Prostaglandin E2 (PGE2) Exerts Biphasic Effects on Human Tendon Stem Cells

Prostaglandin E2 (PGE2) has been reported to exert different effects on tissues at low and high levels. In the present study, cell culture experiments were performed to determine the potential biphasic effects of PGE2 on human tendon stem/progenitor cells (hTSCs). After treatment with PGE2, hTSC proliferation, stemness, and differentiation were analyzed. We found that high concentrations of PGE2 (>1 ng/ml) decreased cell proliferation and induced non-tenocyte differentiation. However, at lower concentrations (<1 ng/ml), PGE2 markedly enhanced hTSC proliferation. The expression levels of stem cell marker genes, specifically SSEA-4 and Stro-1, were more extensive in hTSCs treated with low concentrations of PGE2 than in cells treated with high levels of PGE2. Moreover, high levels of PGE2 induced hTSCs to differentiate aberrantly into non-tenocytes, which was evident by the high levels of PPARγ, collagen type II, and osteocalcin expression in hTSCs treated with PGE2 at concentrations >1 ng/ml. The findings of this study reveal that PGE2 can exhibit biphasic effects on hTSCs, indicating that while high PGE2 concentrations may be detrimental to tendons, low levels of PGE2 may play a vital role in the maintenance of tendon homeostasis in vivo.


Introduction
Tendons transmit muscular forces to bone and, as a result, they are subjected to large mechanical loads in vivo. Consequently, tendons are frequently injured, especially during intense sport activities. Tendon injuries are generally difficult to treat; tendinopathy, a chronic tendon disorder involving tendon inflammation and/or degeneration, is a particularly significant challenge in orthopaedics and sports medicine. Thus far, strategies that stimulate the complete regeneration of tendons after injury have not been developed. To this end, a better understanding of tendon cell biology is essential to devise improved treatment options for tendon injuries such as tendinopathy [1].
One of the major causative factors that contribute to the development of tendinopathy is excessive mechanical loading (or overuse and over-loading) placed on tendons [2,3]. Such excessive mechanical loading has been shown to increase the production of prostaglandin E 2 (PGE 2 ) in cultures of human tendon fibroblasts (tenocytes) in vitro [3,4]. In addition, PGE 2 production was shown to increase after exercise in the peritendinous space of Achilles tendons in vivo [5].
Although PGE 2 levels increase after mechanical loading, baseline levels of PGE 2 are present in the patellar and Achilles tendons of mice under normal conditions without mechanical loading such as treadmill running [6]. This suggests that PGE 2 could have an impact on the tendon stem/progenitor cells (TSCs) that reside in tendons [6][7][8] and could play an important physiological role in the maintenance of tendon homeostasis. Therefore, PGE 2 may have biphasic effects depending on its concentration. A better understanding of the concentrationdependent effects of PGE 2 on tendon cells, particularly TSCs, may shed new light on tendon physiology and pathology. Thus, in this study we hypothesized that lower concentrations of PGE 2 increase TSC proliferation and decrease non-tenocyte differentiation of TSCs, while higher concentrations produce the opposite effects. To test this hypothesis, we carried out cell culture experiments by treating human TSCs (hTSCs) with low and high levels of PGE 2 . We also performed in vivo implantation experiments to determine the differentiation fate of hTSCs after treatment with various concentrations of PGE 2 in vitro.

Ethics Statement
The Gift of Hope Organ and Tissue Donor Network (Elmhurst, IL) provided normal human knee tissues, after obtaining written consent from donors' families and approval from the local ethics committee (Gift of Hope Organ and Tissue Donor Network). The University of Pittsburgh IRB also approved the study protocol for using human tendon tissues in the cell culture and animal studies performed in this study. These specimens were used for investigation only and no human subjects were involved in this project. Data obtained for the study was not through intervention or interaction with individuals and does not have any identifiable private information. Further, the University of Pittsburgh IACUC approved the protocol for the use of rats in the in vivo implantation experiments.
hTSC Culture hTSCs were isolated from the patellar tendons of six human donors (20 to 44 years old) using our previously published protocol [8]. Briefly, after removing the paratenons, the core portions of the patellar tendons were cut into small pieces and digested with collagenase type I (3 mg/ml) and dispase (4 mg/ml) at 37uC for 1 hr. After centrifugation at 3,000 rpm for 15 min and removal of the enzyme-containing supernatant, a single-cell suspension was obtained, which was cultured in growth medium (DMEM plus 20% FBS) at 37uC with 5% CO 2 . After 8 to 10 days in culture dishes, hTSCs formed colonies. The stem cell colonies were then isolated and cultured in DMEM with 20% FBS. These hTSCs at passage 1 were used in the following experiments.

Verification of the Stemness of hTSCs
The stemness of human tendon stem cells (hTSCs) from the patellar tendon used in this study was verified by immunocytochemical analysis of three stem cell markers, including octamerbinding transcription factor 4 (Oct-4), Nanog, and nucleostemin (NS). hTSCs were first seeded into 12-well plates at a density of 20,000 cells/well with 1.5 ml medium and cultured for 3 days. Then, the hTSCs were fixed in 4% paraformaldehyde in PBS for 20 min at room temperature and washed in 0.5% Triton-X-100 in PBS for 15 min. Subsequently, the fixed cells were incubated with either mouse anti-human Oct-4 (1:500), rabbit anti-human Nanog (1:500), or goat anti-human nucleostemin (1:500) overnight at 4uC. After washing three times with PBS, the cells were again incubated for 2 hrs at room temperature with either Cy3-conjugated goat anti-mouse IgG antibodies (1:1000) for Oct-4, Cy3-conjugated goat anti-rabbit IgG (1:500) for Nanog, or Cy3-conjugated donkey anti-goat IgG antibodies (1:500) for Nucleostemin. Nuclei were stained with Hoechst fluorochrome 33342 (1 mg/ml; Sigma, St. Louis, MO). Stained cells were then examined using fluorescence microscopy. All antibodies were obtained from Chemicon International (Temecula, CA), BD Biosciences (Franklin Lakes, NJ), Neuromics (Edina, MN), or Santa Cruz Biotechnology Inc. (Santa Cruz, CA).

Measurement of Proliferation of hTSCs Treated with PGE 2
hTSCs were seeded in 6-well plates (6610 4 /well) and six different concentrations of PGE 2 (0, 0.01, 0.1, 1, 10, and 100 ng/ ml) were added to the culture. Three replicates were maintained for each concentration. The medium was changed every day and PGE 2 was replenished. After 6 days, cell number was measured using a digital cellometer (Nexcelcom Bioscience, Lawrence, MA), and the population doubling time (PDT), which is a measure of cell proliferation, was calculated based on the formula: log 2 [Nc/ N 0 ], where Nc is the total number of cells at confluence, and N 0 is the initial number of cells seeded [8].

Determination of the Effect of PGE 2 Treatment on hTSC Stemness
Stemness of hTSCs was determined by immunocytochemistry and FACS analysis. For immunocytochemistry, hTSCs were seeded in 12-well plates (3610 4 /well) and treated with six different PGE 2 concentrations ranging from 0 to 100 ng/ml for 5 days, with three replicates for each concentration. The effect of PGE 2 treatment on hTSC stemness was then determined by performing immunocytochemistry for stem cell markers SSEA-4 and Stro-1. Briefly, cells were fixed in 4% paraformaldehyde in PBS for 30 min at room temperature. After washing with PBS, the cells were incubated at room temperature with mouse anti-human SSEA-4 (1:350; Invitrogen, Cat. # 414000) for 3 hrs or mouse anti-human Stro-1 (1:200; Invitrogen, Cat. # 398401) for 4 hrs. The cells were then washed three times with PBS, followed by incubation with Cy3-conjugated goat anti-mouse IgG (1:500; Invitrogen, Cat. # A10521) secondary antibody at room temperature for 2 hrs. After a final wash with PBS, the nuclei were stained with Hoechst fluorochrome 33342, as described above. Stained cells were examined and images of cells were obtained using a fluorescence microscope (Nikon eclipse microscope, TE2000-U).

Semi-quantification of Positively-stained hTSCs
For the semi-quantification of stem cell markers in vitro, seven random images were captured from each well at a magnification of 20x under the Nikon eclipse microscope. The positively-stained cells in each picture were manually identified and analyzed using SPOT TM imaging software (Diagnostic Instruments, Inc., Sterling Heights, MI). The positive staining percentage was calculated by dividing the number of positively-stained cells by the total number of cells under the microscopic field. The average value of all seven images from each well represented the percentage of positive staining, which indicates the stemness of hTSCs in the respective PGE 2 concentrations.

Fluorescence Activated Cell Sorting (FACS) Analysis of hTSCs
To determine the effect of PGE 2 treatment on hTSC stemness by FACS analysis, hTSCs (1.5610 6 in 50 ml PBS) were incubated with 20 ml of the appropriate serum in a centrifuge tube at 4uC for 30 min. Subsequently, 0.4 mg of mouse anti-human SSEA-4 (Cell Signaling, Cat. #4755S) or mouse anti-human Stro-1 (Millipore, Cat. #MAB4315) primary antibody was added and incubated at 4uC overnight. The cells were then washed three times with 2% FBS-PBS, followed by centrifugation at 500 g for 5 min/each time. Then the cells were treated with 1 mg Cy3 conjugated goat anti-mouse IgG secondary antibody at room temperature for 2 hrs. The cells treated with the second antibody only were used as a staining negative control. Finally, the cells were washed twice with PBS and fixed in 1% paraformaldehyde, followed by FACS analysis on a BD LSR II Flow Cytometer (BD Biosciences).

Determination of hTSC Differentiation in vitro by qRT-PCR
To determine the effect of PGE 2 treatment on the differentiation of hTSCs, we performed quantitative RT-PCR (qRT-PCR) to measure gene expression using a QIAGEN QuantiTect SYBR Green PCR Kit (QIAGEN). Briefly, total RNA was isolated from hTSCs using the RNeasy Mini Kit with an on-column DNase I digest (Qiagen, Valencia, CA). Then first-strand cDNA was reverse transcribed using SuperScript II (Invitrogen, Grand Island, NY) in a 20 ml reaction containing 1 mg total RNA. Conditions for the cDNA synthesis included 65uC for 5 min, 4uC for 1 min, 42uC for 50 min, and finally 72uC for 15 min. qRT-PCR was performed in a 25 ml PCR reaction mixture with 2 ml cDNA (,100 ng RNA) in a Chromo 4 Detector (MJ Research, Maltham, MA) by incubating at 94uC for 5 min, followed by 30 to 60 cycles of a three temperature program consisting of 1 min at 94uC, 40 sec at 57uC, and 40 sec at 72uC. The PCR reaction was terminated after a 10 min extension at 70uC and stored at 4uC until further analysis. Expression of stem cell markers (Oct-4 and Nanog), tenocyte markers (collagen type I and tenascin C), adipocyte marker (PPARc), chondrocyte marker (Sox9), and osteocyte marker (Runx2) were measured using the primers listed in Table 1. GAPDH was used as an internal control. All primers were synthesized by Invitrogen. Expression of each target gene was normalized to GAPDH gene expression and the relative gene expression levels were calculated from the formula 2 2DDCT . Details of the calculation are described in our previous study [9]. The mean and standard deviation (SD) of the CT values were determined from at least three replicates.

Determination of hTSC Differentiation in vivo by Implantation
To verify the effect of PGE 2 treatment on the differentiation of hTSCs in vivo, eight female nude rats (10 weeks old; 200-250 g) were used. hTSCs at passage 2 were seeded into 24-well plates (8610 6 cells/well) and cultured in DMEM with or without various concentrations of PGE 2 for 6 days, with a change of medium every day. For implantation experiments, the cells were trypsinized from each well and mixed with 0.25 ml Matrigel (BD Scientific) to enable gel formation after implantation. These hTSC-Matrigel mixtures were placed subcutaneously in the back of anesthetized rats. Three pieces of hTSC-Matrigel were positioned in three distinct places on each side of each rat's back. Four weeks after implantation, tissue samples were harvested from the implanted area and placed in pre-labeled base molds filled with frozen section medium (Neg 50; Richard-Allan Scientific; Kalamazoo, MI). The tissue blocks were stored at 280uC until histological analysis.

Immunohistochemical and Histological Analyses
Each frozen tissue block was cut into 10 mm thick sections, placed on glass slides, and then allowed to dry overnight at room temperature. The tissue sections were fixed in 4% paraformaldehyde for 30 min and further washed three times with PBS. They were then incubated at room temperature with mouse anti-human PPARc antibody (Santa Cruz Biotechnology, Inc., Cat. #271392, Santa Cruz, CA) diluted to 1:350 for 2 hrs, mouse anti-collagen type II antibody (1:300; Millipore, Cat. #MAB8887, Temecula, CA) for 2 hrs, or mouse anti-human osteocalcin antibody (1:300; Santa Cruz Biotechnology, Inc., Cat. #74495, Santa Cruz, CA) for 3 hrs. After washing with PBS, Cy3-conjugated goat antimouse IgG (1:500; Santa Cruz Biotechnology) was added as secondary antibody and incubated at room temperature for 1 hr, followed by staining the nuclei with Hoechst fluorochrome 33342 (1 mg/ml; Sigma, St. Louis, MO) at room temperature for 5 min. Additionally, cell morphology and distribution in those tissues that received hTSCs, which had been treated with various concentrations of PGE 2 in culture, were assessed by staining with hematoxylin and eosin (H&E). Finally, all tissue sections were examined under a fluorescence microscope.

Semi-quantification of Positively Stained Tissue Sections
Each tissue section was examined under a microscope (Nikon eclipse, TE2000-U) and five random images were taken for the semi-quantification of hTSC differentiation in vivo. SPOT TM imaging software (Diagnostic Instruments, Inc., Sterling Heights, MI) was used to process positively stained areas, which were manually identified by examining the images taken. The total area viewed under the microscope was divided by the positively stained area to calculate the proportion of positive staining. Five tissue sections were used for each group and five images were obtained per tissue section. These values were averaged to represent the percentage positive staining in all the groups treated with various PGE 2 concentrations, which indicated the extent of cell differentiation.

Statistical Analysis
Data are expressed as mean 6 standard deviation (mean 6 SD). Unless otherwise indicated, at least three replicates were used for each experimental condition. For statistical analysis of data, oneway ANOVA or a student t-test was used wherever appropriate. All comparisons were between each PGE 2 -treated group and the respective control. A P-value less than 0.05 was considered to indicate statistically-significant differences between the groups compared. Prior to using hTSCs for cell culture experiments in this study, we first verified the stemness of these tendon cells. Microscopic examination of hTSCs revealed the typical cobblestone-shaped morphology of tendon stem cells under phase contrast microcopy (Fig. 1A). Further, cells in culture also showed robust expression of all three stem cells markers, Oct-4 (Fig. 1B), Nanog (Fig. 1C), and nucleostemin (Fig. 1D), in immunohistochemical analyses. These characteristics indicated that the cells derived from the human patellar tendons were indeed tendon-specific stem cells.

Effect of PGE 2 on the Proliferation of hTSCs
After establishing that the cells in culture were hTSCs, we investigated cell proliferation after PGE 2 treatment of hTSCs by determining their population doubling time (PDT). Treatment of hTSCs with a lower concentration (0.01 ng/ml) of PGE 2 significantly increased cell proliferation, as evidenced by decreased PDT when compared to the control (Fig. 2). PGE 2 treatment at a higher concentration (0.1 ng/ml) also induced similar proliferative effects, although to a smaller extent. At concentrations of 1 and 10 ng/ml, the proliferation of hTSCs was not significantly different from the control without PGE 2 treatment. At the highest concentration (100 ng/ml), TSC proliferation was significantly decreased.
Additionally, FACS analysis of the stem cell markers also corroborated the immunocytochemical findings. Specifically, as PGE 2 concentrations increased from 0 to 0.01 ng/ml, more cells positively stained with SSEA-4 and Stro-1 (Fig. 5, blue dots) were evident; however, when PGE 2 concentrations were further increased to 1 and 100 ng/ml, few positively-stained cells were detected. Quantification of the results from two independent FACS experiments also confirmed these observations (Fig. 6).
To further characterize the stemness of hTSCs after treatment with PGE 2 , we examined the expression of stem cell genes using qRT-PCR. We found that the gene expression levels of Nanog and Oct-4 were significantly (p,0.05) up-regulated in hTSCs treated with lower concentrations (0.01 and 0.1 ng/ml) of PGE 2 (Fig. 7). Notably, the expression level of Oct-4 was twice as high as that of Nanog at 0.01 ng/ml PGE 2 concentration. When treated with higher concentrations (1, 10, and 100 ng/ml) of PGE 2 , expression levels of both Nanog and Oct-4 were down-regulated and almost reached the levels of controls without PGE 2 treatment.

Effect of PGE 2 on the Differentiation of hTSCs
We next examined the effects of PGE 2 on hTSC differentiation by determining the expression of tenocyte and non-tenocyte related genes. Treatment of hTSCs with lower concentrations Figure 3. Expression of the stem cell marker SSEA-4 by hTSCs cultured in vitro in various concentrations of PGE 2 . A: without PGE 2 treatment; B: 0.01 ng/ml PGE 2 ; C: 0.1 ng/ml PGE 2 ; D: 1 ng/ml PGE 2 ; E: 10 ng/ml PGE 2 ; and F: 100 ng/ml PGE 2 . hTSCs were seeded in 12-well plates, cultured with six different concentrations of PGE 2, incubated with mouse anti-human SSEA-4 primary antibody, and detected with Cy3-conjugated goat anti-mouse IgG. Nuclei were stained with Hoechst (Blue). Expression of SSEA-4 (red) is dose-dependent, with more robust expression seen in hTSCs treated with low levels of PGE 2 (A-D) than expression levels seen in those treated with high levels (E, F). Positively stained cells were also counted to calculate percentage staining (G) (*p,0.05 with respect to hTSCs not treated with PGE 2 ). Bar: 100 mm. doi:10.1371/journal.pone.0087706.g003 (0.01, 0.1, and 1 ng/ml) of PGE 2 significantly (p,0.05) enhanced the expression of both collagen type I and tenascin C, two tenocyte-associated genes (Fig. 8A). However, at these lower concentrations, the expression levels of non-tenocyte associated genes PPARc, Sox9, and Runx2 were lower or only marginally higher than the control (Fig. 8B). On the other hand, treatment of hTSCs with higher concentrations (10 and 100 ng/ml) of PGE 2 significantly (p,0.05) up-regulated PPARc, Sox9, and Runx2 genes associated with adipogenic, chondrogenic, and osteogenic differentiation, respectively (Fig. 8B). This up-regulation corresponded with the down-regulation of collagen type I and tenascin C at 10 and 100 ng/ml PGE 2 concentrations (Fig. 8A).

Non-tendinous Tissue Formation after Implantation of PGE 2 -treated hTSCs
To determine whether PGE 2 -treated hTSCs underwent nontenogenic differentiation, we subcutaneously implanted PGE 2treated hTSCs into nude rats. We found that 4 weeks after implantation, non-tenocyte differentiation of hTSCs was more extensive in the cells treated with higher concentrations (10 and 100 ng/ml) of PGE 2 (Fig. 9E-G, and Fig. 9I-K) when compared to the hTSCs that received the lowest concentration of PGE 2 (0.1 ng/ml) (Fig. 9A-C), as evidenced by higher amounts of PPARc, collagen type II and osteocalcin (stained in red/pink). It appeared that more cells (black dots) were present in tissues that received hTSCs treated with high PGE 2 concentrations (Fig. 9H-L) than those that were treated with low PGE 2 concentrations (Fig. 9D). Specifically, at 100 ng/ml (Fig. 9L), numerous cells were concentrated in a specific region (triangle). The immunohis- A: without PGE 2 treatment; B: 0.01 ng/ml PGE 2 ; C: 0.1 ng/ml PGE 2 ; D: 1 ng/ml PGE 2 ; E: 10 ng/ml PGE 2 ; and F: 100 ng/ml PGE 2 . hTSCs were seeded in 12-well plates, cultured with six different concentrations of PGE 2, incubated with mouse anti-human Stro-1, and detected with Cy3-conjugated goat anti-mouse IgG. Hoechst was used to stain nuclei (blue). Expression of Stro-1 (red) is higher in hTSCs treated with low PGE 2 concentrations (A, B) than hTSCs treated with high concentrations (E-F). Similar to SSEA-4, expression of Stro-1 is also dose-dependent. Positively stained cells were also counted to calculate percentage staining (G) (*p,0.05 in comparison with control hTSCs not treated with PGE 2 ). Bar: 100 mm. doi:10.1371/journal.pone.0087706.g004 tochemical observations were also confirmed by semi-quantification, which showed a significant (P,0.001) dose-dependent increase in the staining of non-tenocyte associated genes with increasing amounts of PGE 2 (Fig. 9M). When compared to hTSCs treated with 0.1 ng/ml PGE 2 , those treated with 10 ng/ml had a ,2 fold increase in PPARc, collagen type II, and osteocalcin. These increases were higher when hTSCs were treated with 100 ng/ml PGE 2 (,3 fold for PPARc, ,4 fold for collagen type II, and ,4 fold for osteocalcin).
Discussion PGE 2 is one of the most abundant prostaglandins in the body, and an important causative factor of inflammation that results from tissue damage or infection. Since our previous study showed that high levels of PGE 2 (1, 10, and 100 ng/ml) decrease proliferation and induce differentiation of mouse TSCs into nontenocytes [6], in the present study we investigated the effects of comparable and lower doses of PGE 2 (0.01 to 100 ng/ml) on hTSC proliferation and differentiation by performing cell culture and cell implantation experiments. Our results revealed a concentration-dependent biphasic effect of PGE 2 on the proliferation and differentiation of hTSCs. PGE 2 treatment of hTSCs increased cell proliferation at lower concentrations, but decreased it at higher concentrations. In particular, low levels of PGE 2 promoted the stemness of hTSCs, as evidenced by the extensive expression of stem cell markers SSEA-4 and Stro-1 in hTSCs treated with low concentrations of PGE 2 . The range of PGE 2 concentrations used in this study also includes the in vivo physiological concentrations of PGE 2 reported in human Achilles tendons (0.860.2 ng/ml, [10] or 54624 pg/ml [11]). It should be noted that these values are likely lower due to two reasons: i) patients in these studies were at rest during these measurements and did not undergo intensive exercise, and ii) these values are average microdialysis measurements of PGE 2 concentrations over a large portion of the tendon instead of at a local site, where PGE 2 concentrations could be much higher.
The biphasic effects of PGE 2 on various tissue properties have been reported in previous studies. For example, PGE 2 has been shown to exert biphasic effects on vascularity [12]; it elicits vasodilation at low concentrations and reverses this effect at higher concentrations. Similarly, PGE 2 treatment reduced proliferation of mesenchymal stem cells (MSCs) in a dose-dependent manner (0.25 mM to 25 mM PGE 2 , or 88 ng/ml to 8.8 mg/ml), with the two lowest concentrations (0.25 nM and 2.5 nM PGE 2 , or 88 pg/ ml to 880 pg/ml) slightly increasing MSC proliferation over baseline levels [13]. In this study, the authors demonstrated that the biphasic effect of PGE 2 was executed by differential activation of two types of protein kinase A (PKA). At low concentrations, PGE 2 activated PKA II, leading to a cascade of events that resulted in cell proliferation; at high concentrations, PGE 2 caused PKA I activation, resulting in cell cycle arrest which reduced MSC proliferation. In addition, PGE 2 was reported to have a biphasic influence on injured esophagus: at low doses PGE 2 was protective, but at high doses it damaged the esophagus, with this effect being mediated by the EP1 receptor [14]. Interestingly, the biphasic effects of PGE 2 were also reported to depend on the growth state of the tissue type. For example, PGE 2 promoted proliferation of quiescent smooth muscle cells indicated by an increase in both DNA and RNA synthesis with increasing levels of PGE 2 (10 210 -10 25 M, or 3.5 ng/ml -3.5 mg/ml). However, when proliferating smooth muscle cells were treated with the same concentrations of PGE 2 , DNA synthesis decreased by 48%, indicating that PGE 2 had an inhibitory effect [15].
In this study, we established the stemness of hTSCs based on three characteristics described previously for human tendon stem/ progenitor cells: a) the ability to form colonies in culture; b) Figure 5. FACS analysis of SSEA-4 and Stro-1 expression in hTSCs treated with various concentrations of PGE 2 . hTSCs in culture were treated with various concentrations of PGE 2 . FACS analysis was performed on these cells (for details, see Methods section). It is evident that when cells were treated with 0.01 ng/ml of PGE 2 (PGE 2 -0.01), more cells positively stained with SSEA-4 and Stro-1 were detected (blue dots in the P2 area) compared to control cells without PGE 2 treatment (PGE 2 -0). When PGE 2 concentration increased to 1 ng/ml (PGE 2 -1) and 100 ng/ml (PGE 2 -100), fewer cells were actually stained with SSEA-4 and Stro-1 when compared to control cells (PGE 2 -0). doi:10.1371/journal.pone.0087706.g005  expression of stem cell markers Oct-4, Nanog, and nucleostemin; and c) multi-differentiation potential [7,8]. In addition, these hTSCs assumed a cobblestone shape when grown to confluence [8]. Further, we used two stem cell markers, SSEA-4 and Stro-1, to measure the stemness of hTSCs treated with various concentrations of PGE 2 . SSEA-4 and Stro-1 are highly expressed in undifferentiated stem cells and therefore are used as markers for stem cell identification. However, after differentiation, SSEA-4 is down-regulated in human embryonic stem cells [16]. Our results showing higher expression of SSEA-4 in cells treated with low levels of PGE 2 indicate that stemness is enhanced in these cells, but not in cells treated with higher levels of PGE 2 . Additionally, we also found that cells treated with low levels of PGE 2 produced higher levels of stem cell-related genes (Oct-4 and Nanog) than cells treated with high levels of PGE 2 . Oct-4 and Nanog are both required for the self-renewal and maintenance of stem cells in an un-differentiated state [17]. These genes were reported to downregulate the expression and activity of lineage specific factors, thereby promoting pluripotency [18]. Their downregulation, however, increased differentiation and thereby decreased the capacity of mouse embryonic stem cells for self-renewal [19][20][21].
This study found that higher expression levels of both Nanog and Oct-4 and corresponded low levels of non-tenocyte related genes, particularly in cells treated with low levels of PGE 2 (0.01, and 0.1 ng/ml). The results indicate maintenance of hTSCs in an undifferentiated state, at least in part through Nanog and Oct-4 suppression of adipocyte-(PPARc), chondrocyte-(Sox9), and osteocyte-(Runx2) related genes. Further, lower expression levels of Nanog and Oct-4, especially in cells treated with high concentrations of PGE 2 (10 and 100 ng/ml), also corresponded to higher expression levels of non-tenocyte related genes. This effect, however, was not observed in the control cells (those without PGE 2 treatment), indicating the role high PGE 2 levels have in promoting non-tenocyte differentiation of hTSCs, which in turn reduces their stemness. Taken together, these results strongly suggest that the beneficial effects of the constitutively maintained low levels of PGE 2 may be critical for the maintenance of homeostasis in tendons in vivo.
hTSCs treated with higher concentrations of PGE 2 exhibited extensive expression of non-tenocyte related genes. In the in vivo experiment, non-tenocyte proteins PPARc, collagen type II, and osteocalcin were up-regulated after implantation of hTSCs treated qRT-PCR was performed on total RNA collected from cultured hTSCs treated with PGE 2 . Expression levels of tenocyte related genes, collagen type I (Collagen I) and Tenascin C, were higher in hTSCs treated with low concentrations of PGE 2 (0.01, 0.1 and 1 ng/ml) than in those treated with high concentrations (10 and 100 ng/ml) (A). However, expression levels of non-tenocyte related genes, PPARc, Sox9, and Runx2, were more reduced in hTSCs treated with low (0.01, 0.1 and 1 ng/ml) than with high concentrations of PGE 2 (10 and 100 ng/ml) (B) (*p,0.05 with respect to corresponding controls that did not receive PGE 2 treatment). doi:10.1371/journal.pone.0087706.g008 with high levels of PGE 2 . These findings suggest that PGE 2 at high concentrations could cause differentiation of TSCs into nontenocytes; this could lead to impaired tendon healing and the formation of non-tendinous tissues in affected tendons, which would consequently reduce tendon strength. Indeed, it has been suggested that PGE 2 , as a local hormone in tendons, may contribute to the development of tendinopathy [2,[22][23][24]. In addition, prostaglandins (PGs) are known to play a pathophysiological role in the skeletal system, including contributing to the pathology of osteoporosis by enhancing bone resorption [25]. However, in the same milieu, PGs also exert a physiological role by stimulating bone formation through increased osteoblast proliferation and differentiation. These functions of PGs are consistent with the biphasic effects of PGE 2 that maintain tendon homeostasis and lead to tendon pathology or tendinopathy.
It should be noted that when hTSCs were treated with low levels of PGE 2 , tenocyte-related genes, including collagen type I and tenascin C, were highly expressed (Fig. 7A). These results suggest that PGE 2 at low concentrations may exert its effects on TSCs in two ways: promoting the stemness of TSCs, and inducing TSCs to differentiate towards tenocytes (or progenitor cells for tenocytes). TSCs in our cultures presumably consisted of two subpopulations of cells: one population consisted of early-stage stem cells expressing stem cell markers, such as Nanog, Oct-4, SSEA-4, Figure 9. In vivo expression of non-tenocyte markers PPARc, collagen type II, and osteocalcin in rats implanted with hTSCs treated with various concentrations of PGE 2 and their respective hematoxylin and eosin (H&E) stained tissue sections. hTSCs cultured with three concentrations (0.1, 10, and 100 ng/ml) of PGE 2 were implanted subcutaneously into rats; later, immunohistochemical and histological analyses were performed on tissue sections. For the immunohistochemical staining, fixed tissue sections were incubated with mouse anti-human PPARc antibody, mouse anti-collagen type II (Collagen II) antibody, or mouse anti-human osteocalcin antibody. Cy3-conjugated goat anti-mouse IgG was then used to detect primary binding. Nuclei were stained with Hoechst (blue). Expression levels of PPARc, collagen type II, and osteocalcin (red) were lower in cells treated with 0.1 ng/ml PGE 2 (A-C) than those treated with the higher concentrations (10 and 100 ng/ml) of PGE 2 (E-G, I-K). H&E staining was also performed on tissue sections (D, H, L). More cells (black dots; see insets in D, H, and L) were observed in tissues implanted with hTSCs that had been treated with high concentrations of PGE 2 in culture (H, L). Specifically, at 100 ng/ml (L), cells were concentrated in a specific region (triangle). Additionally, semi-quantification of the stained cells was performed by counting immuno-positive cells and calculating percentage staining (M) (*p,0.05 in comparison with control hTSCs not treated with PGE 2 ). Bar: 100 mm. doi:10.1371/journal.pone.0087706.g009 and Stro-1, and the other population consisted of progenitor cells, which have differentiated towards tenocytes and expressed collagen type I and tenascin C, as demonstrated in this study. In other words, low levels of PGE 2 not only promote TSC selfrenewal, but also promote the differentiation of TSCs into progenitor cells for tenocytes, suggesting that low concentrations of PGE 2 cause hTSCs to undergo asymmetric differentiation. Endogenous PGE 2 has also been shown to stimulate the proliferation of human MSCs [26], protect mouse embryonic stem cells from apoptosis through EP receptor activation [27], and enhance homing, survival, and proliferation of mouse and human hematopoietic stem cells that lead to increased numbers of repopulating cells and units [28]. As tendon-specific stem cells, TSCs play an important role in the repair of injured tendons by proliferating and differentiating in vivo. When tendons are injured, more tenocytes are needed, and TSCs must be activated to effectively repair injured tendons. Our results indicate that the constitutive baseline levels of PGE 2 , which are low, may be used to effectively expand TSCs for cell therapy of injured tendons by promoting proliferation and maintaining tendon homeostasis.
The beneficial effects of low PGE 2 levels on TSCs have several potential applications in tendon tissue engineering. Since PGE 2 at low levels can promote the stemness of TSCs, it may be used to maintain TSCs in culture. In addition, because low PGE 2 levels can accelerate TSC proliferation, they could be used to quickly expand TSC populations for the use of cell therapy to treat injured tendons. Moreover, in vivo tendon injuries could be potentially treated by injecting low levels of PGE 2 at the site of injury. This could enhance the healing of injured tendons because of the ability of low levels of PGE 2 to stimulate self-renewal of TSCs and promote tenogenesis. A recent study showed that low levels of PGE 2 injected into rat patellar tendons enhanced their structural properties (the ultimate load, stiffness, and elastic modulus) [29].
While this is the first study to demonstrate the biphasic effects of PGE 2 on hTSCs, the molecular mechanisms responsible for these biphasic effects are yet to be investigated. PGE 2 is known to exert its diverse biological effects through the EP receptors [14,15,30] and by differential activation of PKA types [13]. Hence, the biphasic response of hTSCs to PGE 2 observed in this study may also involve multiple EP receptor subtypes and/or differential activation of PKA types. Also, while we have shown the beneficial effects of low PGE 2 levels on hTSCs, one limitation of the study is the use of static culture without mechanical loading applied to hTSCs. However, tendons, and therefore the TSCs in vivo, are constantly subjected to mechanical loading, which regulates the expression levels of collagen type I, PPARc, collagen type II, Sox9, and Runx2 genes. In addition, mechanical loading also increases PGE 2 levels in both patellar and Achilles tendons [9], indicating a potential interaction between mechanical loading and PGE 2 . Additional studies are required to reveal the mechanisms behind this interaction. Further, we investigated only the long term effects (up to 6 days) of PGE 2 treatment on hTSCs. It is known that exercise increases PGE 2 levels in human blood only transiently, with maximum levels observed 2 hrs after exercise [31]. Therefore, it would be of interest to study the short term effects of PGE 2 on hTSCs both in vitro and in vivo.
In summary, we showed in this study that PGE 2 exerted biphasic effects on hTSCs: at low concentrations, PGE 2 enhanced their proliferation and expression of stem cell markers, whereas high concentrations of PGE 2 were detrimental to hTSCs, because they reduced their proliferation and induced non-tenocyte differentiation. These results suggest that, on one hand, low levels of PGE 2 promote tendon homeostasis by maintaining hTSCs and tenogenesis; on the other hand, high levels of PGE 2 in tendons may induce differentiation of hTSCs into non-tenocytes and thus lead to the development of the degenerative tendinopathy often seen in clinical settings.