Inhibition of PAI‐1 limits chemotherapy resistance in lung cancer through suppressing myofibroblast characteristics of cancer‐associated fibroblasts

Abstract Plasminogen activator inhibitor‐1 (PAI‐1) promotes pulmonary fibrosis through increasing myofibroblast (MF) characteristics, expressing alpha‐smooth muscle actin (α‐SMA) in fibroblasts. Fibroblasts in the tumour stroma are called cancer‐associated fibroblasts (CAFs). Some CAFs have MF characteristics and substantially promote tumour progression and chemotherapy resistance. This study determined whether inhibition of PAI‐1 suppressed MF characteristics of CAFs and limited chemotherapy resistance in lung cancer. To investigate cellular PAI‐1 expression and its correlation with α‐SMA expression of CAFs, 34 patients’ paraffin‐embedded lung adenocarcinoma tissue sections were immunohistochemically stained for PAI‐1 and α‐SMA. Immunohistochemical analysis of lung adenocarcinoma tissues showed that PAI‐1 expression was correlated with that of α‐SMA (r = 0.71, p < 0.001). Furthermore, in vitro, α‐SMA expression of CAFs was limited by PAI‐1 inhibition, and apoptosis of CAFs was increased. In addition, the effectiveness of cisplatin on lung cancer cells co‐cultured with CAFs was increased by suppressing α‐SMA expression using PAI‐1 inhibitor. In lung adenocarcinoma tissues, PAI‐1 expression was associated with T factor and TNM stage. Our data suggest that inhibition of PAI‐1 increased the chemotherapeutic effect on lung cancer through suppressing the MF characteristics of CAFs. Hence, PAI‐1 might be a promising therapeutic target for patients with chemotherapeutic‐resistant lung cancer with CAFs.

Fibroblasts in the tumour stroma, called cancer-associated fibroblasts (CAFs), have been reported to induce chemotherapeutic resistance in lung cancer. Actually, a large number of studies showed that CAFs promoted the metastatic ability, invasion and proliferation of lung cancer cells, which induce resistance to chemotherapy by secreting a number of growth factors and inflammatory chemokines. 4,5 Furthermore, a part of CAFs has myofibroblast (MF) characteristic with the expression of alpha-smooth muscle actin (α-SMA), which substantially promotes chemotherapeutic resistance by expressing high levels of inflammatory factors and chemokines. 6,7 Thus, the suppression of the MF characteristics of CAFs can be a novel treatment strategy for a chemotherapeutic-resistant lung cancer with myofibroblastic CAFs.
Plasminogen activator inhibitor-1 (PAI-1), which is produced by endothelial cells of blood vessels, inflammatory cells and fibroblasts, is a glycoprotein with 47-kDa molecular size and inhibits the activation of plasminogen activator. 8 PAI-1 was reported to promote pulmonary fibrosis by blocking fibrinolysis in the lung parenchyma, 9 and also its association with TGF-β induces the differentiation of fibroblasts into MFs. 10 PAI-1 has been also reported to be involved in resistance to chemotherapy and tumourigenesis in various types of cancers. [11][12][13] In vivo studies revealed the significance of PAI-1 in regulating tumour angiogenesis. [14][15][16] Furthermore, several in vitro studies have shown that PAI-1 has a direct pro-proliferative 17 and anti-apoptotic 18 effect on the cancer cells.
From these observations, we suggested that the inhibition of PAI-1 could increase the efficacy of chemotherapy on lung cancer by suppressing the MF characteristics of CAFs. In this study, we aimed to elucidate this hypothesis using PAI-1 inhibitor.

| Materials
Murine Lewis lung carcinoma cells (LLC) and mouse lung fibroblasts Cell Bank (Tokyo, Japan). Original human fibroblasts were obtained from carcinomatous pleural effusions in patients with lung adenocarcinoma. We called these fibroblasts CAF, because these fibroblasts existed with cancer cells in the pleural effusion, similar to fibroblasts in the tumour stroma. Briefly, the method to purify these fibroblasts was as follows. The cells in the pleural effusion were collected by centrifugation at 1500 r.p.m for 3 minutes.
After that, the cells were resuspended with 20 mL of Roswell Park Memorial Institute (RPMI) medium. To divide the tumour cells and fibroblasts from other cells, the cells were centrifuged at 1000 r.p.m for 10 seconds. This final procedure was repeated three times. In the passage process, only cells with spindle shape survived. To confirm that these cells were fibroblasts, we investigated the mRNA expression of fibroblast activation protein (FAP), which is a specific marker of fibroblasts, and α-SMA by quantitative real-time PCR

| Cell culture and treatment
A549, PC-9 and LLC cells were cultured in RPMI supplemented with 10% foetal bovine serum (FBS) and 1% penicillin-streptomycin. MRC-5 cells, MLFs and CAFs were cultured in EMEM. These cells were incubated at 37°C in a 5% CO 2 incubator and used within 6 months after resuscitation. MRC-5 cells, MLFs and CAFs were seeded at a density of 1 × 10 5 cells/well in six-well plates for qPCR, ELISA, quantitative proteomic analysis and phospho-kinase array. In addition, these fibroblasts were seeded at a density of 1.5 × 10 4 cells/insert well in 24-well plates for chemotherapy effect, in the 24-well plates for proliferation and cell cycle assays. For apoptosis assay, these cells were seeded at a density of 1 × 10 4 cells/well in 96-well plates. After the plating, these fibroblasts were cultured in EMEM supplemented with 10% FBS for 12 hours.
Thereafter, MLFs and MRC-5 cells were pre-incubated with or without SK-216, a PAI-1 inhibitor, (20 or 50 μmol/L) in a serum-free medium for 1 hour followed by stimulation with TGF-β1 (mouse or human recombinant TGF-β1, 5 ng/mL, R&D Systems, Minneapolis, MN). On the other hand, CAFs were cultured with or without SK-216 (100, 250, or 500 μmol/L) in the medium with FBS. These cells were used for various analyses, 36 hours after the treatment.

| Quantitative real-time PCR
Total RNA was isolated with RNeasy Mini Kits (Qiagen, Valencia, CA, USA). The isolated total RNA was reverse transcribed into cDNA using a High Capacity RNA-to-cDNA ™ Kit (Applied Biosystems, Framingham, MA, USA) following the manufacturer's instructions.

| Quantification of PAI-1 and TGF-β protein
Total PAI-1 or TGF-β secreted into serum-free culture medium for 36 hours and in the serum were measured using an ELISA kit (R&D systems, Minneapolis, MN) following the manufacturer's instructions. Immunofluorescence staining for the detection of α-SMA was performed following our previous report. 10 To detect α-SMA, α-SMA antibodies (R&D systems) were used. Nuclei were stained with 4', 6-diamidino-2-phenylindole (Vector Laboratories). All images were captured using a microscope at a magnification of 40 or 200× (model BZ-9000; Keyence, Osaka, Japan). The stromal area, excluding cancer cells, was manually identified and calculated using Dynamic cell count software BZ-HIC (Keyence). Afterwards, the PAI-1 or α-SMA positive area was measured using the software. Finally, the percentage of PAI-1 or α-SMA positive area in the stromal area was determined.

| iTRAQ proteomics
A commercial iTRAQ analysis system (Filgen, Nagoya, Japan) was used with mass spectrometry. Briefly, the total protein from whole cells was purified, and 100 μg of protein samples was reduced and alkylated prior to trypsin digestion, and the resulting peptides were lyophilized and reconstituted before labelling with sham control-iTRAQ 113 and TAC 8W-iTRAQ 115 according to the manufacturer's instructions (AB SCIEX). Labelled digests were combined into sample mixtures, and protein identification and relative iTRAQ quantitation were performed with an AB SCIEX TripleTOF 5600 mass spectrometer with ProteinPilot™ software version 4.5 using the Paragon™ Algorithm 4.5.0.0.

| Phospho-mitogen-activated protein kinase (MAPK) array
The total protein from whole cells were purified. Cell lysates were incubated with the Human Phospho-MAPK Array Kit (R&D Systems) according to the manufacturer's instructions. Phospho-MAPK Array data was visualized using a chemiluminescence detection system (WSE-6100; ATTO, Tokyo, Japan) and measured using ImageSaver6 (ATTO).

| Apoptosis and cell cycle assay
Apoptosis was investigated using ApoLive-Glo Multiplex Assay

| Proliferation assay
The proliferation of cells was analysed using the cell counting kit-8 (DOJINDO, Kumamoto, Japan) following the manufacturer's instructions.

| In vitro analysis of efficacy of chemotherapy drugs on cancer cells co-cultured with fibroblasts
As described in the Cell culture and treatment section, after 36 hours, the fibroblasts were treated with SK-216 and TGF-β1, the insert well with the fibroblasts was transferred into a new 24-well plate. Thereafter, A549, PC-9 or LLC cells were seeded at a density of 4 × 10 4 cells/well in the 24-well plate, and co-culture of cancer cells with fibroblast was started. Figure S2 shows the scheme of this in vitro model. After 36 hours, cisplatin (10 μmol/L) or afatinib (1 μmol/L) was added to the medium. The proliferation of cancer cells was determined after 36 hours.

| Statistical analysis
Statistical analyses were undertaken using SPSS 17 (SPSS Japan). All the results are expressed as mean ± standard deviation, and the Student t test or Mann-Whitney U test was used to evaluate statistical differences between the groups. Correlations were analysed with Pearson's correlation coefficient test. A P > 0.05 was considered statistically significant.

| The level of PAI-1 was correlated with the proportion of MF characteristics of CAFs
Because α-SMA is considered as a specific marker of the MF characteristics of fibroblast, 20 Figure 1K). On the other hand, this PAI-1 expression level was not significantly correlated with that in the serum ( Figure 1L).

| PAI-1 was associated with the maintenance of MF characteristics of fibroblasts
First, to confirm that the cells with spindle shape obtained from the carcinomatous pleural effusion were fibroblasts, we investigated the mRNA expression of α-SMA and FAP by qPCR. qPCR revealed that the cells expressed these mRNA similar to MRC-5 cells (Figure 2A, B). Therefore, we confirmed these cells were CAFs.
Next, to examine whether PAI-1 had an association with maintaining the MF characteristics of CAFs, we investigated whether PAI-1 inhibition limited the α-SMA expression of CAFs. As Figure 2C and D showed, the α-SMA expression of CAFs was limited by SK-216, a specific PAI-1 inhibitor in dose-dependent manner. This result showed that PAI-1 was involved in maintaining the MF characteristics of CAFs. The main progenitors of CAFs seem to be the local fibroblasts and these fibroblasts obtain MF characteristics by TGF-β stimulation in the tumour stroma. 21,22 Therefore, to investigate whether PAI-1 had association with normal fibroblasts obtaining MF characteristics in the tumour microenvironment, we examined whether the α-SMA expression of MLF and MRC-5 cell lines after TGF-β stimulation was limited by PAI-1 inhibition. As shown in Figure 2E and F, PAI-1 inhibition limited the expression of α-SMA.
These results indicated that PAI-1 was associated with normal fibroblasts obtaining the MF characteristics after TGF-β stimulation.
The difference in the concentrations of PAI-1 inhibitor required to limit α-SMA expression between CAFs, MRC-5 cells and MLFs was observed. The concentration of PAI-1 secreted by CAFs in the culture medium was not significantly different from that of MRC-5 cells ( Figure S3A). On the other hand, CAFs were cultured in the medium with FBS, but MRC-5 and MLF cells were cultured in the serum-free medium. Therefore, the discrepancy was likely due to several growth factors in the FBS that promoted α-SMA expression.
Next, we performed siRNA knockdown experiments to confirm the association of PAI-1 and MF characteristics of fibroblasts. PAI-1 siRNA suppressed the PAI-1 mRNA expression level of CAFs to 10% of NC-siRNA ( Figure S2B). PAI-1 concentrations in the medium were also shown to be significantly reduced by using PAI-1-siRNA compared with NC-siRNA ( Figure S2C). In addition, the α-SMA expression of CAFs was significantly suppressed by the siRNA (Figure S2D). In the experiments using MRC-5 cells, similar results were observed ( Figure S2E-G).  Table S1.

| PAI-1 maintained the MF characteristics of CAFs by inhibiting apoptosis
The iTRAQ analysis revealed that the expression level of many apoptosis-related proteins was elevated in the CAFs by PAI-1 inhibition. Thus, we investigated whether PAI-1 maintained the MF characteristics by inhibiting apoptosis of CAFs. We showed that PAI-1 inhibition in CAFs by treatment with SK-216 significantly decreased cell viability ( Figure 3B) and increased the caspase activity and the amount of single strand DNA ( Figure 3C and D). In addition, cell cycle analysis revealed the proportion of S phase cells decreased and G0/G1 phase cells significantly increased in the SK-216-treated groups than in the control. However, the degree of phosphorylation of MAPK protein was not different between CAFs treated with SK-216 and untreated control ( Figure   S4A and B).

| Inhibition of PAI-1 increased the effect of chemotherapy on lung cancer cells co-cultured with fibroblasts through suppressing the MF characteristics of fibroblasts
We investigated whether the suppression of the MF characteristics of CAFs by PAI-1 inhibition increased the effect of chemotherapy ( Figure S2). As shown in the Figure 4A ( Figure 4D and E). This result suggested that PAI-1 and TGF-β secreted by CAFs could influence the effect of chemotherapy. TGF-β is reported to be a strong inducer of epithelial mesenchymal transition (EMT), 23 and we showed previously that PAI-1 had association with EMT as a downstream effector of TGF-β. 10,24 Therefore, we investigated the level of the EMT marker in the A549 cells co-cultured with CAF treated with SK-216 and untreated control.
qPCR showed the expression of mesenchymal markers, α-SMA and fibronectin was significantly higher in control than that in SK-216 treated group. On the other hand, the expression of the epithelial marker, e-cadherin was not changed ( Figure 4F).

| The level of PAI-1 expression correlated with α-SMA expression of CAFs and was associated with lung cancer progression
In this study, we investigated the association between PAI-1 and chemotherapy resistance. Finally, to investigate whether PAI-1 We showed that PAI-1 expression was significantly higher in the patients with T 3/4, stage 2/3 than in the patients with T 1/2, stage 1 respectively (Table 2).
F I G U R E 4 Inhibition of plasminogen activator inhibitor-1 (PAI-1) increased the effect of chemotherapy on lung cancer cells co-cultured with fibroblasts through suppressing the MF characteristics of fibroblasts. (A) Cancer-associated fibroblasts (CAFs) were cultured with or without SK-216 (500 μmol/L) in the insert well on 24-well plates for 36 hours. The insert well with CAFs was transferred into a new 24-well plate. A549 cells were seeded at a density of 4 × 10 4 cells/well in the 24-well plate. After 12 hours, cisplatin (10 μmol/L) was added to the medium. Thereafter, the proliferation of cancer cells was determined after 36 hours. (B) MRC-5 (C) MLF were pre-incubated with or without SK-216 (50 μmol/L) in the medium for 1 hour followed by stimulation with TGF-β1 in the insert well on 24-well plates. After incubation for 36 hours, the insert well with the MRC-5 or MLF was transferred into a new 24-well plate. Thereafter, (B) PC-9 or (C) LLC cells were seeded at a density of 4 × 10 4 cells/well in the 24-well plate. After 12 hours, (B) cisplatin (10 μmol/L) (C) afatinib (1 μmol/L) were added to the medium. Thereafter, the proliferation of cancer cells was determined after 36 hours. (D), (E) CAFs were cultured with or without SK-216 (100, 250, or 500 μmol/L) in the medium. ELISA assay of PAI-1 and TGF-β level secreted by CAFs in the medium. (F) A549 were co-cultured with CAFs with or without SK-216 (100, 250, or 500 μmol/L) in the medium for 36 hours. The cells were harvested, and total RNA was extracted. The expression of α-SMA and fibronectin as mesenchymal marker, e-cadherin as epithelial marker was examined by quantitative RT-PCR. The data represent the mean values of four or six samples (±SD) and were analysed with the student t test. *P < 0.05; **P < 0.01

| D ISCUSS I ON
In this study, we showed that the level of PAI-1 was correlated with the proportion of MF characteristics of CAFs in lung adenocarcinoma.
Further, we revealed that PAI-1 was associated with maintaining the MF characteristics of fibroblasts via inhibition of apoptosis. Furthermore, the inhibition of PAI-1 suppressed the MF characteristics of fibroblasts and this suppression increased the effect of chemotherapy on lung cancer cells. Finally, we showed that the level of PAI-1 in the lung adenocarcinoma was also associated with the tumour progression.
A large number of studies showed that cancer stromal cells, such as inflammatory cells, endothelial cells and fibroblasts, induced resistance to chemotherapy in lung cancer. 25 In the stromal cells, fibroblasts have been reported to be a main player in this phenomenon. 4,26 In the present study, we revealed that PAI-1 inhibition limited the chemotherapy resistance through suppressing the MF characteristics of CAFs. In addition, we showed that the level of PAI-1 produced by CAFs was higher than that produced after suppression of the MF characteristics of CAFs by PAI-1 inhibition. This result indicates that there could be a positive feedback loop between PAI-1 and the MF characteristics of CAFs ( Figure S5). Thus, PAI-1 could be a reasonable therapeutic target to inhibit resistance to chemotherapy in the lung cancer with CAFs.
In this study, PAI-1 maintained the MF characteristics of CAFs through inhibiting apoptosis. PAI-1 is reported to be involved in various cancer cell properties such as migration, invasion and proliferation. 8,[27][28][29] Several studies showed PAI-1 was involved in these processes through association with apoptosis, 18 PI3k/Akt, MAPK 17,30 and TGF-β signalling including EMT. 10,24 In this study, when CAFs were treated with a PAI-1 inhibitor, the greater number of proteins related to apoptosis were up-regulated as compared to that of MAPK or TGF-β signalling proteins as observed by quantitative proteomic analysis. On the other hand, the degree of phosphorylation of MAPK protein was not different between CAFs treated with SK-216 and untreated control ( Figure S4A and B). Furthermore, we previously revealed that PAI-1 inhibition in fibroblasts was not associated with TGF-β inducing Smad signalling pathway and non-Smad signalling pathways, including ERKs. 10 PAI-1 was also shown to be involved in the apoptosis of cancer cells, 18 endothelial cells, and neurons. 31 From these observations, we conclude that apoptosis is an important mechanism by which PAI-1 maintains the MF characteristics of CAFs.
Several reports showed that CAFs promote tumour progression and resistance to chemotherapy by secreting various growth factors and inflammatory chemokines. 4,5 In this study, the level of PAI-1 and TGF-β secreted after suppression of the MF characteristics of CAFs by PAI-1 inhibition was significantly lower than that from CAFs without PAI-1 inhibition. Thus, we considered the PAI-1 and TGF-β secreted by CAFs could be associated with resistance to chemotherapy in cancer cells in this study. PAI-1 and TGF-β were reported to be associated with EMT of cancer cells in the absence of CAFs. 10 In this study, the expression of mesenchymal genes was elevated in the cancer cells co-cultured with CAFs compared with CAFs treated with PAI-1 inhibitor. This result suggested that EMT induced by PAI-1 and TGF-β in cancer cells was associated with decreasing the effect of chemotherapy in this study, as previously described. 10,17,18 Previous studies revealed that PAI-1 produced by cancer cells was associated with cancer progression and prognosis in various malignancies. [32][33][34] On the other hand, we previously reported that host, but not tumour, PAI-1 is important in lung cancer progression in vitro and in vivo. 15 In this present study, we showed that the high level of host PAI-1 expression in the tumour stroma was associated with tumour size and the TNM stage using immunohistochemical TA B L E 2 The correlation between PAI-1, α-SMA expression and patient characteristics analysis of lung cancer tissue. Our findings suggest that PAI-1 could be a suitable treatment target for progressive tumour with abundant stroma cells that produce PAI-1.
Our data suggest that PAI-1 may influence the tumour progression and effectiveness of chemotherapy through its association with the MF characteristics of CAFs. Hence, PAI-1 might become a promising target for anti-tumour therapy in the lung cancer with CAFs.

CO N FLI C T S O F I NTE R E S T
The authors confirm that there are no conflicts of interest.