Dose-dependent effects of gamma radiation sterilization on the collagen matrix of human cortical bone allograft and its influence on fatigue crack propagation resistance.

Fatigue crack propagation resistance and high-cycle S-N fatigue life of cortical bone allograft tissue are both negatively impacted in a radiation dose-dependent manner from 0 to 25 kGy. The standard radiation sterilization dose of 25-35 kGy has been shown to induce cleavage of collagen molecules into smaller peptides and accumulation of stable crosslinks within the collagen matrix, suggesting that these mechanisms may influence radiation-induced losses in cyclic fracture resistance. The objective of this study was to determine the radiation dose-dependency of collagen chain fragmentation and crosslink accumulation within the dose range of 0-25 kGy. Previously, cortical bone compact tension specimens from two donor femoral pairs were divided into four treatment groups (0 kGy, 10 kGy, 17.5 kGy, and 25 kGy) and underwent cyclic loading fatigue crack propagation testing. Following fatigue testing, collagen was isolated from one compact tension specimen in each treatment group from both donors. Radiation-induced collagen chain fragmentation was assessed using SDS-PAGE (n = 5), and accumulation of pentosidine, pyridinoline, and non-specific advanced glycation end products were assessed using a fluorometric assay (n = 4). Collagen chain fragmentation increased progressively in a dose-dependent manner (p < 0.001). Crosslink accumulation at all radiation dose levels increased relative to the 0 kGy control but did not demonstrate dose-dependency (p < 0.001). Taken together with our previous findings on fatigue crack propagation behavior, these data suggest that while collagen crosslink accumulation may contribute to reduced notched fatigue behavior with irradiation, dose-dependent losses in fatigue crack propagation resistance are mainly influenced by radiation-induced chain fragmentation.

Conserved sequence in the aggrecan interglobular domain modulates cleavage by ADAMTS-4 and ADAMTS-5.

BACKGROUND: Cleavage of aggrecan by ADAMTS proteinases at specific sites within highly conserved regions may be important to normal physiological enzyme functions, as well as pathological degradation. METHODS: To examine ADAMTS selectivity, we assayed ADAMTS-4 and -5 cleavage of recombinant bovine aggrecan mutated at amino acids N-terminal or C-terminal to the interglobular domain cleavage site. RESULTS: Mutations of conserved amino acids from P18 to P12 to increase hydrophilicity resulted in ADAMTS-4 cleavage inhibition. Mutation of Thr, but not Asn within the conserved N-glycosylation motif Asn-Ile-Thr from P6 to P4 enhanced cleavage. Mutation of conserved Thr residues from P22 to P17 to increase hydrophobicity enhanced ADAMTS-4 cleavage. A P4' Ser377Gln mutant inhibited cleavage by ADAMTS-4 and -5, while a neutral Ser377Ala mutant and species mimicking mutants Ser377Thr, Ser377Asn, and Arg375Leu were cleaved normally by ADAMTS-4. The Ser377Thr mutant, however, was resistant to cleavage by ADAMTS-5. CONCLUSION: We have identified multiple conserved amino acids within regions N- and C-terminal to the site of scission that may influence enzyme-substrate recognition, and may interact with exosites on ADAMTS-4 and ADAMTS-5. GENERAL SIGNIFICANCE: Inhibition of the binding of ADAMTS-4 and ADAMTS-5 exosites to aggrecan should be explored as a therapeutic intervention for osteoarthritis.

Contributions of Chondroitin Sulfate, Keratan Sulfate and N-linked Oligosaccharides to Inhibition of Neurite Outgrowth by Aggrecan.

The role of proteoglycans in the central nervous system (CNS) is a rapidly evolving field and has major implications in the field of CNS injury. Chondroitin sulfate proteoglycans (CSPGs) increase in abundance following damage to the spinal cord and inhibit neurite outgrowth. Major advances in understanding the interaction between outgrowing neurites and CSPGs has created a need for more robust and quantitative analyses to further our understanding of this interaction. We report the use of a high-throughput assay to determine the effect of various post-translational modifications of aggrecan upon neurite outgrowth from NS-1 cells (a PC12 cell line derivative). Aggrecan contains chondroitin sulfate, keratan sulfate, and N-linked oligosaccharides (N-glycans), each susceptible to removal through different enzymatic digestions. Using a sequential digestion approach, we found that chondroitin sulfate and N-glycans, but not keratan sulfate, contribute to inhibition of neurite outgrowth by substrate-bound aggrecan. For the first time, we have shown that N-linked oligosaccharides on aggrecan contribute to its inhibition of neuritogenesis.

Expression of two novel alternatively spliced COL2A1 isoforms during chondrocyte differentiation.

Alternative splicing of the type II procollagen gene (COL2A1) is developmentally regulated during chondrogenesis. Type IIA procollagen (+ exon 2) is synthesized by chondroprogenitor cells while type IIB procollagen (- exon 2) is synthesized by differentiated chondrocytes. Here, we report expression of two additional alternatively spliced COL2A1 isoforms during chondrocyte differentiation of bone marrow-derived mesenchymal stem cells (MSCs). One isoform, named IIC, contains only the first 34 nucleotides of exon 2 by the use of an alternative 5' splice site, resulting in a premature termination codon and possible nonsense-mediated decay of IIC mRNA. Low levels of the IIC isoform were detected by RT-PCR and Southern analysis of COL2A1 cDNA amplified from differentiating rabbit and human MSCs. A second novel transcript, named IID, arises by the use of another 5' alternative splice site in intron 2. The IID isoform contains exon 2 plus 3 nucleotides, resulting in the insertion of an additional amino acid. The IID isoform was co-expressed with the IIA isoform during chondrogenesis, and was approximately one-third as abundant. Deletion of the IIC alternative 5' splice site from a COL2A1 mini-gene construct resulted in a significant increase in the IIA:IIB ratio. A mutant mini-gene that inhibited production of the IID isoform, however, had differential effects on the production of the IIA and IIB isoforms: this was apparently related to the differentiation status of the cell type used. These data suggest that COL2A1 mRNA abundance and other aspects of chondrocyte differentiation may be regulated by the use of these previously undetermined alternative splice sites.

Mammalian expression of full-length bovine aggrecan and link protein: formation of recombinant proteoglycan aggregates and analysis of proteolytic cleavage by ADAMTS-4 and MMP-13.

Aggrecan, a large chondroitin sulfate (CS) and keratan sulfate (KS) proteoglycan, has not previously been expressed as a full-length recombinant molecule. To facilitate structure/function analysis, we have characterized recombinant bovine aggrecan (rbAgg) and link protein expressed in COS-7 cells. We demonstrate that C-terminally truncated rbAgg was not secreted. Gel filtration chromatography of rbAgg and isolated glycosaminoglycan (GAG) chains, and their susceptibility to chondroitinase ABC digestion indicate that the GAG chains are predominantly CS, which likely occupy fewer serine residues than native aggrecan. To confirm functionality, we determined that rbAgg bound hyaluronan and recombinant link protein to form proteoglycan aggregates. In addition, cleavage of rbAgg by ADAMTS-4 revealed that the p68 form of ADAMTS-4 preferentially cleaves within the CS-2 domain, whereas the p40 form only effectively cleaves within the interglobular domain (IGD). MMP-13 cleaved rbAgg within the IGD, but cleaved more rapidly at a site within the CS domains, suggesting a role in C-terminal processing of aggrecan. Our results demonstrate that recombinant aggrecan can be used for in vitro analyses of matrix protease-dependent degradation of aggrecan in the IGD and CS domains, and both recombinant aggrecan and link protein can be used to study the assembly of proteoglycan aggregates with hyaluronan.

Effects of covalently attached chondroitin sulfate on aggrecan cleavage by ADAMTS-4 and MMP-13.

Aggrecan is degraded by several aggrecanase-1 (ADAMTS-4) isoforms differing in the number of sulfated glycosaminoglycan (sGAG)-binding motifs. ADAMTS-4 and MMPs cleave aggrecan more efficiently within the chondroitin sulfate (CS)-rich region than the interglobular domain (IGD). We investigated the influence of CS on aggrecan core protein cleavage by ADAMTS-4 (p68) and (p40) as well as MMP-13, which has no recognizable GAG-binding sites. Chondroitinase ABC-treated cartilage aggrecan was cleaved with ADAMTS-4 (p68) less efficiently than CS-substituted aggrecan within the CS-2 domain. Keratanase-treated aggrecan exhibited reduced IGD cleavage, but when both CS and KS were removed, the IGD cleavage was restored. This result suggests that KS in the IGD may compete with CS for ADAMTS-4 (p68) binding. In the absence of KS, however, p68 binding was shifted to the CS-2 domain. CS-deficient full-length recombinant aggrecan (rbAgg) was produced by chondroitinase ABC treatment, or by expression in the xylosyltransferase-deficient CHO-pgsA745 cell line. When digested with the ADAMTS-4 (p68), each of these preparations exhibited reduced CS-2 domain cleavage compared to CS-substituted CHO-K1 cell-derived aggrecan. Additionally, CS-deficient rbAgg showed increased IGD scission prior to cleavage within the CS-2 domain. ADAMTS-4 (p40) readily cleaved both rbAggs within the IGD, but cleaved poorly within the CS-2 domain, indicating little CS dependence. MMP-13, in contrast, cleaved the CS region and the IGD of both CS-substituted and CS-deficient rbAgg equally well. These data indicate that covalently bound CS enhances ADAMTS-4-mediated cleavage within the CS-rich region. MMP-13 also cleaves preferentially within the CS-region, but by an apparently CS-independent mechanism.

Chondrogenic potential of progenitor cells derived from human bone marrow and adipose tissue: a patient-matched comparison.

PURPOSE: Stem cell-based tissue engineering represents a possible alternative for the repair of cartilage defects. Both bone marrow and adipose tissue contain pluripotential cells capable of chondrogenesis. This study was a qualitative and quantitative comparison of the chondrogenic potential of progenitor cells isolated from bone marrow aspirates and adipose tissue. METHODS: Bone marrow aspirates (BM) and matching adipose tissue (AD) overlying the posterior superior iliac crest were obtained from patients undergoing elective spine surgery. Chondrogenesis was induced using an established aggregate culture technique. Qualitative analysis was performed by histology and immunohistochemistry. DNA and glycosaminoglycan (GAG) quantitative assays were performed. Quantitative RT-PCR analysis was performed to compare expression of type II collagen between BM and AD aggregates. Osteogenic and adipogenic assays were also performed to confirm pluripotentiality of both AD-derived progenitor cells (ADPC) and BM-derived progenitor cells (BMPC). RESULTS: Toluidine blue metachromasia and type II collagen immunohistochemical staining were more extensive in the aggregates formed by BMPC. Quantitative RT-PCR showed a 500-5000 fold higher expression of type II collagen in the BMPC aggregates. The DNA content was 68% higher in the AD aggregates (p<0.02) but proteoglycan deposition per cell was 120% greater for BM-derived cell aggregates as measured by GAG assays (p<0.05). CONCLUSIONS: The tissue formed by the aggregate culture of the expanded ADPC population was less cartilaginous. It is unclear whether this is because there are fewer chondroprogenitor cells or if the monolayer expansion culture favors cells with higher proliferative rates but without differentiation potential. Under the conditions described in this study, BMPCs may represent a better choice for progenitor cell-based strategies for cartilage repair.

Proteoglycans of reactive rat cortical astrocyte cultures: abundance of N-unsubstituted glucosamine-enriched heparan sulfate.

"Reactive" astrocytes and other glial cells in the injured CNS produce an altered extracellular matrix (ECM) that influences neuronal regeneration. We have profiled the glycosaminoglycan (GAG) component of proteoglycans (PGs) produced by reactive neonatal rat cortical astrocytes, and have quantified their neurite-outgrowth inhibitory activity. PGs extracted from cell layers and medium were fractionated on DEAE-Sephacel with a gradient of NaCl from 0.15 to 1.0 M. Monosaccharide analysis of the major peaks eluting at 0.6 M NaCl indicated an excess of GlcNH2 to GalNH2, suggesting an approximate HS/CS ratio of 6.2 in the cell layer and 4.2 in the medium. Chondroitinase ABC-generated disaccharide analysis of cell and medium PGs showed a >5-fold excess of chondroitin 4-sulfate over chondroitin 6-sulfate. Heparin lyase-generated disaccharides characteristic of the highly sulfated S-domain regions within HS were more abundant in cell layer than medium-derived PGs. Cell layer and medium HS disaccharides contained ~20% and ~40% N-unsubstituted glucosamine respectively, which is normally rare in HS isolated from most tissues. NGF-stimulated neurite outgrowth assays using NS-1 (PC12) neuronal cells on adsorbed substrata of PGs isolated from reactive astrocyte medium showed pronounced inhibition of neurite outgrowth, and aggregation of NS-1 cells. Cell layer PGs from DEAE-Sephacel pooled fractions having high charge density permitted greater NGF-stimulated outgrowth than PGs with lower charge density. Our results indicate the synthesis of both inhibitory and permissive PGs by activated astrocytes that may correlate with sulfation patterns and HS/CS ratios.

Forskolin Stimulates Aggrecan Gene Expression in Cultured Bovine Chondrocytes.

Prostaglandins are autacoids that elevate intracellular 3prime prime or minute:5prime prime or minute-cyclic adenosine monophosphate (cAMP) levels in chondrocytes and other cells in culture. To facilitate intracellular cAMP accumulation, bovine chondrocytes were incubated with forskolin alone or forskolin and isobutylmethylxanthine. Both significantly increased proteoglycan synthesis, which was inhibited by the cAMP-dependent protein kinase inhibitor H89. Sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS/PAGE) on 3--16% gels revealed the presence of two large proteoglycan core proteins which migrated more slowly than the 200-kDa marker protein and two small proteoglycan core proteins which migrated slightly slower than the 46-kDa marker. Northern blot hybridization, employing (32)P-labeled cDNA probes, showed that aggrecan steady-state mRNA levels were increased by forskolin and isobutylmethylxanthine after 1 h and 5 h incubation. Decorin and type II collagen mRNA levels were not altered under these conditions. Link protein mRNA levels were slightly elevated, but only at the 5-h time point. These results indicated that stimulation of intracellular cAMP accumulation by forskolin or forskolin and isobutylmethylxanthine resulted in augmented proteoglycan synthesis via increased steady-state aggrecan mRNA levels. Suppression of proteoglycan synthesis by the cAMP-dependent protein kinase inhibitor H89 suggested that cAMP-dependent protein kinase may also play a role in regulating the synthesis and completion of newly synthesized proteoglycans.

Changes in type II procollagen isoform expression during chondrogenesis by disruption of an alternative 5' splice site within Col2a1 exon 2.

This study describes a new mechanism controlling the production of alternatively spliced isoforms of type II procollagen (Col2a1) in vivo. During chondrogenesis, precursor chondrocytes predominantly produce isoforms containing alternatively spliced exon 2 (type IIA and IID) while Col2a1 mRNA devoid of exon 2 (type IIB) is the major isoform produced by differentiated chondrocytes. We previously identified an additional Col2a1 isoform containing a truncated exon 2 and premature termination codons in exon 6 (type IIC). This transcript is produced by utilization of another 5' splice site present in exon 2. To determine the role of this IIC splicing event in vivo, we generated transgenic mice containing silent knock-in mutations at the IIC 5' splice site (Col2a1-mIIC), thereby inhibiting production of IIC transcripts. Heterozygous and homozygous knock-in mice were viable and display no overt skeletal phenotype to date. However, RNA expression profiles revealed that chondrocytes in cartilage from an age range of Col2a1-mIIC mice produced higher levels of IIA and IID mRNAs and decreased levels of IIB mRNAs throughout pre-natal and post-natal development, when compared to chondrocytes from littermate control mice. Immunofluorescence analyses showed a clear increase in expression of embryonic type II collagen protein isoforms (i.e. containing the exon 2-encoded cysteine-rich (CR) protein domain) in cartilage extracellular matrix (ECM). Interestingly, at P14, P28 and P56, expression of embryonic Col2a1 isoforms in Col2a1-mIIC mice persisted in the pericellular domain of the ECM in articular and growth plate cartilage. We also show that persistent expression of the exon 2-encoded CR domain in the ECM of post-natal cartilage tissue may be due, in part, to the embryonic form of type XI collagen (the α3 chain of which is also encoded by the Col2a1 gene). In conclusion, expression of the Col2a1 IIC splice form may have a regulatory function in controlling alternative splicing of exon 2 to generate defined proportions of IIA, IID and IIB procollagen isoforms during cartilage development. Future studies will involve ultrastructural and biomechanical analysis of the collagen ECM to determine the effects of persistent mis-expression of embryonic collagen isoforms in mature cartilage tissue.

Comparison of sensory neuron growth cone and filopodial responses to structurally diverse aggrecan variants, in vitro.

Following spinal cord injury, a regenerating neurite encounters a glial scar enriched in chondroitin sulfate proteoglycans (CSPGs), which presents a major barrier. There are two points at which a neurite makes contact with glial scar CSPGs: initially, filopodia surrounding the growth cone extend and make contact with CSPGs, then the peripheral domain of the entire growth cone makes CSPG contact. Aggrecan is a CSPG commonly used to model the effect CSPGs have on elongating or regenerating neurites. In this study, we investigated filopodia and growth cone responses to contact with structurally diverse aggrecan variants using the common stripe assay. Using time-lapse imaging with 15-s intervals, we measured growth cone area, growth cone width, growth cone length, filopodia number, total filopodia length, and the length of the longest filopodia following contact with aggrecan. Responses were measured after both filopodia and growth cone contact with five different preparations of aggrecan: two forms of aggrecan derived from bovine articular cartilage (purified and prepared using different techniques), recombinant aggrecan lacking chondroitin sulfate side chains (produced in CHO-745 cells) and two additional recombinant aggrecan preparations with varying lengths of chondroitin sulfate side chains (produced in CHO-K1 and COS-7 cells). Responses in filopodia and growth cone behavior differed between the structurally diverse aggrecan variants. Mutant CHO-745 aggrecan (lacking chondroitin sulfate chains) permitted extensive growth across the PG stripe. Filopodia contact with the CHO-745 aggrecan caused a significant increase in growth cone width and filopodia length (112.7% ± 4.9 and 150.9% ± 7.2 respectively, p<0.05), and subsequently upon growth cone contact, growth cone width remained elevated along with a reduction in filopodia number (121.9% ± 4.2; 72.39% ± 6.4, p<0.05). COS-7 derived aggrecan inhibited neurite outgrowth following growth cone contact. Filopodia contact produced an increase in growth cone area and width (126.5% ± 8.1; 150.3% ± 13.31, p<0.001), and while these parameters returned to baseline upon growth cone contact, a reduction in filopodia number and length was observed (73.94% ± 5.8, 75.3% ± 6.2, p<0.05). CHO-K1 derived aggrecan inhibited neurite outgrowth following filopodia contact, and caused an increase in growth cone area and length (157.6% ± 6.2; 117.0% ± 2.8, p<0.001). Interestingly, the two bovine articular cartilage aggrecan preparations differed in their effects on neurite outgrowth. The proprietary aggrecan (BA I, Sigma-Aldrich) inhibited neurites at the point of growth cone contact, while our chemically purified aggrecan (BA II) inhibited neurite outgrowth at the point of filopodia contact. BA I caused a reduction in growth cone width following filopodia contact (91.7% ± 2.5, p<0.05). Upon growth cone contact, there was a further reduction in growth cone width and area (66.4% ± 2.2; 75.6% ± 2.9; p<0.05), as well as reductions in filopodia number, total length, and max length (75.9% ± 5.7, p<0.05; 68.8% ± 6.0; 69.6% ± 3.5, p<0.001). Upon filopodia contact, BA II caused a significant increase in growth cone area, and reductions in filopodia number and total filopodia length (115.9% ± 5.4, p<0.05; 72.5% ± 2.7; 77.7% ± 3.2, p<0.001). In addition, filopodia contact with BA I caused a significant reduction in growth cone velocity (38.6 nm/s ± 1.3 before contact, 17.1 nm/s ± 3.6 after contact). These data showed that neuron morphology and behavior are differentially dependent upon aggrecan structure. Furthermore, the behavioral changes associated with the approaching growth cone may be predictive of inhibition or growth.

Quantification of type II procollagen splice forms using alternative transcript-qPCR (AT-qPCR).

During skeletal development, the onset of chondrogenic differentiation is marked by expression of the α1(II) procollagen (Col2a1) gene. Exon 2 of Col2a1 codes for a cysteine-rich von Willebrand factor C-like domain. Chondroprogenitors express the exon 2-containing IIA and IID splice forms by utilizing adjacent 5' splice sites separated by 3 base pairs. There is a shift to expression of the shorter, exon 2-lacking IIB splice form with further differentiation. Alternative splicing analysis of Col2a1 splice forms has often relied upon semi-quantitative PCR, using a single set of PCR primers to amplify multiple splice forms. We show that this widely used method is inaccurate due to mismatched amplification efficiency of different-sized PCR products. We have developed the TaqMan®-based AT-qPCR (Alternative Transcript-qPCR) assay to more accurately quantify alternatively spliced mRNA, and demonstrate the measurement of Col2a1 splice form expression in differentiating ATDC5 cells in vitro, and in wild type mouse embryonic and postnatal cartilage in vivo. The AT-qPCR assay is based on the use of a multiple-amplicon standard (MAS) plasmid, containing a chemically synthesized cluster of splice site-spanning PCR amplicons, to quantify alternative splice forms by standard curve-based qPCR. The MAS plasmid designed for Col2a1 also contained an 18S rRNA amplicon for sample normalization, and an amplicon corresponding to a region spanning exon 52-53 to measure total Col2a1 mRNA. In mouse E12.5 to P70 cartilages, we observed the expected switch between the IIA and IIB splice forms; no IID or IIC splice products were observed. However, in the ATDC5 cultures, predominant expression of the IIA and IID splice forms was found at all times in culture. Additionally, we observed that the sum of the IIA, IIB and IID splice forms comprises only a small fraction of Col2a1 transcripts containing the constitutive exon 52-53 junction. We conclude from our results that the majority of ATDC5 cells in the assay described in this study remained as chondroprogenitors during culture in standard differentiation conditions, and that additional Col2a1 transcripts may be present. The validity of this novel AT-qPCR assay was confirmed by demonstrating the expected Col2a1 isoform expression patterns in vivo in developing mouse cartilage. The ability to measure true levels of procollagen type II splice forms will provide better monitoring of chondrocyte differentiation in other model systems. In addition, the AT-qPCR assay described here could be applied to any gene of interest to detect and quantify known and predicted alternative splice forms and can be scaled up for high throughput assays.

Changes of chondrocyte expression profiles in human MSC aggregates in the presence of PEG microspheres and TGF-ß3.

Biomaterial microparticles are commonly utilized as growth factor delivery vehicles to induce chondrogenic differentiation of mesenchymal stem/stromal cells (MSCs). To address whether the presence of microparticles could themselves affect differentiation of MSCs, a 3D co-aggregate system was developed containing an equal volume of human primary bone marrow-derived MSCs and non-degradable RGD-conjugated poly(ethylene glycol) microspheres (PEG-µs). Following TGF-ß3 induction, differences in cell phenotype, gene expression and protein localization patterns were found when compared to MSC aggregate cultures devoid of PEG-µs. An outer fibrous layer always found in differentiated MSC aggregate cultures was not formed in the presence of PEG-µs. Type II collagen protein was synthesized by cells in both culture systems, although increased levels of the long (embryonic) procollagen isoforms were found in MSC/PEG-µs aggregates. Ubiquitous deposition of type I and type X collagen proteins was found in MSC/PEG-µs cultures while the expression patterns of these collagens was restricted to specific areas in MSC aggregates. These findings show that MSCs respond differently to TGF-ß3 when in a PEG-µs environment due to effects of cell dilution, altered growth factor diffusion and/or cellular interactions with the microspheres. Although not all of the expression patterns pointed toward improved chondrogenic differentiation in the MSC/PEG-µs cultures, the surprisingly large impact of the microparticles themselves should be considered when designing drug delivery/scaffold strategies.

Sequence, gene structure, and expression pattern of CTNNBL1, a minor-class intron-containing gene--evidence for a role in apoptosis.

We have identified and characterized a cDNA designated CTNNBL1 (catenin (cadherin-associated protein), beta-like 1) coding for a protein of 563 amino acids having predicted structural homology to beta-catenin and other armadillo (arm) family proteins. CTNNBL1 is expressed in multiple human tissues, and its sequence is conserved across widely divergent species. The human CTNNBL1 gene on chromosome 20q11.2 contains 16 exons spanning > 178 kb. Intron 4 is a minor-class intron bearing AT at the 5' splice site and AC at the 3' splice site. An acidic domain, as well as a putative bipartite nuclear localization signal, a nuclear export signal, a leucine-isoleucine zipper, and phosphorylation motifs are present in the protein sequence. Transient expression of CTNNBL1 in CHO cells results in localization to the nucleus and apoptosis. The rate of cell death was higher when cells were transfected with a carboxy-terminal fragment of CTNNBL1, suggesting that the apoptosis-inducing activity is a function of this region.

Characterization and chondrocyte differentiation stage-specific expression of KRAB zinc-finger protein gene ZNF470.

As part of a study to identify novel transcriptional regulators of chondrogenesis-related gene expression, we have cloned and characterized cDNA for zinc-finger protein 470 (ZNF470), the human ortholog of which encodes a 717 amino acid residue protein containing 17 Cys(2)His(2) zinc-finger domains, as well as KRAB-A and KRAB-B motifs. The cDNA library used to isolate the initial ZNF470 clone was prepared from human bone marrow-derived mesenchymal progenitor cells at an intermediate stage of chondrogenic differentiation. We have determined the intron-exon structure of the human ZNF470 gene, which has been mapped to a zinc-finger cluster in a known imprinted region of human chromosome 19q13.4. ZNF470 is expressed at high levels in human testis and is expressed at low or undetectible levels in other adult tissues. Human ZNF470 expressed in mammalian cells as an EGFP fusion protein localizes predominantly to the nucleus, consistent with a role in transcriptional regulation. ZNF470, analyzed by quantitative real time PCR, was transiently expressed before the maximal expression of COL2A1 during chondrogenic differentiation in vitro. We have also characterized the bovine ortholog of human ZNF470, which encodes a 508 amino acid residue protein having 10 zinc-finger domains. A bovine ZNF470 cDNA clone was used to examine expression of ZNF470 in bovine articular chondrocytes treated with retinoic acid to stimulate dedifferentiation. Bovine ZNF470 expression was undetectable in freshly isolated bovine articular chondrocytes, but was dramatically upregulated in dedifferentiated retinoic acid-treated chondrocytes. These results, in two model systems, suggest a possible role for ZNF470 in the regulation of chondrogenesis-specific gene expression.