Deletion of Tgfbr2 in Prx1-cre expressing mesenchyme results in defects in development of the long bones and joints

Abstract

In this study, we address the function of Transforming Growth Factor beta (TGF-beta) and its type II receptor (Tgfbr2) in limb development in vivo. Mouse embryos were generated in which the Tgfbr2 gene was deleted in early limb mesenchyme using Prx1Cre-mediated LoxP recombination. A high level of Tgfbr2 gene deletion was verified in limb mesenchyme by PCR between E9.5 and E10.5 days in Cre expressing mice. RT-PCR assays indicated a significant depletion of Tgfbr2 mRNA by E10.5 days as a result of Cre mediated gene deletion. Furthermore, limb mesenchyme from Cre(+);Tgfbr2(f/f) mice placed in micromass culture did not respond to exogenously added TGF-beta1 confirming the functional deletion of the receptor. However, there was an unexpected increase in the number and intensity of Alcian blue stained chondrogenic nodules in micromass cultures derived from Tgfbr2-deleted limbs relative to cultures from control limbs suggesting that Tgfbr2 normally limits chondrogenesis in vitro. In vivo, early limb development and chondrocyte differentiation occurred normally in Tgfbr2-depleted mice. Later in development, depletion of Tgfbr2 in limb mesenchyme resulted in short limbs and fusion of the joints in the phalanges. Alteration in the length of the long bones was primarily due to a decrease in chondrocyte proliferation after E13.5 days. In addition, the transition from prehypertrophic to hypertrophic cells was accelerated while there was a delay in late hypertrophic differentiation leading to a reduction in the length of the marrow cavity. In the joint, cartilage cells replaced interzone cells during development. Analysis of markers for joint development indicated that the joint was specified properly and that the interzone cells were initially formed but not maintained. The results suggest that Tgfbr2 is required for normal development of the skeleton and that Tgfbr2 can act to limit chondrogenesis in mesenchymal cells like the interzone.

Figure 1. Tissue specific deletion of Tgfbr2 by Prx1-Cre

The wild type (Wt), floxed, and deleted alleles of Tgfbr2 were detected by PCR using the primers indicated in (A). (B) Genomic DNA was isolated from whole limb buds of E9.5, E10.5, E11.5 and E12.5 day embryos of the genotypes Prx1Cre⁺/Tgfbr2^f/w (a), Prx1Cre⁺/Tgfbr2^f/f (b), and Prx1Cre⁻/Tgfbr2^f/w (c). DNA was PCR amplified to detect the relative levels of deleted (del) (610 bp), floxed (flox) (540bp), and wild type (WT) (420bp) DNA. (C) DNA was isolated from condensed (cond) and interdigit (non) regions from Prx1Cre⁺/Tgfbr2^f/w (a) and Prx1Cre⁺/Tgfbr2^f/f (b) E12.5 day limbs using laser capture microdissection. The relative levels of floxed (flox) and deleted (del) alleles within each compartment were determined. (D) RNA was isolated from whole limbs buds of control (c), Prx1Cre⁺/Tgfbr2^f/w (a), and Prx1Cre⁺/Tgfbr2^f/f (b) E10.5 day embryos and the relative levels of Tgfbr2 mRNA were determined using semi-quantitative RT-PCR. Primer sets directed within exon 2 (e2) and primer sets encompassing exon2 and exon3 (e2–e3) were used. Product formation is shown within the linear range. Expression of 18S was used as a normalization control.

Figure 2. Functional assay for the deletion of Tgfbr2

Limb mesenchyme from control Cre-negative (A,B) and Prx1Cre⁺/Tgfbr2^f/f (C,D) E11.5 day (Theiler stage 19) embryos was placed in micromass culture in the absence (A,C) or presence (B,D) of 5ng TGF-β1/ml for 3 days at which time the cells were stained with Alcian blue. In control, untreated Tgfbr2 positive cultures, discreet nodules of condensed mesenchyme stained with Alcian blue were observed. As previously described, TGF-β1 treated cells showed diffuse Alcian blue staining through the whole culture so that discreet nodules were difficult to detect. Cultures from Prx1Cre⁺/Tgfbr2^f/f limbs did not respond to TGF-β1 and an increased number of nodules as well as more intense staining were observed relative to cultures from control mice.

Figure 3. Early skeletal development in Prx1Cre+/Tgfbr2f/f mice

(A–C) Alcian blue stained hindlimbs (A–C), H&E stained (D–F), and Sox9 stained (G–I) sections were prepared from E12.5 day control Prx1Cre−/Tgfbr2f/w (A, D, G), Prx1Cre+/Tgfbr2f/w (B, E, H) and, experimental Prx1Cre+/Tgfbr2f/f (C, F, I) embryos. Initiation of skeletal development occurred in all three cases. Sox9 staining was comparable in all three genotypes (G–I).

Figure 4. Analysis of the skeletal phenotype of in Prx1Cre⁺/Tgfbr2^f/f embryos

(A) Gross appearance of newborn mice. Mutant newborns had very short limbs and were missing the parietal bone. (B–D) Alizarin red/alcian blue stained skeletal preparations. Gross appearance of whole E18.5 day embryos is shown in (B). The sternal elements were split in the Prx1Cre⁺/Tgfbr2^f/f mice (arrows, C). The deltoid tuberosity was missing, the joints in the autopod were fused, and there was reduced mineralization by E15.5 days (D). The parietal bones and frontal bone were missing or reduced in size in the skull of Prx1Cre⁺/tgfr2^f/f embryos (top view E, side view F).

Figure 5. Characterization of the growth plate phenotype in Prx1Cre⁺/Tgfbr2^f/f mice

(A) Alizarin/Alcian blue skeletal preparations of E18.5 day control and Prx1Cre⁺/Tgfbr2^f/f limbs showing the reduced size of the limb. (B,C) Chondrocyte proliferation. Incorporation of BrdU into E13.5 (B) and E15.5 (C) day long bones was measured by immunoflourscence (red). Cells were counterstained with YoPro (green). Proliferation was dramatically reduced by E15.5 days. (D, E) Organization of the growth plate. Close up of the growth plate from the distal radius and ulna in alizarin red/alcian blue E18.5 day control and mutant mice (D). The hypertrophic zone is bracketed by small arrows. H&E stained sections from the ulna of E16.5 day control and mutant limbs (E). The total length of the bone is marked with the double arrows (a). The distance from the top of the bone to the start of the hypertrophic zone (b), the hypertrophic zone (c) and the marrow cavity/trabecular bone (d) are marked.

Figure 6. Characterization of the joint phenotype in Prx1Cre⁺/Tgfbr2^f/f mice

(A–C) Whole mount skeletal preparations of the forelimb autopod in control and Prx1Cre⁺/Tgfbr2^f/f mice at E13.5 (A), E15.5 (B), and E18.5 (C) days. The area denoting the developing joint is marked with an arrow. (D–F) H&E stained sections of the forelimb autopod from control and mutant mice at E13.5 (D), E14.5 (E), and E16.6 (F) days. The interzone is marked with the large arrow.

Figure 7. GDF5 and Wnt9a expression in Prx1Cre⁺/Tgfbr2^f/f mice

Sections of phalanges of E13.5, 14.5, 15.5 day control (A, C, E, G) and Prx1Cre+/Tgfbr2f/f (B, D, F, H) mice were hybridized to an anti-sense probe for GDF5 (A–F) and Wnt9a (G, H) mRNA. Arrows and arrowheads indicate staining in the interzone cells and perichondrium, respectively. GDF5 expression was first detected at E12.5 (data not shown). Expression increased over time and defined the interzone in control limbs (A, C, E). In mutant limbs, GDF5 expression in interzone cells was reduced over time (arrows, B, D, F). In addition, ectopic GDF5 expression was detected in the perichondrium of the developing phalanges (arrowheads, B, D, F). Wnt9a showed a same expression pattern with GDF5 (G, H and data not shown).

Figure 8. Cartilage replaces interzone in E15.5 day Prx1Cre⁺/Tgfbr2^f/f embryos

Sections from control (A, D) and Prx1Cre⁺/Tgfbr2^f/f (B, C, E, F) E15.5 day limbs were stained with H&E (A–C) or PNA (D–F). PNA stained sections were counterstained with YoPro, a green nuclear stain. The red channel, indicating PNA staining, is shown alone (red) as well as merged with the green nuclear stain (merge). Insets in A–C show close-up images of the joint region. Asterisks are next to the joint region shown in the inset.

Figure 9. Reduced apoptosis in Prx1Cre⁺/Tgfbr2^f/f joints

Apoptotic cells were measured by TUNEL assay on sections from control (A, C) and Prx1Cre⁺/Tgfbr2^f/f embryos (B, D) at E14.5 (A,B) and E15.5 (C,D) days. Apoptosis was seen in the interzone of the developing joint in control limbs. No apoptotic cells were detected in the presumptive joint area of mutant E14.5 and E15.5 day embryos. We did not detect apoptotic cells in control or mutant joints at E13.5 days (not shown).