Inactivation of Patched1 in Murine Chondrocytes Causes Spinal Fusion Without Inflammation

Authors


Abstract

Objective

During development of the vertebrate skeleton, chondrocytes form a cartilage template that is gradually replaced by bone. Hormones of the Hedgehog (HH) family have been implicated in the ossification process, but their exact relationship to normal or pathogenic bone formation is unclear. This study was undertaken to establish a genetic tool that allows the discrete inactivation of genes in spinal chondrocytes, and to investigate in vivo how chondrocyte-specific ablation of the inhibitory HH receptor Patched 1 (Ptch1) affects skeleton integrity.

Methods

A Cre-deleter mouse strain, mb1-Cre, for selective gene recombination in spinal chondrocytes was identified by in situ hybridization and histologic analysis. The mb1-Cre+/− animals were crossed with mice that harbor a loxP-flanked Ptch1 gene (Ptch1flox/flox) to abrogate the inhibition of the HH signaling pathway in chondrocytes. The skeletal integrity of F1 mice was characterized by high-resolution flat-panel–based volume computed tomography and histologic staining procedures.

Results

During the first weeks after birth, all mb1-Cre+/−/Ptch1flox/flox mice developed progressive spinal fusion with malformation of the vertebrae. This phenotype was caused by aberrant chondrocyte proliferation in the intervertebral discs that blocked endochondral ossification. Importantly, the disease pattern occurred in an inflammation-independent manner.

Conclusion

Our findings indicate that chronic activation of the HH signal pathway in spinal chondrocytes can trigger an ankylosing spine morphology without immune cell contributions. Hence, the destruction of cartilage and loss of axial joint integrity can result from chondrocyte-intrinsic defects of monogenic origin.

Development of the axial and appendicular skeleton of vertebrates is initiated by local condensation of mesenchymal cells and their differentiation into chondrocytes, which form cartilage anlagen ([1, 2]). Proper ossification of the anlagen is associated with a complex chondrocyte maturation program ([3]). Small, highly proliferating chondrocytes produce extracellular matrix proteins that comprise types II, IX, and XI collagen as well as the proteoglycan aggrecan.

Indian hedgehog (IHH), a member of the hedgehog (HH) family of secreted morphogens, and parathyroid hormone–related protein (PTHrP) have been reported to maintain chondrocytes in a proliferative state ([4-7]). Other growth factors, such as thyroid hormone or bone morphogenic proteins, antagonize the actions of IHH and PTHrP, thereby promoting cell cycle exit. This is followed by terminal differentiation of chondrocytes into hypertrophic cells, which secrete type X collagen (ColX) and ultimately die by apoptosis. Following its calcification, the cartilage matrix provides a template for the balanced action of hematopoietic osteoclasts and mesenchyme-derived osteoblasts to finalize bone formation. The long bones and the vertebrae of the spinal column are connected by synovial joints and intervertebral discs, respectively.

Chondrogenesis and endochondral ossification have been studied in vivo by conditional gene targeting in mice using the Cre/loxP recombination system ([8]). Several promoter/enhancer combinations were used to drive the expression of Cre recombinase in mesenchymal and osteochondroprogenitor cells, and hence to delete loxP-flanked gene segments in skeletal cell types ([8]). Chondrocytic gene functions were targeted using the ColIIa-Cre or ColX-Cre deleter strains, which revealed an important role of IHH and PTHrP signaling in chondrocyte differentiation ([7, 9-11]). The aggrecan-Cre deleter turned out to be useful for gene targeting in the adult cartilage ([12]). These studies contributed significantly to our understanding of the chondrocyte maturation program and skeletal development. However, Cre activity in the described deleter strains was detected in all cartilaginous elements of the organism, and moreover, some recombination activity was evident in osteoblasts and perichondrial cells ([8]). Herein we show that the mb1-Cre deleter strain is an ideal tool to inactivate loxP-engineered genes selectively in spinal chondrocytes. When applied to delete expression of the HH inhibitor Patched 1 (PTCH-1), the resulting mouse mutant showed a severe spinal malformation but developed in the absence of any signs of inflammation. The fusion of lumbar and thoracic vertebrae resulted from sustained proliferation of spinal chondrocytes in the intervertebral discs. Hence, spinal fusion and inflammation are not necessarily linked.

MATERIALS AND METHODS

Generation and genotyping of mice

The mb1-Cre+/− mice ([13]) were crossed with Ptchflox/flox mice ([14]) or with the ROSA26 Cre reporter strain (R26R) ([15]). Mouse genotyping as well as detection of inactive Ptchdel alleles was performed as previously described ([13-15]). All animal studies were undertaken with consideration of the legal requirements. Primary chondrocytes were obtained from the tail and spine of 3-day-old mice following the removal of skin and muscles, and incubated with 2 mg/ml Pronase and 3 mg/ml collagenase D (Roche). Isolated chondrocytes were incubated in 1 mM Tris HCl, 5 mM KCl, 0.45% Nonidet P40, and 0.45% Tween supplemented with 10 mg/ml proteinase K (Promega) overnight at 56°C, followed by proteinase K inactivation for 10 minutes at 95°C. Genotyping of genomic chondrocyte DNA was performed as previously described ([16]).

Skeletal and histologic analyses

To analyze spine morphology, mice were killed with carbon dioxide. Dissected spines were incubated overnight without shaking at 56°C in 1 mM Tris HCl, 5 mM KCl, 0.45% Nonidet P40, and 0.45% Tween with 10 mg/ml proteinase K (Promega). Vertebrae were washed in deionized H2O and ethanol to remove the remaining tissue. Skin and soft tissue were removed, and mouse skeletons were fixed in 95% ethanol. Serial sections of 5 μm were stained with Alcian blue and alizarin red S.

To verify Cre recombinase activity in mb1-Cre+/−/R26R+/− mice, embryos were first fixed overnight in 4% paraformaldehyde (Roth) at 4°C. Next, they were dehydrated with increasing ethanol concentrations and embedded in paraffin wax (Roth). To monitor β-galactosidase expression, LacZ stainings were performed on frozen sections that were fixed by incubation in ice-cold phosphate buffered saline (PBS) containing 0.2% glutaraldehyde for 10 minutes. After washing 3 times for 5 minutes at room temperature with LacZ buffer (PBS containing 100 mM MgCl2, 1% sodium desoxycholate, and 10% Nonidet P40), sections were incubated with LacZ staining solution (LacZ buffer containing 20 mg/ml of X-Gal, 250 mM potassium ferrocyanide, and 250 mM potassium ferricyanide) overnight at 30°C. Slides were washed 3 times in PBS and mounted onto coverslips with Glycergel Mounting Medium (Dako).

For Safranin O–Weigert stainings, spinal preparations were fixed overnight in 4% paraformaldehyde (Roth) at 4°C. To discriminate between cartilage and bone, samples were treated with Weigert's hematoxylin (Roth), 0.1% fast green (Sigma), and 0.1% Safranin O (Sigma).

For hematoxylin and eosin staining, samples were fixed in 4% paraformaldehyde overnight at 4°C, embedded in paraffin, and sectioned at 5–7 μm.

For in situ hybridization, antisense riboprobes were radiolabeled with [P33]-UTP (Hartman Analytic). Hybridization was performed in 50% formamide (Roth) at 70°C. Developed slides were counterstained with 0.2% toluidine blue O (Sigma) in 1% sodium borate (Roth). In situ hybridization probes have been described previously for type II collagen ([17]), type X collagen ([18]), Gli-1 ([19]), and IHH ([20]). To detect expression of mb1 an 845-bp fragment of the coding region was amplified by polymerase chain reaction (PCR) using the primer pair 5′-TAC-CAA-GAA-CCG-CAT-CAT-CA-3′ and 5′-AGG-AGG-GTG-AGG-CCC-TAT-AA-3′, and ligated into the pCRII-TOPO vector.

Proliferation analyses

Staining of paraffin sections with 5-bromo-2′-deoxyuridine (BrdU) was performed as previously described ([21]). Briefly, 6-week-old mice were killed 2 hours after receiving an intraperitoneal injection of 100 μg/gm body weight of BrdU. Proliferating cells were detected with a rat monoclonal BrdU antibody (Dako) and an Alexa 568–labeled secondary antibody (Invitrogen). Fluorescence images were taken on a Zeiss Axiovert 200 microscope with a Spot 23.0 camera (Diagnostik Instruments) and processed using MetaMorph imaging software (Visitron Imaging Systems).

Imaging techniques

Microscopic images were recorded with a Spot 14.2 or 23.0 camera (Diagnostik Instruments) using the MetaMorph imaging software (Visitron Imaging Systems). Hybridization signals were visualized using darkfield microscopy (Intralux 5000-1), and the expression domains were determined using Spot analysis software. Monitoring of bone morphology by noninvasive flat-panel volume computed tomography was performed on a laboratory animal flat-panel volume computed tomography system from GE Global Research as previously described ([22]).

Quantitative reverse transcriptase–PCR

Splenic B cells of 8-week-old mice were separated using a B cell isolation kit (Miltenyi Biotec). Total RNA was extracted using TRIzol reagent (Invitrogen) and reverse transcribed using random hexamers and SuperScript II reverse transcriptase (Invitrogen). SYBR Green–based assays were performed to detect transcripts of wild-type Ptch1 (5′-AAA-GCC-GAA-GTT-GGC-CAT-GGG-TAC-3′/5′-TGC-TTG-GGA-GTC-ATT-AAC-TGG-A-3′) or Ptch1del (5′-AAA-GCC-GAA-GTT-GGC-CAT-GGG-TAC-3′/5′-TTA-AAC-AGG-CAT-AGG-CAA-GCT-GAC-3′). The primer pair 5′-GGT-CAT-CTA-CGA-GAC-CAA-CTG-C-3′/5′-GTG-TCT-TCA-GGT-TCT-CCA-GGC-3′ was used to quantify Gli2 transcripts. Quantification of Gli1 transcripts was performed with the primer pair 5′-TGC-ACC-AAG-CGC-TAC-ACA-GAT-CCC-3′/5′-AGC-TGA-TGC-AGC-TGA-TCC-AGC-CTA-3′. Amplification of 18S ribosomal RNA was used to normalize for the amount of sample complementary DNA. All samples were measured in triplicate on an ABI Prism 7900 sequence detection system (Applied Biosystems) by real-time PCR.

Blood cell analyses

Peripheral blood samples were collected from the retro-orbital plexus and incubated for 15 minutes on ice with different combinations of the following antibodies: fluorescein isothiocyanate (FITC)–conjugated anti-CD4, phycoerythrin (PE)–Cy7 (PE–Cy7)–conjugated anti-CD8, FITC-conjugated anti-CD19, PE–Cy7–conjugated anti-B220, PE-conjugated anti–Gr-1, and PE–Cy7–conjugated anti-CD11b (all from BD PharMingen). To analyze B cell development, single-cell suspensions of bone marrow from femur and tibia of hind limbs and spleen of adult mice were prepared. Cells were incubated with the following antibodies: FITC-conjugated anti-CD21, PE-conjugated anti-CD24, PE–Cy7–conjugated anti-CD45R/B220, FITC-conjugated anti-IgD (all from BD PharMingen), and RPE-conjugated anti-IgM (Southern Biotechnology Associates). Cells were analyzed on an LSR II flow cytometer (BD PharMingen). Data acquisition and analysis were performed using BD FACSDiva software (BD PharMingen) and FlowJo software (Tree Star), respectively. For flow cytometric Ca2+ flux analysis, splenocytes were isolated and incubated with fluorescence-labeled antibodies directed against B220, CD21, and CD24 to distinguish B220-positive B lymphocytes of transitional type 1 (CD24+CD21−) from transitional type 2 B cells (CD24+CD21+) and mature B cells (CD24lowCD21low). Ratiometric Ca2+ flux analyses of individual B cell subpopulations were performed by flow cytometry as previously described ([23]).

RESULTS

Conditional ablation of the HH receptor PTCH-1 in spinal chondrocytes via the Cre-deleter strain mb1-Cre.

The mb1 gene encodes the Igα signaling subunit of the B cell antigen receptor and was originally described to be expressed in a B lymphocyte–restricted manner ([24]). However, in situ hybridization with mb1 antisense riboprobes revealed mb1 transcription in the developing spine of embryonic day (ED) 12.5 mouse embryos (Figure 1A). Mice with a Cre expression cassette integrated into one mb1 allele have been generated previously (mb1-Cre+/−) ([13]). We mated mb1-Cre+/− mice with LacZ reporter mice (R26R) harboring in the ROSA26 locus a LacZ expression cassette that only becomes transcribed upon Cre-mediated excision of a loxP-flanked neomycin cassette ([15]). Histologic staining for lacZ-encoded β-galactosidase in spinal sections from mb1-Cre+/−/R26R+/− embryos or neonates exclusively labeled mesenchyme-derived chondrocytes surrounding the intervertebral discs between the spinal vertebrae (Figure 1B). These data showed that, in addition to B cells, spinal chondrocytes exhibit robust mb1 activity, which suffices to drive Cre/LoxP-mediated gene recombination in these cells via the mb1-Cre+/− deleter strain.

Figure 1.

Induction of recombination of loxP-flanked gene segments in chondrocytes by mb1-driven Cre expression. A, In situ hybridization of longitudinal spine sections from embryonic day (ED) 12.5 wild-type mouse embryos with antisense mb1 riboprobes. Arrows indicate high chondrocytic mb1 expression. Bars = 100 μm. B, Chondrocytic β-galactosidase activity in spinal cross sections from mb1-Cre+/−/R26R+/− mouse embryos on ED 16.5, neonates on postnatal day 2 (P2), and neonates on postnatal day 4. Bars = 50 μm. C, Polymerase chain reaction analysis of spinal chondrocytes from 3-day-old mouse neonates for the presence of an mb1-inserted Cre expression cassette (Cre; 600 bp) (top row), Ptch1 alleles containing the loxP-flanked exons 8 and 9 (Ptch1flox; 1,735 bp) (second row) or lacking exons 8 and 9 (Ptch1del; 950 bp) (third row), or wild-type Ptch1 (wt Ptch1; 1,401 bp) (bottom row). Note that the loxP-engineered Ptch1 allele was detected even after successful deletion of exons 8 and 9 owing to specific primer combinations (see Materials and Methods for details). The last lane shows the no-template control (ntc). D, In situ hybridization of longitudinal spine sections from ED 16.5 mb1-Cre+/−/Ptch1flox/flox mouse embryos, using antisense riboprobes that detected the expression of Gli, Ihh, or type X collagen (ColX). Bars = 100 μm.

Since lymphocytes (which have known mb1 activity) can be easily distinguished from chondrocytes, mb1-Cre+/− mice provided an ideal tool to investigate candidate regulators of skeletal development by conditional gene targeting in the spine. We crossed mb1-Cre+/− animals with mice harboring 2 loxP-engineered Ptch1 alleles (Ptch1flox/flox) ([14]). PTCH-1 functions as a cell surface receptor for HH family members ([25]), and these interactions are critical for proper vertebral skeletogenesis ([26]). Ligation of PTCH-1 releases its inhibition of the transmembrane signaling partner Smoothened (Smo), which in turn activates intracellular signaling cascades, most notably through activation of transcription factors of the Gli family ([25]). In the resulting mouse model, Cre expression and inactivation of Ptch1 alleles was confirmed by PCR analyses on isolated spinal chondrocytes from 3-day-old mice that were heterozygous (mb1-Cre+/−/Ptch1flox/+) or homozygous (mb1-Cre+/−/Ptch1flox/flox) for the conditional Ptch1 allele (Figure 1C).

Lack of chondrocytic Ptch1 caused activation of the HH signaling cascade, as indicated by increased Gli1 expression in the vertebral column of mb1-Cre+/−/Ptch1flox/flox ED 18.5 embryos compared to control mice (Figure 1D). Expression of the chondrocyte markers Ihh and ColX was detected in the vertebral growth plate of wild-type and mb1-Cre+/−/Ptch1flox/flox animals (Figure 1D). Consistent with the findings of previous studies ([27]), we did not detect Ihh or Gli1 expression in the intervertebral discs of ED 18.5 mouse embryos. These data show that mb1-driven Cre expression in Ptch1flox/flox spinal chondrocytes efficiently abrogated Ptch1 expression and resulted in tonic HH signaling.

PTCH-1 deficiency driven by mb1-Cre induces vertebral fusion resulting from defective endochondral ossification

Heterozygous and homozygous mice were born at the expected Mendelian ratios and appeared grossly normal during the first week after birth. After 2–3 weeks, mb1-Cre+/−/Ptch1flox/flox mice showed aberrant tail development. Specifically, the tail was significantly shortened, the body-proximal part was sharply bent to the right or left side of the animal, and the whole tail became progressively inflexible until it finally was almost entirely stiff (see Supplementary Movies S1-S3, available online at http://www.immunologie.uni-goettingen.de/Wienands/Movies/MovieS1.mov, http://www.immunologie.uni-goettingen.de/Wienands/Movies/MovieS2.mov, and http://www.immunologie.uni-goettingen.de/Wienands/Movies/MovieS3.mov, respectively). This phenotype was inherited with 100% penetrance. All heterozygous mb1-Cre+/−/Ptch1flox/+ littermates and Ptch1flox/flox control animals remained phenotypically normal throughout their lives.

To analyze tail and skeletal morphology of mb1-Cre+/−/Ptch1flox/flox animals in more detail, we performed flat-panel–based volume computed tomography, which allows noninvasive high-resolution imaging of the live animal in 3 dimensions ([22]). The resulting images revealed a striking spine pathology in adult Ptch1 mutant mice, as shown in Figure 2A and Supplementary Movies S4 and S5, available online at http://www.immunologie.uni-goettingen.de/Wienands/Movies/MovieS4.mov and http://www.immunologie.uni-goettingen.de/Wienands/Movies/MovieS5.mov, respectively. Most of the lumbar and thoracic vertebrae, as well as the intervertebral discs, were damaged or absent, which collectively resulted in severe spinal fusion (Figure 2A and Supplementary Movies S4 and S5). Cervical vertebrae were slightly less affected, while the remaining skeleton, including limbs, developed normally (Supplementary Movies S4 and S5). False color display of flat-panel volume computed tomography images demonstrated an enhanced relative spinal bone density in the mutant mice compared to healthy controls (Figure 2A).

Figure 2.

Chondrocytic inactivation of Ptch1 induces spinal malformation in mice due to the fusion of abnormally structured vertebrae. A, Volume rendering of in vivo acquired flat-panel–based volume computed tomography 3-dimensional data sets (resolution 150 μm) depicting skeletal morphology, including relative bone densities (bottom) of 6-week-old wild-type and Ptch1 mutant mice. Yellow indicates high bone density; gray indicates low bone density. B, Macroscopic images of vertebrae dissected from the cervical spine (top), thoracic/lumbar spine (middle), or tail (bottom) of 12-week-old wild-type and Ptch1 mutant mice. C, Alcian blue/alizarin red–stained spines from 4-week-old wild-type and Ptch1 mutant mice. The top and middle rows show longitudinal views; the bottom row shows cross-sections. Bars = 1,000 μm.

Isolated vertebrae of Ptch1 mutant mice displayed dramatic deformities. The transverse, spinous, and superior processes were abnormal for cervical vertebrae, and almost absent for thoracic and lumbar vertebrae (Figure 2B). Compared with wild-type mice, the tail vertebrae of Ptch1 mutants were much shorter, irregularly shaped, and severely lacerated (Figure 2B). Staining spinal preparations with Alcian blue and alizarin red further confirmed the ankylosing spine morphology and malformations of individual spine vertebrae, both of which are characteristic features of the Ptch1 mutant phenotype (Figure 2C). Importantly, adult mb1-Cre+/−/Ptch1flox/flox animals were otherwise normal and healthy, without symptoms of weakness (Supplementary Movie S1).

The observations that mb1-Cre+/−/Ptch1flox/flox mice showed spinal fusion, and that the HH pathway was constitutively active in Ptch1-negative chondrocytes, suggested that the animals had defective endochondral ossification. This hypothesis was directly confirmed by hematoxylin and eosin staining of spine sections (Figure 3A) and labeling using the proliferative markers BrdU (Figure 3B) and proliferating cell nuclear antigen (data not shown). These experiments revealed massive proliferation of chondrocytes and an abnormal growth plate with an absence of hypertrophic chondrocytes in the hyperproliferative cell fraction. Furthermore, staining with Safranin O–Weigert dye to discriminate between bone and cartilage revealed the presence of proliferating chondrocytes in the intervertebral discs and a marked loss of fibrous cartilage, which surrounded the nucleus pulposus in the control animals (Figures 3C and D). Ptch1 mutant mice showed a characteristic columnar alignment of growth plate chondrocytes, but the columns were shortened and distorted due to the massive enrichment of cells in the adjacent intervertebral cartilage (Figures 3C and D). Abnormal endochondral bone formation extended to the cortical bone in the Ptch1 mutant mice. The nucleus pulposus seemed to be displaced by the proliferating chondrocytes. Importantly, no infiltration of leukocytes could be detected. These data demonstrated ongoing chondrocyte proliferation in the intervertebral discs of adult Ptch1 mutants, which in turn blocked chondrocyte maturation and proper endochondral ossification.

Figure 3.

Ptch1-negative chondrocytes remain in a hyperproliferative state. A, Histologic analysis of hematoxylin and eosin–stained longitudinal tail sections from 4-week-old wild-type and Ptch1 mutant mice. Broken lines indicate the boundary between the growth plate and the intervertebral disc. Bars = 50 μm. B, Histologic analysis of chondrocytes in longitudinal tail sections from 6-week-old wild-type and Ptch1 mutant mice. Sections were labeled with 5-bromo-2′-deoxyuridine (top) and stained with Safranin O–Weigert dyes to discriminate between cartilage (red) and bone (blue) (middle and bottom). Bars = 200 μm in top and middle panels; 100 μm in bottom panels. C, Histologic analysis of Safranin O–Weigert dye–stained longitudinal tail sections from 4-week-old wild-type and Ptch1 mutant mice. Cartilage is shown in red, and bone is shown in blue. Bars = 1,000 μm in top panels; 100 μm in middle and bottom panels. D, Safranin O–Weigert staining of longitudinal lumbar spine sections from 4-week-old wild-type and Ptch1 mutant mice. Bars = 1,000 μm in top panels; 100 μm in bottom panels.

Spinal fusion in Ptch1 mutant mice results from a chondrocyte-autonomous effect

Dysregulated proliferation of spinal chondrocytes and the concomitant distortion of the intervertebral discs could explain the spinal fusion seen in the mb1-Cre+/−/Ptch1flox/flox animals. However, mb1-driven Cre expression efficiently recombines loxP-flanked gene segments in B lymphocytes, as demonstrated by the genetic analyses for which the mb1-Cre deleter strain was originally developed and subsequently successfully used ([13]). Furthermore, human spondylopathies are frequently associated with spinal inflammation. We thus considered the possibility that in mb1-Cre+/−/Ptch1flox/flox mice, B lymphoid inactivation of Ptch1 dysregulated B cell activation, and that this contributed to the spinal Ptch1 deficiency symptoms. In fact, the loxP-engineered B cell alleles of Ptch1 were effectively inactivated in the mb1-Cre+/−/Ptch1flox/flox mice although this did not increase B lymphoid Gli expression (Figure 4A). Normal numbers of B cells were detected for all B cell compartments in the bone marrow and spleen in mb1-Cre+/−/Ptch1flox/flox mice (Figure 4B). Importantly, the PTCH-1 mutant B cells responded to B cell receptor stimulation like their wild-type counterparts (Figure 4C). Hence, the development of B cells, their distribution, and their function appear to be unaffected by the absence of PTCH-1. This is consistent with the normal B cell phenotype of mice that lack the HH activator Smo ([28]).

Figure 4.

B cells in mb1-Cre+/−/Ptch1flox/flox mice develop and function normally. A, Expression analyses of wild-type (WT) Ptch1, Ptch1del, Gli1, and Gli2 transcripts in selected splenic B cells from Ptch1flox/flox mice and mb1-Cre+/−/Ptch1flox/flox mice. Gene expression was normalized to the expression of 18S ribosomal RNA. Measurement of each sample was performed in triplicate. Bars show the mean ± SD (n = 3 mice per group). B, Representative flow cytometric analyses of bone marrow and spleen cells from control and mb1-Cre+/−/Ptch1flox/flox mice. Bone marrow cells were analyzed using antibodies against B220 and IgM to distinguish pro-B/pre-B cells (I) from immature B cells (II) and recirculating B cells (III). Spleen cells were analyzed using antibodies against B220, CD21, and CD24 to distinguish T1, T2, and mature (M) B lymphocytes in the B220+ cell fraction. C, Analysis of B cell antigen receptor–induced Ca2+ flux in mouse splenic B cells. Splenic B cell subsets were analyzed as described in B and simultaneously loaded with the ratiometric Ca2+ chelator Indo-1.

To finally rule out a possible B cell contribution to the PTCH-1 deficiency syndromes in mb1-Cre+/−/Ptch1flox/flox mice, we generated mice with a Cre expression cassette in both mb1 alleles (mb1-Cre+/+/Ptch1flox/flox), thereby abrogating mb1 expression and Igα protein production. Since Igα is indispensable for B cell development ([29]), homozygous mb1-Cre+/+/Ptch1flox/flox mice completely lacked peripheral B cells (Figure 5). Yet, these B cell–negative mice had fused spines with symptoms that were indistinguishable from those in heterozygous mb1-Cre+/−/Ptch1flox/flox mice (Figure 5). Likewise, the developmental onset and progression phase of the spinal malformations were identical between the two Ptch1 mutant strains. These data unambiguously excluded the possibility that B cells contributed to the spinal Ptch1 deficiency symptoms. Furthermore, the T cell compartments, monocytes, and granulocytes were present and contained normal cell numbers in mb1-Cre+/−/Ptch1flox/flox mice (Table 1). These data, taken together with the complete lack of leukocyte infiltration into the intervertebral discs as well as the absence of any signs of inflammation (Figure 3), rule out a major contribution of immune cells to the PTCH-1 deficiency syndrome in our mice. We conclude that the spinal fusion developed as a direct result of Ptch1 deletion in chondrocytes of the intervertebral discs, which caused chronic HH signaling and thereby sustained chondrocytic proliferation.

Figure 5.

B cells in mb1-Cre+/−/Ptch1flox/floxmice are dispensable for the spinal PTCH-1 deficiency syndromes. Photographs of 8-week-old mice with the indicated genotype and results of flow cytometric analysis of B lymphocytes in peripheral blood using antibodies directed against B220 and IgD are shown.

Table 1. Absence of leukocytosis in mb1-Cre+/−/Ptch1flox/flox mice*
 Control micemb1-Cre+/−/Ptch1flox/flox mice
  1. Values are the mean ± SD percent. Relative cell numbers in peripheral blood samples were determined by fluorescence-activated cell sorting analyses using anti-CD19/anti-B220 (for B cells), anti-CD4/anti-CD8 (for T cells), and anti–Gr-1/anti-CD11b (for monocytes and granulocytes) from control mice (n = 3) and mb1-Cre+/−/Ptch1flox/flox mice (n = 3).
B cells6.41 ± 1.545.71 ± 2.10
CD4+ T cells0.83 ± 0.520.70 ± 0.30
CD8+ T cells0.87 ± 0.181.04 ± 0.34
Monocytes0.58 ± 0.280.60 ± 0.33
Granulocytes2.44 ± 0.651.86 ± 0.20

DISCUSSION

Diseases of the spine with destruction of cartilage and loss of axial joint integrity are believed to have a multifactorial etiology with significant contributions from inflammatory processes. The clinical picture of these diseases is heterogeneous and varies from patient to patient. In this study, we showed that HH/PTCH-1 signaling acts as a chondrocytic master regulator for the proper development and maintenance of skeletal integrity. Chronic activation of the chondrocytic HH pathway following PTCH-1 ablation suffices to induce spinal fusion without aberrant activation of the immune system. It has previously been documented that IHH signaling is important for normal chondrocyte growth and differentiation and is involved in joint formation ([4, 30-32]). In the growth plate of the limbs, forced expression of either sonic HH or IHH activates chondrocyte proliferation and impairs joint formation ([4, 11, 33]). In the lumbar spine, chondrocyte proliferation and HH responsiveness measured by a Gli-1 reporter ceases during aging. Importantly, proliferation is abolished after 3 weeks postnatally ([34]). However, the impact of the HH signaling pathway on spinal development in the adult living organism has not been investigated in detail to date. This has now been made possible by our finding that the mb1-Cre deleter strain can be used to conditionally inactivate loxP-modified genes such as Ptch1 in spinal chondrocytes. Within the first weeks after birth, mb1-Cre+/−/Ptch1flox/flox mice developed severe spinal malformation with 100% penetrance.

In vivo imaging and histologic stainings revealed a pronounced spinal malformation with fused and malformed vertebrae. This phenotype was caused by a massive proliferation of Ptch1-negative chondrocytes that blocked chondrocyte maturation and consequently resulted in defective endochondral ossification. Constitutive activation of HH signaling in the vertebrae of mb1-Cre+/−/Ptch1flox/flox mutant mice keeps the chondrocytes in a proliferative state, while these cells cease proliferation in adult wild-type mice ([34]). The invasive growth prevents regular formation of the intervertebral disc and, ultimately articulation, thereby leading to the fusion of vertebrae. Importantly, the spinal malformation in Ptch1 mutant mice occurred in the absence of any signs of inflammation, and no leukocyte infiltration into intervertebral discs was observed. No alterations in the composition of lymphocyte populations were detected. The total absence of B cells in homozygous mb1- Cre+/+/Ptch1flox/flox mice changed neither the course nor the severity of spinal fusion. We conclude that the spinal fusion in our mice developed as a direct result of Ptch1 deletion in chondrocytes surrounding the intervertebral discs.

These findings raise questions about the impact of inflammatory events in the development of human spondyloarthritic diseases. The available mouse model demonstrates that abnormal bone development and inflammation can be experimentally uncoupled. These mice represent features of several human diseases that are directly or indirectly associated with spinal malformation. Nevertheless, the type of spinal fusion with affected intervertebral discs is in contrast to spondyloarthritis, where the spinal fusion occurs in the paravertebral ligaments with unaffected intervertebral disc morphology ([35]). Ptch1 mutant mice are in many aspects phenotypically related to the noninflammatory human disease diffuse idiopathic skeletal hyperostosis (DISH). These patients have calcification and ossification of the axial skeleton, but in contrast to our mouse model, peripheral entheses are also involved ([36, 37]).

Although the genetic link in DISH is missing, other human spinal malformations are associated with a genetic defect. Spinal deformities, together with the life-threatening basal cell carcinomas, are clinical manifestations of the human Gorlin-Goltz syndrome ([38, 39]). Affected patients are heterozygous for Ptch1-null alleles, and this predisposes them to complex disease patterns upon secondary mutations in somatic cells. Similar to our Ptch1 mutant mice, patients with Gorlin-Goltz syndrome usually possess irregularly shaped vertebrae that are tightly fused without overt symptoms of spinal inflammation ([40, 41]). It is thus possible that this clinical feature of Gorlin-Goltz syndrome is caused by secondary mutations in spinal chondrocytes that completely abrogate PTCH-1 function, leading to chondrocytic hyperproliferation and a concomitant block of endochondral ossification. Alternatively, no additional somatic mutations are required because the inhibition of the HH pathway in human chondrocytes is gene dose–dependent as it requires the presence of two wild-type Ptch1 alleles. In both cases, however, chondrocytic dysregulation causes inflammation-independent fusion of vertebrae.

In summary, our mouse model supports the view that skeletal malformations can occur without immune cell contribution. Rather, proliferating chondrocytes can progressively distort the anatomy of joints or bones, thereby lowering skeletal mobility and causing mechanical tissue damage. In diseases with altered skeletal morphology, inflammation might not be causative but contributes to disease progression. Our mouse model, and the possibility to study chondrocyte-intrinsic aspects of bone formation by gene targeting, will contribute to future investigations that are needed to ultimately dissect the pathogenesis of skeletal disorders.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Wienands had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Dittmann, Wuelling, Hahn, Wienands.

Acquisition of data. Dittmann, Wuelling, Uhmann, Dullin, Schweyer.

Analysis and interpretation of data. Dittmann, Wuelling, Dullin, Schweyer, Vortkamp, Wienands.

Acknowledgments

We thank Drs. Michael Reth and Elias Hobeika for providing mb1-Cre deleter mice, Dr. Philippe Soriano for providing ROSA26-LacZ mice, and Gabriele Sonntag for expert technical assistance.

Ancillary