Exosomal miR-19a from adipose-derived stem cells suppresses differentiation of corneal keratocytes into myofibroblasts

In this study, we investigated the effects of exosomal microRNAs (miRNAs) from adipose-derived stem cells (ADSCs) on the differentiation of rabbit corneal keratocytes. Keratocytes grown in 10% FBS differentiated into myofibroblasts by increasing HIPK2 kinase levels and activity. HIPK2 enhanced p53 and Smad3 pathways in FBS-induced keratocytes. Keratocytes grown in 10% FBS also showed increased levels of pro-fibrotic proteins, including collagen III, MMP9, fibronectin, and α-SMA. These effects were reversed by knocking down HIPK2. Moreover, ADSCs and exosomes derived from ADSCs (ADSCs-Exo) suppressed FBS-induced differentiation of keratocytes into myofibroblasts by inhibiting HIPK2. Quantitative RT-PCR analysis showed that ADSCs-Exos were significantly enriched in miRNA-19a as compared to ADSCs. Targetscan and dual luciferase reporter assays confirmed that the HIPK2 3’UTR is a direct binding target of miR-19a. Keratocytes treated with 10% FBS and ADSCs-Exo-miR-19a-agomir or ADSCs-Exo-NC-antagomir showed significantly lower levels of HIPK2, phospho-Smad3, phospho-p53, collagen III, MMP9, fibronectin and α-SMA than those treated with 10% FBS plus ADSCs-Exo-NC-agomir or ADSCs-Exo-miR-19a-antagomir. Thus, exosomal miR-19a derived from the ADSCs suppresses FBS-induced differentiation of rabbit corneal keratocytes into myofibroblasts by inhibiting HIPK2 expression. This suggests their potential use in the treatment of corneal fibrosis.


INTRODUCTION
The cornea is the outermost part of the eye that acts as a barrier against infections and provides a clear path for light [1]. The cornea is made up of three layers: epithelium, stroma, and endothelium [2]. Nearly 90% of the corneal volume is made up of the stroma, which is primarily responsible for clarity and ocular immunity [3]. The corneal stromal tissue is primarily made up of collagen fibers and extracellular matrix [4]. Keratocytes are the major cells of the stroma that produce collagen and matrix metalloproteinases [5]. However, wound healing response to corneal injury, infections, and surgery decrease corneal transparency and visual acuity [6]. During stromal wound healing, keratocytes differentiate into fibroblasts and myofibroblasts [7]. Moreover, deposition of extracellular matrix and decreased crystallin protein expression by the keratocytes causes scar formation and reduces corneal transparency [8]. Therefore, effective methods are necessary to inhibit Homeodomain-interacting protein kinase 2 (HIPK2) is a serine/threonine kinase that is primarily located in the nucleus of eukaryotic cells [23]. HIPK2 is a pro-fibrotic gene that plays an important role in kidney fibrosis [24]. Previous studies show that HIPK2 regulates fibrosis by acting upstream of p53, Transforming Growth Factor β (TGF-β), SMAD family member 3 (Smad3), and INT-1 (Wnt)/β-catenin pathways [24,25]. Hu et al showed that exosomal miR-1229 promotes angiogenesis of colorectal cancer cells by targeting HIPK2 [26]. The relationship between exosome-derived miRNAs secreted by ADSCs and HIPK2 is not known. Therefore, the aim of this study was to investigate if the exosome-derived miRNAs secreted by ADSCs regulated differentiation of keratocytes into myofibroblasts using the rabbit corneal keratocytes and ADSCs.

FBS induces differentiation of rabbit corneal keratocytes into myofibroblasts
Previous studies have reported that vimentin and CK12 are specifically expressed in the keratocytes and corneal epithelium, respectively [27,28]. Therefore, we examined the expression of stromal and epithelial markers in primary rabbit keratocytes by immunofluorescence assay. The primary rabbit keratocytes showed positive expression of vimentin, but did not express CK12 ( Figure 1A). This confirmed that the rabbit corneal stroma cells were keratocytes and not epithelial cells.
Keratocytes can be differentiated into myofibroblasts when grown in presence of FBS [29]. We observed that keratocytes grown in serum-free medium showed dendritic morphology, whereas, keratocytes cultured with 10% FBS for 7 days exhibited a fibroblast phenotype ( Figure 1B). Furthermore, we performed western blot analysis of the expression of fibroblast-related proteins, such as, keratocan, collagen I, collagen III, MMP9, fibronectin and α-SMA in the cultured keratocytes. Keratocytes grown in 10% FBS showed significantly reduced expression of keratocan and collagen I and increased levels of collagen III, MMP9, fibronectin and α-SMA compared to the controls ( Figure 1C, 1D). These data demonstrate that FBS induced differentiation of keratocytes into myofibroblasts.

Characterization of ADSCs and ADSCs-Exo
Next, we characterized the ADSCs isolated from rabbit adipose tissues. Flow cytometry analysis showed positive surface expression of CD29 and CD90 and absence of CD34 and CD45 expression in the primary ADSCs (Figure 2A, 2B). This confirmed successful isolation of ADSCs from the rabbit adipose tissues.
Furthermore, we analyzed the exosomes isolated from the ADSCs. Nanoparticle tracking analysis (NTA) showed that the ADSCs-Exo were approximately 100 nm in diameter with typical cup-shaped morphology ( Figure  2C). Western blot analysis showed higher expression of exosomal markers, namely, CD9, CD81 and flotillin-1 in the ADSCs-Exo compared with the ADSCs (Figure 2D, 2E). These data confirmed isolation of purified ADSCs-Exo from the ADSC culture supernatants.

ADSCs-Exo inhibits FBS-induced differentiation of keratocytes into myofibroblasts
We then characterized the effects of ADSCs-Exo on the keratocytes that were grown in DMEM/F12 medium AGING containing 10% FBS for 7 days. CCK-8 cell proliferation assay showed that 10% FBS significantly  induced proliferation of corneal myofibroblasts at 24,  48, 72, and 96 h time points, whereas, FBS-induced  keratocyte cell proliferation was significantly inhibited  by ADSCs and ADSCs-Exo at 48, 72 and 96 h (Figure  3A). Moreover, keratocytes grown in DMEM/F12 medium containing 10% FBS showed significantly reduced keratocan and collagen I protein expression, and increased collagen III, MMP9, fibronectin and α-SMA protein levels compared with those grown in serum-free medium, but, these FBS-induced changes were inhibited when keratocytes were grown in presence of ADSCs or ADSCs-Exo ( Figure 3B-3D). These data suggest that ADSCs and ADSCs-Exo inhibit differentiation of keratocytes into myofibroblasts and the proliferation of myofibroblasts.

HIPK2 is the direct binding target of miR-19a
Next, we investigated the effects of ADSCs-Exo on the differentiation of keratocytes into myofibroblasts. Exosomes act as key mediators of intercellular communication by delivering miRNAs such as miR-19a-3p, miR-18a-5p and miR-30c-5p to recipient cells [30,31]. QRT-PCR analysis showed that miR-19a-3p levels were significantly higher in the ADSCs-Exo compared with the ADSCs ( Figure 5A).
TargetScan analysis suggested that HIPK2 is a potential target of miR-19a-3p ( Figure 5B). Moreover, miR-19a levels were significantly upregulated in miR-19a agomir transfected keratocytes cultured in medium containing AGING 10% FBS compared with the controls ( Figure 5C). Dual luciferase reporter assay results showed that miR-19a suppressed the luciferase activity of the psiCHECK-2-HIPK2-WT construct, but did not affect the luciferase activity of the psiCHECK-2-HIPK2-MUT construct ( Figure 5D). These results confirmed that miR-19a directly targeted the 3'-UTR of HIPK2.

DISCUSSION
Previous investigations show that corneal keratocytes incubated with 10% FBS differentiate into myofibroblasts [32]. Moreover, ADSCs can be induced to differentiate into functional keratocytes under specific growth conditions [1,33]. ADSCs show immense therapeutic potential because they can mediate changes in cellular functions and signaling by secreting exosomes [34]. In this study, we demonstrate that ADSCs-Exo inhibit FBS-induced differentiation of corneal keratocytes by upregulating the levels of keratocan and collagen I, and downregulating the levels of collagen III, MMP9, fibronectin and α-smooth muscle action (α-SMA). Functional keratocytes express cornea-specific proteoglycans, such as keratocan and collagen I, but do not express collagen III [2,35]. The myofibroblasts are characterized by high expressions of α-SMA, fibronectin and some ECM components [36]. Verhoekx et al showed that ADSCs inhibit myofibroblasts in Dupuytren's disease by downregulating α-SMA [37]. These data are consistent with our results, which show that ADSCs-Exo inhibit AGING FBS-induced keratocyte differentiation, thereby suggesting the potential to regenerate the corneal stroma.
Exosomes are small vesicles that are released by all cells, and carry lipids, proteins, DNA, and RNAs, including mRNAs and miRNAs [38]. ADSCs-Exos are enriched with miRNAs that can modulate cellular functions in recipient cells [39]. Fang et al demonstrated that exosome-derived miRNA-21 and miR-23a suppress myofibroblast differentiation during wound healing by
The results of the dual luciferase reporter assay confirmed that HIPK2 was a direct binding target of miR-19a. MiR-19a is enriched in the exosomes derived from the mesenchymal stromal cells [42]. In this study, we demonstrate that miR-19a is enriched in the exosomes derived from the ADSCs. Moreover, miR-19a in the ADSCs-Exo inhibits the expression of HIPK2 in the keratocytes cultured with 10% FBS. Souma et al showed that the miR-19a-19b-20a sub-cluster inhibits TGF-β-induced activation of fibroblasts in patients with pulmonary fibrosis [43]. Furthermore, miR-133b repairs corneal stroma by downregulating α-SMA and prevents scar formation [44]. Our data indicates that exosomal miR-19a derived from the ADSCs inhibits fibrosis by suppressing HIPK2.

Figure 7. Putative mechanism by which ADSCs-Exo-miR-19a suppress the FBS-induced differentiation of rabbit corneal keratocytes into myofibroblasts. ADSCs-Exo-miR-19a suppresses FBS-induced differentiation of keratocytes into myofibroblasts by
inhibiting HIPK2 expression. Reduced HIPK2 levels suppress the TGF-β/Smad3 and p53 signaling pathways, and reduce the expression of profibrotic markers and ECM components. Overall, these events decrease cell viability and ECM degradation. AGING Moreover, ADSCs-Exo-miR-19a decreased the levels of α-SMA and ECM-related proteins in keratocytes cultured with 10% FBS. This suggests that ADSCs-Exo-miR-19a can potentially regenerate corneal stroma by inhibiting keratocyte differentiation. Yin et al showed that ADSCs-Exo-miR-181-5p inhibits liver fibrosis by downregulating α-SMA [46]. Extracellular vesicles in the human serum contain miRNAs that inhibit liver fibrosis by decreasing the expression of pro-fibrotic genes [47]. The findings of our study are in agreement with these reports.
In conclusion, our study shows that ADSCs-Exo-miR-19a inhibits the differentiation of corneal keratocytes into myofibroblasts by suppressing HIPK2 expression. The downregulation of HIPK2 inhibits the TGFβ/Smad3 and p53 pathways, which results in reduced expression of pro-fibrotic markers and ECM components, thereby decreasing cell viability and ECM degradation. Therefore, our data indicates the therapeutic potential of ADSCs-Exo-miR-19a in repairing damaged corneal stromas.

Isolation and culturing of primary rabbit corneal keratocytes
Ten week-old New Zealand male rabbits weighing 2.3 -2.5 kg were purchased from the Zhenlin Biotechnology Co. Ltd (Jiangsu, China). They were housed under standard conditions (temperature: 18 -22 °C; relative humidity, 50% -70%; noise level: 60 dB; 12 h light and dark cycle) and fed a standard rabbit diet and normal water ad libitum. The animal study was approved by the Institutional Ethics Committee of the Zhejiang Hospital. Corneal keratocytes were obtained from the rabbit eyes as previously described [48,49]. Briefly, the corneal stroma layer was dissected into small fragments, digested with collagenase type II (Thermo Fisher Scientific, Waltham, MA, USA), and cultured in DMEM/F12 medium (Thermo Fisher Scientific) containing 10% FBS (Thermo Fisher Scientific) at 37°C for 7 days to generate myofibroblasts. The cellular phenotype of the myofibroblasts was monitored using a laser scanning confocal microscope (Olympus CX23 Tokyo, Japan).

Western blotting
Total protein extracts were prepared by lysing the keratocytes and other cultured cells cells using the RIPA buffer (Beyotime, Shanghai, China) and the protein concentrations were measured using the BCA Protein Assay Kit (Thermo Fisher Scientific). Then, equal amounts of protein samples (30 μg per lane) were separated on 10 % SDS-PAGE gels, transferred onto PVDF membranes (Thermo Fisher Scientific), and blocked with 5% skimmed milk at room temperature. This was followed by incubation with primary antibodies against Keratocan (1:1000), p-p53 (1:1000), p53 (1:1000) at 4°C overnight. Then, the membranes were incubated with the secondary goat anti-rabbit IgG antibody (1: 5000) at room temperature for 1 h. The blots were developed using ECL detection reagents (Thermo Fisher Scientific). The protein bands were scanned using the Odyssey infrared scanner (LICOR Biosciences, Lincoln, NE, USA), and analyzed with the Odyssey v2.0 software. All antibodies were obtained from Abcam.

Isolation of rabbit ADCSs from adipose tissue
ADSCs were isolated from the subcutaneous adipose tissue obtained from the groin of the rabbits as previously described [50]. Briefly, the subcutaneous adipose tissue fragments were digested by incubation with collagenase type II and then treated with 10% FBS (Thermo Fisher Scientific) to terminate the enzyme reaction. The primary ADSCs were cultured for 15 days in DMEM/F12 medium containing 10% FBS and 100 U/mL streptomycin/penicillin at 37°C and 5% CO 2 .

Flow cytometry
The primary ADSCs were stained with fluoresceinconjugated antibodies against CD29, CD90, CD34, and CD45 (Thermo Fisher Scientific) and analyzed using a BD flow cytometer (BD Biosciences, Mountain View, CA, USA). Briefly, the ADSCs were incubated on ice for 30 min with each antibody (1: 100 dilution), washed with brilliant stain buffer (BD Biosciences, Franklin Lake, NJ, USA), centrifuged to remove unbound antibodies in the supernatant, resuspended in brilliant stain buffer, and analyzed by flow cytometry.

Isolation and characterization of rabbit ADSC exosomes
A total of 5 x 10 6 ADSCs were grown in complete culture medium for 24 h. The medium was then replaced with serum-free DMEM/F12 medium and the cells were cultured for another 24 h. The exosomes were isolated from this cell culture medium using the Exosome isolation kit (Thermo Fisher Scientific) according to the manufacturer's protocol. The ADSC exosomes (ADSCs-Exo) were pelleted by ultracentrifugation. The exosome pellet was resuspended in PBS and stored at -80°C. The ADSCs-Exos were characterized by nanoparticle tracking analysis (NTA), and western blotting. In the NTA assay, the size, distribution, and the number of particles in the ADSCs-Exo were evaluated using a nanoparticle tracking analyzer (v3.1, Malvern Instruments, Ltd., Worcestershire, UK). The ADSCs-Exo were analyzed by western blotting using the following primary antibodies: anti-CD9 (1:1000, Abcam), anti-CD81 (1:1000, Abcam), and antiflotillin-1 (1:1000, Abcam).

Cell proliferation assay
Cell proliferation was determined using the CCK-8 kit (Beyotime Biotechnology, Suzhou, China) according to the manufacturer's instructions. Briefly, rabbit corneal keratocytes were grown in DMEM/F12 medium containing 10% FBS at 37°C for 7 days. Then, 5 x 10 3 rabbit corneal keratocytes per well were incubated for 0, 24, 48, 72 and 96 h at 37°C in the presence of 10% FBS. The control keratocytes were grown in serum-free DMEM/F12 medium. At the defined time points, cells were incubated with 10 μL of the CCK8 reagent at 37°C for another 2 h. Then, the optical density (OD) was determined at 450 nm using a microplate reader (Bio-Tek Instruments Inc., Winooski, VT, USA).

HIPK2 kinase activity
Total protein lysates were prepared by cellular lysis using the RIPA buffer (Beyotime, Shanghai, China). HIPK2 activity in the samples was detected as previously described by Millipore (Calbiochem-Merck-Millipore, Darmstadt, Germany) [30]. Recombinant HIPK2 was used to construct a standard curve and the myelin basic protein (MBP) was used as the substrate.

Lentivirus production and cell transfection
The pHBAd-CMV HIPK2 cDNA and lentiviral vector plasmids were obtained from GenePharma. The HIPK2 plasmids were co-transfected into 293T cells with the backbone plasmid (pHBAd-BHG). The lentiviral particles were collected from the supernatant at 72 h after transfection at 32°C and concentrated by centrifugation. The rabbit keratocytes (4 x 10 5 cells / well) were grown in 60 mm cell plates at 37°C overnight. Then, the cells were transfected with HIPK2 cDNA-containing lentiviral supernatants for 24 h. The medium containing the virus was then replaced with fresh complete medium, and the positively transfected cells were selected in medium containing 2.5 μg/mL puromycin (Thermo Fisher Scientific) for 3 days. The qRT-PCR assay was used to assess the levels of HIPK2 in different experimental groups of keratocytes.

Statistical analysis
All data were analyzed using the GraphPad Prism software version 7 for windows (GraphPad Software, La Jolla, CA, USA) and presented as mean ± SD. The differences between two experimental groups were analyzed using the Student's t-test, whereas, comparisons between multiple experimental groups were estimated using the one-way analysis of variance (ANOVA) followed by Tukey's test. All experiments were performed at least thrice and P < 0.05 was considered statistically significant.