The investigation conformed with the principles outlined in the Declaration of Helsinki. Informed consent was obtained from all of the subjects before participation. Human right ventricular outflow tracts (RVOTs) were obtained from a total of 19 patients who had cardiac operations for congenital heart diseases in our department. Seven patients with an SpO₂ (pulse oxygen saturation) of less than 85% were assigned to the cyanotic group, seven patients whose SpO₂ was above 95% were assigned to the acyanotic group, and the other 5 patients were assigned to the middle group. The clinical data of these patients is shown in Table I. Samples of human stenotic right outflow tracts were immediately preserved in 4% poly formaldehyde or liquid nitrogen. All procedures in our study were approved by the Human Ethical Committee of Xinqiao Hospital.

All animal protocols were approved by the Institutional Animal Care and Use Committee of the Army Medical University. All of the experiments were performed on age-matched male mice. A hypoxia workstation (BAKER RUSKINN's InvivO2-1000) maintaining a 10% O₂ environment content at 24°C was used to raise the 8-week-old mice for four weeks to mimic chronic hypoxia conditions (hypoxic group, n=10). A group of age-, weight- and sex-matched C57 BL/6 mice were exposed to 21% O2 condition for 4 weeks (normoxic group, n=10; Fig. 1). There were 10 mice in each group.

Adult cardiac fibroblasts were isolated from 12-week-old mice. Briefly, 3–5 adult mouse hearts were obtained from freshly euthanized mice, cut into 1-mm³ granules and digested with a mixture of 0.1% trypsin and 0.1% collagenase II in a 37°C water bath for 1 h. Then, the digested products were filtered through a 200-mesh screen filter to remove undigested cellular aggregates, non-cellular tissue components and debris and were centrifuged at 1,000 rpm for 5 min. The cell pellet was resuspended in DMEM/F12 with 10% FBS (Fetal Bovine Serum), and then the cells were plated and incubated in 95% air and 5% CO₂ at 37°C for 1 h. The supernatant was discarded, and new culture medium was added. We cultured the cardiac fibroblasts in a 1% O₂ and 5% CO₂ cell incubator (Thermo Fisher Scientific, Inc., Waltham, MA, USA) for the hypoxia treatment for 0, 6, 12 and 24 h.

Adult cardiac fibroblasts were cultured to 80% confluence in 6-well plates and transfected with three short hairpin RNAs (shRNAs) targeting TUG1 (sh-TUG1 1, sh-TUG1 2 and sh-TUG1 3) and the negative control shRNA were inserted into GV102 vectors (sh-NC), which were all synthesized by GeneChem Corporation (Shanghai, China). The miR-29c mimics and the negative controls were purchased from Ribobio Corporation (Guangzhou, China). The miR-29 mimics, sh-TUG1s or their respective negative controls were then transfected into cells. The predicted binding sites of TUG1 and miR-29c and their corresponding mutations were respectively recombined into a vector containing the luciferase gene. Each of these vectors or negative controls were cotransfected with the miR-29c mimic or its negative control into HEK293 cells for 48 h using the Fugene^®6 Transfection Reagent (Promega Corporation, Madison, WI, USA) according to the manufacturer's protocol. Transfected cells were cultured for an additional 24 h prior to hypoxia treatment.

We performed a Cell Counting Kit-8 (CCK8) assay (Beyotime, China) to analyze cell proliferation. The CCK-8 reagent was added to each well in their own condition with or without hypoxia culture, and the cells were incubated for 2 h. The optical density (OD) of each well was recorded at 450 nm on a microplate reader (Thermo Fisher Scientific, Inc.).

Masson's trichrome was used to stain all forms of collagen in the heart. All of the tissues maintained in 4% poly formaldehyde were embedded in paraffin and cut into 5-µm sections. After dewaxing the sections, stained the sections with Mayer-s Hematoxylin, Ponceau S, 1% phosphomolybdic acid aqueous solution and aniline blue (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) successively, and dehydrated in 95% ethyl alcohol (Chuandong Chemical Industry, Chongqing, China) several times, followed by anhydrous alcohol (Chuandong Chemical Industry). Finally, the samples were hyalinized with dimethylbenzene (Beijing Solarbio Science & Technology Co., Ltd.) and sealed with neutral balsam (ZSGB-Bio, Beijing, China). The stain for the collagen fibers was blue, the cytoplasm, muscle fibers and red blood cells were red, and the nuclei were black.

The staining sections were performed on a Leica inverted microscope. The images were captured 5 different vision from one section and we used average fibrotic area ratio of one section as the average fibrotic level from one sample The fibrotic area ratio was quantified by calculating the percentage of collagen staining (blue staining) area on the whole section area (all the staining tissue areas) using Image-pro Plus v.6.0.

The frozen tissues were embedded in OCT and sectioned to 6 µm thickness. After being fixed with 4% paraformaldehyde, the sections were blocked with normal goat serum for 1 h and then incubated with primary antibodies overnight at 4°C. And the sections were incubated with the corresponding secondary antibodies conjugated to Alexa Fluor 488 or 647 (Beyotime Institute of Biotechnology, Haimen, China). The following primary antibodies were used: Anti-Vimentin (1:100; BosterBio Co., Ltd., Wuhan, China) and anti-α-SMA (1:200, Sigma-Aldrich; Merck KGaA, Darmstadt, Germany). Immunofluorescence was performed on a Leica confocal microscopy system.

Total RNA was extracted from human tissue, mouse tissue and cells using TRIzol (Invitrogen; Thermo Fisher Scientific, Inc.). A total of 1 ml TRIzol was added to 100 mg of tissue or to 10⁶ cells, and the subsequent steps were performed according to the manufacturer's instructions. RNA concentration and purity were estimated using a NanoDrop 1000 (Thermo Fisher Scientific, Inc.). Reverse transcription was carried out using the RT PrimeScript™ RT reagent kit (Perfect Real Time; Takara Bio, Inc., Otsu, Japan) according to the manufacturer's instructions. Amplification of the cDNA products was performed with SYBR^® Premix Ex Taq™ II (Perfect Real Time; Takara Bio, Inc.) on an Applied Biosystems 7500 Fast Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA). The amplification kit used U6 snRNA as the internal control for the microRNAs and used β-actin (ACTB) as the internal control for the lncRNAs. The primers for the miR-29 families and U6 detection were purchased from Guangzhou RiboBio Co., Ltd, (Guangzhou, China), and the primers for ACTB and TUG1s were synthesized by Shengong Biotech Co., Ltd., (Shanghai, China) and are shown in Table II.

Total protein was extracted from tissues and cell cultures using Cell Extraction Buffer (Thermo Fisher Scientific, Inc.) and the Mini Protease Inhibitor Cocktail (Roche Diagnostics, Basel, Switzerland), and proteins were separated via SDS-PAGE. The gels were wet transferred to PVDF membranes (Roche Diagnostics), incubated overnight with primary antibodies, and then incubated with secondary antibodies. Immunoreactivity was revealed by chemiluminescence using a western blot imaging system (GE Healthcare, Chicago, IL, USA), and the gray value was analyzed by ImageJ 1.50i. β-actin was used as the internal reference. The normoxic group was used as the control for the protein level analysis in the mouse tissue protein, and the 0-h hypoxia cultured cells, untransfected and normoxic cultured cells served as controls for their own groups. All of the samples were run at least in triplicate. The antibodies utilized were α-SMA (1:1,000), collagen I (1:1,000) and beta-actin (internal reference; 1:1,000; Boster Co., Ltd).

The predicted targeted sequence wide type (WT) of TUG1 and a mutated (MUT) sequence of TUG1 with mutations at the predicted binding site of miR-29c were inserted into the GV272 control vector in the XbaI site, which was purchased from GeneChem, Inc., (Shanghai, China). For the dual-luciferase assay, HEK293 cells were transfected with the WT vector and the miR-29c mimics or the MUT vector and the miR-29c mimics. Following transfection, the activities of Firefly luciferase and Renilla luciferase were measured with a dual-luciferase reporter assay kit (Promega E1910; Promega Corporation) on a luminometer (i3×; Molecular Devices, LLC, Sunnyvale, CA, USA) according to the manufacturer's instructions.

Data are presented as the mean ± SEM. Statistical significance was evaluated using unpaired Student's T-tests with Welch's correction for comparing two groups. For comparing three or more groups, the means were compared using one-way analysis of variance accounting for differences in variance among the groups. Post hoc pairwise comparisons were performed using either the LSD or Tukey-Kramer method. Data were analyzed using GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA, USA) or SPSS v.22.0 (IBM Corp., Armonk, NY, USA) statistical software.

To evaluate whether the degree of fibrosis was correlated with oxygen saturation, we collected (RVOT) muscle from 19 patients (7 cyanotic, 7 acyanotic and 5 middle). The clinical characteristics are summarized in Table I. We divided the patients into three groups according to the SpO₂, with the cyanotic group being below 85%, the acyanotic group being above 95% and the rest of the patients was the middle group. Age, sex and body weight were all matched between the cyanotic and acyanotic groups. We evaluated cardiac fibrosis via Masson's trichrome labeling (Fig. 2A). An analysis of the Masson's trichrome-stained tissue sections from the cyanotic patients showed a significant deterioration of fibrosis area ratio (collagen area ratio) compared with that of the acyanotic patients (Fig. 2B). We also analyzed the SpO₂ and degree of fibrosis in the sections of all 19 patients, and found that the degree of fibrosis was negatively correlated with the oxygen saturation (Fig. 2C). Given that the fibroblast-to-myofibroblast transition (FMT) is a critical component of cardiac fibrosis, we examined the myofibroblast ratio of these sections via confocal microscopy (Fig. 2D). The myofibroblast/fibroblast ratio was consistent with the Masson's trichrome staining, with the cyanotic group showing a higher myofibroblast ratio than the acyanotic group (Fig. 2E). The ratio was negatively correlated with the oxygen saturation in the tissue samples of all 19 patients (Fig. 2F), while it was positive correlated to the collagen area (Fig. 2G). Recently, lncRNAs were reported to participate in the physiological and pathological progression of the heart (18), and we found that TUG1 lncRNA expression was increased in the cyanotic patients compared with the acyanotic patients (Fig. 2H). These data indicated that cardiac fibrosis was enhanced in heart tissues with lower oxygen saturation and that TUG1 expression was coincident with the level of cardiac fibrosis.

Compared with the normoxic group, the hypoxic mice weighed less but had a greater heart/body weight ratio (Fig. 3A). We measured the degree of cardiac fibrosis by Masson's trichrome staining. The Masson's trichrome staining showed that the fibrotic area ratio (collagen area ratio) were significantly greater in the hypoxic group than in the normoxic group (Fig. 3B and C). We also evaluated the myofibroblast ratio via confocal microscopy, and the myofibroblast ratio of the hypoxic group was significantly greater than that in the normoxic group (Fig. 3D and E). In addition, western blotting revealed that the cardiac tissue had higher relative expression of collagen I (COLI) and α-SMA than in the normoxic group (Fig. 3F and G), and one of these (COLI) is a major component of fibrotic lesions and the other (α-SMA) is a marker of the differentiation of fibroblasts to myofibroblasts. Meanwhile, the level of TUG1 was increased (1.84±0.48-fold) in the hypoxic group compared with that in the normoxic group. These results indicated that hypoxia is strongly correlated with FMT and fibrosis, and TUG1 might play a role in this process.

To further confirm the role of TUG1 in cardiac FMT induced by hypoxia, we detected the cardiac fibroblast proliferation and the expression of COLI, α-SMA and TUG1 in the cardiac fibroblasts under normoxic and hypoxic conditions. The CCK-8 results showed that cardiac fibroblast proliferation increased with the hypoxic time (Fig. 4A). The COLI and α-SMA levels increased with the hypoxic time, which was coincidence with the HIF-1α expression (Fig. 4B and C). Furthermore, TUG1 expression was also increased with the hypoxic time (Fig. 4D). We next knocked down the expression of TUG1 by transfecting cardiac fibroblasts with vectors containing the TUG1 shRNAs (shRNA1, shRNA2 and shRNA3) or the vehicle control (empty vector). All three TUG1 shRNAs decreased TUG1 expression, with the greatest decrease observed with shRNA1 (Fig. 4E). After the transfected cells were cultured under hypoxic conditions for 24 h, the levels of TUG1 were increased. Although the expression of TUG1 increased under hypoxia, the levels in the knockdown groups (using the three TUG1 shRNAs) still remained lower than in those in the blank and negative groups (Fig. 4F). We also found that proliferation was significantly decreased in the TUG1 shRNA-transfected cells, especially those transfected with shRNA1 (Fig. 4G). Cell proliferation was still increased under hypoxia though the levels were lower in the three knockdown groups. The COLI and α-SMA levels were also decreased in the shRNAs-transfected cells compared with those in the blank and vehicle control cells (Fig. 4H and I). Besides, COLI and α-SMA levels were all increased after hypoxia but the levels were lower in the three knockdown groups, which was consistent with the TUG1 level (Fig. 4H and I). Therefore, we concluded that TUG1 facilitated cardiac FMT in hypoxia.

Some researchers report that miR-29c shows reduced expression in TOF patients according to microarray analysis (19) and is related to hypoxia. Increasing evidence indicates that the miR-29s family plays an important role in cardiac fibrosis (20,21). We found that TUG1 expression was correlated with the miR-29s family in human breast, kidney and lung cancer tissues from the starBase V2.0 database, and the predicted target sites of miR-29c in human TUG1 are shown in Fig. 5A. Considering the conservation of lncRNAs and microRNAs, we blasted the TUG1 and miR-29c target sites in the mice genome and found they were conserved (Fig. 5A). The information of recombined vectors showed in the Fig. 5B. The dual-luciferase reporter system results showed that the relative luciferase expression of the WT group with miR-29c mimics was significantly lower than those of the negative control group with miR-29c mimics and the MUT group with miR-29c mimics (Fig. 5C), confirming the predicted binding sites of TUG1 and miR-29c.

To further examine whether TUG1 participated in the cardiac FMT induced by hypoxia via miR-29c, we examined the miR-29c levels of the clinical samples and hypoxic mice. We found that the expression of miR-29c was lower in the cyanotic group and hypoxic group, compared with the acyanotic and normoxic group, respectively (Fig. 5D). The expression of miR-29c in cardiac fibroblasts appeared to decrease with the hypoxic time (Fig. 5E). All the human and mice tissues and the cell data indicated that the expression of TUG1 and miR-29c were negatively correlated (Fig. 5F). We then examined the miR-29c level after TUG1 knockdown. MiR-29c expression was significantly increased in the cardiac fibroblasts transfected with TUG1 shRNAs, especially the cells transfected with shRNA1, whereas miR-29c expression decreased slightly with hypoxia (Fig. 5G and H). COLI is a target gene of miR-29c, and α-SMA is also regulated by miR-29c in cardiac fibrosis (22). In conclusion, TUG1 might promote the cardiac FMT in hypoxia via miR-29c.

To further confirm the function of miR-29c in cardiac fibroblasts, we overexpressed miR-29c in adult mouse cardiac fibroblasts by treating them with an miR-29c sequence mimic and its negative control (Fig. 6A). Using the CCK8 assay, we found that the hypoxia group had higher OD values, whereas the mimic-treated cardiac fibroblasts had lower OD values in both of their own groups. Hypoxia stimulated the proliferation of cardiac fibroblasts, which was restrained to some extent by the overexpression of miR-29c (Fig. 6B). Confocal imaging revealed that α-SMA expression was reduced in the normoxic group compared with that in the hypoxic group, and the mimic group under hypoxic conditions had lower α-SMA expression than the control and negative control groups (Fig. 6C). In addition, western blotting indicated that the hypoxic group had greater α-SMA and collagen I levels than the normoxic group, whereas the mimic-treated groups had lower α-SMA and collagen I levels under both normoxic and hypoxic conditions (Fig. 6D and E). These results suggested that overexpressing miR-29c alleviates the elevated α-SMA and collagen I levels induced by hypoxia, that is the overexpression of miR-29c indeed ameliorated activation of FMT induced by hypoxia. And higher level of TUG1 induced by hypoxia might restrain inhibiting effect of miR-29c on cardiac FMT. Collectively, we conclude the lncRNA TUG1 might function as the sponge of miR-29c to promote FMT in hypoxia (Fig. 7).