intensity of expression. Three separate cases were analyzed for each genotype. Bar in (a, c) is 200 μm.

Figure 6

Analysis of FGF signaling components in the Fgf15 mutants. (a-l) In situ hybridization on coronal sections: Fgf8 at E12.5 (a, b) and E14.5 (c, d); Spry2 at E12.5 (e, f, i, j) and E14.5 (g, h, k, l). Arrows in (e-h) indicate the change in Spry2 gene expression in the lateral pallium; arrowheads in (e, f) indicate the change in the medial pallium. CP: cortical plate; Se: septum. Bars in (e, g, i, k) are 200 μm.

Figure 7

Activation of the FGF15 and FGF8 downstream cytosolic effectors. Comparison of phosphorylation levels of four proteins that are modified in response to FGF15 and FGF8. E12.5 primary cortical cultures were starved for 24 hours and then treated with recombinant FGF15 or FGF8. Cell lysates were analyzed after 0, 5, 15, 30 and 60 minutes by immunoblottingblotting to detect phosphorylated forms of pERK 42/44 (a, b, top panels), pAKT (a, b, middle panels) and pS6 (c, d, top panels), and pGSK3 (c, d, middle panels). The α-Tubulin (α-Tub) antibody (a-d, bottom panels) was used for normalization (these results are characteristic of what we observed the three times these experiments performed). The numbers under the bands indicate the fold-induction or reduction, with respect to T = 0, and are normalized with respect to the α-Tubulin level. (e-j) Immunofluorescence for β-III-Tubulin (red), phosphohistone-3 (PH3) (green) and Hoechst (blue) on E12.5 primary cortical cultures that were either not treated, treated with 50 ng/ml recombinant FGF8 or FGF15 for 24 and 48 hours (n = 4 for each experiment; p = 0.01, Student's t-test). (k) Graph showing the mitotic index calculated before starting the treatment (T0) and after 24 and 48 hours of treatment with recombinant FGF8 and FGF15. Non-treated cells were used as a control. The mitotic index was calculated by dividing the number of PH3⁺ cells with the total number of cells (Hoechst labeled nuclei). Bars in the graph represent the standard deviation. The average number of nuclei and PH3⁺ cells for each sample/350 μm² were as follow. T0: 113.5 ± 5 nuclei and 6.25 ± 2.5 PH3⁺ cells. At 24 hours: control, 226.75 ± 4 nuclei and 20.5 ± 3.8 PH3⁺ cells; FGF8, 297.5 ± 12.5 nuclei and 48.3 ± 1.5 PH3⁺ cells; FGF15, 113 ± 10 nuclei and 8.25 ± 2.5 PH3⁺ cells. At 48 hours: control, 79.6 ± 3 nuclei and 8.25 ± 1 PH3⁺ cells; FGF8, 164 ± 11 nuclei and 29 ± 3 PH3⁺ cells; FGF15, 44 ± 3 nuclei and 3.4 ± 0.5 PH3⁺ cells.

Figure 8

Model of genetic interactions upstream and downstream of Fgf15 and Fgf8 within embryonic telencephalon. Shh promotes Fgf15 expression (Addition file 2) [6,20,35], and maintains Fgf8 expression [68]. Fgf8 is required for Shh induction [22]. Fgf15 activates, whereas Fgf8 represses, expression of CoupTF1 (among other genes), which represses proliferation and promotes differentiation and caudal fate [37,38].