ADAMTS metalloproteases generate active versican fragments that regulate interdigital web regression

Abstract

We show that combinatorial mouse alleles for the secreted metalloproteases Adamts5, Adamts20 (bt), and Adamts9 result in fully penetrant soft-tissue syndactyly. Interdigital webs in Adamts5(-/-);bt/bt mice had reduced apoptosis and decreased cleavage of the proteoglycan versican; however, the BMP-FGF axis, which regulates interdigital apoptosis was unaffected. BMP4 induced apoptosis, but without concomitant versican proteolysis. Haploinsufficiency of either Vcan or Fbln1, a cofactor for versican processing by ADAMTS5, led to highly penetrant syndactyly in bt mice, suggesting that cleaved versican was essential for web regression. The local application of an aminoterminal versican fragment corresponding to ADAMTS-processed versican, induced cell death in Adamts5(-/-);bt/bt webs. Thus, ADAMTS proteases cooperatively maintain versican proteolysis above a required threshold to create a permissive environment for apoptosis. The data highlight the developmental significance of proteolytic action on the ECM, not only as a clearance mechanism, but also as a means to generate bioactive versican fragments.

Figure 1. Combinatorial mutant Adamts alleles develop STS with greater penetrance than single Adamts mutants

A. Histogram of penetrance of STS in mice of the indicated genotype. The number of mice evaluated for each genotype is indicated above each bar. Note the general trend towards a higher penetrance with increasing dosage of inactivated (Adamts5, Adamts9) or mutant (bt) alleles. B. Representative types of STS seen in Adamts5−/−; bt/bt forelimbs and hindlimbs compared to wild-type limbs. Arrows indicate persistent interdigital webs. C. Three-dimensional reconstruction following mCT analysis of 3 week-old Adamts5−/−; bt/bt paws shows normal patterning and no radiological anomalies. Wild-type limbs are shown as controls.

Figure 2. Expression of Adamts genes in the developing autopod

A and B. β-gal staining (blue-green) was used to profile Adamts5 (A) and Adamts9 (B) expression in forelimb and hindlimb autopods from E11.5 through E14.5. Note strong interdigital staining at E13.5 and E14.5 (arrows). The asterisk in the E11.5 Adamts5−/− autopod indicates staining of the brachial plexus. C. In situ hybridization of Adamts20 in E11.5, E12.5, and E13.5 autopods. Diffuse expression was seen in the IDT at E12.5 and E13.5, and in the medial and lateral edges of the autopod (arrows). D. β-gal staining of Adamts4−/− autopods shows weak expression in forelimb IDT at E13.5. E. RT-PCR of RNA extracted from E13.5 wild-type hindlimbs. Specific amplicons are indicated by yellow asterisks. Neg: control PCR using water in place of cDNA.

Figure 3. Adamts5−/−; bt/bt autopods have decreased apoptosis but a normal BMP-FGF axis

A. TUNEL staining of a section from wild-type E14.5 hindlimb autopod (control) shows the expected distribution of apoptoptic cells in the IDT (green fluorescence); TUNEL positive cells are absent in the IDT between digits 2 and 3 as well as between digits 3 and 4 of Adamts5−/−;bt/bt mice. The image is representative of 3 separately stained hindlimbs. Nuclei are stained blue using DAPI. Scale bar equals 50 μM. B. Acridine orange staining of E14.5 hindlimbs showing staining in the IDT of Adamts5−/−; bt/+ mice, whereas staining is absent in the indicated Adamts5−/−; bt/bt IDT (white asterisks). Similar results were obtained in forelimbs of Adamts5−/−; bt/bt mice at E13.5. C. Acridine orange staining was scored in each hindlimb interdigital space in a blinded fashion according to the scoring scheme shown on the right. In Adamts5−/−; bt/bt hindlimbs (n=6), the IDT between hindlimb digits 2 and 3 and digits 3 and 4 consistently had fewer stained cells than the control genotype (Adamts5−/−, bt/+) hindlimbs (n=6). D. Hindlimb autopods stained by whole mount in situ hybridization with the indicated probes shows comparable mRNA expression in E13.5 Adamts5−/−; bt/bt mice and controls. Similar expression was obtained in forelimbs. E. pSmad 1/5/8 immunohistochemistry demonstrates nuclear pSmad staining in both mutant and wild-type E14.5 hindlimb webs. Scale bar equals 50 μM.

Figure 4. Reduced versican processing in Adamts5−/−; bt/bt IDT, and increased STS in bt/bt; Vcan+/− mice

A. An E16.5 hindlimb from an Adamts5−/−; bt/bt mouse shows lack of IDT resorption between digits 3 and 4 (boxed area, left-hand panel). The boxed area is enlarged in the center panel to illustrate a low cell density and loose ECM in the unresorbed IDT, suggestive of proteoglycan accumulation. The non-regressed ECM stains strongly with an anti-versican GAG antibody (right-hand panel, green signal). Scale bar equals 50 μm. B. Cleaved versican identified by anti-DPEAAE staining (green) is present in the IDT of an E14.5 wild-type hindlimb (left-hand panel). In contrast, greatly reduced anti-DPEAAE staining is seen in E14.5 Adamts5−/−; bt/bt hindlimb IDT (right-hand panel). Scale bar equals 50 μm. C. STS in bt/bt; hdf/+ mice. Data indicating penetrance in Adamts5−/−; bt/bt mice, and bt/bt mice (black bars), are taken from Fig. 1. Note the lack of syndactyly in compound heterozygotes of bt and hdf and the greatly increased penetrance in bt/bt; hdf+ mice (gray bar).

Figure 5. Fibulin-1 and versican co-localize in the IDT and fibulin-1 acts as a co-factor for Adamts5-mediated versican proteolysis

A. Intact versican (GAGß, green, left-hand panel) has a similar localization as fibulin-1 (green, right-hand panel) in E14.5 wild-type hindlimb perichondrium (P). A weak signal is present in IDT. These data are representative of 3 stained autopods. Nuclei are stained blue by DAPI. Scale bar equals 50 μM. B. Fibulin-1 enhances ADAMTS5 processing of versican. Western blots show expression of fibulin-1C, and fibulin-1D (upper panel) in cell lysates from HEK293F cells co-transfected with full-length Adamts5 (detected by a propeptide specific antibody RP-2, center panel). Anti-DPEAAE was used to detect the cleaved 70 kDa versican species in the medium (lower panel). C. Quantification of anti-DPEAAE reactive molecular species in medium of cells expressing ADAMTS5 alone, or ADAMTS5 plus fibulin-1C or fibulin 1-D. Levels of anti-DPEAAE-reactive peptide obtained by densitometry were normalized to levels of Adamts5 propeptide and are representative of four independent experiments (data show mean ± S.E.M.). There is a statistically significant enhancement of versican processing by ADAMTS5 in the presence of both fibulin-1 isoforms (Student t-test, *p <0.05). D. Penetrance of syndactyly is increased in bt/bt; Fbln1/+ mice compared to bt/bt and Fbln1+/−; bt/+ mice, suggesting a genetic interaction between Adamts5 and Fbln1 (Note: Fbln1 haploinsufficiency leads to STS in mice lacking ADAMTS20, thus uncovering a functional association with another ADAMTS protease, most likely ADAMTS5, as evidence by findings in panel B). Data for penetrance in Adamts5−/−; bt/bt mice and bt/bt mice (black bars) are taken from Fig. 1.

Figure 6. Versican expression and localization during autopod development and the effect of local BMP4 application on apoptosis and versican proteolysis

A. Whole-mount ISH with a Vcan probe shows strong expression in the precartilaginous mesenchyme of digital rays at E12.5 (arrows), with forelimb expression anticipating the hindlimb. At later stages, Vcan is strongly expressed in perichondrium (arrows at E13.5), with weak expression in the IDT. B. Serial sections from a wild-type E14.0 hindlimb autopod were immunostained with indicated antibodies to show the precise relationship between versican, cleaved versican, fibulin-1 and ADAMTS5. The boxed regions in the upper panels are enlarged in the lower panels. Cleaved versican is located in the IDT whereas the strongest signals for intact versican and fibulin-1 are present in the perichondrium (P). ADAMTS5 is located at the boundary between perichondrium and IDT. C. Implantation of a BMP4-soaked bead in E13.75 Adamts5−/−; bt/bt hindlimb IDT leads to apoptosis (TUNEL) but not to versican cleavage, whereas a bead soaked in BSA did not induce cell death. The result is representative of three independent experiments.

Figure 7. Versican G1-DPEAAE⁴⁴¹ fragment induces apoptosis

A. Western blotting shows production and secretion of the G1-DPEAAE⁴⁴¹ fragment in conditioned medium (CM) of stably transfected cells. The absence of immunoreactivity in cell lysate (CL) may reflect the relative abundance of accumulated G1-DPEAAE⁴⁴¹ fragment in CM. B. A bead soaked in G1-DPEAAE⁴⁴¹ fragment induced greater apoptosis over a wider distance compared to the control bead. The result is representative of six independent experiments for both the G1-DPEAAE⁴⁴¹ fragment and control media. C. A proposed model for the role of ADAMTS proteases and versican in IDT regression, suggesting that ADAMTS-generated versican fragments may work synchronously with BMPs and create a permissive environment for BMP induction of apoptosis. D. Cooperative model of ADAMTS action. Cooperative maintenance of the required threshold of versican proteolysis is provided by spatial and temporal overlap of versican-degrading ADAMTS proteases in the IDT. Because two additional versican-degrading proteases (Adamts1 and Adamts9) are implicated, we speculate on the existence of a second threshold below which more severe syndactyly, or STS affecting all web spaces may be seen (gray bar). Alteration of the level of fibulin-1 alters the proteolytic threshold by reducing ADAMTS5 activity.