The endothelial-to-mesenchymal transition (EndMT) contributes to cancer, fibrosis, and other pathologic processes. However, the underlying mechanisms are poorly understood. Endothelial HSP1 (HSPB1) protects against cellular stress and has been implicated in cancer progression and pulmonary fibrosis. In this study, we investigated the role of HSPB1 in mediating the EndMT during the development of pulmonary fibrosis and lung cancer. HSPB1 silencing in human pulmonary endothelial cells accelerated emergence of the fibrotic phenotype after treatment with TGFβ or other cytokines linked to pulmonary fibrosis, suggesting that HSPB1 maintains endothelial cell identity. In mice, endothelial-specific overexpression of HSPB1 was sufficient to inhibit pulmonary fibrosis by blocking the EndMT. Conversely, HSPB1 depletion in a mouse model of lung tumorigenesis induced the EndMT. In clinical specimens of non–small cell lung cancer, HSPB1 expression was absent from tumor endothelial cells undergoing the EndMT. Our results showed that HSPB1 regulated the EndMT in lung fibrosis and cancer, suggesting that HSPB1-targeted therapeutic strategies may be applicable for treating an array of fibrotic diseases. Cancer Res; 76(5); 1019–30. ©2016 AACR.

Fibrosis is characterized by extracellular matrix abnormalities, mesenchymal transition, and the differentiation of activated fibroblasts, also known as myofibroblasts (1–3). These cells are derived from various cell types, such as resident stromal fibroblasts and bone marrow–derived fibrocytes, and may also arise via mesenchymal transition [i.e., the epithelial-to-mesenchymal transition (EMT) and endothelial-to-mesenchymal transition (EndMT)] (3, 4). The EndMT is characterized by the loss of endothelial marker expression and the acquisition of mesenchymal or fibroblastic phenotypes consisting of the production of smooth muscle actin (SMA), fibroblastic protein-1, and type I collagen (COLI), resulting in cells that have invasive and migratory potential (5, 6). The EndMT is a critical process for embryonic cardiac development (7, 8) and occurs during the progression of cancer and cardiac, intestinal, and renal fibrosis (6, 8–10). Thus, modulation of the EndMT may provide an effective therapeutic strategy for various fibrotic diseases. The EndMT was recently shown to occur in bleomycin-induced pulmonary fibrosis (11), suggesting that this process may be a source of myofibroblasts. The EndMT has also been reported to play a critical role in pulmonary hypertension, with accumulation of mesenchymal-like cells in obstructive pulmonary vascular lesions (12).

Cancer-associated fibroblasts (CAF) facilitate tumor cell proliferation by inducing changes in the tumor microenvironment. However, the origins of these cells are unclear (13–15). The EndMT is an important source of CAFs in the Rip-Tag2 pancreatic carcinoma model (16). Myofibroblasts also accumulate in solid tumors (17). In addition, Kaposi sarcoma promotes the EndMT to enhance the invasiveness of infected endothelial cells, contributing to malignant progression (18). The EndMT was shown to be necessary for metastatic extravasation in brain endothelial cells (19).

TGFβ-related signals have been linked to the EndMT in both normal fibrosis and cancer (20); however, the molecular mechanisms are not fully understood. Several studies have shown a critical role for NOTCH signaling (6, 18). β-Catenin and canonical Wnt signaling have been implicated in the EndMT following cardiac diseases (21, 22). During cardiac fibrosis, bone morphogenetic protein 7 suppresses the EndMT (23). Moreover, cluster of differentiation (CD) 171 regulates the EndMT via JAK–STAT signaling in pancreatic tumors, suggesting that the EndMT is crucial for tumor vasculogenesis (24). Different phenotypes of the TGFβ-induced EndMT may be related to the plasticity and heterogeneity of the tumor vasculature (25).

Endothelial HSP1 (HSPB1; also known as HSP27 in humans or HSP25 in rodents) protects against cellular stress (26, 27). HSPB1 also modulates cellular movement and morphologic changes via interaction with cytoskeletal actin (28) and has been linked to various diseases (26, 29, 30). In addition, HSPB1 overexpression has been reported in various cancers (31), with higher levels detected in metastatic tissues (22), and HSPB1 increases drug resistance of several cancer cells (31). However, the mechanisms through which HSPB1 modulates tumor progression and metastasis are unknown. HSPB1 also regulates pulmonary fibrosis via regulation of the EMT (32); however, the role of HSPB1 in the development of tissue fibrosis is controversial (33, 34).

Hyperphosphorylation of HSPB1 induces actin stress fiber formation and resistance to hypoxia- or TGFβ-induced permeability of pulmonary artery microvascular endothelial cells (35, 36). We previously showed that HSPB1 is localized to endothelial cells of normal tissues and regulates angiogenesis in sarcoma and lung metastatic tumors (37).

Here, we report a novel function of HSPB1 as a specific regulator of the EndMT and clarify the effects of endothelial HSPB1 on the development of pulmonary fibrosis and lung tumor progression in a mouse model. We also show that HSPB1-dependent tumor EndMT is a common phenomenon in human lung cancer.

Cell culture and treatment

Human pulmonary arterial endothelial cells (HPAEC) and human pulmonary microvascular endothelial cells (HPMEC) were obtained from PromoCell and used within nine passages. All endothelial cells were validated by testing positive expression of vWF and CD31.

siRNAs against HSPB1, VEGFR1, and VEGFR2 and a control siRNA were purchased from Santa Cruz Biotechnology and were transfected into cells using Lipofectamine 2000 (Invitrogen). Recombinant human IL1β, stromal cell–derived factor (SDF)-1α, and plasminogen activator inhibitor (PAI)-1 were purchased from Sigma-Aldrich. Recombinant human TGFβ1 was purchased from PeproTech. Cells were irradiated with gamma rays from a [137Cs] source (Atomic Energy of Canada) at a dose rate of 3.81 Gy per minute.

Immunoblotting and immunocytochemistry

Immunoblotting and immunocytochemistry were performed as previously described (37) using antibodies against HSPB1, VEGFR1, VEGFR2, CD31, VE-cadherin, TIE2, and TGFβ1 (Santa Cruz Biotechnology); α-SMA (Abcam); p-VEGFR2 (Tyr1175), p-Smad2/3, and Smad2/3 (Cell Signaling Technology); and β-actin (Sigma-Aldrich). The band intensities were analyzed using ImageJ.

Microarray analysis

Three days after HSPB1 siRNA transfection, total RNA was isolated from HPMECs, and RNA quality was assessed using an Agilent 2100 bioanalyzer (Agilent Technologies). Gene-expression profiling was performed using a Human GE 4 × 44 K v2 Microarray kit (Agilent Technologies). The microarray data were deposited as GSE72485 in the GEO repository.

Transgenic mice

cDNA encoding mouse Hspb1 was subcloned into the pTRE-Tight-BI-ZsGreen1 vector (Clontech Laboratories, Inc), and the doxycycline-dependent induction of HSPB1 and ZsGreen1 expression was tested in vitro using CT26 and B16F10 cell lines (kindly provided Dr. Sam S. Yoon, Massachusetts General Hospital, Boston, MA). FVB-TRE-Hspb1,-ZsGreen1 transgenic mice were generated by microinjecting the transgene into fertilized eggs of FVB mice; the resulting offspring were crossed with FVB-Tg(Cdh5-tTA) mice (The Jackson Laboratory; Stock# 013585) to obtain double-transgenic mice. Doxycycline (Sigma-Aldrich) was added to the drinking water (2 mg/mL with 5% sucrose). Continuous doxycycline administration was initiated before mating of the parental mice, and the offspring were administered doxycycline until 1 week before treatment. Primer sequences and PCR conditions for genotyping are listed in Supplementary Table S1.

Radiation-induced fibrosis model

All procedures were approved by the Institutional Animal Care and Use Committee of the Korea Institute of Radiological and Medical Sciences and Yonsei University Medical School. Radiation was delivered using an X-RAD 320 platform (Precision X-ray) as previously described (38). The left main bronchi of 9-week-old mice were irradiated at 90 Gy using a 3-mm diameter field to mimic an ablative dose. Mouse serum proteins were analyzed using a Mouse Angiogenesis Array (R&D Systems).

Primary tumor model

LSL-KrasG12D;Trp53fl/fl mice (The Jackson Laboratory) were described previously (39). Lung adenocarcinomas were induced via intranasal administration of Ade-CRE (Cell Biolabs). At 2 weeks, mice were intranasally administered control or Hspb1 shRNA-expressing lentiviral particles (Santa Cruz Biotechnology). At 14 weeks, lung tissue was harvested, and fibrotic regions of lung tumors were analyzed using a Mouse Angiogenesis Array (R&D Systems).

Histology and IHC

Lung tissue was harvested and fixed in 10% (v/v) neutral-buffered formalin before preparing paraffin sections. For IHC, primary antibodies against HSPB1, TGFβ, CD31 (1:100; Santa Cruz Biotechnology), α-SMA, or fibroblast-specific protein (FSP)-1 (1:100; Abcam) were used as previously described (40).

For human cancer tissue analysis, a lung cancer tissue Accumax array (catalog no. A202) comprising 40 cases was purchased from ISU ABXIS Co., Ltd., and a non–small cell lung cancer (NSCLC) tissue array (catalog no. LC10012) comprising 45 cases was purchased from US Biomax.

Statistical analysis

Student t tests and ANOVA were used to evaluate the statistical significance of differences between experimental groups. Statistical analyses were performed using GraphPad Prism version 5.0 (GraphPad Software, Inc.). A P value of <0.05 was considered significant

Additional methods are described in the Supplementary Information.

HSPB1 deficiency induced the EndMT

We previously showed that HSPB1 is coexpressed with endothelial cell markers in several tissues (37). To investigate the role of HSPB1 in endothelial cells, we evaluated whether HSPB1 deficiency led to phenotypic changes in HPMECs (Fig. 1A). Cells treated with HSPB1 siRNA showed decreased expression of the endothelial cell marker CD31 and vascular endothelial (VE)-cadherin and increased expression of the fibroblast marker α-SMA, consistent with the EndMT phenotype (Fig. 1A). HSPB1 deficiency significantly reduced VEGFR2 protein but not mRNA levels from day 3, resulting in increases in α-SMA and TGF1β1 expression 6 days later (Fig. 1B). Therefore, we examined whether HSPB1 knockdown increased α-SMA and TGF1β1 levels in response to decreased VEGFR2 expression because VEGFR2 is involved in maintaining the characteristic activity of endothelial cells (41, 42). However, VEGFR2 knockdown did not regulate HSPB1 and did not significantly increase α-SMA and TGF1β1 expression compared with that observed after HSPB1 knockdown (Fig. 1B and Supplementary Fig. S1A). In addition, HSPB1 knockdown increased VEGFR1 (FLT-1) mRNA levels, whereas VEGFR1 knockdown increased COL, ZEB, and TWIST mRNA levels (Fig. 1B and Supplementary Fig. S1A and S1B). Furthermore, HSPB1 knockdown reduced VEGF165-induced VEGFR2 phosphorylation and endothelial cell proliferation, caused the loss of endothelial cell–specific characteristics, and reduced angiogenic effects such as tubulogenesis (Supplementary Fig. S2). Collectively, these data suggest that the increase in α-SMA and TGF1β1 expression induced by HSPB1 knockdown was not caused by decreased VEGFR2 expression. However, loss of HSPB1 caused only minor changes in SNAIL mRNA levels and significant changes in ZEB and TWIST mRNA levels (Fig. 1B, left). In contrast, COLI mRNA levels were significantly increased. Furthermore, the HSPB1 knockdown–mediated increase in α-SMA expression was reversed by overexpression of Hspb1, confirming the role of HSPB1 in the EndMT (Supplementary Fig. S2D).

Figure 1.

HSPB1 deficiency induced the EndMT in HPMECs. A, effects of HSPB1 knockdown on the expression of endothelial and mesenchymal cell markers in HPMECs; scale bar, 100 μm. B, expression of genes encoding endothelial and mesenchymal cell markers and HSPB1 was analyzed by qRT-PCR and Western blotting after knockdown of HSPB1 or VEGFR2 in HPMECs. Transcript levels are shown as fold changes relative to the expression in control cells (n = 3); error bars, SD; ***, P < 0.001; **, P < 0.01; *, P < 0.05 (vs. SiControl). C, effects of HSPB1 knockdown on the TGFβ1-induced EndMT. HPMECs were treated with TGFβ1 for 4 days starting 2 days after HSPB1 siRNA transfection, and marker expression was assessed by Western blotting and immunocytochemistry. Graphs show relative α-SMA expression normalized to cell numbers for four independent experiments; scale bar, 20 μm; error bars, SD; ****, P < 0.0001; **, P < 0.01; *, P < 0.05. D, effects of HSPB1 overexpression on the expression of marker proteins indicating the TGFβ1-induced EndMT. HPMECs were transduced with a lentiviral vector encoding HSPB1 and treated with TGFβ1 for 4 days; scale bar, 20 μm; data are the mean ± SD in four independent experiments; ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05 (D, vs. Ade-GFP).

Figure 1.

HSPB1 deficiency induced the EndMT in HPMECs. A, effects of HSPB1 knockdown on the expression of endothelial and mesenchymal cell markers in HPMECs; scale bar, 100 μm. B, expression of genes encoding endothelial and mesenchymal cell markers and HSPB1 was analyzed by qRT-PCR and Western blotting after knockdown of HSPB1 or VEGFR2 in HPMECs. Transcript levels are shown as fold changes relative to the expression in control cells (n = 3); error bars, SD; ***, P < 0.001; **, P < 0.01; *, P < 0.05 (vs. SiControl). C, effects of HSPB1 knockdown on the TGFβ1-induced EndMT. HPMECs were treated with TGFβ1 for 4 days starting 2 days after HSPB1 siRNA transfection, and marker expression was assessed by Western blotting and immunocytochemistry. Graphs show relative α-SMA expression normalized to cell numbers for four independent experiments; scale bar, 20 μm; error bars, SD; ****, P < 0.0001; **, P < 0.01; *, P < 0.05. D, effects of HSPB1 overexpression on the expression of marker proteins indicating the TGFβ1-induced EndMT. HPMECs were transduced with a lentiviral vector encoding HSPB1 and treated with TGFβ1 for 4 days; scale bar, 20 μm; data are the mean ± SD in four independent experiments; ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05 (D, vs. Ade-GFP).

Close modal

TGFβ is known to trigger the EndMT. HSPB1-deficient cells exhibited the well-organized actin cytoskeleton typically observed in mesenchymal cells treated with TGFβ1, which also induced upregulation of α-SMA and downregulation of CD31 expression more prominently in HSPB1-deficient cells than in control cells (Fig. 1C). However, TGFβ1-induced Smad2/3 phosphorylation was not increased in HSPB1-deficient cells, indicating that Smad2/3 activity was not related to the HSPB1 knockdown–mediated EndMT. Moreover, Smad3 inhibition did not attenuate HSPB1 knockdown–induced α-SMA expression (Supplementary Fig. S2E). TGFβ1-induced α-SMA expression and cytoskeletal changes associated with the mesenchymal transition were not observed in Hspb1-overexpressing endothelial cells (Fig. 1D and Supplementary Fig. S2F). Taken together, these results suggest that HSPB1 expression was required to maintain the characteristics of endothelial cells; HSPB1 deficiency, therefore, increased susceptibility to the effects of TGFβ1 and facilitated the transition to the mesenchymal phenotype.

HSPB1 deficiency increased the susceptibility to fibrosis-related cytokines

We previously showed that the EndMT in human aortic endothelial cells was associated with radiation-induced vascular fibrosis (43). Here, irradiation of HPMECs increased α-SMA levels in HSPB1-deficient cells to a greater degree than in control cells. Moreover, HSPB1 overexpression inhibited the radiation-induced increase in α-SMA levels (Fig. 2A). Also, we examined the HSPB1 knockdown–mediated EndMT in HPAECs and human umbilical vein endothelial cells (HUVEC). HSPB1 deficiency increased the TGFβ1- and radiation-induced EndMT in HPAECs and HUVECs (Supplementary Fig. S3).

Figure 2.

HSPB1 deficiency increased susceptibility to the effects of fibrosis-related cytokines and promoted EndMT progression. A, effects of HSPB1 on the radiation-induced EndMT. HPMECs were transduced with HSPB1 siRNA or lentiviral vector encoding GFP (control) or HSPB1 and then irradiated with 10 Gy. α-SMA and CD31 expression were assessed. Data are the means of four independent experiments. B, serum protein levels from C57BL/6 mice irradiated at 90 Gy were compared with those from nonirradiated mice 2 weeks after irradiation using an angiogenesis antibody array; error bars, SD (n = 4 mice/group). Lung fibrosis was analyzed by Masson's trichrome staining (top); scale bar, 100 μm. C, effects of recombinant human IL1β, SDF-1, PAI-1, and TGFβ1 on the radiation-induced EndMT in HSPB1-deficient HPMECs. Data are representative of four independent experiments; scale bar, 40 μm. D, clustering of genes differentially expressed in HSPB1 siRNA-transfected HPMECs versus control siRNA-transfected HPMECs (lane marked as normalized). Differential expression of genes (marked in colors other than gray) in HSPB1-deficient HPMECs is shown as a heat map showing fold changes in the expression of upregulated genes, grouped on the basis of their relationships to the indicated genes shown in the bottom of the bar graph using a color-coding system. E, qRT-PCR analysis in control- and HSPB1 siRNA-transfected HPMECs. Data are shown as fold changes relative to the expression in control cells (n = 3); error bars, SD; ***, P < 0.001; **, P < 0.01; *, P < 0.05 (A, B, and E).

Figure 2.

HSPB1 deficiency increased susceptibility to the effects of fibrosis-related cytokines and promoted EndMT progression. A, effects of HSPB1 on the radiation-induced EndMT. HPMECs were transduced with HSPB1 siRNA or lentiviral vector encoding GFP (control) or HSPB1 and then irradiated with 10 Gy. α-SMA and CD31 expression were assessed. Data are the means of four independent experiments. B, serum protein levels from C57BL/6 mice irradiated at 90 Gy were compared with those from nonirradiated mice 2 weeks after irradiation using an angiogenesis antibody array; error bars, SD (n = 4 mice/group). Lung fibrosis was analyzed by Masson's trichrome staining (top); scale bar, 100 μm. C, effects of recombinant human IL1β, SDF-1, PAI-1, and TGFβ1 on the radiation-induced EndMT in HSPB1-deficient HPMECs. Data are representative of four independent experiments; scale bar, 40 μm. D, clustering of genes differentially expressed in HSPB1 siRNA-transfected HPMECs versus control siRNA-transfected HPMECs (lane marked as normalized). Differential expression of genes (marked in colors other than gray) in HSPB1-deficient HPMECs is shown as a heat map showing fold changes in the expression of upregulated genes, grouped on the basis of their relationships to the indicated genes shown in the bottom of the bar graph using a color-coding system. E, qRT-PCR analysis in control- and HSPB1 siRNA-transfected HPMECs. Data are shown as fold changes relative to the expression in control cells (n = 3); error bars, SD; ***, P < 0.001; **, P < 0.01; *, P < 0.05 (A, B, and E).

Close modal

We recently found that the EndMT occurs during the development of radiation-induced pulmonary fibrosis (RIPF; ref. 40). Therefore, we next observed fibrotic changes at 2 weeks after focal irradiation (90 Gy, 3 mm) of the left lungs of mice (Fig. 2B). Late-stage fibrosis was observed by 4 weeks after irradiation. To identify the cytokines associated with RIPF, we performed cytokine immunoassays with sera from mice showing RIPF at 2 weeks after irradiation (Fig. 2B and Supplementary Fig. S4A). IL1β, SDF-1, and PAI-1 were detected at higher levels than those in control mice (Fig. 2B, bottom). These cytokines enhanced the fibrotic phenotype in HSPB1-deficient cells compared with that in control cells and induced FSP-1–positive cytoskeletal changes, and these effects were potentiated by irradiation (Fig. 2C and Supplementary Fig. S4B). From the differentially regulated cytokines, we selected IL1β, SDF-1, and PAI-1 according to the observed upregulation of receptor proteins in HSPB1-deficient endothelial cells (Fig. 2D). These results revealed that loss of HSPB1 promoted the EndMT and increased the susceptibility of endothelial cells to fibrotic cytokines.

To investigate the molecular basis for the increased fibrotic phenotype caused by the loss of HSPB1 following cytokine treatment, we performed a microarray analysis, comparing control and HSPB1-deficient endothelial cells (Fig. 2D). In cells lacking HSPB1, various molecules associated with EndMT progression were upregulated (Fig. 2D, left), including IL1 receptor 1/2, C–X–C chemokine receptor 7 (CXCR7), and very low density lipoprotein receptor/low density lipoprotein receptor–related protein 1 (LRP1), which encode receptors for IL1β, SDF-1, and PAI-I, respectively, and also the signaling molecules activated through these receptors (Fig. 2D and E). From the microarray results, given that HSPB1 deficiency in endothelial cells resulted in upregulation of inhibitor of kappa light polypeptide gene enhanced in B cells kinase β (IKBKB; related to IL1β), JAK2STAT5 (related to SDF-1), and STAT1 (related to LRP1), we hypothesized that these cytokines activated JAK–STAT signaling in the absence of HSPB1. Indeed, of several inhibitors of signaling pathways linked to HSPB1, only the JAK2 inhibitor blocked the increase in α-SMA expression caused by HSPB1 deficiency, irrespective of radiation treatment, and inhibited STAT3 phosphorylation (Supplementary Fig. S5A and S5B). In addition, the JAK2 inhibitor blocked the radiation-dependent increase in α-SMA expression (Supplementary Fig. S5C). HSPB1 overexpression also inhibited the radiation-induced increase in STAT3 phosphorylation (Supplementary Fig. S5D). Thus, HSPB1-dependent JAK–STAT signaling may be responsible for the radiation-induced EndMT.

Endothelial HSPB1 overexpression suppressed pulmonary fibrosis by regulating the EndMT

To analyze the role of endothelial HSPB1 in pulmonary fibrosis in vivo, we generated double transgenic mice that conditionally overexpressed HSPB1 (Cdh5-tTA;TRE-Hspb1,-ZsGreen1) in which the Cdh5-tTA construct drove HSPB1 expression after discontinuation of tetracycline administration (Fig. 3A). ZsGreen-HSPB1 expression driven by the VE-cadherin Cdh5 promoter was detected in pulmonary vascular endothelial cells (Fig. 3B). We previously reported that focal irradiation of mouse lungs led to significant collagen deposition in vessels after 2 weeks and late-stage fibrosis at 4 weeks (38). In the current study, in mice overexpressing HSPB1 in endothelial cells, radiation-induced collagen deposits and fibrotic regions were reduced compared with those in controls (Fig. 3C and D).

Figure 3.

Overexpression of endothelial HSPB1 blocked the EndMT and reduced RIPF. A, schematic illustration of the Tet-Off system and constructs used for generating TRE-Hspb1,-ZsGreen1 lines. B, frozen lung tissue sections of mice. ZsGreen1 (corresponding to HSPB1 overexpression) and CD31 colocalization) is shown in yellow. C–E, irradiated lung tissue was analyzed 2 weeks later. C and D, effects of endothelial HSPB1 overexpression on RIPF (hematoxylin and eosin, Masson trichrome staining); scale bar, 50 μm. Data are the relative levels of collagen deposition in lung tissue as the average of five fields (magnification, ×100; n = 6). E, costaining with HSPB1 and CD31 antibodies. Data are the relative levels of colocalization (yellow pixels) of CD31 and α-SMA in vessels as the average of five fields (magnification, ×100; n = 6); error bars, SEM; **, P < 0.01; *, P < 0.05 (D and E). All data are representative of three independent experiments.

Figure 3.

Overexpression of endothelial HSPB1 blocked the EndMT and reduced RIPF. A, schematic illustration of the Tet-Off system and constructs used for generating TRE-Hspb1,-ZsGreen1 lines. B, frozen lung tissue sections of mice. ZsGreen1 (corresponding to HSPB1 overexpression) and CD31 colocalization) is shown in yellow. C–E, irradiated lung tissue was analyzed 2 weeks later. C and D, effects of endothelial HSPB1 overexpression on RIPF (hematoxylin and eosin, Masson trichrome staining); scale bar, 50 μm. Data are the relative levels of collagen deposition in lung tissue as the average of five fields (magnification, ×100; n = 6). E, costaining with HSPB1 and CD31 antibodies. Data are the relative levels of colocalization (yellow pixels) of CD31 and α-SMA in vessels as the average of five fields (magnification, ×100; n = 6); error bars, SEM; **, P < 0.01; *, P < 0.05 (D and E). All data are representative of three independent experiments.

Close modal

Colocalization of α-SMA and CD31 was evident by immunocytochemistry in control mice 2 weeks after irradiation (Fig. 3E). In severely fibrotic regions, α-SMA–positive vessels were detected in control mice, and there was no colocalization with CD31, consistent with a transition to the fibroblast phenotype and progression of the EndMT (Supplementary Fig. S6A). In endothelial HSPB1-overexpressing mice, the radiation-induced EndMT significantly decreased compared with that in controls (Fig. 3E). As shown in Supplementary Fig. S6B, the EMT (i.e., colocalization of Pro-SPC and α-SMA) in alveolar type II epithelial cells decreased in endothelial HSPB1-overexpressing mice compared with that in wild-type mice, suggesting that the EndMT may affect fibrotic changes in neighboring cells. Long-term tetracycline treatment and subsequent HSPB1 induction did not affect the vascular phenotype (Fig. 3C and E).

To determine whether JAK–STAT signaling was activated during the HSPB1-mediated EndMT and fibrosis development, we evaluated the phosphorylation of STAT3 by IHC (Supplementary Fig. S6C). Phospho-STAT3 immunoreactivity was higher in vascular endothelial cells than in other cell types in the radiation-induced fibrotic tissue of control mice, but decreased in endothelial cells of HSPB1-overexpressing mice. Because STAT3 phosphorylation in radiation-induced fibrotic tissues was observed in cell types other than endothelial cells, JAK–STAT3 signaling may be partly responsible for the HSPB1-regulated EndMT during RIPF. Therefore, we next analyzed whether a JAK2 inhibitor blocked RIPF in vivo (Supplementary Fig. S6D and S6E). The JAK2 inhibitor alone significantly reduced radiation-induced collagen deposition and colocalization of α-SMA and CD31. Thus, our data suggest that JAK–STAT3 signaling regulates the EndMT and may represent a new target in RIPF.

HSPB1 knockdown increased the radiation-induced EndMT in lung tissues

Next, we examined whether HSPB1 suppression alone induced the EndMT in normal murine lungs (Supplementary Fig. S7). HSPB1 knockdown by lentiviral Hspb1 shRNA induced the vascular EndMT in normal lung tissues. We observed significant increases in the thickness of α-SMA–positive lesions around CD31- and α-SMA–positive vascular endothelial cells in HSPB1-deficient lung vessels. However, we did not find increased collagen deposition following HSPB1 knockdown alone (Supplementary Fig. S7C). In contrast, irradiation increased collagen deposition in Hspb1 shRNA-treated lungs. Thus, these data suggest that HSPB1 deficiency in endothelial cells was sufficient to induce the EndMT in normal lungs and that the EndMT may cause rapid development of pathologic disease under specific conditions, such as RIPF or tumorigenesis.

In addition, we examined whether the cytokines secreted by endothelial cells during the radiation-induced EndMT affected other pulmonary cells. The conditioned medium (CM) from irradiated HPAECs increased fibroblastic marker expression in a radiation dose–dependent manner in HPASMCs, HPFs, and HPSAEpCs (Supplementary Fig. S8A). In addition, the proliferation of HPASMCs and HPSAEpCs was increased by CM from irradiated endothelial cells in a dose-dependent manner. CM from irradiated HSPB1-deficient endothelial cells significantly increased collagen 1 and α-SMA expression in HPFs and HPASMCs compared with CM from irradiated siControl endothelial cells (Supplementary Fig. S8B), suggesting that paracrine effects during the EndMT may modulate fibrotic changes and subsequent collagen production by other cell types.

To determine whether the HSPB1-regulated EndMT was limited in RIPF, we examined the effects of bleomycin-induced fibrosis and found that the EndMT and collagen deposition markedly decreased in endothelial HSPB1-overexpressing mice compared with control mice (Supplementary Fig. S9). Taken together with our in vitro data, these findings indicate that EndMT was regulated by endothelial HSPB1 overexpression and that HSPB1 played a critical role in the maintenance of endothelial cells during fibrotic stress, modulating the EndMT and the subsequent development of fibrosis.

HSPB1 deficiency induced the EndMT in lung tumorigenesis

Fibroblastic changes in endothelial cells via the EndMT contribute to normal tissue fibrosis and to the occurrence of CAFs in tumors. We found that HSPB1 was localized in the tumor endothelium rather than in tumor cells in spontaneous lung cancer tissues collected from LSL-KrasG12D;Trp53fl/fl mice (KP mice; ref. 44), which express oncogenic K-ras and lack p53 (Fig. 4A).

Figure 4.

HSPB1 deficiency promoted fibrosis in spontaneous lung cancer. A, HSPB1 expression (brown) in tumor vessels of lung cancer tissue from KP mice (n = 8); scale bar, 20 μm. B, spontaneous lung tumor size after HSPB1 knockdown. A lentivirus-expressing Hspb1 shRNA was intranasally injected into the lungs of mice 2 weeks after induction of spontaneous lung cancer. Tumor nodules were visualized after 14 weeks by hematoxylin and eosin staining. Data are the average sizes of tumor nodules per field (magnification, ×100; n = 8). C, costaining with HSPB1 and CD31 antibodies. Data are the relative levels of colocalization of HSPB1 and CD31 in vessels, determined as the average of the five fields (magnification, ×100; n = 6). D, effects of HSPB1 knockdown on collagen deposition and TGFβ1 and FSP-1 expression in lung tumors. Immunocytochemistry for HSPB1 and CD31, Masson trichrome staining, and IHC were performed; scale bar, 20 μm. Graphs show quantitative analyses as the average of five fields (magnification, ×100; n = 8); error bars, SEM; ****, P < 0.0001 (B–D). Data are representative of three independent experiments.

Figure 4.

HSPB1 deficiency promoted fibrosis in spontaneous lung cancer. A, HSPB1 expression (brown) in tumor vessels of lung cancer tissue from KP mice (n = 8); scale bar, 20 μm. B, spontaneous lung tumor size after HSPB1 knockdown. A lentivirus-expressing Hspb1 shRNA was intranasally injected into the lungs of mice 2 weeks after induction of spontaneous lung cancer. Tumor nodules were visualized after 14 weeks by hematoxylin and eosin staining. Data are the average sizes of tumor nodules per field (magnification, ×100; n = 8). C, costaining with HSPB1 and CD31 antibodies. Data are the relative levels of colocalization of HSPB1 and CD31 in vessels, determined as the average of the five fields (magnification, ×100; n = 6). D, effects of HSPB1 knockdown on collagen deposition and TGFβ1 and FSP-1 expression in lung tumors. Immunocytochemistry for HSPB1 and CD31, Masson trichrome staining, and IHC were performed; scale bar, 20 μm. Graphs show quantitative analyses as the average of five fields (magnification, ×100; n = 8); error bars, SEM; ****, P < 0.0001 (B–D). Data are representative of three independent experiments.

Close modal

To clarify the role of HSPB1 in the tumor endothelium, we evaluated spontaneous tumorigenesis after HSPB1 knockdown. Fibrotic regions markedly increased in tumor tissue after treatment with Hspb1 shRNA as compared with that after treatment with control shRNA (Fig. 4B, top). The average size of tumor nodules was larger in Hspb1 shRNA-treated mice than in control shRNA-treated mice (Fig. 4B, bottom). Costaining images of HSPB1 and CD31 showed endothelial cell–specific HSPB1 expression in lung tumors of KP mice (Fig. 4C and D). Moreover, HSPB1 knockdown reduced HSPB1 expression in the endothelium of spontaneous lung tumors and increased collagen deposition and TGFβ1 and FSP-1 expression around tumor vessels (Fig. 4D). The EndMT occurred more frequently in tumors from Hspb1 shRNA–treated mice than in tumors from control shRNA–treated mice (Fig. 5A). HSPB1 and CD31 were expressed in the fibrotic tissue of HSPB1-deficient lung tumors (Supplementary Fig. S10A). Moreover, in KP mice, severe hemorrhage occurred during tumorigenesis 4 weeks after HSPB1 knockdown (Supplementary Fig. S10B) likely owing to vascular leakage caused by the EndMT in HSPB1-deficient endothelial cells. As shown in Supplementary Fig. S10C and S10D, at week 13 after treatment with Hspb1 shRNA without induction of lung adenocarcinoma in KP mice, vascular EndMT lesions and collagen deposition were increased.

Figure 5.

HSPB1 deficiency induced the EndMT in spontaneous lung cancer. A, effects of HSPB1 deficiency on the EndMT in autonomous lung adenocarcinoma. Lentivirus-expressing Hspb1 shRNA was intranasally injected into the lungs of mice 2 weeks after induction of lung adenocarcinoma; scale bar, 20 μm. The graph shows the relative levels of colocalization of CD31 and α-SMA in vessels as the average of five fields (magnification, ×100; n = 8); error bars, SEM. Data are representative of three independent experiments. B, effects of HSPB1 knockdown on protein expression in lung adenocarcinoma. Data show the fold changes of differentially expressed proteins in Hspb1 shRNA-treated lung adenocarcinoma tissues; error bars, SD (n = 3 mice/group); ****, P < 0.0001; **, P < 0.01; *, P < 0.05 (vs. control shRNA). Representative blots of protein spots on the array are shown (right).

Figure 5.

HSPB1 deficiency induced the EndMT in spontaneous lung cancer. A, effects of HSPB1 deficiency on the EndMT in autonomous lung adenocarcinoma. Lentivirus-expressing Hspb1 shRNA was intranasally injected into the lungs of mice 2 weeks after induction of lung adenocarcinoma; scale bar, 20 μm. The graph shows the relative levels of colocalization of CD31 and α-SMA in vessels as the average of five fields (magnification, ×100; n = 8); error bars, SEM. Data are representative of three independent experiments. B, effects of HSPB1 knockdown on protein expression in lung adenocarcinoma. Data show the fold changes of differentially expressed proteins in Hspb1 shRNA-treated lung adenocarcinoma tissues; error bars, SD (n = 3 mice/group); ****, P < 0.0001; **, P < 0.01; *, P < 0.05 (vs. control shRNA). Representative blots of protein spots on the array are shown (right).

Close modal

Furthermore, to clarify the effects of the HSPB1 deficiency–mediated EndMT on lung tumors in KP mice, we examined the effects of HSPB1 knockdown on tumor endothelial cells and tumor cells isolated from lung tumors (Supplementary Fig. S11). Isolated tumor endothelial cells exhibited higher expression of HSPB1 than tumor cells, and HSPB1 knockdown upregulated α-SMA and downregulated CD31, VEGFR2, and VE-cadherin in isolated tumor endothelial cells. Tumor cells did not express HSPB1; thus, Hspb siRNA did not change α-SMA expression. However, as shown in Supplementary Fig. S12A, in Hspb1 shRNA–treated tumors, colocalization of FSP-1 and RasN17 (a tumor cell marker in KP mice) was increased compared with that in control shRNA-treated tumors. Because HSPB1 was not expressed in lung tumor cells from KP mice, Hspb1 shRNA may primarily affect tumor endothelial cells. Therefore, the increases in FSP-1 expression in tumor cells may be affected by fibrotic changes in HSPB1-deficient tumor endothelial cells and subsequent tumor fibrotic changes. In Hspb1 shRNA–treated tumors, the population of PCNA-positive tumor cells significantly increased compared with that in control shRNA-treated tumors, consistent with the increased tumor burden (Supplementary Fig. S12B).

In addition, we found that HSPB1 deficiency increased the phosphorylation of STAT3 in tumor endothelial cells in lung tumors from KP mice (Supplementary Fig. S12C). These results imply that JAK–STAT signaling plays an important role in the EndMT mediated by HSPB1 deficiency in tumors.

To examine the role of HSPB1-deficient tumor endothelial cells in the tumor microenvironment, we performed immunoassays for angiogenic cytokines in Hspb1- and control shRNA-treated fibrotic tumor tissues. We selected cytokines showing the same pattern in multiple replicates (Fig. 5B). The expression of angiopoietin-3, insulin-like growth factor–binding protein 3, matrix metalloproteinase 9, and proliferin increased, whereas many angiogenic cytokines were downregulated. Thus, we hypothesized that tumor angiogenic effects were reduced during the substantial fibrotic phase (Fig. 5B). Changes in the levels of selected cytokines were confirmed by ELISA (Supplementary Fig. S12D). In addition, the deregulated cytokines shown in Fig. 2B exhibited different patterns. However, only proliferin was significantly upregulated in both Fig. 2B and Fig. 5B. Collectively, our data suggest that HSPB1 deficiency in lung tumors in KP mice induced the EndMT and thus may cause tumor fibrosis.

HSPB1 was expressed in vessels of human lung tumors

Given the observed localization of HSPB1 in the vascular endothelium of spontaneous lung tumors in KP mice, we examined whether HSPB1 was also expressed in tumor vessels. HSPB1 immunoreactivity was observed in tumor vascular endothelial cells in small-cell, large-cell, and squamous cell carcinoma and adenocarcinoma (Fig. 6). Tumor cells from lepidic adenocarcinoma arising from terminal bronchioles had high HSPB1 expression, and some HSPB1-positive tumor vessels were also detected in these tumor stroma (Fig. 6, vii, viii, ix). Therefore, HSPB1 expression in tumor vascular endothelial cells may be a general feature of human lung cancers.

Figure 6.

HSPB1 was expressed in tumor endothelial cells in various human lung cancers. A human lung cancer tissue microarray comprising 40 cases was examined for HSPB1 expression. Representative HSPB1-positive tumor vessels are indicated by black arrowheads. Small cell carcinoma (i–ii); large cell carcinoma pneumoconiosis (iii); large cell neuroendocrine carcinoma (iv); squamous cell carcinoma (v); adenocarcinoma (vi–viii); and bronchioloalveolar carcinoma (ix); scale bar, 100 μm.

Figure 6.

HSPB1 was expressed in tumor endothelial cells in various human lung cancers. A human lung cancer tissue microarray comprising 40 cases was examined for HSPB1 expression. Representative HSPB1-positive tumor vessels are indicated by black arrowheads. Small cell carcinoma (i–ii); large cell carcinoma pneumoconiosis (iii); large cell neuroendocrine carcinoma (iv); squamous cell carcinoma (v); adenocarcinoma (vi–viii); and bronchioloalveolar carcinoma (ix); scale bar, 100 μm.

Close modal

The HSPB1-dependent EndMT was observed in human lung cancers

Next, we examined whether endothelial HSPB1 was associated with the EndMT in human NSCLC tissue (Fig. 7). Colocalization of HSPB1 and CD31 but not α-SMA was observed in human tumor endothelial cells (Fig. 7A, i, ii, iii) and was markedly higher in nonfibrotic regions without collagen deposits than in fibrotic regions (Fig. 7A and B). HSPB1-negative, CD31-positive tumor endothelial cells generally expressed α-SMA, indicating that HSPB1-deficient tumor vessels were typical of the EndMT (Fig. 7A, iv, v, vi). Furthermore, α-SMA–positive tumor vessels were primarily observed in the fibrotic regions of lung tumors, which were negative for CD31 and HSPB1 expression (Fig. 7A, vii, viii, and 7B). endothelial cells with a myofibroblastic phenotype that were detached from the tumor vascular endothelium were also detected in α-SMA–positive regions (Fig. 7A, vi, vii, viii). These data provide evidence that the EndMT was associated with the loss of HSPB1 expression in tumor endothelial cells of human NSCLCs.

Figure 7.

The HSPB1-dependent EndMT was observed in human lung cancers. A, a human NSCLC tissue microarray comprising 45 cases was analyzed by Masson trichrome staining and immunocytochemistry for expression of HSPB1, CD31, and α-SMA. The areas enclosed by the broken line in trichrome staining images (black scale bar, 100 μm) are shown, enlarged, to the right as fluorescence micrographs. Representative images are shown; white scale bar, 40 μm. Pathologic information on cancer tissues is shown in the right. B, graphs show quantitative analyses of the colocalization of HSPB1 and CD31 (top) or α-SMA and CD31 (bottom) in tumor vessels per field (magnification, ×100); error bars indicate SEM for 45 cases; ***, P < 0.001; *, P < 0.05 (B).

Figure 7.

The HSPB1-dependent EndMT was observed in human lung cancers. A, a human NSCLC tissue microarray comprising 45 cases was analyzed by Masson trichrome staining and immunocytochemistry for expression of HSPB1, CD31, and α-SMA. The areas enclosed by the broken line in trichrome staining images (black scale bar, 100 μm) are shown, enlarged, to the right as fluorescence micrographs. Representative images are shown; white scale bar, 40 μm. Pathologic information on cancer tissues is shown in the right. B, graphs show quantitative analyses of the colocalization of HSPB1 and CD31 (top) or α-SMA and CD31 (bottom) in tumor vessels per field (magnification, ×100); error bars indicate SEM for 45 cases; ***, P < 0.001; *, P < 0.05 (B).

Close modal

In our previous studies, we showed that HSPB1 was expressed in lung vascular endothelial cells and spontaneous lung cancer tissues of KP mice (37). Here, we found, for the first time, that HSPB1 is a key regulator of the EndMT in lung cancer. HSPB1 was required to maintain endothelial cell identity upon exposure to various fibrosis-inducing cytokines, such as TGFβ; in the absence of HSPB1, endothelial cells showed a higher susceptibility to their effects but a reduced response to angiogenic cytokines, such as VEGF. This was discrepancy, likely resulted from the downregulation of an endothelial cell–specific receptor (VEGFR2) and upregulation of cytokine receptors and related signaling pathways in HSPB1-deficient endothelial cells compared with control endothelial cells. In agreement with our in vitro findings, another study found that overexpression of a phospho-mimic of HSPB1 caused changes in the actin cytoskeleton and reduced cytoskeletal plasticity, thereby increasing the pulmonary endothelial barrier (36). Other studies have shown that HSPB1 strongly colocalizes with cortical actin and is involved in endothelial cell adhesion and motility (45). Recently, we reported that the EndMT might be an early crucial step during RIPF, contributing to vascular collagen deposition appearing before tissue fibrosis (40). Thus, HSPB1-overexpressing endothelial cells may reduce early vascular collagen deposition via the radiation-induced EndMT. Furthermore, we suggest that HSPB1-overexpressing endothelial cells may reduce paracrine effects on other neighboring lung cells through the radiation-induced EndMT, subsequently blocking paracrine signaling-induced fibrotic changes. We also found that the bleomycin-induced EndMT and fibrosis were reduced in HSPB1-overexpressing mice, consistent with the results of other studies showing that HSPB1 can inhibit tissue fibrosis and injury in vivo (30, 33, 34). Our data suggest that HSPB1 deficiency in endothelial cells was sufficient to induce the EndMT in normal lungs and that this occurrence of the EndMT may contribute to the development of pathologic diseases in specific environments, such as RIPF or tumorigenesis. In addition, our data suggest that JAK–STAT signaling is a new target for the regulation of the EndMT, demonstrating the relationship between JAK–STAT signaling and the HSPB1-mediated EndMT in lung fibrosis and tumorigenesis.

HSPB1 is overexpressed in various types of cancer, including lung cancer (26, 27); however, the role of HSPB1 in cancer progression has been controversial. For example, HSPB1 expression has been reported to be lower in malignant tumors than in benign tumors, whereas HSP70, HSP86, and HSP84 levels are higher in salivary gland tumor tissues (46). Our report is the first to show that HSPB1 is highly expressed in the tumor endothelium in human lung tissue. In tumor endothelial cells of KP mice, HSPB1 deficiency promoted the EndMT and increased the number of tumor fibroblasts and the size of the fibrotic region. Indeed, many human lung cancer tissues exhibited severe collagen deposition, which has been linked to tumor metastasis or resistance to anticancer drugs (44, 47).

An unexpected finding of our study was that HSPB1-positive tumor vessels coexpressed CD31 in nonfibrotic regions of human NSCLC tissues. However, HSPB1-negative tumor vessels showed an EndMT-like phenotype and coexpressed CD31 and α-SMA. From these analyses of human tissues, it was not possible to establish a link between the EndMT resulting from the loss of HSPB1 and tumor metastasis or progression; however, HSPB1 deficiency in tumor vascular endothelial cells was closely related to the occurrence of the EndMT in human lung cancer tissues, whereas HSPB1-positive tumor vascular endothelial cells were primarily observed in nonfibrotic regions. Thus, current experiments in our laboratory are focusing on determining whether the HSPB1-regulated EndMT directly regulates tumor metastasis or progression by crossing HSPB1-overexpressing mice onto the LSL-KrasG12D;Trp53fl/fl background.

HSPB1 has been reported to mediate the EMT by promoting the stabilization of SNAIL in idiopathic pulmonary fibrosis and enhancing the EMT in lung cancer, leading to tumor invasion and metastasis (32, 48). These findings, which contradict our results, may be explained by the different roles of the protein in normal and pathologic cells or cell-/tissue-specific patterns of expression (33). Under normal conditions, HSPB1 protects cells against stress via modulation of actin cytoskeleton dynamics, oxidative stress, and apoptosis (27). However, under pathologic conditions, HSPB1 may cause deleterious effects by allowing cancer cells to evade immune surveillance or the effects of anticancer drugs (49).

In summary, our findings strongly suggest that the mesenchymal transition of vascular endothelial cells caused by endothelial HSPB1 deficiency contributes to lung fibrosis and tumorigenesis. We propose that HSPB1 in vascular endothelial cells may play a key regulatory role in the maintenance of the vasculature by inhibiting the EndMT after stimulation with various fibrotic cytokines during these pathologic processes. Targeting of endothelial HSPB1 may thus represent a potential strategy for treating fibrotic diseases and cancer.

No potential conflicts of interest were disclosed.

Conception and design: S.-H. Choi, Y.H. Ji, J. Cho, Y.-J. Lee

Development of methodology: S.-H. Choi, J. Jang, Y.H. Ji, Y.-J. Lee

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.-H. Choi, J.-K. Nam, J. Jang, H.-J. Lee, S. Park, J. Cho, Y.-J. Lee

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.-H. Choi, J.-K. Nam, B.-Y. Kim, Y.-B. Jin, Y.-J. Lee

Writing, review, and/or revision of the manuscript: S.-H. Choi, J.-K. Nam, B.-Y. Kim, J. Cho, Y.-J. Lee

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.-K. Nam, B.-Y. Kim, Y.-J. Lee

Study supervision: J. Cho, Y.-J. Lee

This work was supported by the Nuclear Research and Development Program (grant nos. NRF-2012M2A2A7012377, NRF-2011-0031697, and NRF-2013M2A2A7043580) through the National Research Foundation of Korea, funded by the Ministry of Science, Information and Communications Technology, and Future Planning.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Lopez-Novoa
JM
,
Nieto
MA
. 
Inflammation and EMT: an alliance towards organ fibrosis and cancer progression
.
EMBO Mol Med
2009
;
1
:
303
14
.
2.
Li
J
,
Bertram
JF
. 
Review: endothelial–myofibroblast transition, a new player in diabetic renal fibrosis
.
Nephrology
2010
;
15
:
507
12
.
3.
Saito
A
. 
EMT and EndMT: regulated in similar ways?
J Biochem
2013
;
153
:
493
5
.
4.
Wynn
TA
. 
Cellular and molecular mechanisms of fibrosis
.
J Pathol
2008
;
214
:
199
210
.
5.
Potenta
S
,
Zeisberg
E
,
Kalluri
R
. 
The role of endothelial-to-mesenchymal transition in cancer progression
.
Br J Cancer
2008
;
99
:
1375
9
.
6.
Lin
F
,
Wang
N
,
Zhang
TC
. 
The role of endothelial–mesenchymal transition in development and pathological process
.
IUBMB Life
2012
;
64
:
717
23
.
7.
Moore
CS
,
Mjaatvedt
CH
,
Gearhart
JD
. 
Expression and function of activin beta A during mouse cardiac cushion tissue formation
.
Dev Dyn
1998
;
212
:
548
62
.
8.
Piera-Velazquez
S
,
Li
Z
,
Jimenez
SA
. 
Role of endothelial–mesenchymal transition (EndoMT) in the pathogenesis of fibrotic disorders
.
Am J Pathol
2011
;
179
:
1074
80
.
9.
Sato
Y
,
Nakanuma
Y
. 
Role of endothelial-mesenchymal transition in idiopathic portal hypertension
.
Histol Histopathol
2013
;
28
:
145
54
.
10.
Nie
L
,
Lyros
O
,
Medda
R
,
Jovanovic
N
,
Schmidt
JL
,
Otterson
MF
, et al
Endothelial-mesenchymal transition in normal human esophageal endothelial cells cocultured with esophageal adenocarcinoma cells: role of IL-1beta and TGF-beta2
.
Am J Physiol Cell Physiol
2014
;
307
:
C859
77
.
11.
Hashimoto
N
,
Phan
SH
,
Imaizumi
K
,
Matsuo
M
,
Nakashima
H
,
Kawabe
T
, et al
Endothelial-mesenchymal transition in bleomycin-induced pulmonary fibrosis
.
Am J Respir Cell Mol Biol
2010
;
43
:
161
72
.
12.
Ranchoux
B
,
Antigny
F
,
Rucker-Martin
C
,
Hautefort
A
,
Péchoux
C
,
Bogaard
HJ
, et al
Endothelial-to-mesenchymal transition in pulmonary hypertension
.
Circulation
2015
;
131
:
1006
18
.
13.
Kalluri
R
,
Zeisberg
M
. 
Fibroblasts in cancer
.
Nat Rev Cancer
2006
;
6
:
392
401
.
14.
Augsten
M
. 
Cancer-associated fibroblasts as another polarized cell type of the tumor microenvironment
.
Front Oncol
2014
;
4
:
62
.
15.
Cirri
P
,
Chiarugi
P
. 
Cancer associated fibroblasts: the dark side of the coin
.
Am J Cancer Res
2011
;
1
:
482
97
.
16.
Zeisberg
EM
,
Potenta
S
,
Xie
L
,
Zeisberg
M
,
Kalluri
R
. 
Discovery of endothelial-to-mesenchymal transition as a source for carcinoma-associated fibroblasts
.
Cancer Res
2007
;
67
:
10123
8
.
17.
Erez
N
,
Truitt
M
,
Olson
P
,
Arron
ST
,
Hanahan
D
. 
Cancer-associated fibroblasts are activated in incipient neoplasia to orchestrate tumor-promoting inflammation in an NF-kappaB-dependent manner
.
Cancer Cell
2010
;
17
:
135
47
.
18.
Gasperini
P
,
Espigol-Frigole
G
,
McCormick
PJ
,
Salvucci
O
,
Maric
D
,
Uldrick
TS
, et al
Kaposi sarcoma herpesvirus promotes endothelial-to-mesenchymal transition through Notch-dependent signaling
.
Cancer Res
2012
;
72
:
1157
69
.
19.
Krizbai
IA
,
Gasparics
A
,
Nagyoszi
P
,
Fazakas
C
,
Molnar
J
,
Wilhelm
I
, et al
Endothelial–mesenchymal transition of brain endothelial cells: possible role during metastatic extravasation
.
PLoS ONE
2015
;
10
:
e0123845
.
20.
van Meeteren
LA
,
ten Dijke
P
. 
Regulation of endothelial cell plasticity by TGF-beta
.
Cell Tissue Res
2012
;
347
:
177
86
.
21.
Liebner
S
,
Cattelino
A
,
Gallini
R
,
Rudini
N
,
Iurlaro
M
,
Piccolo
S
, et al
Beta-catenin is required for endothelial–mesenchymal transformation during heart cushion development in the mouse
.
J Cell Biol
2004
;
166
:
359
67
.
22.
Aisagbonhi
O
,
Rai
M
,
Ryzhov
S
,
Atria
N
,
Feoktistov
I
,
Hatzopoulos
AK
. 
Experimental myocardial infarction triggers canonical Wnt signaling and endothelial-to-mesenchymal transition
.
Dis Model Mech
2011
;
4
:
469
83
.
23.
Zeisberg
EM
,
Tarnavski
O
,
Zeisberg
M
,
Dorfman
AL
,
McMullen
JR
,
Gustafsson
E
, et al
Endothelial-to-mesenchymal transition contributes to cardiac fibrosis
.
Nat Med
2007
;
13
:
952
61
.
24.
Magrini
E
,
Villa
A
,
Angiolini
F
,
Doni
A
,
Mazzarol
G
,
Rudini
N
, et al
Endothelial deficiency of L1 reduces tumor angiogenesis and promotes vessel normalization
.
J Clin Invest
2014
;
124
:
4335
50
.
25.
Xiao
L
,
Kim
DJ
,
Davis
CL
,
McCann
JV
,
Dunleavey
JM
,
Vanderlinden
AK
, et al
Tumor endothelial cells with distinct patterns of TGFbeta-driven endothelial-to-mesenchymal transition
.
Cancer Res
2015
;
75
:
1244
54
.
26.
Calderwood
SK
,
Gong
J
. 
Molecular chaperones in mammary cancer growth and breast tumor therapy
.
J Cell Biochem
2012
;
113
:
1096
103
.
27.
Arrigo
AP
,
Simon
S
,
Gibert
B
,
Kretz-Remy
C
,
Nivon
M
,
Czekalla
A
, et al
Hsp27 (HspB1) and alphaB-crystallin (HspB5) as therapeutic targets
.
FEBS Lett
2007
;
581
:
3665
74
.
28.
Doshi
BM
,
Hightower
LE
,
Lee
J
. 
The role of Hsp27 and actin in the regulation of movement in human cancer cells responding to heat shock
.
Cell Stress Chaperones
2009
;
14
:
445
57
.
29.
Lu
X
,
Kakkar
V
. 
The role of heat shock protein (HSP) in atherosclerosis: pathophysiology and clinical opportunities
.
Curr Med Chem
2010
;
17
:
957
73
.
30.
Ferns
G
,
Shams
S
,
Shafi
S
. 
Heat shock protein 27: its potential role in vascular disease
.
Int J Exp Pathol
2006
;
87
:
253
74
.
31.
Ciocca
DR
,
Calderwood
SK
. 
Heat shock proteins in cancer: diagnostic, prognostic, predictive, and treatment implications
.
Cell Stress Chaperones
2005
;
10
:
86
103
.
32.
Wettstein
G
,
Bellaye
PS
,
Kolb
M
,
Hammann
A
,
Crestani
B
,
Soler
P
, et al
Inhibition of HSP27 blocks fibrosis development and EMT features by promoting Snail degradation
.
FASEB J
2013
;
27
:
1549
60
.
33.
Vidyasagar
A
,
Reese
S
,
Acun
Z
,
Hullett
D
,
Djamali
A
. 
HSP27 is involved in the pathogenesis of kidney tubulointerstitial fibrosis
.
Am J Physiol Renal Physiol
2008
;
295
:
F707
16
.
34.
Vidyasagar
A
,
Wilson
NA
,
Djamali
A
. 
Heat shock protein 27 (HSP27): biomarker of disease and therapeutic target
.
Fibrogenesis Tissue Repair
2012
;
5
:
7
.
35.
Hirano
S
,
Rees
RS
,
Yancy
SL
,
Welsh
MJ
,
Remick
DG
,
Yamada
T
, et al
Endothelial barrier dysfunction caused by LPS correlates with phosphorylation of HSP27 in vivo
.
Cell Biol Toxicol
2004
;
20
:
1
14
.
36.
Liu
T
,
Guevara
OE
,
Warburton
RR
,
Hill
NS
,
Gaestel
M
,
Kayyali
US
. 
Modulation of HSP27 alters hypoxia-induced endothelial permeability and related signaling pathways
.
J Cell Physiol
2009
;
220
:
600
10
.
37.
Lee
YJ
,
Lee
HJ
,
Choi
SH
,
Jin
YB
,
An
HJ
,
Kang
JH
, et al
Soluble HSPB1 regulates VEGF-mediated angiogenesis through their direct interaction
.
Angiogenesis
2012
;
15
:
229
42
.
38.
Hong
ZY
,
Eun
SH
,
Park
K
,
Choi
WH
,
Lee
JI
,
Lee
EJ
, et al
Development of a small animal model to simulate clinical stereotactic body radiotherapy-induced central and peripheral lung injuries
.
J Radiat Res
2014
;
55
:
648
57
.
39.
Kirsch
DG
,
Dinulescu
DM
,
Miller
JB
,
Grimm
J
,
Santiago
PM
,
Young
NP
, et al
A spatially and temporally restricted mouse model of soft tissue sarcoma
.
Nat Med
2007
;
13
:
992
7
.
40.
Choi
SH
,
Hong
ZY
,
Nam
JK
,
Lee
HJ
,
Jang
J
,
Yoo
RJ
, et al
A hypoxia-induced vascular endothelial-to-mesenchymal transition in development of radiation-induced pulmonary fibrosis
.
Clin Cancer Res
2015
;
21
:
3716
26
.
41.
Olsson
AK
,
Dimberg
A
,
Kreuger
J
,
Claesson-Welsh
L
. 
VEGF receptor signalling—in control of vascular function
.
Nat Rev Mol Cell Biol
2006
;
7
:
359
71
.
42.
Shibuya
M
. 
Vascular endothelial growth factor and its receptor system: physiological functions in angiogenesis and pathological roles in various diseases
.
J Biochem
2013
;
153
:
13
9
.
43.
Kim
M
,
Choi
SH
,
Jin
YB
,
Lee
HJ
,
Ji
YH
,
Kim
J
, et al
The effect of oxidized low-density lipoprotein (ox-LDL) on radiation-induced endothelial-to-mesenchymal transition
.
Int J Radiat Biol
2013
;
89
:
356
63
.
44.
Spivey
KA
,
Chung
I
,
Banyard
J
,
Adini
I
,
Feldman
HA
,
Zetter
BR
. 
A role for collagen XXIII in cancer cell adhesion, anchorage-independence and metastasis
.
Oncogene
2012
;
31
:
2362
72
.
45.
Trott
D
,
McManus
CA
,
Martin
JL
,
Brennan
B
,
Dunn
MJ
,
Rose
ML
. 
Effect of phosphorylated hsp27 on proliferation of human endothelial and smooth muscle cells
.
Proteomics
2009
;
9
:
3383
94
.
46.
Wang
G
,
Gu
X
,
Chen
L
,
Wang
Y
,
Cao
B
,
E
Q
. 
Comparison of the expression of 5 heat shock proteins in benign and malignant salivary gland tumor tissues
.
Oncol Lett
2013
;
5
:
1363
9
.
47.
Sethi
T
,
Rintoul
RC
,
Moore
SM
,
MacKinnon
AC
,
Salter
D
,
Choo
C
, et al
Extracellular matrix proteins protect small cell lung cancer cells against apoptosis: a mechanism for small-cell lung cancer growth and drug resistance in vivo
.
Nat Med
1999
;
5
:
662
8
.
48.
Mizutani
H
,
Okano
T
,
Minegishi
Y
,
Matsuda
K
,
Sudoh
J
,
Kitamura
K
, et al
HSP27 modulates epithelial to mesenchymal transition of lung cancer cells in a Smad-independent manner
.
Oncol Lett
2010
;
1
:
1011
6
.
49.
Yang
L
,
Cao
L
,
Yang
M
,
Tang
D
,
Kang
R
,
Min
X
, et al
Hsp27: a novel therapeutic target for pediatric M4/M5 acute myeloid leukemia
.
Oncol Rep
2013
;
29
:
1459
66
.

Supplementary data