Zinc-finger protein 145, acting as an upstream regulator of SOX9, improves the differentiation potential of human mesenchymal stem cells for cartilage regeneration and repair

Authors


  • A PCT (international) patent application has been submitted by Genome Institute of Singapore for the use of transcription factor ZNF145 for the repair and regeneration of cartilage.

Abstract

Objective

Human mesenchymal stem cells (hMSCs) represent one of the most promising stem cell therapies for traumatic injury and age-related degenerative diseases involving cartilage. However, few genetic factors regulating chondrogenesis of MSCs have been identified. One study showed that zinc-finger protein 145 (ZNF145), a transcription factor, was up-regulated during 3-lineage differentiation of hMSCs. The present study was undertaken to validate whether this novel transcription factor is useful for the repair and regeneration of cartilage.

Methods

Human MSCs were transfected with lentiviral short hairpin RNA (for small interfering RNA knockdown of ZNF145) and a lentiviral vector for overexpression of ZNF145, and the effects of ZNF145 on chondrogenesis were studied using quantitative polymerase chain reaction and immunostaining. Microarray and transient expression analyses were used to determine whether ZNF145 is a factor operating upstream of SOX9. Allogeneic transplantation of hMSCs into osteochondral defects in rats was performed to determine the effects of ZNF145 on repair of cartilage in vivo.

Results

Small interfering RNA–mediated gene silencing of ZNF145 slowed down chondrogenesis, whereas overexpression of ZNF145 enhanced chondrogenesis. Global gene expression profiling showed up-regulated gene expression in ZNF145-overexpressing MSCs, and transient overexpression of ZNF145 enhanced the expression of SOX9, suggesting that ZNF145 acts as a factor upstream of SOX9, the master regulator of chondrogenesis. Moreover, allogeneic transplantation of hMSCs into osteochondral defects of rat knees showed that ZNF145-overexpressing MSCs repaired cartilage defects better and earlier than empty control MSCs.

Conclusion

These findings suggest that ZNF145 gene therapy may be a very useful strategy for improving the quality of cartilage regeneration and repair.

Chondrogenesis is a tightly regulated process in which multipotent mesenchymal stem cells (MSCs) differentiate into chondrocytes to form cartilage. Although some gain in the understanding of the molecular basis of cartilage differentiation has been made in recent years, relatively few genetic factors involved in chondrogenesis have been identified. Identification of novel molecules regulating chondrogenesis will increase our understanding of the molecular basis of chondrogenesis and may have great potential for clinical application in repairing damaged cartilage.

To understand whether MSCs derived from different tissue sources represent fundamentally similar or different cell types, we compared gene expression profiles of human bone marrow–derived MSCs and adipose tissue–derived MSCs during differentiation toward 3 lineages (osteoblasts, chondrocytes, and adipocytes), and found that ZNF145 (the zinc-finger protein 145 gene, also known as PLZF) was among the most highly and commonly up-regulated genes during 3-lineage differentiation of MSCs (1). ZNF145 encodes a zinc-finger–type transcription factor, comprising 9 Kruppel-like C2H2 zinc fingers located in the C-terminus of the protein (2) and a BTB (bric-a-brac, tram track, broad complex)/POZ (poxvirus, zinc finger) domain in the N-terminal end, which is required for dimerization, transcriptional repression, formation of high molecular weight DNA–protein complexes, nuclear sublocalization, and growth suppression (3).

ZNF145 was first identified in acute promyelocytic leukemia, in which a reciprocal chromosomal translocation, t(11;17)(q23;q21), resulted in a fusion with the RARA gene encoding retinoic acid receptor α (2). ZNF145 is expressed during the development of the central nervous system and limb buds and in the perinatal kidney, liver, and heart (4). ZNF145 also affects myeloid cell growth, differentiation, and apoptosis (5). ZNF145−/− mice display patterning defects (6) and impaired spermatogenesis (7). Although ZNF145 has been shown to regulate osteogenesis as a factor acting upstream of core-binding factor α1 (8), little is known about its role in chondrogenesis. In this study, we present functional and biochemical evidence to indicate that ZNF145 is a novel cartilage-promoting gene that functions as an upstream regulator of SOX9. The findings of our study will provide a very useful strategy for treatment of traumatic joint injuries and degenerative diseases related to cartilage.

MATERIALS AND METHODS

Experiments with human MSCs (hMSCs).

Human bone marrow–derived MSCs were harvested from the iliac crest of human subjects after informed consent had been obtained, in accordance with the guidelines of the Institutional Review Board of the National University Hospital of Singapore. The hMSCs were cultured as described previously (9). To prevent spontaneous differentiation, the cells were maintained at subconfluent levels. MSCs were induced to differentiate toward adipocytes for 14 days in adipogenic medium, and to differentiate toward osteoblasts for 14 days in osteogenic medium, as described previously (1). In addition, a previously described pellet culture system (1, 10) was used to induce differentiation of chondrocytes for 28 days. The medium was replaced every 3–4 days.

Differentiation of MSCs was evaluated by real-time reverse transcription–polymerase chain reaction (RT-PCR) and immunostaining. For the immunostaining analyses of MSCs in each differentiation state, the following materials were used: oil red O stain to detect lipid deposits during adipogenesis, alizarin red S stain to detect calcium deposits and alkaline phosphatase (AP) stain to detect expression of AP during osteogenesis, type II collagen immunostaining to detect the major collagen of cartilage during chondrogenesis, and alcian blue stain to detect cartilage proteoglycans during chondrogenesis. To process the immunostaining of hMSCs (a method also used for the processing of cartilage histologic staining in the rodent studies described below), pellets were fixed in 10% buffered formalin for 2 hours at room temperature, and then cut into 5-μm sections. Staining of the pellet cultures was performed with hematoxylin and eosin, type II collagen immunostaining, and alcian blue stain.

Plasmid construction.

For construction of the short hairpin RNA (shRNA) targeting ZNF145, gene-specific regions, 19 basepairs in length, were designed on the basis of a previously described algorithm (11). The sequence targeted by the shRNA is 5′-GATGTTTGAGATCCTCTTC-3′. The ZNF145 shRNA was cloned into pLentilox3.7.

For overexpression studies, full-length ZNF145 was amplified from complementary DNA (cDNA) of MSCs during osteogenesis for 14 days, and then cloned into pEntry3C (Invitrogen). The SOX9 ultimate open-reading frame clone was from Invitrogen. Using LR reaction between pEntry3C containing the genes of interest and pLenti6/V5 (Invitrogen), the pLentiviral vectors for overexpressing ZNF145 or SOX9 were created.

Generation of stable knockdown of ZNF145 and overexpression of ZNF145 in MSCs by lentiviral infection.

For the generation of MSCs stably overexpressing shRNA against ZNF145 or overexpressing ZNF145 and SOX9, lentiviral plasmids were cotransfected with packaging vectors (Invitrogen) into 293FT cells. Supernatants were harvested after 48 hours.

MSCs were infected with the respective lentiviral supernatants, containing 8 μg/ml Polybrene, to achieve stable knockdown or overexpression. MSCs infected with lentivirus for overexpression of ZNF145 and SOX9 were selected with 5 μg/ml blasticidin for 7 days. An empty vector was used as control.

RT-PCR analyses.

To quantify the effects of ZNF145 overexpression or knockdown on differentiation of MSCs, quantitative real-time PCR was performed with a TaqMan expression assay using the ABI Prism 7900 Sequence Detection System (Applied Biosystems). Briefly, 0.3 μg of total RNA was converted to cDNA using a high-capacity cDNA archive kit in 30 μl, which was then diluted to 300 μl. Quantitative real-time PCR was carried out as follows: initial denaturation for 2 minutes at 50°C and then 10 minutes at 95°C, followed by 40 cycles of PCR (95°C for 15 seconds, 60°C for 1 minute) using 5 μl of 2× Master Mix, 0.5 μl of TaqMan probe, and 4.5 μl of cDNA. All probes were designed with a 5′ fluorogenic 6-FAM and a 3′ quencher TAMRA. The expression of human GAPDH was used to normalize gene expression levels.

To detect messenger RNA (mRNA) expression levels, 1 μg total RNA was reverse transcribed into cDNA using SuperScript II reverse transcriptase (Invitrogen) for subsequent amplification.

Immunofluorescence microscopy.

Cells growing on chambers were washed with phosphate buffered saline (PBS) and fixed with 10% neutral formalin for 15 minutes at room temperature. After 2 washes with Rinse buffer (1× Tris buffered saline plus 0.05% Tween 20 [TBST]), cells were permeabilized with 0.1% Triton X-100/PBS for 10 minutes. The cells were treated with 4% goat serum (blocking buffer) for 30 minutes, and then incubated for 1 hour with a primary antibody against ZNF145 diluted 1:50 in blocking buffer. After 3 washes with Rinse buffer, the cells were incubated with a fluorescein isothiocyanate–conjugated secondary antibody diluted 1:150 in PBS for 45 minutes. After 3 washes, immunolocation was examined with an Olympus fluorescence microscope.

Western blot analysis.

Cells were collected by centrifugation, and cell pellets were resuspended in lysis buffer (25 mM Tris, pH 7.5, 150 mM NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate) containing proteinase inhibitors, after which the pellets were incubated on ice for 30 minutes. Following centrifugation at 16,000g for 15 minutes at 4°C, the supernatant containing total cell extract was collected and kept at −80°C. Protein from cell extracts in the gel was electrophoretically transferred onto a Hybond PVDF membrane (Amersham Biosciences). The membrane was incubated for 1 hour at room temperature in blocking buffer (TBST containing 5% skim milk) to block nonspecific protein binding, and then incubated at room temperature for 1 hour with the primary antibody against ZNF145 (Santa Cruz Biotechnology) or SOX9 (Chemicon) diluted 1:300 in blocking buffer for 1 hour. Following 4 washes with TBST, the membrane was incubated for 1 hour with a horseradish peroxidase–conjugated secondary antibody diluted 1:3,000 in blocking buffer for 1 hour. Antibody binding was visualized with an enhanced chemiluminescence Western blotting detection system (Amersham Biosciences).

Microarray analysis of cDNA.

To determine the targets of ZNF145 in undifferentiated MSCs, we overexpressed ZNF145 and analyzed its gene expression profile in undifferentiated MSCs using microarrays. Total RNA was isolated from ZNF145-overexpressing MSCs and empty control MSCs using RNeasy mini kits (Qiagen), in accordance with the manufacturer's protocol. Briefly, 3.5 μg total RNA was used to synthesize double-stranded DNA, using a one-cycle cDNA synthesis kit. The cDNA was purified using the Sample Cleanup Module. In vitro transcription was performed to produce biotin-labeled complementary RNA (cRNA), using a GeneChip in vitro transcription labeling kit. Biotinylated cRNA was cleaned and fragmented to 50–200 nucleotides with the Sample Cleanup Module and hybridized for 16 hours at 45°C to Affymetrix HG U133 Plus 2, containing more than 54,675 probe sets. After washing, the array was stained with streptavidin–phycoerythrin (Molecular Probes). The staining signal was amplified using biotinylated antistreptavidin (Vector), followed by streptavidin–phycoerythrin staining, and then scanned on a GCOS 3000 (Affymetrix).

The data were analyzed using Software Genespring. A t-test based on the normalized values for intensity followed by the ratio change in intensity (defined as a ratio of normalized intensity greater than or equal to 2 or less than or equal to −2) was used to generate the list of genes that displayed a significant change in expression profile. In this study, MSCs from 2 different patients were used, with samples tested in duplicate.

Animal experiments.

Tumorigenicity in SCID mice and NOG mice.

SCID and NOG mice at 6–8 weeks of age were housed under pathogen-free conditions in a temperature-controlled room, with food and water provided ad libitum. All procedures involving animals were conducted in accordance with national and international regulations. Immortalized MSCs (5 × 106) were injected subcutaneously into the lower back of SCID and NOG mice. Five mice were used for each condition. Tumorigenicity was observed for 12 weeks.

Transplantation of hMSCs into rats, and histologic evaluation of rat cartilage tissue.

Male Sprague-Dawley rats (weighing 500 gm) were anesthetized using an intraperitoneal injection of a mixture of ketamine (10 mg/100 gm) and xylazine (1 mg/100 gm). An anterior midline incision was made through the skin of the knee. The knee joints were opened via a medial parapatellar approach and the patella was everted. An osteochondral defect (1.5 mm in diameter and 1.5 mm in depth) was created in the patellar groove of the distal femur. The pellet, with a diameter of ∼1.5–2.0 mm, was pressed to fit into a defect. The pellets were retained in the defects after relocation of the patella and during free movement of the animal postsurgery. Pellets from ZNF145-overexpressing hMSCs were transplanted into the right knee of rats, and pellets from empty control hMSCs were transplanted into the left knee of rats. The pellets from 3 × 105 ZNF145-overexpressing hMSCs or empty hMSCs were induced into chondrocyte differentiation in vitro, as described previously (1, 10), for 1 week before transplantation. The recipient animals received daily subcutaneous injections of cyclosporine (14 mg/kg; Novartis Pharma) immediately after surgery.

At 6 weeks and 12 weeks after surgery, rats from each group were killed. The distal femurs with defects were collected and fixed in 10% buffered formalin, and the tissue samples were decalcified and cut into 5-μm sections. Staining was performed with hematoxylin and eosin, type II collagen immunostaining for major collagen of cartilage, and alcian blue stain for sulfated proteoglycan matrix of cartilage. Each sample was graded according to a scoring system for histologic features, as described previously (12). Five categories of histologic features were assessed in the rat cartilage tissue, as follows: cell morphology, matrix staining, surface regularity, thickness of cartilage, and integration of donor with host cartilage. The histologic scores ranged from 0 (normal articular cartilage) to 14 (no cartilaginous tissue).

Statistical analysis.

Comparisons of histologic scores between the empty control and ZNF145-overexpressing groups were performed using the Mann-Whitney U test for nonparametric analyses. Otherwise, statistical analyses were performed using Student's unpaired 2-tailed t-tests. P values less than 0.05 were considered statistically significant.

RESULTS

Expression pattern of ZNF145 during in vitro chondrogenesis.

Our previous microarray data showed that ZNF145 was commonly up-regulated during 3-lineage differentiation of MSCs at both early and late stages (1). To validate the microarray data, ZNF145 mRNA levels were determined by real-time RT-PCR analysis of MSCs during 3-lineage differentiation in vitro. The results showed that ZNF145 was up-regulated during 3-lineage differentiation of the MSCs, as compared with that in the undifferentiated control (P < 0.01) (Figure 1A).

Figure 1.

Correlation between the expression of zinc-finger protein 145 (ZNF145) and differentiation of human mesenchymal stem cells (MSCs) into 3 lineages. A, Expression of ZNF145 in MSCs was quantified by real-time reverse transcription–polymerase chain reaction during differentiation of MSCs into adipogenesis (ad) and osteogenesis (os) for 14 days and during chondrogenesis (ch) under pellet culture for 21 days, with the results showing an up-regulation of ZNF145 during 3-lineage differentiation as compared with undifferentiated control MSCs. Bars show the mean ± SD fold change in transcript expression in MSCs from 2 patients, tested in duplicate, relative to control (arbitrarily set at 1). ∗ = P < 0.01 versus control. B and C, Expression of ZNF145 was localized in the nuclei of MSCs during adipogenesis and osteogenesis for 7 days (B) and during chondrogenesis under pellet culture for 7 days (C), whereas ZNF145 was not expressed in undifferentiated control MSCs, as revealed by immunostaining with fluorescein isothiocyanate (FITC) and DAPI. D, Reverse transcription–polymerase chain reaction indicated up-regulated expression of ZNF145 and other chondrogenic markers during in vitro chondrogenesis of MSCs over 21 days; actin was used as a positive control. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.

To determine whether expression of ZNF145 is differentiation state–specific, immunostaining for ZNF145 was performed. Results of immunostaining of the MSCs revealed that ZNF145 was expressed in the nuclei during 3-lineage differentiation, whereas ZNF145 was not expressed in undifferentiated MSCs (Figures 1B and C). To relate the expression of ZNF145 to the stage of in vitro chondrocyte differentiation, MSCs were induced toward chondrocyte differentiation under pellet culture. ZNF145 was up-regulated during chondrocyte differentiation, with concomitant up-regulation of the chondrogenic markers COL2A1 (the gene encoding the type II collagen α1 chain), AGC1 (the gene encoding aggrecan 1), and COL10A1 (Figure 1D). These findings suggest that ZNF145 expression depends on the specific environment and may function differently under different environmental conditions. Thus, we postulate that ZNF145 is involved in chondrogenesis.

Effects of ZNF145 knockdown in slowing down the 3 lineages of differentiation.

To understand the molecular role of ZNF145 in the differentiation of MSCs, we investigated the effects of ZNF145 knockdown in MSCs, achieved by stably transducing the MSCs with an shRNA-expressing vector. Lentiviral shRNA targeting ZNF145 was efficiently introduced into MSCs (Figure 2A), and ZNF145 was knocked down during 3-lineage differentiation of MSCs. The shRNA-mediated gene silencing targeting ZNF145 decreased the transcriptional levels of chondrogenic genes (COL2A1, AGC1, SOX9, and COL10A1) as well as osteogenic genes (for osteocalcin [OC], osteopontin [OPN], and alkaline phosphatase [ALP]) and adipogenic genes (for CCAAT/enhancer binding protein α and peroxisome proliferator–activated receptor γ) during differentiation (all P < 0.01 versus control) (Figures 2B–D).

Figure 2.

Short hairpin RNA (shRNA)–mediated gene silencing of ZNF145 slows down differentiation of human mesenchymal stem cells (MSCs). A, Green fluorescent protein (GFP) on fluorescence microscopy (left), with phase contrast (right), revealed that MSCs were infected with high efficiency by shRNA targeting ZNF145, as compared with empty vector control lentivirus. B–D, Gene expression levels were assessed by real-time reverse transcription–polymerase chain reaction in MSCs induced into chondrogenesis for 28 days and into osteogenesis and adipogenesis for 14 days, after the MSCs had been infected with ZNF145 shRNA or an empty vector control, with the results revealing a decrease in transcript levels of ZNF145, in parallel with a decrease in levels of chondrogenic markers (B), osteogenic markers (C), and adipogenic markers (D), after ZNF145 knockdown as compared with empty vector control. Bars show the mean ± SD fold change in transcript expression in MSCs from 2 patients, tested in duplicate, relative to control (arbitrarily set at 1). ∗ = P < 0.01 versus control. E, Staining revealed that ZNF145 knockdown slowed down the 3-lineage differentiation of MSCs. ZNF145-knockdown MSCs and empty vector control MSCs were induced into in vitro adipogenesis for 14 days, into chondrogenesis under pellet culture for 28 days, and into osteogenesis for 14 days. Effects of ZNF145 knockdown on differentiation of MSCs were assessed by oil red O staining for intracellular lipid-filled droplets in adipogenesis, type II collagen immunostaining and alcian blue staining for sulfated proteoglycan matrix in chondrogenesis, and alizarin red S staining for calcium deposits in osteogenesis. Compared to control, ZNF145-knockdown MSCs showed a notably lower staining intensity in all 3 lineage-differentiation pathways.

These results were consistent with our findings from immunostaining of ZNF145-knockdown MSCs during differentiation into the 3 lineages. We found that ZNF145-knockdown MSCs showed decreased lipid deposits during adipogenesis, as revealed with oil red O staining, decreased immunostaining for COL2A1, as revealed with type II collagen immunostaining, decreased sulfated proteoglycan matrix during chondrogenesis, as revealed with alcian blue staining, and decreased calcium deposits during osteogenesis, as revealed with alizarin red S staining (Figure 2E). These results, observed in MSCs during osteogenesis, are consistent with previous observations from a study in which small interfering RNA duplexes were used (8). Thus, the findings show that ZNF145 plays an important role in the 3-lineage differentiation of MSCs.

Effects of ZNF145 overexpression in improving chondrogenesis and osteogenesis in primary MSCs and an MSC cell line.

To assess whether ZNF145 overexpression improves differentiation of MSCs, the MSCs were infected with a lentivirus for stable ZNF145 overexpression. Immunostaining for ZNF145 showed that ZNF145 was overexpressed in the nuclei of lentiviral-infected MSCs, whereas ZNF145 was not expressed in undifferentiated MSCs (Figure 3A). However, there were no overt signs of differentiation of ZNF145-overexpressing MSCs.

Figure 3.

ZNF145 overexpression improves chondrogenesis and osteogenesis of human mesenchymal stem cells (MSCs). A, ZNF145 overexpression was detected in the nuclei of MSCs infected with a lentiviral system, whereas ZNF145 was not detected in undifferentiated MSCs, as revealed by immunostaining with fluorescein isothiocyanate (FITC) and DAPI. B and C, Gene expression levels were assessed by real-time reverse transcription–polymerase chain reaction in ZNF145-overexpressing MSCs and empty vector control MSCs induced into chondrogenesis for 28 days and osteogenesis for 14 days. ZNF145 overexpression enhanced the transcript levels of chondrogenic markers in MSCs during chondrogenesis (B) and osteogenic markers in MSCs during osteogenesis (C) as compared with empty vector control. Bars show the mean ± SD fold change in transcript expression in MSCs from 2 patients, tested in duplicate, relative to control (arbitrarily set at 1). ∗ = P < 0.01; ∗∗ = P < 0.05 versus control. D, ZNF145 overexpression enhanced staining for chondrogenic differentiation of MSCs by type II collagen immunostaining and alcian blue staining for sulfated proteoglycan matrix in cartilage under chondrogenic medium for 28 days, and staining for osteogenic differentiation by alizarin red S staining for calcium deposits in MSCs induced into osteogenesis under osteogenic medium for 14 days, as compared with empty vector control. E and F, Staining for alkaline phosphatase (AP) (E) and an assay for AP activity (F) showed enhanced expression of AP in ZNF145-overexpressing MSCs as compared with empty vector control MSCs during osteogenesis for 14 days. Bars show the mean ± SD arbitrary units of AP activity in MSCs from 2 patients, tested in duplicate. ∗∗ = P < 0.05 versus control. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.

The ZNF145-overexpressing MSCs were then induced into chondrogenesis using pellet cultures and into osteogenesis. Overexpression of ZNF145 in MSCs undergoing chondrogenesis increased the expression of the chondrogenic markers COL2A1, AGC1, COL10A1, and SOX9, as compared with that in empty controls (Figure 3B), showing that ZNF145 enhances chondrocyte differentiation of MSCs during chondrogenesis. This was consistent with the finding that ZNF145 overexpression enhanced type II collagen immunostaining and alcian blue staining for sulfated proteoglycan matrix in cartilage during chondrogenesis, as compared with that in empty controls (Figure 3D).

During in vitro osteogenesis, ZNF145 overexpression increased the expression of the osteogenic markers OPN, OC, and ALP, as compared with that in empty controls (Figure 3C). This finding was consistent with our observations of enhanced calcium deposits in ZNF145-overexpressing MSCs undergoing osteogenesis, as revealed by alizarin red staining (Figure 3D), as well as increased levels of AP, according to AP staining (Figure 3E) and AP assay (Figure 3F), showing that ZNF145 overexpression also leads to improvement of osteogenesis. These findings are consistent with the results from a prior study using a nonviral system (8).

Bone marrow–derived MSCs may pose a problem, because they have a limited lifespan and show variance from donor to donor. To overcome these disadvantages, immortalized MSCs were generated via a retroviral system, with a combination of human telomerase reverse transcriptase and antigen large T from SV40. Immortalized MSCs displayed a similar surface antigen profile as that of bone marrow–derived MSCs and showed potential for normal differentiation toward the 3 lineages (chondrocytes, osteoblasts, and adipocytes) (details available from the corresponding author upon request).

To test the tumorigenesis of immortalized MSCs, 5 × 106 immortalized MSCs per site were subcutaneously transplanted into nude mice and NOG mice. No tumors were observed until 12 weeks posttransplantation (results not shown). Interestingly, ZNF145 overexpression had similar effects on immortalized MSCs as that observed on bone marrow–derived MSCs, in that the overexpression of ZNF145 enhanced the expression of chondrogenic markers under chondrogenesis and osteogenic markers under osteogenesis in immortalized MSCs, as compared with that in empty controls (results not shown). This was consistent with our findings of enhanced type II collagen immunostaining and alcian blue staining for sulfated proteoglycan matrix in cartilage during chondrogenesis, as well as enhanced calcium deposits, as revealed by alizarin red staining, and increased levels of AP, as revealed by AP staining and AP assay, in immortalized MSCs during osteogenesis, as compared with that in empty controls (results available from the corresponding author upon request).

Targets of ZNF145 in MSCs.

To explore the molecular mechanisms underlying the role of ZNF145 in chondrogenesis, we examined the repertoire of genes in ZNF145-overexpressing MSCs. The expression profile revealed that 423 genes were up-regulated, whereas 678 genes were down-regulated by ZNF145 overexpression in undifferentiated MSCs (results not shown). MSCs from 2 patients showed a similar pattern of up-regulated gene expression upon ZNF145 overexpression (Figure 4A).

Figure 4.

Role of ZNF145 as an upstream factor of SOX9 in undifferentiated human mesenchymal stem cells (MSCs). A, The heat map shows the genes found to be up-regulated (red) or down-regulated (green) in ZNF145-overexpressing MSCs. ZNF145-overexpressing MSCs from 2 patients showed similar gene expression profiles. All microarray data are available from the Gene Expression Omnibus (available at http://www.ncbi.nlm.nih.gov/geo) under accession code GSE26272. B, The microarray data were verified by reverse transcription–polymerase chain reaction analysis, showing a similar pattern of up-regulated gene expression in ZNF145-overexpressing MSCs compared to empty vector controls. C and D, Overexpression of ZNF145 was found to up-regulate the expression of SOX9 in undifferentiated MSCs. The MSCs were infected with a lentivirus for overexpressing ZNF145 and SOX9 for 2 days, while empty vector was used as the control. ZNF145 increased SOX9 expression, whereas SOX9 did not increase the expression of ZNF145 mRNA (C) or protein (D), suggesting that ZNF145 is an upstream regulator of SOX9. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.

The expression of selected genes identified in the parallel samples by microarray was subsequently compared using RT-PCR analysis for validation. Results from the RT-PCR assays were consistent with the microarray data (Figure 4B). Intriguingly, the expression of SOX9, the master regulator of chondrogenesis (13), was increased 3.65-fold. Expression of HAPLN1 (the gene for hyaluronan and proteoglycan link protein 1), an extracellular matrix protein in cartilage and important for the formation of proteoglycan aggregates and normal organization of hypertrophic chondrocytes (14), was increased 3-fold. Expression of GDF5 (the gene for growth differentiation factor 5), a mutation of which results in acromesomelic chondrodysplasia Grebe type (a condition characterized by short stature, very short limbs, and hand/foot malformations [15]), was increased 3.5-fold. Expression of ALPL (alkaline phosphatase, liver/bone/kidney), an early marker of bone formation, was up-regulated 2.4-fold. Expression of SFRP1 (the gene for secreted frizzled related protein 1), which regulates Wnt signaling for bone morphogenetic protein 2–induced chondrocyte differentiation (16), was increased 10.9-fold. Expression of TGFB1 (the gene for transforming growth factor β1), which modulates the induction of chondrogenesis, was increased 2-fold. Thus, the up-regulation of genes related to chondrogenesis may be responsible for the improved cartilage differentiation by ZNF145 overexpression.

Regulation of SOX9 by ZNF145 as an upstream factor.

SOX9 is the master regulator during chondrogenesis. It functions by directly binding and activating chondrocyte-specific enhancer elements, including COL2A1 (17–19), AGC1 (20), CRTL1 (the gene for cartilage link protein) (21), COL11A2 (22), and cartilage-derived retinoic acid–sensitive protein (23). To understand how ZNF145 functions in chondrogenesis, it is crucial to determine whether ZNF145 acts as an upstream regulator or a downstream factor of SOX9. Our microarray data showed that SOX9 is one target of ZNF145. To validate this, SOX9 and ZNF145 were transiently introduced into MSCs for 2 days using the lentiviral system. Overexpression of ZNF145 enhanced the expression of SOX9, whereas overexpression of SOX9 did not enhance the expression of ZNF145 mRNA (Figure 4C) or protein (Figure 4D) in undifferentiated MSCs. These findings suggest that ZNF145 is an upstream regulator of SOX9.

Improvement in repair of cartilage defects by ZNF145 in an in vivo rat model.

To assess whether ZNF145-overexpressing MSCs could improve the repair of a cartilage defect in vivo, we compared ZNF145-overexpressing MSCs with control empty vector–transfected MSCs that had been subjected to in vitro chondrogenesis for 7 days under pellet culture and then transplanted into osteochondral defects of rat knees. Six weeks after transplantation, the defects were filled with reparative tissue from transplanted pellets that resembled hyaline cartilage. The superficial layers from ZNF145-overexpressing MSC transplants had more intense matrix staining compared with control MSCs. At higher magnification, the cells resembled well-differentiated chondrocytes and were surrounded by metachromatic matrix.

The ZNF145-overexpressing group showed continuous and similar alcian blue staining for sulfated proteoglycan matrix and type II collagen immunostaining for the major collagen of cartilage, similar to that in adjacent cartilage, whereas empty control MSCs showed discontinuous alcian blue staining and type II collagen immunostaining at the sites of defects. Most importantly, cartilage from ZNF145-overexpressing MSCs integrated well into native cartilage (Figures 5A and B).

Figure 5.

ZNF145 improves osteochondral defect repair in a rat model. ZNF145-overexpressing and empty vector control human mesenchymal stem cells (MSCs) were induced into cartilage differentiation for 7 days under pellet culture, and then the pellets were transplanted into osteochondral defects of rat knees for 6 and 12 weeks. The results of immunostaining (indicated by broken and solid red lines in A, B, D, and E) and histologic scoring of cartilage (C and F) showed that the cartilage transplanted with ZNF145-overexpressing MSCs produced better and earlier repair of the osteochondral defects than did the empty vector control at 6 weeks (A–C) and 12 weeks (D–F). Immunostaining was done using hematoxylin and eosin (HE), type II collagen immunostaining, and alcian blue staining. Scoring of histologic features (lower scores indicate more improved histologic features [see ref.12]) showed significant differences in the repair of cartilage defects between the groups (P < 0.01). In C and F, data are shown as box plots, where each box represents the 25th to 75th percentiles, the lines outside the boxes represent the 10th and 90th percentiles, and the lines inside the boxes represent the median (n = 7 in C and n = 9 in F). Bars in A and D = 1 mm; bars in B and E = 200 μm. Color figure can be viewed in the online issue, which is available at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.

Histologic scores, as described previously (12), showed that the rat knees transplanted with ZNF145-overexpressing MSCs had significantly better scores for histologic features of the cartilage than did the control group (P < 0.01) (Figure 5C). These results showed that ZNF145-overexpressing MSCs repaired cartilage defects much better and earlier than the empty vector control group. Similar to the reparative effects at 6 weeks, reparative cartilage from ZNF145-overexpressing MSCs at 12 weeks integrated well to both edges of the adjacent cartilage, when compared with cartilage from control MSCs, as shown by type II collagen immunostaining and alcian blue staining (Figures 5D and E). Moreover, at 12 weeks, the histologic scores showed that the ZNF145-overexpressing group had much better scores for histologic features of the cartilage than did the control group (P < 0.01) (Figure 5F). These results showed that ZNF145 improved the quality of repair of cartilage defects and was able to do it better and earlier than the control. These findings suggest that manipulation of ZNF145 activity may be a useful therapeutic strategy for cartilage regeneration and repair.

DISCUSSION

Diseases of the cartilage are most often due to trauma or wear and tear in age-related degenerative disease or degenerative joint disease. Cartilage defects are most commonly seen in the knee joint, and treatment is a real challenge for the orthopedic surgeon. Injuries of this tissue, which has limited potential to heal, often cause pain and disability. Treatment of symptomatic osteochondral defects has been especially difficult, and most surgical techniques, such as drilling or microfracture, have resulted, at best, in healing with fibrocartilage. More recently, transfer of hMSCs has been attempted as a method for repair and regeneration of chondral defects (24).

Although zinc-finger proteins represent the most abundant protein superfamily in mammals, relatively few zinc-finger proteins related to chondrogenesis have been functionally characterized (25–27). The results of the present study showed that ZNF145 overexpression greatly enhanced chondrogenesis as well as osteogenesis of bone marrow–derived and immortalized MSCs in vitro. When compared with the effects of genes enhancing only one lineage (chondrogenesis or osteogenesis) of MSCs, we postulated that improvement in the quality of MSCs by ZNF145 would be more comprehensive, and thus would provide a very useful strategy for clinical application. Transplantation of ZNF145-overexpressing and empty control MSCs into an osteochondral defect in rat knee cartilage showed that ZNF145-overexpressing MSCs repaired the cartilage defect better and earlier than did empty control MSCs. These data suggest that genetic modification with a specific gene, for example, ZNF145, or a combination of genes may be a useful strategy for improving the quality of cartilage regeneration and repair.

SOX9 has been identified as a master transcription factor in chondrocyte differentiation (13). One study showed that cartilage formation was sensitive to the dosage of SOX9, in that low levels of SOX9 overexpression enhanced COL2A1 gene transcription in chondrocytes, whereas high levels of SOX9 overexpression induced an inhibition of COL2A1 gene expression, regardless of the differentiation state of the chondrocytes (28). It was reported that adenoviral-overexpressed SOX9 in pellet-cultured MSCs failed to induce sufficient chondrogenesis (29).

Since ZNF145 is involved in chondrogenesis and SOX9 is a master regulator of chondrogenesis, it is crucial to determine whether ZNF145 is a factor that acts upstream or downstream of SOX9. Our microarray and transient overexpression data demonstrated that SOX9 was up-regulated by ZNF145 in undifferentiated MSCs. However, ZNF145-overexpressing MSCs retained the morphologic features of MSCs and underwent 3-lineage differentiation, suggesting that SOX9 up-regulation by ZNF145 is not sufficient to prime for chondrogenesis. ZNF145 overexpression enhanced the expression of SOX9, whereas SOX9 overexpression did not affect the expression of ZNF145. These data suggest that ZNF145 is a factor operating upstream of SOX9 in regulating chondrogenesis.

The precise mechanistic role for ZNF145 in greatly improving chondrogenesis needs more clarification. Our findings suggest that at least one of the important effects of ZNF145 on chondrogenesis from undifferentiated MSCs is the induction of expression of SOX9, a master chondrogenesis gene, as well as other chondrogenic markers, such as HAPLN1, GDF5, and TGFB1.

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. Lim 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. Liu, Lim, Lee.

Acquisition of data. Liu, Guo, Tan, Hui.

Analysis and interpretation of data. Liu, Guo, Lim, Lee.

Acknowledgements

We are grateful to Dr. Wen Cai Zhang for performing the transplantation of immortalized MSCs into NOG mice. We also thank Dr. Xia Fei Ren for assistance in rodent surgeries.

Ancillary