|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TISSUE-SPECIFIC STEM CELLS |
aInstitut für Laboratoriums und Transfusionsmedizin, Herz und Diabeteszentrum NRW, Universitätsklinik der Ruhr-Universität Bochum, Bad Oeynhausen, Germany;
bInstitut für Transfusionsmedizin und Immunologie, DRK-Blutspendedienst Baden-Württemberg-Hessen, Mannheim, Germany
Key Words. Chondrogenesis • Xylosyltransferase I • Proteoglycans • Matrix deposition • Mesenchymal stem cells Differentiation
Christian Götting, Ph.D., Institut für Laboratoriums- und Transfusionsmedizin, Herz- und Diabeteszentrum NRW, Georgstraße 11, 32545 Bad Oeynhausen, Germany. Telephone: +49-5731-972033; Fax: +49-5731-972013; e-mail: cgoetting{at}hdz-nrw.de
Received October 12, 2005;
accepted for publication June 10, 2006.
First published online in STEM CELLS EXPRESS June 15, 2006.
| ABSTRACT |
|---|
|
|
|---|
-4-N-acetylhexosaminyltransferase), GalNAcT (ß-1,4-N-acetylgalactosaminyltransferase), and GlcAC5E (glucuronyl C5 epimerase). All key enzymes showed a similar regulation with temporarily downregulated mRNA levels (up to 87-fold) after chondrogenic induction. In accordance to previous studies, we observed a similar increase in the expression of PG core proteins. In conclusion, we could show that key enzymes for CS, DS, and HS synthesis, especially XT-I, are useful markers for the developmental stages of chondrogenic differentiation.
| INTRODUCTION |
|---|
|
|
|---|
The potential of adult MSCs to differentiate toward chondrocytes is well established. However, knowledge of remodeling of cartilaginous ECM during the differentiation of MSCs into chondrocytes is still incomplete. Particularly, data about the transcriptional regulation of enzymes and core proteins involved in the biosynthesis of proteoglycans (PGs) are insufficient. Here, we analyzed the mRNA expression from genes coding for PG core proteins as well as the regulation of the key enzymes involved in PG assembly.
Xylosyltransferase I (XT-I) is the key enzyme that initiates the biosynthesis of glycosaminoglycan chains [6, 7]. The enzyme catalyzes the transfer of uridinediphosphate-xylose to specific serine residues of the core proteins in PGs and is located in the early Golgi compartments [8]. Previously, we could show that XT-I is secreted from the Golgi apparatus into the extracellular space together with PGs [911]. Therefore, XT-I is a reliable marker for the PG biosynthesis rate, as our group has demonstrated [9]. In that recent study, we analyzed the expression of XT-I and additionally investigated the regulation of xylosyltransferase II (XT-II). XT-II is a phylogenetically conserved protein that is highly homologous to XT-I. Even though the physiological function of this protein is still unclear, a functional role during the assembly of ECM is probable [12].
After the synthesis of the tetrasaccharid-linker initiated by XT-I, the elongation of glycosaminoglycan chains demands a complex interaction of different enzymes. To analyze a differential expression of heparan sulfate (HS), chondroitin sulfate (CS), and dermatan sulfate (DS) chains during ECM assembly, we determined the expression levels of further key enzymes involved in the differentiating pathways.
To differentiate between the pathways of CS, DS, and HS biosynthesis, we analyzed the expression of the key enzymes:
4-N-acetylhexosaminyltransferase (EXTL2), ß1,4-N-acetylgalactosaminyltransferase (GalNAcT), and glucuronyl C5 epimerase (GlcAC5E). Thereby, EXTL2 transfers GalNAc/GlcNAc from uridinediphosphate-GalNAc/GlcNAc to the tetrasaccharide linker and is therefore a key enzyme for the synthesis of HS chains. GalNAcT initiates the synthesis of CS and DS chains by the transfer of the first GalNAc to the core tetrasaccharide in the protein linkage. According to these GAG chains, an irreversible epimerisation of CS to DS is catalyzed by GlcAC5E [13].
To identify marker genes and proteins involved in ECM remodeling during cartilage differentiation, we subjected human BM-derived MSCs to chondrogenic differentiation in high-density pellet culture and analyzed the expression of marker genes, proteins, and enzymes at different time points.
Reverse transcription-polymerase chain reaction (RT-PCR), histochemical, and immunohistochemical analyses revealed the appropriate phenotype of differentiated cells (positive Alcian Blue, Sirius red, and Safranin O staining, upregulation of collagen type II and aggrecan). During chondrogenic differentiation, stem cells expressed XT-I, XT-II, EXTL2, GalNAcT, GlcAC5E, perlecan, decorin, syndecan-2, glypican-3, aggrecan, and type II collagen. The results of XT-I mRNA expression were confirmed by enzymatic activity analysis. Assembly of PGs was detected by histochemical analysis with Alcian Blue and immunohistochemically with specific antibodies against decorin, CS, and XT-I.
| MATERIALS AND METHODS |
|---|
|
|
|---|
The aspirated cells were diluted 1:5 with phosphate-buffered saline (PBS)/2 mM EDTA (Baxter, Unterschleißheim, Germany, http://www.baxter.de, and Merck, Darmstadt, Germany, http://www.merck.de) and centrifuged on Ficoll-Hypaque gradients (GE Healthcare, Little Chalfont, Buckinghamshire, UK, http://www.gehealthcare.com) at 435g for 30 minutes at room temperature. The interface mononucleated cells (MNCs) were collected, washed, and counted.
BM-derived MNCs were cultivated at a density of 1 x 105/cm2 in 75-cm2 tissue culture flasks (NUNC GmbH & Co. KG, Wiesbaden, Germany, http://www.nunc.de) in MSC growth medium (MSCGM BulletKit; Cambrex, St. Katharinen, Germany, http://www.cambrex.com).
After overnight incubation at 37°C in humidified atmosphere containing 5% CO2, all nonadherent cells were discarded, and the adherent culture was maintained with a twice weekly complete exchange of culture medium. A monolayer of approximately 70% confluency was observed after 710 days of initial plating. The cells were then harvested by using 0.04% trypsin/0.03% EDTA (PromoCell, Heidelberg, Germany, http://www.promocell.com). To expand the cells, they were replated at a density of 4,0005,000 cells per cm2 at passage 1 and thereafter.
MSC identity was proven by performing fluorescence-activated cell sorting analysis and mesodermal differentiation assays (osteogenic, adipogenic, and chondrogenic) as shown previously [14].
Chondrogenic Differentiation
Chondrogenesis was induced in passages 24 as described previously [4]. Briefly, 2.5 x 105 cells were gently centrifuged (150g, 5 minutes) in a 15-ml polypropylene tube. Without disturbing the formed pellet, the cells were cultured in complete chondrogenic differentiation medium containing 0.1 µM dexamethasone (Decadron; Merck & Co., Inc., Whitehouse Station, NJ, http://www.merck.com), 1 mM sodium pyruvate, 0.17 mM ascorbic acid-2-phosphate, 0.35 mM proline, 6.25 µg/ml bovine insuline, 6.25 µg/ml transferrin, 6.25 µg/ml selenous acid, 5.35 µg/ml linoleic acid, and 1.25 mg/ml bovine serum albumin (BSA) (Cambrex) supplemented with 10 ng/ml TGF-ß3 (Strathmann Biotec GmbH & Co. KG, Hamburg, Germany, http://www.biotec-ag.de) by feeding twice a week. The conditioned medium was harvested at the indicated time points. At indicated time points, pellets were harvested in RLT buffer (Qiagen, Hilden, Germany, http://www1.qiagen.com) for the isolation of mRNA or cryosections were performed.
Histological Staining
The sections were fixed with icecold acetone (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) and stained with 0.1% aqueous Safranin O solution (Sigma-Aldrich). Cell nuclei were counterstained with Weigert's iron hematoxylin (Sigma-Aldrich).
Alcian Blue staining was used for histological assessment of the total PG content. Cryosections of differentiated chondrogenic pellets were stained for 30 minutes with 1% Alcian Blue in 3% acetic acid. Cell nuclei were counterstained for 5 minutes with 0.1% Kernechtrot in 5% aluminum sulfate.
The total collagen content was detected by Sirius red staining. Cryosections were stained for 60 minutes with 0.1% Sirius red F3B (Sigma-Aldrich) in an aqueous solution of picric acid (1.2%).
Immunohistochemistry and Densitometric Analysis
MSCs undergoing chondrogenic differentiation were subjected to immunohistochemical analysis for XT-I, CS, and decorin. Cryosections from differentiated pellets were fixed with icecold acetone and preincubated with 2% BSA and 3% fetal calf serum in Tris-buffered saline (TBS). For the detection of XT-I, the sections were incubated with a 1:50 dilution (diluent S0809; Dako, Hamburg, Germany, http://www.dako.com) of anti-XT-I antibody (polyclonal rabbit IgG against human XT-I peptide [15]) for 1 hour at room temperature. After rinsing with TBS, the secondary antibody (Universal LSAB Plus-Kit AP; Dako) was applied at room temperature for 15 minutes. After rinsing with TBS, signals were developed using a Fuchsin-Plus staining solution (Dako) to identify bound antibody. The detection of decorin was carried out using anti-decorin antibody (polyclonal rabbit IgG against human decorin core protein [16]). CS-56 (monoclonal mouse IgM anti-CS; Sigma-Aldrich) was used for the detection of CS. Both antibodies were used at a 1:50 dilution (diluent: S0809; Dako) analogue to the immunohistological staining of XT-I.
For the semiquantitative and quantitative analyses of the relative Fuchsin red content in the histological stainings, a densitometric analysis was performed using ImageJ software (National Institutes of Health, Bethesda, MD, http://www.nih.gov) as described below. The relative content of stained red fractions was determined by color-selective conversion of the red-stained areas and the subsequent analysis of stained pixels. Multiple representative sections (at least six) were analyzed by two independent researchers.
RNA Extraction and Reverse Transcription
The total RNA isolates were obtained from cells within the chondrogenic differentiation using a commercial kit with additional on-column DNase I treatment according to the manufacturer's recommendations (Qiagen). Reverse transcription was performed using SuperScript II Reverse Transcriptase (Invitrogen, Karlsruhe, Germany, http://www.invitrogen.com).
LightCycler Real-Time Quantitative RT-PCR Analysis
The mRNA expression of all target genes was analyzed by a fluorogenic RT-PCR assay using the LightCycler System (Roche, Mannheim, Germany, http://www.roche.com). The PCR for the mRNA quantification was performed using an SYBR Green Taq-DNA polymerase mixture (Platinum SYBR Green qPCR SuperMix-UDG; Invitrogen). Thermal cycling conditions included enzymatic degradation of uracil-containing DNA at 50°C for 2 minutes, activation of the DNA polymerase at 95°C for 2 minutes, followed by 45 cycles at 94°C for 5 seconds, at 58°C for 15 seconds, and at 72°C for 15 seconds. Primers used for a specific amplification of the target genes are shown in Table 1. The transcriptional levels of all target genes were normalized to constant mRNA levels of ubiquitin.
|
Statistical Analysis
Statistical evaluation was performed with the Student's t test. A p value of less than .05 was considered significant. Values of the mRNA expression are expressed in arbitrary units as mean ± SD.
| RESULTS |
|---|
|
|
|---|
|
|
|
|
|
The regulation of GalNacT, which is a key enzyme for CS/DS biosynthesis, was similar to EXTL2. RT-PCR for GalNAcT revealed downregulated levels of mRNA on the first day (2.7 ± 0.4-fold) and minimum levels after 3 days (9.6 ± 0.2-fold) after chondrogenic induction. As with EXTL2, the mRNA expression nearly reached the base level at later stages of chondrogenic differentiation.
For GlcAC5E, which is responsible for the conversion of chondroitin to dermatan and heparin to HS, we found an equal mRNA abundance with minimum expression levels 4 days (10.7 ± 0.5-fold) after lineage-specific differentiation.
Differential mRNA Expression of the PGs Decorin, Aggrecan, Perlecan, Glypican-3, and Syndecan-2
The large aggregating PG aggrecan and the small leucin-rich PG decorin are PGs that contain CS side chains. For both targets, chondrogenic induction caused an early response with an augmentation of the mRNA abundance followed by intermediately downregulated expression levels. In the late stages of differentiation, mRNA levels increased up to threefold over base level (Fig. 3). Thereby, decorin revealed a 2.2-fold (SD 0.11) heightened mRNA abundance 4 weeks after induction with TGF-ß3. Levels of aggrecan mRNA showed a maximum increase after 7 days with a 3.1-fold (SD 0.8) elevation.
Glypican-3, syndecan-2, and perlecan are all PGs mostly associated with HS chains and revealed a differential mRNA expression. In parallel with decorin and aggrecan, the chondrogenic induction caused a comparable mRNA transcription for perlecan. Glypican-3 was found to be a differentially regulated HS PG; we identified glypican-3 with low transcriptional levels at the first days of incubation but with a continuous increase after 5 days with a nearly ninefold augmentation after 4 weeks of differentiation. In contrast, syndecan-2 exhibits only marginal transcriptional changes during chondrogenic differentiation with maximum mRNA levels after 28 days of differentiation with up to 1.1-fold increased mRNA levels (data not shown).
Immunohistochemical Staining of XT-I, Decorin, and CS PGs in High-Density Pellets
Immunohistochemistry of slides cut through the center of the differentiated pellets indicated the highest levels of XT-I directly after chondrogenic induction and a weak staining in the following days (Fig. 6A6C). This observation is in parallel with mRNA and enzymatic data. The deposition was detected in discrete areas around the newly synthesized PGs.
|
| DISCUSSION |
|---|
|
|
|---|
Differential Regulation of XT-I and XT-II During Chondrogenic Differentiation
XT-I catalyzes the initial and rate-limiting step in the biosynthesis of glycosaminoglycan chains in CS, DS, and HS PGs and, therefore, represents a key enzyme for ECM assembly. Increased levels of XT-I mRNA (Fig. 4) and enzymatic activity (Fig. 5) during the early stages of chondrogenic induction show the essential function of this enzyme: building up a distinct ECM that is characteristic for cartilage. These results are in parallel with the localization of the protein in chondrogenic pellets, where we found significantly increased levels of XT-I in areas of newly synthesized matrix (Fig. 6A6C).
Regarding mRNA and enzymatic data, we observed shifted maximums which are plausible because of the different half-life periods of both parameters in vivo. XT-I is secreted from the Golgi apparatus into the extracellular space together with PGs. Therefore, the physiologically stable XT-I enzyme could accumulate in the supernatant, which results in increased enzymatic activities (Fig. 5).
For XT-II, a protein that is highly homologous to XT-I and phylogenetically conserved, we found highly increased mRNA levels after 7 days of chondrogenic induction. As is already known, XT-I is involved in the assembling of ECM, whereas XT-II with a differential transcription may exhibit other physiological functions, which will have to be clarified in further experiments.
Regulatory Response of the CS and DS Biosynthetic Pathway on the TGF-ß3 Induced Chondrogenic Differentiation
In addition to investigating aggrecan, which builds up the major PG in chondrocytes, we investigated the regulation of decorin. It belongs to the group of small leucin-rich PGs and is known to modulate collagen fibrillogenesis [20, 21]. Decorin carries one CS or DS side chain that is probably involved in maintaining the fibril to fibril spacing [22].
Immunohistochemical analysis revealed a broad allocation of the decorin core protein in the chondrogenic cryosections. As expected from the mRNA analysis, we found increasing amounts of decorin in the cryosections during the period of chondrogenic differentiation (Fig. 6D6F). Thereby, the uniform distribution was colocalized with the chemically stained ECM. In this context, the mRNA expression of GlcAC5E, which is responsible for the synthesis of DS [23], is an interesting parameter. After chondrogenic induction, we found significantly downregulated mRNA levels of GlcAC5E after up to 1 week of chondrogenesis and increased levels after 4 weeks. The irreversible conversion of the CS side chain into DS [13] observed in later stages of differentiation may therefore have a regulatory effect on the formation of the collagen fibrils.
For GalNAcT, which is a key enzyme for the CS and DS pathway, we found a similar intermediate transcriptional downregulation directly after chondrogenic induction and increasing amounts of mRNA after 5 days of cultivation. These results were in parallel with the major transcriptional increase of decorin at this point in time.
These data show that the mRNA abundance of key enzymes and core proteins involved in the biosynthesis of PGs with CS or HS chains is equally regulated at early chondrogenic stages, especially after 5 days. Increased levels of decorin and GalNAcT in early stages of matrix deposition are in parallel with the mRNA content of type II collagen. As already reported, decorin was shown to stabilize collagen fibrils in the ECM and orientate fibrillogenesis [2427]. Therefore, we could show that, besides collagen, the matrix formation during chondrogenic development can be characterized by mRNA expression of core proteins and key enzymes involved in the biosynthesis of PGs. Those targets, which are involved in the formation of collagen fibers, are potentially interesting markers for early chondrogenic differentiation.
Regulatory Response of the HS Biosynthetic Pathway on the TGF-ß3 Induced Chondrogenic Differentiation
We examined the regulation of glypican-3, syndecan-2, and perlecan according to the PGs with HS side chains. Glypican-3 and syndecan-2 belong to the group of cell surface PGs and are involved in the initiation and support of chondrogenic development [28, 29]. For glypican-3, we detected significantly increased mRNA abundance starting from early chondrogenic stages (day 5) up to the differentiated chondrocytes (day 28). Starting from day 5, the proportion of transcriptional increase of glypican-3 and EXTL2 was very similar, indicating a direct correlation between the key enzyme involved in HS chain elongation and its target. Interestingly, we found only low transcriptional levels with marginal changes during the differentiation process for syndecan-2. These findings have to be clarified in future studies, which should focus on the role of syndecan-2 in MSC differentiation.
Perlecan represents the major HS PG [30] and exhibits important physiological functions during the matrix maturation of chondrocytes, as shown in previous studies [31, 32]. In this case, we observed heightened mRNA levels in the later stages of lineage-specific differentiation. The transcriptional regulation of EXTL2 was in parallel to GalNAcT, with an intermediate downregulation directly after chondrogenic induction and increasedmRNA levels after 7 days. This indicates that during chondrogenic morphogenesis the expression of both modifying enzymes as well as core proteins of the GAGs responds to the same progression.
Summary
In accordance with our previous studies, we could show that the chondrogenic differentiation of MSCs correlates with the assembly of ECM during the different stages of maturation. Besides different types of collagen, core proteins like aggrecan and perlecan are standard markers for the development of chondrocytes [4, 30, 3234]. The present study reveals that the key enzymes involved in the assembly of GAGs are useful for the detection of matrix deposition during chondrogenesis. Increased levels of XT-I, EXTL2, and GalNAcT indicate newly synthesized matrix, whereas increasing levels of GlcAC5E are associated with an irreversible ECM remodeling of CS toward DS.
According to the progression of the chondrogenic differentiation, we could also show that key enzymes of the GAG synthesis as well as the corresponding core proteins are regulated in a very similar coordinated way. In this context, it is interesting to look at the upregulation of type II collagen in early chondrogenic stages. Collagen and PGs exhibit similar characteristics in the trend of mRNA abundance. This indicates the complex interplay of the different matrix-forming components.
XT-I seems to have an exceptional position during this complex process of ECM production. The expression of this first enzyme in GAG synthesis is significantly increased in early chondrogenic stages, suggesting that XT-I is a limiting factor during PG biosynthesis. These findings are in concordance with other studies in which XT-I could also be identified as the rate-limiting step in PG biosynthesis [8, 35]. Previous studies from our group have shown that serum XT-I activity is a biochemical marker for staging and monitoring the progression of articular cartilage damage in osteoarthritis. Additionally, we could show that mutations in the XT-II gene are in correlation with an earlier manifestation of osteoarthritis [36]. This indicates that the key enzymes responsible for chondrogenic maturation, especially XT-I, are potentially interesting genes in the context of matrix buildup, regeneration, and destruction.
Limitations
The present study reveals new data for the complex process of ECM remodeling and deposition during the chondrogenic differentiation of MSCs. We used different methods to detect changes in the expression of selected targets on the mRNA, protein, or enzymatic activity level. According to the mRNA data, the transcriptional level of mRNA and the corresponding total mRNA content, which is isolated from a cell, do not have to be identical. Therefore, our investigations are based on the cellular mRNA content and are not able to distinguish between an extended mRNA half-life and actually increased transcription rates. To differentiate between different transcriptional mRNA levels, it would be necessary to investigate the mRNA half-life for all target genes at all times of mRNA isolation during chondrogenic differentiation. However, we do have first evidence for cytokine-induced alterations of XT-I expression levels that altered mRNA levels are a cause of different transcriptional activity rather than of mRNA half-life alterations (C. Prante, unpublished data). MSCs do have a capacity for osteogenic differentiation. We have taken into account that ECM molecules like XT-I are also influenced by triggers for an osteogenic differentiation. However, first results on XT-I during osteogenic differentiation of MSCs show different transcriptional activity patterns (data not shown).
| DISCLOSURES |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
|
|
|---|
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| STEM CELLS | THE ONCOLOGIST | CME | ALPHAMED PRESS JOURNALS |
