The apical ectodermal ridge is a timer for generating distal limb progenitors

Abstract

The apical ectodermal ridge (AER) is a transient embryonic structure essential for the induction, patterning and outgrowth of the vertebrate limb. However, the mechanism of AER function in limb skeletal patterning has remained unclear. In this study, we genetically ablated the AER by conditionally removing FGFR2 function and found that distal limb development failed in mutant mice. We showed that FGFR2 promotes survival of AER cells and interacts with Wnt/beta-catenin signaling during AER maintenance. Interestingly, cell proliferation and survival were not significantly reduced in the distal mesenchyme of mutant limb buds. We established Hoxa13 expression as an early marker of distal limb progenitors and discovered a dynamic morphogenetic process of distal limb development. We found that premature AER loss in mutant limb buds delayed generation of autopod progenitors, which in turn failed to reach a threshold number required to form a normal autopod. Taken together, we have uncovered a novel mechanism, whereby the AER regulates the number of autopod progenitors by determining the onset of their generation.

Fig. 1. Fgfr2 removal causes loss of the apical ectodermal ridge (AER), reduction of AER-FGF and loss of distal limb skeletal elements. (A–F)

Skeletal preparations of E18.5 embryos, with cartilage stained blue and bone stained red. In Fgfr2^AER-KO embryos, the hindlimb was missing except for the pelvic bones (arrow in D) and the distal autopod (hand-plate) was absent in the forelimb (asterisk in E, F; n=18). Individual digits in control embryos are numbered I–V, from anterior to posterior (C). Scale bars: 1 mm. (G–N) AER histology as examined by CD44 immunofluorescence (green) on vibratome sections. Samples were counterstained with a nuclear dye To-Pro3 (red). Scale bars: 50 μm. (O–V) Levels of AER-Fgf8 expression as detected by whole-mount RNA in situ hybridization assays at the stages indicated. Fgf8-expressing cells were sparser in the mutant AER (R, R′) than in the control AER (Q, Q′) at the 33 s stage. Black arrowheads indicate gaps in the AER that lack Fgf8 expression in the mutant limb buds (T, T′). Arrow indicates a small patch of the limb bud edge that was still expressing Fgf8 (V). Scale bars: 200 μm. Abbreviations: an, anterior; di, distal; do, dorsal; pr, proximal; po, posterior; ve, ventral; s, somite.

Fig. 2. Altered expression of key mesenchymal patterning genes in Fgfr2^AER-KO forelimb buds

(A–V) Whole-mount RNA in situ hybridization assays for the genes indicated. (A–H) Mkp3 expression as a read-out of mesenchymal responses to AER-FGFs. Vertical white lines in C, D indicate the depth of Mkp3-positive domain with black and white arrowheads marking the anterior and posterior boundaries, respectively. Asterisk indicates the Mkp3 expression domain in the sub-AER mesenchyme that was absent in the mutant limb bud (dotted line, H). (I–N′) Expression of Shh (I–L) and Gre (M–N′). The distal strip of sub-AER mesenchyme (broken white lines), which lacks Gre expression, was much smaller in mutant forelimb buds. (M′, N′) The strip outlined by broken white lines, indicating the Gre-negative middle section of the LBM was greatly reduced in mutant limb buds at E11.5 (N′). (O–V) Expression of genes that primarily mark the anterior (Alx4;O, P), proximal (Meis2; Q, R) and middle mesenchyme (Hoxa11; S, T) and Hoxd11 expression (U, V) at E11.5. There are two distinct domains of Hoxd11 expression, marking the proximal and distal (*) mesenchyme. The distal domain of Hoxd11 expression, which gives rise to the autopod, was missing in the mutant limb (broken lines in V). E11.5 is equivalent to ~the 45 s stage. Scale bars: 200 μm. Abbreviations: an, anterior; di, distal; do, dorsal; pr, proximal; po, posterior; ve, ventral; s, somite.

Fig. 3. Normal cell death and cell proliferation in the distal mesenchyme of Fgfr2^AER-KO forelimb buds

(A–D) Cell death as detected by TUNEL assay (green) at the stages indicated. Note that there were more dying cells in the AER and ventral ectoderm of mutant forelimb buds than in controls at the 30 s (B; n=6) and the 36 s (D; n=6) stages. (E–L) Cell proliferation as detected via BrdU incorporation (E–H′, M) and pH3 immunofluorescence (I–L, N), which marks nuclei in cells undergoing mitosis, in the forelimb buds of mutant and control embryos. White circles indicate the areas (φ=200 μm) of sub-AER mesenchyme in which BrdU- (M) or pH3- (N) positive cells were quantified. Values are the mean±s.d. for each data point in M and the mean±s.e.m. for each data point in N. No statistically significant differences in the percentage of BrdU-positive cells or in the number of pH3-positive cells were observed between the control and mutant limb buds at the stages indicated (unpaired, two-tailed Student’s t-test). E11.5 is equivalent to ~45 s stage. Scale bars: 50 μm. Abbreviations: an, anterior; di, distal; do, dorsal; pr, proximal; po, posterior; ve, ventral; s, somite.

Fig. 4. Expression of Hoxd11, Hoxd13 and Hoxa13 in early mouse limb buds

(A–I) Whole-mount in situ hybridization assays for Hoxd11, Hoxd13 and Hoxa13 expression in normal forelimb buds. (A–C) Hoxd11 expression. Its distal domain (asterisk) in the posterior-distal autopod (B, C) marks the future autopod at E12.5 (C). (D–F) Hoxd13 expression. The proximal domain of Hox13 expression (arrowheads) at the 34–35 s stage was gradually lost over time. (G–I) Hoxa13 expression at the stages indicated. E10.5 is equivalent to ~35 s stage. Scale bars: 200 μm. Abbreviations: an, anterior; di, distal; do, dorsal; pr, proximal; po, posterior; ve, ventral; s, somite.

Fig. 5. Delayed generation of autopod progenitors in Fgfr2^AER-KO forelimbs

(A–T) Whole-mount in situ hybridization assays for Hoxa13, Hoxd13 and Sox9 expression. (A–L) Generation and expansion of autopod progenitors as indicated by Hoxa13 expression. Hoxa13 expression was delayed by 2 s stages in the mutant. The outline of mutant limb buds changed between the 45 s (H) and 47 s (J) stages, coinciding with the onset of autopod condensation (S, T). (M–P) Hoxd13 expression in the distal autopod. Asterisk in N, P indicates residual Hoxd13 expression in mutant limb buds. The proximal domain of Hoxd13 expression (arrowheads) was present in control (M) and mutant limb buds (N) at the 47 s stage. (Q–T) Skeletal progenitors at the 40 s and 46 s stages as marked by Sox9 expression. Skeletal condensation of the autopod, yet to start at the 40 s stage (Q, R), occurred by the 46 s stage as indicated by primitive limb elements (S). (U, U′) Quantification of the area of Hoxa13-expression domain during autopod development. Values in parentheses at each data point are the numbers of control and mutant samples examined, and the percentage of the area of Hoxa13-expression domain in the mutant compared with that in the control. Values are the mean±s.d. for each data point in U, U′. Scale bars: 200 μm. Abbreviations: an, anterior; Au, autopod; di, distal; po, posterior; pr, proximal; s, somite; St, stylopod; Ze, zeugopod.

Fig. 6. FGFR2 interacts with Wnt/β-catenin signaling to maintain the AER

(A–D) Reduction of Wnt/β-catenin signaling in Fgfr2^AER-KO forelimb buds. (A, B) lacZ expression in BAT-Gal transgenic limb buds. Broken black lines mark the basement membrane. Arrow in B denotes ectopic Wnt signaling in the distal-dorsal mesenchyme. (C, D) Lef1 expression was moderately reduced in mutant limb buds. (E–T) β-Catenin^GOF prevented premature AER loss and restored the autopod of Fgfr2^AER-KO embryos. (E–H) β-Catenin expression immunofluorescence. Stabilized β-catenin was present in the ventral ectoderm. (I–L) Fgf8 expression on vibratome sections. Ectopic Fgf8-expression in the ventral ectoderm (white arrowheads). Black arrowheads indicate the endogenous AER (K, L). (M, N) Cell death as detected by TUNEL. Dying cells in the ventral ectoderm, which were prominent in Fgfr2^AER-KO limb buds, were absent in Fgfr2^AER-KO; β-cat^GOF limb buds at this stage (Fig. 3D). Broken white lines indicate the basement-membrane. (O, P) Hoxa13 expression (white arrow) was initiated in β-cat^GOF limb buds at 32 s. (Q–T) Sox9 expression and skeletal preparations at the stages indicated. Note that the autopod was restored in Fgfr2^AER-KO; β-cat^GOF embryos (R, T). Asterisk indicates skeletal fusions between digits (syndactyly) in β-cat^GOF (S) and Fgfr2^AER-KO; β-cat^GOF limbs (T). VI′ indicates a post-axial extra digit (S). E10.5 is equivalent to ~35 s stage and E11.75 is equivalent to ~48 s stage. Scale bars: red, 50 μm; black, 200 μm. Abbreviations: an, anterior; di, distal; do, dorsal; pr, proximal; po, posterior; ve, ventral; s, somite.

Fig. 7. A model of AER function in development of the mouse forelimb autopod

(A, B) AER maintenance is essential for AER-FGF signaling and generation of autopod progenitors. Schematic diagrams of early limb buds in transverse views (top row in A, B) and whole-mount dorsal views (bottom row in A, B) to illustrate development of the AER (green) and autopod progenitors, as marked by Hoxa13 expression (red). Red dots indicate dying cells in the ventral ectoderm and AER. (A) In control forelimb buds, the AER is maintained and AER-FGF production (orange) is normal. As a result, FGF signaling in the distal limb bud mesenchyme (LBM) is sustained and autopod progenitors are generated at the 31–32 s stage. The progenitor pool expands subsequently and a sufficient number of skeletal progenitors are available to form a normal autopod when condensation starts at around the 46 s stage. (B) In Fgfr2^AER-KO forelimb buds, the AER is not maintained because of increased cell death and AER-FGF production progressively decreases. As a result, FGF signaling in the distal mesenchyme is reduced and generation of autopod progenitors is delayed by 2-somite stages until the 33–34 s stage. Although the progenitor pool expands grossly normally at later stages, it fails to produce a sufficient number of skeletal progenitors to form a normal autopod at the onset of autopod condensation.