Testis development requires the repression of Wnt4 by Fgf signaling

Abstract

The bipotential gonad expresses genes associated with both the male and female pathways. Adoption of the male testicular fate is associated with the repression of many female genes including Wnt4. However, the importance of repression of Wnt4 to the establishment of male development was not previously determined. Deletion of either Fgf9 or Fgfr2 in an XY gonad resulted in up-regulation of Wnt4 and male-to-female sex reversal. We investigated whether the deletion if Wnt4 could rescue sex reversal in Fgf9 and Fgfr2 mutants. XY Fgf9/Wnt4 and Fgfr2/Wnt4 double mutants developed testes with male somatic and germ cells present, suggesting that the primary role of Fgf signaling is the repression of female-promoting genes. Thus, the decision to adopt the male fate is based not only on whether male genes, such as Sox9, are expressed, but also on the active repression of female genes, such as Wnt4. Because loss of Wnt4 results in the up-regulation of Fgf9, we also tested the possibility that derepression of Fgf9 was responsible for the aspects of male development observed in XX Wnt4 mutants. However, we found that the relationship between these two signaling factors is not symmetric: loss of Fgf9 in XX Wnt4(-/-) gonads does not rescue their partial female-to-male sex-reversal.

Fig. 1

Wnt4, Rspo1 and Foxl2 transcript levels are indistinguishable from XX gonads, whereas Sox9 levels are somewhat lower in XY Fgf9^−/− gonads. ((A)–(D)) qRT-PCR data from XY control (cyan), XY Fgf9^−/− (green), and XX control (magenta) gonads at E11.5, E11.75, and E12.5. n=3 or 4. Mean normalized expression (as described in the methods) is shown, using Canx as the normalizing gene. A significant difference (as determined by a T-test, p-value <0.05) between XX and XY samples is indicated by a black asterisk over the sex with lower expression. A significant difference between XY Fgf9^−/− and XY control samples is indicated by a cyan asterisk, while a significant difference between the XY Fgf9^−/− and XX control samples is indicated by a magenta asterisk. (A) Sox9 was expressed at an intermediate level at E12.5 in XY Fgf9^−/− gonads when compared to both XX and XY controls, even though Sox9 protein was lost by this stage (Kim et al., 2006). While there was no significant difference in expression of Sox9 between XY control and Fgf9^−/− gonads at E11.5 and E11.75, the average expression of Sox9 in XY Fgf9 mutants was lower than that in XY controls, indicating that Fgf9 could play a role in up-regulating Sox9 or fewer Sox9-positive cells may be present. ((B)–(D)) For female associated genes, no significant difference was observed at any time point between the XY Fgf9^−/− and XX control gonads. While the differences between XX and XY control gonads were not always significant before E12.5, elevated expression of these female-associated genes was apparent in both the XX control and XY Fgf9^−/− gonads when compared to the XY control gonads. Thus, loss of Fgf9 resulted in the over-expression of multiple female genes.

Fig. 2. Deleting Wnt4 rescued the partial sex reversal induced by loss of Fgfr2

Immunofluorescence of E13.5 gonads stained with PECAM1 (labeling germ cells and vasculature, blue), SOX9 (red), and FOXL2 (green). XX controls (A) had no testis cords and expressed FOXL2. In contrast, XY controls (B) and XY Wnt4 mutants (C) had SOX9-positive Sertoli cells surrounding the germ cells in testis cords with rare FOXL2-positive cells (arrowheads). (D) XY mice with Fgfr2 deleted in the supporting cells were partially sex reversed. Cells at the poles of the gonad (arrows) expressed FOXL2 and cords were absent, as in the female control. The center of the gonad was more testicular with SOX9-positive cells and testis cord structure. ((E) and (F)) Loss of one or two alleles of Wnt4 substantially rescued the Fgfr2 phenotype. Particularly in the XY Sf1-Cre; Fgfr2^flox/flox; Wnt4^−/− samples (F), there were testis cords throughout the gonad with SOX9 and few FOXL2-positive cells (arrowheads). Images were captured using a 40X objective, assembled to span the entire gonad, and placed on a black background. Scale bar=100 μm. n≥3 of each genotype.

Fig. 3. Deleting Wnt4 rescued the sex reversal induced by loss of Fgf9, further supporting the Wnt4 repression model

Brightfield images of E16.5–E17.5 gonads with mesonephric-derived structures. Control XX ovaries (A) and XY testes (B) had distinct morphologies. Dotted lines surround the ovaries. (C) XY gonads carrying a deletion of Wnt4 developed as testes, but (D) the deletion of Fgf9 resulted in sex reversal and development of an ovary, even with the deletion of one copy of Wnt4. (E) Deleting both Wnt4 and Fgf9 rescued testis development. Scale bar=1 mm. n≥3 of each genotype.

Fig. 4

XY Fgf9/Wnt4 double mutants express molecular markers typical of testis development consistent with a full rescue of sex reversal. Immunostaining of E16.5–E17.5 gonads. DNA is shown in blue. ((A)–(E)) Gonads labeled with the Sertoli-cell markers SOX9 (red) and AMH (green). Sertoli cells labeled with SOX9 and AMH were absent in XX controls (A), but present in XY controls (B) and XY Wnt4 mutants (C). Similar to XX controls, sex reversed XY Fgf9 mutants (D) lacked differentiating Sertoli cells, although a variable number of SOX9 positive cells were observed that did not express AMH (arrows). Sertoli cells were rescued by deleting Wnt4 in XY Fgf9 mutants (E). ((F)–(J)) Gonads labeled with the germ cell marker CDH1 (red) and the Leydig cell marker 3β-HSD (green). In the XX control (F), germ cells were present but not clustered into testis cords and Leydig cells were absent. In the XY control (G) and XY Wnt4 mutant (H), germ cells were clustered into testis cords with Leydig cells positive for 3β-HSD outside of testis cords. Sex reversed XY Fgf9 mutants (I) resembled XX controls, but deleting Wnt4 in an XY Fgf9 mutant (J) rescued Leydig cells and the presence of germ cells in testis cords. The arrowhead indicates occasional background staining with the 3β-HSD antibody in the testis cords. Thus, deleting Wnt4 rescued the Fgf9 phenotype in the XY samples based on the detection of molecular markers characteristic of the prominent testicular cell types. All images shown are maximum intensity projections of 12 images over approximately 12 μm using a 40X objective. For XY Wnt4 ((C) and (H)) and Fgf9 ((D) and (I)) single mutants, the other locus was either homozygous wild type or heterozygous (designated “Fgf9⁺” and “Wnt4⁺”). Scale bars=100 μm. n≥3 of each genotype.

Fig. 5. Germ cells in the XY Fgf9/Wnt4 double mutants adopt the male fate

Immunostaining of E16.5–E17.5 gonads. DNA is shown in blue. ((A)–(E)) Gonads labeled with the meiosis marker SCP3 (red) that is nuclear in meiotic germ cells. Meiotic germ cells were present in XX controls (A), absent in XY controls (B) and XY Wnt4 mutants (C), and present in sex reversed XY Fgf9 mutants (D). The germ cells in the Fgf9/Wnt4 double mutants were not in meiosis (E). Images are maximum intensity projections of 12 images over approximately 12 μm using a 40X objective. Scale bar=100 μm. n=3 of each genotype. ((F)–(J)) Gonads labeled with the male germ cell marker, DNMT3L (red). In E16.5 gonads, DNMT3L is absent in XX control gonads (F), but present in XY controls (G) and XY Wnt4 single mutants (H). Germ cells in Fgf9 mutants (I) do not express DNMT3L, consistent with their failure to enter the male pathway. However, DNMT3L expression is detected in Fgf9/Wnt4 double mutants, suggesting rescue of the male pathway (J). Images are maximum intensity projections of 6 images over approximately 6 μm using a 40X objective. Scale bar=100 μm. n=3 of each genotype. For XY Wnt4 ((C) and (H)) and Fgf9 ((D) and (I)) single mutants, the other locus was either homozygous wild type or heterozygous (designated “Fgf9⁺” and “Wnt4⁺”).

Fig. 6. Deleting Fgf9 did not rescue the phenotypes of XX Wnt4 mutants

DNA is shown in blue. ((A), (C), (E) and (G)) PECAM1 (green) labels both the endothelial cells and the germ cells, but with varying intensity in the germ cells at E16.5–E17.5. ((B), (D), (F) and (H)) E16.5–E17.5 gonads labeled with the steroidogenic marker 3β-HSD (red). Unlike XX control ovaries ((A) and (B)), XX Wnt4 mutants developed an ectopic coelomic vessel ((C), arrows) and steroidogenic cells (D). ((E) and (F)) In the presence of one or more functional copies of Wnt4, Fgf9 mutants did not develop this ectopic vessel or steroidogenic cells. However, both the ectopic vessel ((G), arrows) and steroidogenic cells (H) were present in the XX Wnt4/Fgf9 double mutants. The images shown are maximum intensity projections of 5 images over approximately 10 μm ((A), (C), (E) and (G)) or 10 images over approximately 30 μm ((B), (D), (F) and (H)) taken using a 20X objective. For XX Wnt4 ((C) and (D)) and Fgf9 ((E) and (F)) single mutants, the other locus was homozygous wild type or heterozygous (designated “Fgf9⁺” and “Wnt4⁺”). Scale bar=100 μm. n=3 of each genotype.

Fig. 7. Model of sex determination

A model of the genetic interactions during early gonad development. SRY up-regulates SOX9 in XY gonads, and SOX9 becomes self-regulating (Sekido and Lovell-Badge, 2008). SOX9 up-regulates Fgf9 (Kim et al., 2006) to repress Wnt4, and possibly also reinforce its own expression. A failure to repress Wnt4 (in Fgf9 or Fgfr2 mutants) results in repression of SOX9 by Wnt4 signaling and loss of male development.