Downregulation of MMP1 in MDS-derived mesenchymal stromal cells reduces the capacity to restrict MDS cell proliferation

The role of mesenchymal stromal cells (MSCs) in the pathogenesis of myelodysplastic syndromes (MDS) has been increasingly addressed, but has yet to be clearly elucidated. In this investigation, we found that MDS cells proliferated to a greater extent on MDS-derived MSCs compared to normal MSCs. Matrix metalloproteinase 1(MMP1), which was downregulated in MDS-MSCs, was identified as an inhibitory factor of MDS cell proliferation, given that treatment with an MMP1 inhibitor or knock-down of MMP1 in normal MSCs resulted in increased MDS cell proliferation. Further investigations indicated that MMP1 induced apoptosis of MDS cells by interacting with PAR1 and further activating the p38 MAPK pathway. Inhibition of either PAR1 or p38 MAPK can reverse the apoptosis-inducing effect of MMP1. Taken together, these data indicate that downregulation of MMP1 in MSCs of MDS patients may contribute to the reduced capacity of MSCs to restrict MDS cell proliferation, which may account for the malignant proliferation of MDS cells.

In the present study, the role of MMP1 in the interaction of MSCs and MDS cells was evaluated. MMP1 secreted from MSCs inhibits the growth and induces apoptosis of SKM-1cells and primary CD34 + cells from MDS patients through interaction with PAR1, which further activates p38 MAPK and downstream genes. Thus, downregulation of MMP1 in MDS-derived MSCs is associated with increased MDS cell proliferation.

MDS cells proliferate to a greater extent on MDS-MSCs compared with normal control MSCs.
SKM-1 cells and MDS-derived CD34 + cells were cultivated alone or in the presence of normal MSCs or MDS-MSCs at a ratio of 5:2 and were tested for their proliferative activity after 72 h of culture by the EdU assay. In addition, cell numbers were counted using a haemocytometer at 24 h, 48 h and 72 h of culture. Co-culture with both normal MSCs and MDS-MSCs suppressed the proliferation activity of MDS cells compared with MDS cells cultured alone. Importantly, both the EdU assay and cell counting indicated that MDS cells proliferated to a greater extent on MDS-MSCs compared with normal control MSCs (Fig. 1).  Fig. S1 and Fig. 2a). In addition, high-grade MDS patients possessed lower levels of MMP1 than low-grade MDS patients. MMP1 mRNA expression was further confirmed through a comparison with another house-keeper gene ( Supplementary Fig. S2a). The MMP1 protein levels were also decreased in MDS-derived MSCs, which is consistent with MMP1 mRNA expression (Fig. 2b). To test whether MMP1 is involved in the reduced capacity of MDS-MSCs to restrict the proliferation of MDS cells, we added the MMP1 inhibitor FN439 (5 μ M) to normal MSCs and SKM-1 in co-culture. The addition of FN439 significantly increased the proportion of SKM-1 cells in the S phase (Fig. 2c). However, in the absence of MSCs, FN439 did not show any effects on MDS cell proliferation ( Supplementary Fig. S2b). The above results suggest that MMP1 plays an important role in suppressing MDS cell proliferation in MSCs and MDS cells in co-culture.

MMP1 as an inhibitory factor of MDS cell proliferation.
The inhibitory effect of MSCs on MDS cell proliferation is decreased when MMP1 is knocked down. To further confirm that MMP1 is an important factor involved in the inhibitory effect of MSCs on cell proliferation, we constructed 2 retrovirus-based RNAi vectors that transfect MSCs with high efficiency. Normal MSCs were infected with the retroviral supernatant containing shRNA specific to human MMP1. On average, MMP1 was reduced by approximately 90%, as evaluated by real-time RT-PCR (Fig. 3a) and western blotting Scientific RepoRts | 7:43849 | DOI: 10.1038/srep43849 (Fig. 3b). We then evaluated the overall proliferation rate of MDS cells in the MMP1-knockdown (KD) group and negative control group. Similar to the results obtained from the MMP1 inhibitor assay, the proportion of MDS cells in the S phase was increased in the MMP1-KD group compared with the negative control group (Fig. 3c). In addition, co-culture with MMP1-KD MSCs resulted in decreased numbers of apoptotic MDS cells compared with negative control MSCs (Fig. 3d). Also, the proliferative proportion of CD34 + cells from healthy donors was increased and the apoptotic proportion was slightly decreased in the MMP1-KD group compared with negative control group (Supplementary Fig. S3a and b). In summary, the growth inhibition and apoptosis induction effects of MSCs on MDS cells were reduced when MMP1 was knocked down.

MMP1 affects MDS cell proliferation and apoptosis through interaction with PAR1. PAR1 has
been reported to be the target of MMP1. The proliferation of MDS cells was suppressed when exogenous activated MMP1 was added to MDS-MSCs and MDS cells in co-culture (Fig. 4a). Importantly, the proportion of apoptotic MDS cells, as measured by Annexin-V and PI staining, was significantly increased (Fig. 4b). To explore whether the growth suppressing and apoptosis inducing effects of MMP1 were mediated via PAR1, the PAR1 antagonist RWJ56110 was introduced prior to MMP1 addition. MMP1-induced growth inhibition and apoptosis was blocked by the PAR1 antagonist ( Fig. 4a and b), thereby demonstrating that the effect of MMP1 on MDS cells was PAR1 dependent.

MMP1/PAR1 exerts an apoptotic effect on MDS cells through the p38 MAPK pathway. MAPKs
have been established as downstream components of the MMP1-PAR1-G protein axis, and the phosphorylation of MAPKs in response to MMP1 has been shown to occur in platelets 23 . Therefore, we hypothesized that MMP1 can regulate apoptosis by activating the MAPK pathways upon interaction with PAR1. As predicted, treatment of SKM-1 cells with activated MMP1 caused a rapid and robust induction of p38 MAPK phosphorylation which peaked at 1 h upon stimulation and subsided by 4 h (Fig. 5a). RWJ-56110 inhibited the phosphorylation of p38 MAPK induced by MMP1 (Fig. 5b).
Next, we explored the significance of p38 MAPK signalling in the context of MMP1-induced apoptosis. We observed that the p38 inhibitor SB203580 completely reversed the proportion of apoptotic cells induced by MMP1 (Fig. 5c). In addition, the expression of pro-apoptotic proteins, such as Bax and cytochrome c which were increased in response to MMP1, were also blocked by SB203580 (Fig. 5d). These results strongly suggest that MMP1 confers cytotoxicity by activating the PAR1-p38 MAPK pathway. Thus, downregulation of MMP1 in MDS-derived MSCs leads to reduced apoptosis which may result in increased MDS cell proliferation (Fig. 6).

Discussion
In this study, we demonstrated that MDS cells proliferated to a greater extent on MDS-MSCs compared with normal control MSCs. Downregulation of MMP1 of MDS-MSCs may partly account for this phenomenon. Either inhibition or knock-down of MMP1 in normal MSCs leads to increased MDS cell growth. MMP1 confers cytotoxicity by activating the PAR1-p38 MAPK pathway.
Recently, studies on MDS-derived MSCs mainly focused on their biological characteristics and hematopoietic support capacities. However, the interactions between MSCs and MDS cells are rarely reported. MSCs have been shown to suppress the proliferation of tumour cells by many researchers 27 . We demonstrated that MDS cells proliferated to a greater extent on MDS-MSCs compared with normal control MSCs, which may explain the possible pathogenesis of MDS.
Among the mediators released from MSCs, MMPs have been shown to be important regulators of the tumour microenvironment and various tumour-related processes, such as tumour growth, apoptosis, angiogenesis, invasion and metastasis 15 . MMP1 has been widely reported to be involved in tumour invasion; however, its regulation of cell apoptosis and proliferation has not been well covered in the literature. In this study, we demonstrated that MMP1 played an important role in apoptosis and that downregulation of MMP1 in MDS-MSCs may account for the reduced capacity to restrict proliferation and induce apoptosis of MDS cells. Consistently, Kittang et al. 28 also observed decreased levels of MMP1 in high-grade MDS patients compared with low-grade MDS patients, which may support our findings given that high-grade MDS is characterized by the accumulation of blasts.
PAR1 is a G protein coupled receptor that is classically activated by thrombin 29 . Recently, MMP1 has been discovered to cleave and activate PAR1 at a non-canonical site, triggering Gα 12/13 -MAPK 24 . Our results demonstrate that a PAR1 antagonist is able to reverse the growth inhibition and apoptosis effects induced by MMP1, confirming the role of PAR1 in this process. Moreover, p38 MAPK was activated when MDS cells were treated with MMP1. Consistent with our data, Trivedi et al. also showed that exogenously added MMP-1 activated p38 MAPK and its substrate, MAPK-activated protein kinase-2 (MAPKAP-K2), in platelets 23 .
The role of p38 MAPK in apoptosis depends on the cell type and stimuli 30 . In some cell types, p38 MAPK has pro-apoptotic effects 31,32 . The possible mechanisms may involve the translocation or phosphorylation of Bcl-2 family proteins, resulting in the release of cytochrome c from the mitochondria 33-35 , caspase-8 activation induced by transforming growth factor-β 36 and modulation of membrane blebbing and nuclear condensation 37 . In addition, growth arrest and DNA damage (GADD)-inducible genes also mediate the pro-apoptotic effects of p38 MAPK 38 . In this study, we found that inhibition of p38 MAPK reversed cell apoptosis induced by MMP1, indicating that the apoptosis effect induced by MMP1 was mediated by p38 MAPK. Furthermore, the Bax and cytochrome c protein levels, which were increased by MMP1, were also reversed by p38 MAPK inhibition, suggesting that the Bcl-2 family and cytochrome c may be involved in the mechanism of MMP1-PAR1-p38 MAPK-induced apoptosis.
In summary, our results demonstrate that MMP1 secreted from MSCs exhibits growth inhibition and apoptosis induction effects on SKM-1 cells and MDS-derived CD34 + cells by interacting with PAR1, which further activates p38 MAPK and downstream genes. Thus, reduced expression of MMP1 in MSCs from MDS patients had a decreased capacity to restrict the proliferation of MDS cells, which may account for the malignant proliferation of MDS cells.  in this study. Their characteristics are detailed in Table 1. Patients were classified for the study as "low-grade" (International Prognostic Scoring System (IPSS)-low/int-1) or "high-grade" (IPSS-int-2/high) 40 . A total of 23 healthy volunteers were used as controls and were matched by gender and age.

Isolation and culture of BM-MSCs. Mononuclear cells (MNCs) were isolated from fresh BM aspirates
and separated by a Ficoll-Paque Plus (GE Healthcare, Uppsala, Sweden). MNCs were seeded at an initial concentration of 1 * 10 6 cells/mL and cultured in Human Mesenchymal Stem Cell Growth Medium (Cyagen Biosciences Inc., Guangzhou, China) supplemented with 10% foetal bovine serum (FBS), glutamine, and 100 U/mL Penicillin-Streptomycin at 37 °C with 5% CO 2 in a fully humidified atmosphere. After 72 h, the culture medium was replaced and non-adherent cells were removed. Thereafter, medium was replaced every 3 to 4 d. Upon achieving greater than 80 to 90% confluency, cells were detached with 0.25% trypsin-EDTA (Gibco, Grand Island, NY, USA). At the third passage (P3), adherent BMMSCs were harvested and utilized for experimental analysis. BMMSCs were evaluated by cytometry for the absence of CD34 and CD45 antigens and the presence of CD73, CD90, CD105 and CD166.
Cell lines and culture. MDS cell line SKM-1 cells were gifted from Prof. Nakagawa. Cell lines were maintained in RPMI-1640 with 10% foetal bovine serum and penicillin(100 units/ml)/streptomycin(100 μ g/ml). All cells were maintained in humidified air containing 5% CO 2 at 37 °C.   Western blot analysis. Equal quantities of protein were analysed via 8 to 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to poly (vinylidene difluoride) membranes. After incubation at 4 °C with primary antibodies against MMP1 (Proteintech Group, Rosemont, IL, USA), p-p38, p38, Bax, Cytochrome c and GAPDH (Cell Signalling Technologies, Boston, MA, USA) overnight, the blots were washed, exposed to corresponding HRP-conjugated secondary antibodies for 1 h, and finally detected by chemiluminescence reagents (Millipore, Billerica, MA, USA).