Expression of the Chitinase Family Glycoprotein YKL-40 in Undifferentiated, Differentiated and Trans-Differentiated Mesenchymal Stem Cells

The glycoprotein YKL-40 (CHI3L1) is a secreted chitinase family protein that induces angiogenesis, cell survival, and cell proliferation, and plays roles in tissue remodeling and immune regulation. It is expressed primarily in cells of mesenchymal origin, is overexpressed in numerous aggressive carcinomas and sarcomas, but is rarely expressed in normal ectodermal tissues. Bone marrow-derived mesenchymal stem cells (MSCs) can be induced to differentiate into various mesenchymal tissues and trans-differentiate into some non-mesenchymal cell types. Since YKL-40 has been used as a mesenchymal marker, we followed YKL-40 expression as undifferentiated MSCs were induced to differentiate into bone, cartilage, and neural phenotypes. Undifferentiated MSCs contain significant levels of YKL-40 mRNA but do not synthesize detectable levels of YKL-40 protein. MSCs induced to differentiate into chondrocytes and osteocytes soon began to express and secrete YKL-40 protein, as do ex vivo cultured chondrocytes and primary osteocytes. In contrast, MSCs induced to trans-differentiate into neurons did not synthesize YKL-40 protein, consistent with the general absence of YKL-40 protein in normal CNS parenchyma. However, these trans-differentiated neurons retained significant levels of YKL-40 mRNA, suggesting the mechanisms which prevented YKL-40 translation in undifferentiated MSCs remained in place, and that these trans-differentiated neurons differ in at least this way from neurons derived from neuronal stem cells. Utilization of a differentiation protocol containing β-mercaptoethanol resulted in cells that expressed significant amounts of intracellular YKL-40 protein that was not secreted, which is not seen in normal cells. Thus the synthesis of YKL-40 protein is a marker for MSC differentiation into mature mesenchymal phenotypes, and the presence of untranslated YKL-40 mRNA in non-mesenchymal cells derived from MSCs reflects differences between differentiated and trans-differentiated phenotypes.

YKL-40 protein is present in normal human serum at concentrations in the low nanomolar range [10]. It is elevated in the serum of patients and in the affected cells of a number of noncancerous pathological conditions including rheumatoid arthritis [11], asthma [12], and hepatic fibrosis/cirrhosis [13][14][15]. YKL-40 is also synthesized and secreted by a number of human tumors, including tumors of the breast, bone, colon, thyroid, liver, prostate, ovaries, and lung [3,4,14]. Further, YKL-40 serum protein levels were directly correlated with morbidity and or mortality in patients suffering from cancers of the breast [16][17][18], colon [19,20], ovaries [21], and brain [22]. YKL-40 is expressed in a limited number of brain cancers, and is the most highly upregulated gene in high grade glioblastoma multiformae. YKL-40 expression in these high grade glioblastoma multiformae appears linked to the degree of tumor vascularization [23].
Mesenchymal stem cells (MSCs) are multipotent adult stem cells derived primarily from bone marrow, first described as being capable of differentiating into adipocytes, chondrocytes, myoblasts and osteocytes in vitro under defined culture conditions [24]. MSCs have more recently been shown to be able to transdifferentiate into cells of endodermal and ectodermal lineage [25][26][27][28], as well as functional neurons [29], thus MSCs are a relevant system in which to study the molecular details of lineage-specific differentiation. When injected into animal models, MSCs have facilitated the repair and regeneration of damaged tissue types, which has spurred the use of MSCs in clinical trials for a number of disease states, including myocardial ischemia, myocardial infarction, spinal cord injury, stroke, multiple sclerosis, Type I and II diabetes, cirrhosis, kidney transplants, chronic obstructive pulmonary disease, cartilage defects, bone fractures, and graft versus host disease [30].
MSCs differentiate into a number of somatic tissues that express YKL-40, and changes in YKL-40 expression have been noted in differentiation accompanying fetal development in osteogenic and chondrogenic cell lineages [31]. YKL-40 has also been used as a mesenchymal marker in fully differentiated tissues, so we were interested in how YKL-40 is expressed during the formation of these differentiated phenotypes. To this end we examined YKL-40 expression as undifferentiated MSCs were differentiated into osteogenic and chondrogenic phenotypes, as well as during transdifferentiation into neurons. Our specific interest was how YKL-40 expression patterns changed with the differentiation processes, to gain insight into the potential roles of YKL-40 in the mature phenotypes and perhaps in the differentiation processes.

Reagents and Antibodies
All chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise noted. Polyclonal Rabbit anti-YKL-40 antibody was purchased from Quidel (San Diego, CA). Polyclonal chicken anti-Neuron Specific Enolase, Polyclonal chicken anti-b-Tubulin III, Polyclonal rabbit anti-NeuN, Polyclonal rabbit anti-Collagen II, and rabbit anti-chicken HRP antibodies were purchased from Millipore (Temecula, CA). Monoclonal mouse anti-CD44 antibodies were purchased from R&D Systems (Minneapolis, MN). Monoclonal mouse anti-OC4-30 (Osteocalcin) antibodies were purchased from Abcam (Cambridge, MA). Goat anti-rabbit HRP conjugate antibody was purchased from Bio-Rad Laboratories (Hercules, CA). Horse anti-mouse HRP antibodies were purchased from Cell Signaling Technology (Danvers, MA). Mouse monoclonal anti-b-actin HRP conjugate antibody was purchased from Sigma-Aldrich.

Cell Culture
Mesenchymal stem cells were purchased from Lonza (Allendale, NJ). All other cell lines were purchased from the American Type and Culture Collection (ATCC, Manassas, VA) and cultured in typical culture conditions ''TCC'' of DMEM media Mediatech Inc. (Manassas, VA) containing 10% FBS, Hyclone (South Logan, Utah), and 0.292 mg/ml L-Glutamine, Mediatech (Manassas, VA) unless otherwise stated. All cell cultures were grown in vented tissue culture flasks from CorningH, (Corning, NY), or tissue culture chamber slides from Lab-Tek/Nalge Nunc (Naperville, IL) for staining.

Neuronal Differentiation Protocol 1
MSCs of passage 2-5 were differentiated into neurons by first growing them in TCC to 20% confluence, approximately 100,000 cells per tissue culture dish. Then MSCs were exposed for 7, 14 and 30 days to neuronal induction media: Ham's DMEM/F12, (Mediatech Inc.), 2% FBS, 2% B27 supplement, Invitrogen (Carlsbad, CA), 12.5 ng/ml bFGF (Invitrogen), and 20 mM ATRA, with media changes every 3 to 4 days. This protocol was established by Greco and Rameshwar who demonstrated that this procedure resulted in cells that expressed neural markers, synthesized, packaged and secreted neurotransmitters, and produced post-synaptic currents [29].

Neuronal Differentiation Protocol 2
MSCs of passage 2-5 were differentiated into neurons by first growing them in TCC to 20% confluence, approximately 100,000 cells per tissue culture dish. Cells were then pre-induced with DMEM, 20% FBS and 1 mM b-mercaptoethanol (b-ME) for 24 hours. The pre-inducement media was then removed, the cells washed with PBS, and differentiation media composed of: DMEM and 10 mM b-ME was added. The cells were then incubated for 24 hours [32].

Chondrocyte Differentiation using Micromass Pellet Culture (Micropellet)
MSCs of passage 2-5 were removed from culture using a 5 mM EDTA PBS wash, followed by trypsinization. Cells were resuspended in chondrocyte differentiation media at a density of 100,000 cells/10 ml. 10 ml drops were seeded (one 10 ml drop per chamber) into a tissue culture chamber slide and placed in the cell incubator at 37uC, 5% CO2 for 2.5 hours to form 3D cell aggregates. Then chondrocyte differentiation media was gently added to each chamber. Chondrocyte differentiation media consisted of: a-MEM, Lonza (Verviers, Belgium) with 0% FBS, 100 nM dexamethasone, 10 ng/ml TGF b-3, R&D Systems (Minneapolis, MN) & Cell Signaling (Danvers, MA), SITE liquid media supplement, Sigma-Aldrich (as per manufacturer's instructions 1:100), 100 mg/ml Sodium Pyruvate (Sigma-Aldrich), and 1.5 mg/ml BSA (Sigma-Aldrich). Media was replenished every 3 days, and cells were harvested after 28 days [33].

Osteoblast Differentiation
MSC's of passage 2-5 were grown in TCC until 90% confluent (500,000-600,000 per T75 flask). Then a-MEM (Lonza) with 10% FBS, 50 mg/ml ascorbic acid was added on day '09. After 7 days, 3 mM b-glycerol phosphate was added to the existing protocol and the cells exposed for 23 days, with media changes every 3 to 4 days [34].

Isolation of RNA
The initial RNA isolation utilized the RNeasy Minikit from Qiagen, following the manufacturer's instructions. For all subsequent RNA isolations, Trizol was added and cells were lysed directly in the flask (after media removal). The resulting lysate was transferred to 1.5 ml eppi tubes. Samples were then either used or stored at -80uC. 1 ml of each sample of Trizol lysate was then vortexed, and centrifuged at 12,000 g for 10 minutes at 4uC. Cleared supernatants were then moved to new 1.5 ml eppi tubes. Samples were then incubated for 5 minutes at room temperature to allow nucleoprotein dissociation. For phase separation, 0.2 ml of chloroform was added and the tube shook vigorously for 15 seconds, then incubated at room temperature for 3 minutes. Samples were centrifuged at 12,000 g for 15 minutes at 4uC. The clear aqueous phase was moved to a new 1.5 ml eppi tube and RNA precipitation was performed. For RNA precipitation, 5 mg of RNase free glycogen was added to act as an RNA carrier. Then 0.5 ml of 100% isopropanol was added and the sample was incubated at room temperature for 10 minutes. The samples were then centrifuged at 12,000 g for 10 minutes at 4uC. The supernatant was discarded and the resulting RNA pellet was washed two times with 75% ethanol diluted with DEPC treated water. The pellet was then dried for 5 minutes under a laminar flow hood and resolublized in 20 ml of nuclease free water and assessed by the A260/280 ratio method.

Reverse Transcription-PCR
cDNA was produced following the instructions for the Omniscript Reverse Transcription Kit (Quiagen, Valencia, Ca) utilizing 1 mg of template RNA, 10 mg of Oligo (dT)16 primers (Applied Biosystems, Foster City, CA) and 20 units of RNase Inhibitor (Applied Biosystems) per reaction. (In the initial screening assay, Random Hexamer Primers (Quiagen) were utilized in reverse transcription; all subsequent procedures utilized Oligo (dT)16 primers.) PCR products were then produced using the HotStar Taq Plus PCR kit (Quiagen).

Polymerase Chain Reaction
The HotStar HiFidelity Polymerase Chain Reaction kit (Qiagen) was used to amplify cDNAs produced during the RT reaction. Kit directions were followed using 2-5 ml RT reaction products and 0.5 mM of each primer in a total volume was 50 ml per sample (primer sequences shown in Table 1; all primers were made by the Molecular Resource Facility (MRF) of the New Jersey Medical School). For all primer pairs except GAPDH, the amplification cycles consisted of an activation step of 5 minutes at 95uC, then 35 cycles of: (30 seconds at 94uC, 30 seconds at 55uC, 1 minute at 72uC), followed by a clean up step of 10 minutes at 72uC and a 4uC indefinite hold. For GAPDH, the times and temperatures were the same except the temperature of the annealing step was 62uC. Samples removed from the thermocycler were either refrigerated and analyzed by agarose gel electrophoresis within a few days or stored at 220uC.

Agarose Gel Electrophoresis
Samples (25 ml each) were run on a 2% agarose gel in TAE at 50v, using a 5x DNA loading buffer consisting of glycerol (3% v/v) and Fast Orange dye (1% w/v). The gel was then trimmed and stained with ethidium bromide in 1x TAE for K an hour, destained three times for 5 minutes each, using approximately 200 ml of ddH2O and visualized in UV using a Syngene transilluminator and Genesnap software from Synoptics (Frederick, MD).

Western Blot Analysis
Cells were washed three times in ice-cold phosphate buffered saline (PBS) and were lysed with cell lysis buffer (Cell Signaling, Danvers, MA), containing Roche Complete protease inhibitor (Roche Diagnostics, Mannheim, Germany). The cell lysate was then spun at 10,000 g for 20 minutes to remove insoluble cellular materials. The supernatant was then removed to fresh 1.5 ml eppendorf tubes and the protein concentration assessed using the Bio-Rad Protein Assay. SDS-PAGE sample buffer with b-mercaptoethanol was then added to the protein extract and the combination was incubated for 5 minutes at 100uC. Samples were then used immediately or stored at -80uC. SDS-PAGE electrophoresis was then performed using precast 4-20% gradient Tris-HCl gels (Bio-Rad). All gels were pre-run in the presence of a running buffer containing SDS for one hour just prior to use. After gel electrophoretic protein separation, the gel was then blotted overnight onto a PVDF membrane (Bio-Rad), then incubated in a solution of 5% milk in TBST (Tris buffered saline with Tween: 50 mM Tris-HCl, 137 mM NaCl, 0.1% Tween 20) to prevent nonspecific binding. The blot was then incubated with a rabbit anti-YKL-40 (1:1,000) antibody (Quidel), followed by Goat anti-rabbit HRP conjugate (1:5,000) secondary antibody (Sigma). A Coomassie blue gel stain was performed to confirm equal total protein loading. Occasionally mouse monoclonal anti-b-actin HRP conjugate antibody (1:10,000) blotting was performed on a membrane subsequent to stripping with Restore Western Blot Stripping Buffer, Pierce/Thermo Scientific (Rockford, IL) to further confirm protein loading.

Alcian Blue Stain
Chondrocyte like cells produced in the microdot method were first treated with 4% paraformaldehyde/PBS (without magnesium

Alizarin Red S Stain
Osteoblast like cells were first treated with 4% paraformaldehyde/PBS (without magnesium or calcium). 100 ml of 4% paraformaldehyde/PBS was added to the existing 200 ml of osteoblast differentiation media and incubated at room temperature for 5 minutes. Then all fluid was gently aspirated away, and 200 ml of 4% paraformaldehyde/PBS was added to the cell culture and incubated at room temperature for 1 hour. The culture was then washed 2x in PBS and incubated with Alizarin Red S stain solution for 45 minutes in the dark at room temperature. The stain solution was then aspirated and the culture was washed 2x in PBS taking care to not remove any crystals. The culture was then imaged. Alizarin Red S stain solution consists of: 0.2 g of Alizarin Red S dissolved in 10 ml of ddH2O. Just prior to use, the pH of the Alizarin Red S solution was adjusted to pH 5 using NaOH.

Undifferentiated MSCs make YKL-40 mRNA but not YKL-40 Protein
Undifferentiated MSCs have been reported to express YKL-40 mRNA, so initial experiments were performed to establish the levels of YKL-40 mRNA and the corresponding levels of YKL-40 protein synthesized by undifferentiated mesenchymal stem cells [35,36]. As shown in Figure 1, MSCs transcribe significant quantities of YKL-40 mRNA, along with mRNA for the related chitinase family proteins chitinase and chitotriosidase, as well as for MIF (Macrophage migration inhibitory factor). Surprisingly, western blots revealed the absence of YKL-40 protein in both the undifferentiated MSC cell lysate and cell culture media (Figure 2A). To insure that YKL-40 was not being translated and then rapidly degraded by the proteasome, undifferentiated MSCs were incubated with the proteosome inhibitor MG132 for up to 6 hours, however western blot analysis demonstrated that YKL-40 protein was not present in either the cell lysate or media supernatant at any time within this 6 hour period ( Figure 2B). Thus undifferentiated MSCs transcribe measurable amounts of YKL-40 mRNA, but that message is not translated into protein.

MSCs Differentiated into Osteoblasts and Chondrocytes
Synthesize YKL-40 Protein. MSCs Trans-differentiated into Neurons using bFGF and ATRA do not Synthesize YKL-40 Protein but Still Transcribe YKL-40 mRNA Upon differentiation into osteoblast and chondrocyte lineages, significant amounts of YKL-40 protein appeared in both the cell lysates and media supernatants, demonstrating that YKL-40 protein was both synthesized and secreted by these two differentiated phenotypes ( Figure 3A). Protein marker expression was consistent with previous reports: type II collagen is a strong chondrocyte marker but is also expressed in osteocyte progenitors, CD44 is a cartilage marker but is also expressed by osteocytes and osteoclasts, and the bone marker osteocalcin is also expressed in arthritic chondrocytes. However, the differentiated chondrocytes and osteocytes both demonstrated classic phenotype-specific staining with Alcian Blue and Alizarin Red, respectively ( Figure 3B). However, MSCs trans-differentiated into neurons using ATRA and bFGF (protocol I) did not synthesize or secrete YKL-40 protein even after 30 days ( Figure 3C). A number of neuronal marker proteins were upregulated during this procedure, in agreement with previous studies that established significant parallels between these trans-differentiated neurons and normal neurons, including neuronal marker expression, neurotransmitter synthesis and release, and the presence of post-synaptic currents [29]. YKL-40 protein was similarly absent in the neuronal cell line HCN2 ( Figure 3D). Thus neither the trans-differentiated neuronal cells nor the cell line derived from normal neurons express YKL-40 protein, in agreement with the general absence of YKL-40 protein in most CNS tissues. However, RT-PCR demonstrated that the level of YKL-40 mRNA in the bFGF/ATRA transdifferentiated MSCs remained essentially identical to that seen in undifferentiated MSCs, even two weeks after the initiation of differentiation (Figure 4). This suggests that the mechanism which suppresses the translation of YKL-40 mRNA in the undifferentiated MSCs remains in force when MSCs are trans-differentiated into neurons using this protocol.
MSCs Trans-differentiated using ßME Express but do not Secrete YKL-40 Protein ß-mercaptoethanol has been used as a differentiating agent and can stimulate MSC differentiation into a phenotype that has some neural characteristics [32]. MSCs differentiated using BME (protocol II) express YKL-40 protein in the cell lysate at a level comparable to that seen in the cell lysates of the differentiated osteoblasts and chondrocytes, but no YKL-40 protein was observed in the media supernatant ( Figure 5). Thus MSCs differentiated using this protocol synthesize YKL-40 protein but do not secrete it, an unusual behavior not seen in normally differentiated cells.

Discussion
YKL-40 has been utilized as a mesenchymal marker in numerous gene expression array studies, including those of stem cell differentiation and epithelial to mesenchymal transition events in normal and tumor cells [2,23,37]. The presence of YKL-40 mRNA in undifferentiated MSCs has been previously documented [35,36], however the absence of YKL-40 protein translation by undifferentiated MSCs was unexpected and represents an interesting new finding.
One obvious reason why undifferentiated MSCs would store but not translate an mRNA is the protein coded by that mRNA is required for processes in the very early stages of differentiation, and thus having readily available mRNA would allow cells to quickly express that particular protein. Since cells can easily synthesize new proteins within an hour of an initial stimulus, untranslated mRNAs that are stored in cells usually encode genes whose products are needed very early in a stimulus-driven process such as in activation, differentiation or cell stress, thus the metabolic cost of storing the dormant mRNA can be justified.
However, the YKL-40 mRNA did not appear to be immediately translated in the MSCs stimulated to differentiate into osteocytes or chondrocytes -it took at least 24 hours following the addition of either osteoblast or chondrocyte differentiation media for detectable levels of YKL-40 protein to appear (Hoover, unpublished). While we cannot rule out the requirement for YKL-40 in very early differentiation at levels below our detection threshold, our results suggest that YKL-40 is not needed early in these two differentiation pathways that lead to mature YKL-40-synthesizing phenotypes. YKL-40 protein might be required early in one of the MSC differentiation pathways not examined in this study, such as differentiation into adipocytes, although it would be surprising if YKL-40 played a role in one of those pathways but not in the pathways examined considering the well-documented expression of YKL-40 in bone and cartilage. It is also possible that the YKL-40 mRNA may have important roles completely separate from any roles of the protein, for example, in undifferentiated MSCs the untranslated YKL-40 mRNA may function as a ''competing endogenous RNA (ceRNA) and participate in the regulation of the transcriptome as has been recently proposed [38]. These possibilities are intriguing and merit further study.
Our finding that MSCs do not express YKL-40 protein when differentiated into a neuronal phenotype using ATRA and bFGF correlates with the observation that YKL-40 protein is not usually found in normal human CNS neurons, nor generally in normal brain tissues with the exception of low level expression in a subpopulation of microglia and reactive astrocytes [39][40][41]. The ATRA/bFGF/B27 differentiation protocol was previously dem- Figure 1. Screening RT-PCR analysis of transcribed mRNAs in undifferentiated MSCs. Total RNA was isolated from undifferentiated mesenchymal stem cells cultured in tissue culture plates using Trizol. Following reverse transcription using random hexamer primers, the resulting cDNAs were amplified in a series of PCR reactions using primers for YKL-40 and other members of the human chitinase family, as well as primers for two control mRNAs (beta actin and GAPDH), the cytokine MIF, and the NF-kB inhibitor IkB (primer sequences shown in Table 1). PCR reaction products were separated using agarose gel electrophoresis and visualized using EtBr and UV light. doi:10.1371/journal.pone.0062491.g001   onstrated to differentiate MSCs into a phenotype that expressed glial and neuronal progenitor markers (many also assessed in this study in Figure 3C), synthesized, packaged and released neurotransmitters and produced spontaneous post-synaptic currents [29], thus further corroborating the absence of YKL-40 protein production with the neuronal phenotype. While very low levels of YKL-40 mRNA have been reported in homogenates of normal brain tissues [42], the continued presence of high levels of YKL-40 mRNA following ATRA/bFGF/B27 neuronal differentiation suggests that despite the many similarities to mature neurons, some differences exist between normal neurons differentiated from neural stem cells and neurons trans-differentiated from MSCs using this protocol. This also indicates that the mechanisms suppressing YKL-40 translation in undifferentiated MSCs remain active in ATRA/bFGF/B27 differentiated MSCs.
The presence of YKL-40 protein in MSCs following treatment with ß-mercaptoethanol as a trans-differentiating agent (protocol II) suggests that MSCs differentiated using this protocol represent even less of a true neuronal phenotype than those transdifferentiated using the ATRA/bFGF/B27 protocol. While MSCs differentiated using small molecule chemical inducers such as ßmercaptoethanol and DMSO show increases in the expression of some neuronal markers, they also show increased apoptosis and do not show the electrophysiological properties of neurons in patch clamp experiments [43]. It should be noted that any in vitro transdifferentiation protocol may push MSCs to assume a phenotype that is beyond their normal differentiation limits in vivo. Transdifferentiation may result in cells with significant epigenetic differences as compared to the normal phenotype, and care must be taken both in interpreting results of studies that push cells well beyond the natural limits of differentiation, and in utilizing transdifferentiated cells for therapeutic applications [44].
The mechanism of suppression of YKL-40 mRNA translation in undifferentiated MSCs remains unknown. A common mechanism of translational suppression is miRNA binding to mRNA 39 sequences, and reports have documented significant changes in the levels of specific miRNAs in MSC differentiation, including miR-27a, miR-148b and miR-489 in MSC differention into osteoblasts and miR-145 in MSC differentiation into chondrocytes [45,46]. None of these miRNAs were identified by the major miRNA target search algorithms as being able to target YKL-40 mRNA, and the miRNAs that were listed as being able to target YKL-40 mRNA sequences were not among those identified as being significantly up or downregulated during MSC differentiation. Thus if YKL-40 mRNA is being silenced in undifferentiated MSCs by miRNAs, it is being silenced either by miRNA species that bind in a manner that is not recognized by the current algorithms, or by miRNA species whose total levels do not change appreciably during differentiation but whose available levels change due to significant changes in the levels of other binding partners during osteogenic and chondrogenic differentation.
These results further support the utility of YKL-40 protein expression as a marker for MSC differentiation into mature mesenchymal phenotypes as well as a mesenchymal marker in general. In addition, they suggest that the presence of YKL-40 mRNA is indicative of a mesenchymal or a pre-mesenchymal phenotype, or of a trans-differentiated phenotype with at least one residual mesenchymal characteristic. Although these bFGF/ ATRA trans-differentiated neurons appear much like normal neurons -they express multiple neuronal markers, they synthesize, package and secrete neurotransmitters, and they produce postsynaptic currents -the presence of YKL-40 mRNA demonstrates that the trans-differentiated neurons are in at least one way different from normal neurons, which do not contain YKL-40 mRNA. Such differences must be screened and evaluated when trans-differentiated cells are utilized in place of their normally differentiated counterparts.