Long Non-Coding RNA H19 Prevents Lens Fibrosis through Maintaining Lens Epithelial Cell Phenotypes

The integrity of lens epithelial cells (LECs) lays the foundation for lens function and transparency. By contrast, epithelial-mesenchymal transition (EMT) of LECs leads to lens fibrosis, such as anterior subcapsular cataracts (ASC) and fibrotic forms of posterior capsule opacification (PCO). However, the underlying mechanisms remain unclear. Here, we aimed to explore the role of long non-coding RNA (lncRNA) H19 in regulating TGF-β2-induced EMT during lens fibrosis, revealing a novel lncRNA-based regulatory mechanism. In this work, we identified that lncRNA H19 was highly expressed in LECs, but downregulated by exposure to TGF-β2. In both human lens epithelial explants and SRA01/04 cells, knockdown of H19 aggravated TGF-β2-induced EMT, while overexpressing H19 partially reversed EMT and restored lens epithelial phenotypes. Semi-in vivo whole lens culture and H19 knockout mice demonstrated the indispensable role of H19 in sustaining lens clarity through maintaining LEC features. Bioinformatic analyses further implied a potential H19-centered regulatory mechanism via Smad-dependent pathways, confirmed by in vitro experiments. In conclusion, we uncovered a novel role of H19 in inhibiting TGF-β2-induced EMT of the lens by suppressing Smad-dependent signaling, providing potential therapeutic targets for treating lens fibrosis.


Introduction
The ocular lens is a unique transparent organ whose homeostasis is maintained mainly by a single layer of lens epithelial cells (LECs) underneath the anterior lens capsules [1]. Without proper functioning of LECs, lens transparency cannot be sustained; instead, opacification of the lens canH lead to cataracts, which are the leading cause of vision loss across the globe [2].
Upon exposure to external insults, such as ocular trauma or surgeries, transforming growth factor-beta (TGF-β), which is usually latent [3] and tightly regulated in the lens, increases dramatically to a great extent [4,5]. In particular, TGF-β2 predominates in human ocular media [6][7][8][9]. The abnormally activated TGF-β signaling leads to the pathological transformation of LECs to mesenchymal cells with matrix deposition and capsule wrinkling, contributing to epithelial-mesenchymal transition (EMT). During this process, LECs lose their apicobasal cell polarity and intercellular junctions, acquire mesenchymal characteristics, and transdifferentiate into multi-layered spindle-shaped myofibroblasts with an accumulation of α-smooth muscle actin (αSMA) [10,11]. These cells are prone to proliferate and migrate aberrantly, accompanied by an abnormal secretion of fibronectin (FN) and collagen I/III, not typically expressed in the lens, into the extracellular matrix (ECM), driving the contraction of the lens capsules [12,13]. Importantly, EMT lays the foundation for fibrotic cataracts, including fibrotic forms of posterior capsular opacification (PCO) and anterior subcapsular cataracts (ASC) [4,5,8,9,14].
Accumulative studies have pointed out that TGF-β2-induced EMT is a crucial feature manifested in embryonic development and various pathological phenomena, particularly wound healing, fibrotic diseases, and some invasive cancers [15][16][17]. EMT triggers lens fibrosis and can be targeted at multiple levels [18][19][20]. Long non-coding RNAs (LncRNAs) are a group of RNAs whose length exceeds 200 nucleotides but cannot be translated into functional proteins [21]. Instead of serving as genetic background noises, they emerge as a breakthrough for deciphering the molecular mechanism of both biological and pathological processes, including EMT [22][23][24]. More importantly, their mechanisms of action are diverse and still up-and-coming: lncRNAs can epigenetically regulate gene expression through binding to DNA or chromatin modifiers [25,26]. They can also sequestrate microRNAs (miRNAs) or act as a scaffold, decoy, or guide for vital proteins [27]. Notably, H19 is the first discovered eukaryote lncRNAs with a relatively high evolutionary conservancy in humans and mice [28,29], and plays an essential role in human embryonic development and diverse conditions, including invasive cancers and fibrotic diseases [30,31]. However, its role in the pathogenesis of fibrotic cataracts remains unclear.
Using cutting-edge RNA-sequencing, this study first identified that H19 was highly expressed in LECs but downregulated upon TGF-β2 treatment. Then, both loss-and gainof-function studies conducted in human lens epithelial explants and cell lines revealed that H19 could help maintain LECs properties and reverse TGF-β2-induced EMT in the lens. In addition, semi-in vivo whole lens culture studies showed that when H19 was forcibly knocked down, lens transparency could not be sustained following the loss of LEC features. Lens opacities similar to ASC were also observed in homozygous H19 knockout mice. Furthermore, we explored possible downstream mechanisms responsible for its functions, highlighting the potential role of the H19/miR-675-3p/SMAD4 axis in fibrotic cataracts. To conclude, this study revealed H19 as an indispensable modulator of TGF-β2-induced EMT in the lens, implying that H19 could be a novel prophylactic or therapeutic target for tackling human fibrotic cataracts.

Human Lens Epithelial Explants Collection, Cell Culture, and Treatment
All human samples collected from organ donors were provided by the Guangdong Provincial Eye Bank. We first obtained written informed consent from the family members of organ donors, for each human subject involved. This study was approved by the institutional review board of the Zhongshan Ophthalmic Center (Project identification code: 2019KYPJ118; Date of approval: 20 September 2019) and conducted following the principles of the Declaration of Helsinki. The ages of donors ranged from 30 to 50 years old. With capsulotomy scissors dissecting the lenses at the equatorial regions and all lens fibers removed, the remaining anterior lens capsules with underlying lens epithelial cells were then collected and annotated as human LEEs. Immediately after collection, they were cultured in Minimum Essential Medium (MEM, Thermo Fisher Scientific, Waltham, MA, USA) containing 1% fetal bovine serum (FBS, Thermo Fisher Scientific, Thornton, AU) with 1%NEAA (Life Technologies, Carlsbad, CA, USA), as previously described [32]. For the TGF-β2-induced EMT model, the human lens epithelial explants were first cultured for at least 6 h to ensure the excellent condition of the LECs, and then treated with 10 ng/mL TGF-β2 (R&D Systems, Minneapolis, MN, USA) for 48 h. To minimize variations caused by sample selection, we used match-paired epithelial explants of each eye from the same donor as treatment and control. Samples from age-matched donors were also adopted as biological replicates for statistical analysis. In total, 6 samples from three age-matched donors (annotated 48y_Con, 48y_T2, 52y_Con, 52y_T2, 56y_Con, and 56y_T2) were used for RNA-Sequencing. The human lens epithelial cell line SRA01/04 was kindly offered by Professor Fu Shang of Tufts University (Boston, MA, USA) and cultured in Dulbecco's Modified Eagle's medium (DMEM, Thermo Fisher Scientific, Waltham, MA, USA) with 10% FBS, with or without TGF-β2 (5 ng/mL).

In Vitro Transfection with siRNA and Infection with Lentiviruses
The human epithelial explants and SRA01/04 cells were transfected with silencing RNAs (siRNAs) targeting H19 and nontargeting siRNAs as a negative control (Ribobio, CHN): si-h-H19-001: CCCACAACATGAAAGAAAT; si-h-H19-002: GCCTTTGAATCCG-GACACA; si-h-H19-003: GCACTACCTGACTCAGGAA. Following the manufacturer's instructions, siRNAs were mixed with the OPTI-MEM buffer and transfection reagents (Lipofectamine 3000; Invitrogen, Waltham, MA, USA) to a final concentration of 100 nM. Cells at 70% confluency were used for transfection. At least 24 h of transfection was required for maximal transfection efficiency. For the gain-of-function studies, a specific lentivirus (LV, HANBIO., Shanghai, CHN) carrying a full-length sequence of H19 (oH19) was added into the culture medium at a multiplicity of infection (MOI) of 30 for at least 24 h. Its specific ZsGreen expression detected under fluorescent microscopy guaranteed the transfection efficiency.

RNA Isolation and Analysis
Total RNA from the human lens epithelial explants was collected and purified using the Arcturus PicoPure RNA isolation kit (KIT0204, Applied Biosystems, Waltham, MA, USA) following the manufacturer's manual. Bio-Rad digital droplet Polymerase Chain Reaction (ddPCR) was then used for RNA analysis. Based on water-oil emulsion droplet technology, each sample was fractionated at the endpoint of amplification into nearly 20,000 nanoliter-sized droplets enclosing RNA inside or not, annotating 'true' or 'false' events, and PCR amplification took place in each droplet. Absolute RNA concentrations were then calculated (copies/µL).
RNA extraction from SRA01/04 cells was performed with an RNeasy Mini kit (74106, Qiagen, Hilden, Germany), complying with the manufacturer's protocol. For proteincoding genes and H19, PrimeScript IV 1st strand cDNA Synthesis Mix (6215A, Takara, Canton, MA, USA) was used to synthesize first-strand cDNA from total RNA. Quantitative Reverse Transcription-PCR (qRT-PCR) was conducted using QuantiNova SYBR Green PCR (208054, Qiagen, Hilden, Germany) on a QuantStudio 7 Flex system in default thermal cycling settings. Regarding validation of miR-675-3p expression, the microRNA Reverse Transcription Kit was used for adding Poly (A) tail and reverse transcription (EZB-miRT4). EZ-probe qPCR Master Mix for miRNA was used for quantification (EZB-miProbe-R2). The RNA expression levels were calculated using the 2ˆ(-delta delta CT) method. The sequences of primers used in this study are listed in Table 1. (Roche Diagnostics, Mannheim, Germany. A BCA assay measured protein concentrations in each sample to normalize the protein loading. To assess the protein levels in lens epithelial explants more accurately, capillary electrophoresis immunoassay, or Simple Western (Wes), was performed on an automated capillary electrophoresis-based platform (ProteinSimple), according to the manufacturer's protocol (Protein Simple, San Jose, CA, USA). This cuttingedge method is considered more sensitive and accurate than traditional Western blotting, generating highly reproducible results. The Wes 66-440 kDa separation module (Cat. No.: SM-W006) was used. Briefly, protein extracts from each sample were diluted with sample buffer to a final concentration of 2 ug/uL and added to a master mix containing dithiothreitol (DTT) and fluorescent molecular weight markers, then heated at 95 • C for 5 min. After plate loading with the preheated samples, fully automated electrophoresis running, antibody incubation, and immuno-signal detection took place in the capillary system. FN, N-Cad, and α-Actinin (loading control) were detected by primary antibodies ab2413 (1:300, Abcam, Cambridge, UK), ab18203 (1:200, Abcam), and Rabbit anti-α-actinin (1:200, CST 6487S), respectively. Goat Anti-Rabbit horseradish peroxidase (HRP) conjugated secondary antibodies (ready-to-use, DM-001) were used according to the manufacturer's instructions (ProteinSimple). Compass software (ProteinSimple) generated the digital image visualizing virtual gels and data analysis. Relative protein quantification was generated using α-actinin as an internal control, normalized to the control group.
Traditional Western blotting was performed for protein analysis in SRA01/04 cells. An equal amount of proteins were mixed with 2× SDS-PAGE loading buffer, separated with SurePAGE gels in gradient (4-12%) concentrations (GenScript Biotech Corporation, Nanjing, China), and transferred onto PVDF membranes with pore size 0.45 µm (Merck Millipore, Burlington, MA, USA). The membranes were then blocked in 5% bovine serum albumin (BSA) for 1 h at room temperature (RT) and incubated with primary antibodies (1:1000) overnight at 4 • C. After incubation with secondary HRP-conjugated antibodies (CST, Danvers, MA, USA) at a concentration of 1:2000 for 2 h at RT, protein levels could be visualized with chemiluminescence detection systems (Merck Millipore, IRL). The intensity of each band was determined using Image J software (NIH, Bethesda, MD, USA) and further normalized to GAPDH protein levels. The primary antibodies used in this study are listed in Table 2.

RNAscope/LncRNA In Situ Hybridization
RNAscope kit (Advanced Cell Diagnostics) was used for the in situ detection of m transcripts with specifically designed probes (RNAscope probe_Ms-H19 and RNAscope probe_Hm-H19, catalog no.: 423751 and 400771, San Jose, CA, USA). Freshly cut sections from formalin-fixed, paraffin-embedded (FFPE) human and mouse lens tissues, wholemount of lens epithelial explants, and slides for adherent SRA01/04 cells were pretreated differently following the manufacturer's instructions. Housekeeping gene PPIB was used as a positive control to ensure RNA integrity of the slides (RNAscope probe_Hm-PPIB and RNAscope probe_Ms-Ppib, catalog no.: 18302A and 18346A, San Jose, CA, USA). In contrast, dihydrodipicolinate reductase in Bacillus subtilis (dap B) was adopted as a negative control to rule out the possibilities of false-positive signals. A horseradish peroxidase-based system (RNAscope 2.0 HD Detection Kit-Brown, catalog 310035, San Jose, CA, USA) was adopted for signal amplification, and hematoxylin was used for counterstaining (BA-4097, BASO diagnostics, Zhuhai, CHN). Positive staining was visualized as brown punctate dots.

Immunofluorescence Staining on the Whole Mounts and SRA01/04 Cells, Hematoxylin and Eosin (H&E) Staining on Sections from the Whole Lenses
After treatment, the human lens epithelial explants and SRA01/04 cells were collected. The whole-mount explants were first flattened on a glass slide with tweezers and left half-desiccated. They were fixed with 4% paraformaldehyde for 30 min at RT and washed 3 times with clean PBS. The samples were then saturated with a blocking solution of PBS containing 0.3% Triton X-100 and 3% BSA for 30 min at RT, followed by overnight incubation with primary antibodies at 4 • C. The next day, after three washes in PBS with 0.1% Tween-20 (PBST), lens epithelial cells and explants were incubated with secondary antibodies for 1 h at RT and counterstained with DAPI. The primary antibodies applied are listed in Table 2.
Human and mouse whole lenses were fixed by 10% neutral buffered formalin, dehydrated by gradient ethanol, and embedded into paraffinized blocks. Freshly cut sections were first deparaffinized and rehydrated. Then the nuclei were stained with hematoxylin solution, rinsed with running tap water, and then counterstained with Eosin solution to delineate the cellular cytoplasm (BA-4097 and BA-4098, BASO diagnostics, Zhuhai, CHN).

Cell Scratch Assay
SRA01/04 cells were seeded into 6-well plates and transfected with H19 silencing reagents or H19 overexpression lentiviruses. After their concentration reached a 100% confluence, cells were starved for at least 12 h before scratch. With 200 µL pipette tips, the monolayered cells were scratched vertically, leaving a cell-free area (wound) in each well, and detached cells were removed with a PBS wash. Cells were cultured in a serum-free medium to exclude the effects of proliferation, with or without TGF-β2. The migration of cells into the gap was observed and imaged at the start time of 0 h, 12 h, 16 h, and 24 h using a transmitted light microscope. Lines at regular intervals were drawn on the bottom of the plate before starting the experiment to ensure that the exact locations were chosen for observation at distinct time points. The healed areas were measured and quantified using Image J software.

Transwell Migration Assay
After transfection with H19-targeting silencing RNAs, or infection with lentiviruses overexpressing H19 for at least 24 h, SRA01/04 cells were then resuspended at a density of 1 × 10 5 /mL. Cells in 100 µL serum-free medium with or without TGF-β2 were seeded onto the individual upper chamber of 24-well Transwell plates (1 × 10 4 /well) and allowed to migrate vertically through the 8.0 µm pores (Corning, NY, USA). Every lower chamber contained 600 µL DMEM with 10% FBS as the chemoattractant. After 36 h of incubation, the remaining cells on the top of the inserts (non-migrated) were removed with a cotton swab. Cells that migrated to the lower bottom were then fixed with 4% paraformaldehyde (PFA) and detected with Crystal Violet (Beyotime, Shanghai, China). Migrated cells of five randomly selected fields in each well were counted.

Semi-In Vivo Whole Lens Culture
All procedures involving animals were performed following the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research, and were approved by the Animal Use and Care Committee of Zhongshan Ophthalmic Center (Permit number: 2020-039; Approval date: 3 September 2020). Whole lenses of 21-day-old mice maintained in the C57BL/6J background were collected and cultured in serum-free M199 medium supplemented with 0.1% BSA, 0.1 µg/mL L-glutamine, 100 IU/mL penicillin, and 100 IU/mL streptomycin [33,34]. Whole lenses from human donors were also freshly collected and cultured with the same media. After isolation, lenses were kept in the media for at least 24 h before any treatment. Those with cloudiness caused during surgical collection were excluded. After transfection or infection for 24 h, TGF-β2 (R&D Systems) was added to the medium (10 ng/mL). The medium was changed every other day. Mouse whole lenses were cultured for up to 6 days and photographed under a stereoscope before being harvested. Human whole lens explants were cultured for up to 10 days. RNA levels of H19 in lens epitheliums were further detected to ensure the efficiency of silencing or overexpressing agents.

RNA Sequencing (RNA-Seq)
Human lens explants were isolated and cultured with or without 10 ng/mL TGF-β2 for 48 h. SRA01/04 cells were first infected with lentiviruses for at least 24 h and then treated with 5 ng/mL TGF-β2 for another 24 h. Total RNA was isolated, and RNA integrity was assessed using a commercial kit and system (RNA Nano 6000 Assay Kit of the Bioanalyzer 2100, Santa Clara, CA, USA). Sequencing libraries were generated using a prep kit for Illumina (NEBNext Ultra RNA Library Prep Kit; NEB). High throughput bulk RNA-sequencing was conducted with a high sequence platform (NovaSeq 6000; Illumina, San Diego, CA, USA) to produce the raw read. Raw sequencing data reported in this paper have been deposited (PRJCA010973) in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in the National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences, under the Accession codes HRA002775 for human lens epithelial explants and HRA002785 for SRA01/04 cells, and are accessible upon request at https: //ngdc.cncb.ac.cn/gsa-human (accessed on 10 August 2022). The Perl script filtered the original data (Raw Reads) and generated Clean Reads to guarantee the data quality used for further analysis. As for the alignment, the reference genomes and the annotation file were first downloaded from the ENSEMBL database. Bowtie2 v2.2.3 was used for building the genome index, and TopHat v2.0.12 was adopted for gene mapping. Reads Count for each gene in each sample was counted by HTSeq v0.6.0. in RPKM (Reads Per Kilobase Million Mapped Reads), which was then calculated to estimate the expression level of genes to eliminate the effects of sequencing depth and gene length.
DESeq2 was used for differential gene expression analysis between two samples with biological replicates, using a model based on the negative binomial distribution. The p-value was adjusted by Benjamini and Hochberg's approach for controlling the false discovery rate. Genes with adjusted p ≤ 0.05 and |log2_FC| ≥ 1 were identified as differentially expressed genes (DEGs). Metascape with default parameters [35] was applied for GO enrichment analysis. An Official GSEA software (version 4.0.3, a joint project of UC San Diego and Broad Institute, San Diego, CA and Cambridge, MA) was adopted with weighted enrichment statistics and a t-test ranking metric for Gene Set Enrichment Analysis (GSEA) [36]. Gene Set Variation Analysis (GSVA) R package was used to identify the differential signal pathway enrichment [37].

In Vivo Photography with Slit Lamp and Whole Lens Imaging with Stereoscope
H19 knockout mice were kindly offered by Professor Zhao Meng Lab (Sun Yat-Sen University) [38,39]. The mice were maintained in the C57BL/6J background. The BX 900 slit lamp (Haag-Streit, Mason, OH, USA) was used for a thorough ocular examination of the wild-type and H19 knockout mice. Immediately after safe anesthesia with intraperitoneal injections of chloral hydrate (5 mL/Kg), their pupils were dilated with 0.5% tropicamide/0.5% phenylephrine mixed eye drop (Santen, Osaka, Japan). A diffusing filter with a wide-open slit-lamp beam at 45 degrees (diffuse illumination) was used to overview the eye appearances. For more detailed information, a narrow focal slit beam (<0.5 mm) and a moderate slit (2-4 mm) was projected at a 60 • angle to the lens which directly illuminated the lens pathology. An indirect form of retro-illumination mode was also used: A decentred wide slit beam, adjusted to be square-like, was projected near the pupil margin to reflect light from the fundus, showing the defects in the reflective media along the visual axis. To prevent acute lens opacity induced by anesthesia, we adopted carbomer eye drops, and each eye was examined within 15 min after confirmed anesthesia [40].
Mouse and human lenses were dissected from freshly enucleated eyeballs and washed in warm PBS with antibiotics three times. While still immersed in PBS, the whole lenses were imaged under a Leika stereoscope (S8AP0, Teaneck, NJ, USA) with a high-fidelity digital camera.

Statistical Analysis
All experiments were performed at least three times independently. Data were presented as the mean ± standard deviation (SD). The significance of differences between the two groups was determined by the student's t-test, using GraphPad Prism software version 9.0. A one-way analysis of variance (ANOVA) was performed for differences among more than two groups. All statistical tests were two-tailed, and p-values less than 0.05 were considered statistically significant (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).

LncRNA H19 Was Highly Expressed in the Normal Lens Epithelium While Downregulated by Exposure to TGF-β2
Staining LECs in situ on the whole-mount of human lens epithelial explants (LEEs) could better recapitulate lens epithelial features in vivo. In this model, treatment with TGF-β2 resulted in increased extracellular, intracellular, and membranous mesenchymal markers, including FN, α-SMA, and Neural (N)-cadherin (N-Cad), compared to the control group (LEE-Con). At the same time, we observed a loss of tight junction proteins zonula occludens-1 (ZO-1) with cytoplasmic or nuclear translocation of β-catenin, which is typically anchored by Epithelial (E)-cadherin (E-Cad) on cell membranes (Figure 1a). To identify differentially expressed lncRNAs that might play a crucial role in dictating lens epithelial cell fate, we collected the LEEs with or without exposure to TGF-β2 and conducted a high-throughput bulk RNA sequencing. Principal co-ordinates analysis (PCoA) demonstrated dramatic transcriptomic changes in human LEEs upon exposure to TGF-β2 ( Figure  S1a). Over 5000 genes were differentially expressed with statistical significance, including  Figure S1b). Genes decreased by TGF-β2 in LEEs are enriched in GO terms vital to lens development and LEC integrity, such as molecular metabolism, membrane spectrin or channel activities, and oxidation/reduction reactions ( Figure S1c). Therefore, the competence of LECs was negatively affected by TGF-β2. By contrast, TGF-β2 treatment activated both classical Smad-mediated and non-Smaddependent downstream signaling pathways. Increased cell proliferation and migration were also implied ( Figure S1d).

TGF-β2-Reduced H19 Expression in Lens Epithelial Cells in a Time-Dependent Manner
To better understand the role of H19 during EMT in the ocular lens, the well-established human lens epithelial cell line SRA01/04 was also adopted in this study. Immunofluorescent staining of SRA01/04 cells showed TGF-β2 also led to a more significant amount of mesenchymal markers, including FN, α-SMA, and N-Cad, and drastic mor- LncRNA H19 was highly expressed in normal lens epithelium while downregulated by exposure to TGF-β2. (a) Immunofluorescent staining of mesenchymal markers (FN, α-SMA, and N-Cad), and epithelial markers (ZO-1, β-catenin), on human lens epithelial explants (LEEs), when exposed to TGF-β2 (10 ng/mL, 48 h; LEE-T2), in comparison to control group (LEE-Con). Scale bars, 20 µm. (b) RNA analysis of LEEs-Con and LEEs-T2 (age-matched, n = 6) was performed using β-actin as the internal control. (c) H19 probes were hybridized in situ. Scale bars, 20 µm. (d,e) Image J was used to quantify the absolute value of H19 positive areas (d) and the ratio of H19 positive to total area within each cell (e). At least three experiments were repeated, and data were shown as mean ± SD. *** p < 0.001, **** p < 0.0001 vs. LEE-Con. (f) RNAscope targeting H19 on the freshly cut sections from human and mouse whole lenses was performed. The lower two panels showed immunostaining of mesenchymal markers FN and N-Cad. Scale bars, 20 µm. Furthermore, RNA-Seq also identified the top 10 lncRNAs with the most remarkable fold changes across sample replicates ( Figure S1e). We uncovered that H19, the first-discovered lncRNA, was the most highly expressed in human LEEs ( Figure S1f). Quantitative real-time PCR (qRT-PCR) analysis of lens epithelial explants (age-matched, n = 6) further confirmed that H19 levels decreased significantly with TGF-β2 treatment (Figure 1b). Using RNAscope (RNA in-situ hybridization) with H19-targeting probes, we could delineate the subcellular distribution of H19 and visualize its relative expression levels. It was self-evident that H19 expression in the cytoplasm of LECs was significantly reduced under the influence of TGF-β2 compared to the control group (Figure 1c). Quantification data were displayed as absolute values of H19-positive areas and relative ratio (Figure 1d,e).
There was a consistently high accumulation of H19 in the cytoplasm of LECs underneath the anterior capsules from human to adult mouse ocular lenses. Meanwhile, we also demonstrated a conspicuous decrease in H19 foci and loss of epithelial phenotypes upon exposure to TGF-β2 in semi-in vivo cultures of both human and adult mouse lenses (Figure 1f).
These results indicated that H19 was indispensable in maintaining the finely-tuned lens epithelial properties. At the same time, the downregulation of lncRNA H19 might aid in the EMT processes and accelerate the development of lens opacities.

TGF-β2-Reduced H19 Expression in Lens Epithelial Cells in a Time-Dependent Manner
To better understand the role of H19 during EMT in the ocular lens, the well-established human lens epithelial cell line SRA01/04 was also adopted in this study. Immunofluorescent staining of SRA01/04 cells showed TGF-β2 also led to a more significant amount of mesenchymal markers, including FN, α-SMA, and N-Cad, and drastic morphological changes. Loss of the sub-membranous scaffold protein ZO-1 was observed with an apparent localization shift of β-catenin from cell membranes to the cytoplasm or nucleus ( Figure S2a). Activation of vital EMT-related transcription factors, such as Snail and Slug, which belonged to the Snail superfamily of zinc-finger transcription factors, were also observed ( Figure S2a). With RNA probes targeting H19 hybridized in situ, it was clear that H19 expression in the cytoplasm of SRA01/04 cells was significantly reduced with TGF-β2 (5 ng/mL, 48 h) in comparison to the control group (con, 0 h) ( Figure S2a). We further quantified H19-positive foci using Image J ( Figure S2e,f). Next, we employed EMT markers to validate these alterations at ribonucleic acid (RNA) and protein levels. Treatment of SRA01/04 cells with TGF-β2 (5 ng/mL) increased messenger RNA (mRNA) levels of mesenchymal markers, ranging from FN, and α-SMA, to collagen type 1 (COL-1), along with a reduction in the epithelial junction protein ZO-1, in a time-dependent manner ( Figure S2b). A duration-dependent decrease in H19 RNA expression was also observed in SRA01/04 cells in response to TGF-β2 ( Figure S2c). Protein analysis confirmed that TGF-β2 treatment of SRA01/04 cells gradually increased mesenchymal markers, including FN, α-SMA, N-Cad, and essential transcription factors, Snail and Slug ( Figure S2d).
Since we had already managed to establish TGF-β2-induced EMT in both SRA01/04 cells and lens epithelial explants, during which a significant reduction in H19 was observed, the next step was to overexpress or knockdown H19 forcibly to establish a cause-andeffect relationship.

Knockdown of LncRNA H19 Aggravated TGF-β2-Induced EMT in Both SRA01/04 Cells and Human Lens Epithelial Explants
We first examined the effects of H19 knockdown in the lens epithelial explants. Due to the enormous variation between each human sample, it is unreliable to pool all capsule samples from different donors and analyze their RNA or protein levels together. However, the amount of RNA or protein collected from every capsule is relatively insufficient for precise and reproducible analysis. Hence, new techniques with higher sensitivity and specificity are needed. The Bio-Rad QX200 droplet digital PCR (ddPCR) system and an automated Simple Western platform (Wes) were thus introduced for RNA and protein analysis for human LEEs. With ddPCR, we measured the absolute RNA copy numbers of H19 and FN. It was self-evident that FN mRNA levels increased significantly and inversely as H19 expression dropped ( Figure S3a,b). The Wes platform corroborated this at the protein levels of mesenchymal markers FN and N-Cad (Figure 2a-d). Immunofluorescent staining of human LEEs revealed consistent findings between mesenchymal markers FN, α-SMA, N-Cad, and epithelial markers ZO-1 and β-catenin (Figure 2e). More importantly, all these experiments consistently showed that silencing of H19 alone led to EMT changes comparable to TGF-β2 treatment, which implied that H19 was necessary for maintaining lens epithelial phenotypes. Immortalized human lens epithelial cells were also used for functional studies to depict H19 properties. LncRNA H19 was forcedly downregulated in SRA01/04 cells by transfection with siH19 RNAs, among which siH19-002 showed the most outstanding silencing effects (Figure 3a). Knockdown of H19 in cells exposed to TGF-β2 (siH19-002+TGF-β2) further increased mRNA levels of FN than TGF-β2 treatment alone (siNC+TGF-β2) (Figure 3b). Western blots confirmed that the downregulation of H19 led to more production of mesenchymal markers, FN, α-SMA, N-Cad, and abolished the expression of epithelial marker ZO-1 (Figure 3c,d). Consistently, immunofluorescent staining of mesenchymal markers, including FN, α-SMA, N-Cad, and epithelial markers ZO-1 and β-catenin in SRA01/04 cells revealed that silencing of H19 intensified TGF-β2-induced alterations (Figure 3e). Notably, similar to the observations in epithelial explant studies, silencing of H19 could also lead to EMT-like alterations, even without the presence of TGF-β2. Furthermore, silencing the lncRNA H19 plus TGF-β2 treatment group displayed the most signif- Immortalized human lens epithelial cells were also used for functional studies to depict H19 properties. LncRNA H19 was forcedly downregulated in SRA01/04 cells by transfection with siH19 RNAs, among which siH19-002 showed the most outstanding silencing effects (Figure 3a). Knockdown of H19 in cells exposed to TGF-β2 (siH19-002+TGF-β2) further increased mRNA levels of FN than TGF-β2 treatment alone (siNC+TGF-β2) (Figure 3b). Western blots confirmed that the downregulation of H19 led to more production of mesenchymal markers, FN, α-SMA, N-Cad, and abolished the expression of epithelial marker ZO-1 (Figure 3c,d). Consistently, immunofluorescent staining of mesenchymal markers, including FN, α-SMA, N-Cad, and epithelial markers ZO-1 and β-catenin in SRA01/04 cells revealed that silencing of H19 intensified TGF-β2-induced alterations (Figure 3e). Notably, similar to the observations in epithelial explant studies, silencing of H19 could also lead to EMT-like alterations, even without the presence of TGF-β2. Furthermore, silencing the lncRNA H19 plus TGF-β2 treatment group displayed the most significant migrative potential in SRA01/04 cells. It covered almost all of the scratchinduced wound areas within 24 h after the initial scratch (Figure 3f,g). We also used a Transwell assay to confirm this observation, which detected that cells migrated vertically over a certain period (Figure 3h,i). It also demonstrated that knockdown of H19 alone resulted in a more remarkable migration, either with or without TGF-β2 (Figure 3f-i). In summary, forced downregulation of H19 accelerated TGF-β2-induced EMT in human lens epithelial cells with an abnormally active migration. It is worth noting that the knockdown of H19 could also lead to greater migration capacities and the loss of normal epithelial phenotypes even without TGF-β2 treatment. These results implied that H19 was necessary for maintaining proper lens epithelial phenotypes. μm. (f-i) After transfection with siH19-002, the migration capacities of SRA01/04 cells induced by TGF-β2 was detected by cell scratch assay (f,g) horizontally and Transwell assay (h,i) vertically. Image J was used for further quantification and calculation. Scale bars, 50 μm. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control group(siNC). In summary, forced downregulation of H19 accelerated TGF-β2-induced EMT in human lens epithelial cells with an abnormally active migration. It is worth noting that the knockdown of H19 could also lead to greater migration capacities and the loss of normal epithelial phenotypes even without TGF-β2 treatment. These results implied that H19 was necessary for maintaining proper lens epithelial phenotypes.

Overexpression of LncRNA H19 Partially Reversed TGF-β2-Induced EMT and Restored Lens Epithelial Phenotypes
Since H19 was downregulated during TGF-β2-induced EMT, our next step was to upregulate H19 expression during EMT and explore its functional roles. We first used a lentivirus vector carrying a full-length sequence of H19 (oH19) to overexpress H19 in human lens epithelial explants. In contrast, an empty lentivirus vector was applied as a negative control for overexpression (oNC). Human epithelial explants were first infected with lentiviruses, followed by TGF-β2 induction or medium alone.
Then RNA levels of H19 and FN were measured and quantified by ddPCR. In contrast to what we observed in H19 silencing studies, upregulation of H19 partially reversed the transcription of the mesenchymal marker FN, despite the presence of TGF-β2 ( Figure S3c,d). We also confirmed EMT changes at protein levels by the Wes platform: H19 overexpression prevented TGF-β2-induced EMT (Figure 4a-d). Immunofluorescent staining on LEEs confirmed that H19 reduced the induction of mesenchymal markers FN, α-SMA, and N-Cad, and partially restored membranous ZO-1 and β-catenin (Figure 4e).  Consistent with explant studies, H19 overexpression in SRA01/04 cells inhibited TGF-β2-induced production in FN and α-SMA mRNA levels (Figure 5a,b). Western blots confirmed the upregulation of H19, the attenuated increase in mesenchymal markers, FN, α-SMA, and N-Cad, and the restoration of epithelial marker ZO-1 (Figure 5c,d). Accordingly, immunofluorescent staining of mesenchymal and epithelial markers in SRA01/04 cells proved that oH19 reversed TGF-β2-induced EMT (Figure 5e). Furthermore, overexpression of H19 reduced TGF-β2-induced migration in SRA01/04 cells by the scratch assay (Figure 5f,g). Findings from the Transwell assay could draw a similar conclusion (Figure 5h,i). SMA, and N-Cad, and the restoration of epithelial marker ZO-1 (Figure 5c,d). Accordingly, immunofluorescent staining of mesenchymal and epithelial markers in SRA01/04 cells proved that oH19 reversed TGF-β2-induced EMT (Figure 5e). Furthermore, overexpression of H19 reduced TGF-β2-induced migration in SRA01/04 cells by the scratch assay (Figure 5f,g). Findings from the Transwell assay could draw a similar conclusion ( Figure  5h,i). Briefly, forced upregulation of H19 maintained LEC phenotypes and partially reversed or prevented TGF-β2-induced EMT in SRA01/04 cells and human lens epithelial explants.  Briefly, forced upregulation of H19 maintained LEC phenotypes and partially reversed or prevented TGF-β2-induced EMT in SRA01/04 cells and human lens epithelial explants.

H19 Was Required to Maintain LEC Phenotypes and Lens Clarity
We then moved on to verifying the role of H19 in the semi-in vivo whole lens studies, which could reflect on its functions more directly. Lenses from 3-week-old mice were first transfected with H19 silencing agents for 24 h and then incubated with 10 ng/mL TGF-β2 for at least 24 h. Prominent opacities beneath the lens capsules developed, and the knockdown of H19 extensively aggravated TGF-β2-induced lens fibrosis. Notably, the silencing of H19 alone without TGF-β2 treatment also induced moderate opacities in the anterior polar region of the lens (Figure 6a). RNA analysis of the epithelial cells collected from mouse whole lenses confirmed the effective knockdown of H19 and the highest accumulation of FN mRNA in the siH19 plus TGF-β2 group (Figure S3e,f). Histological examinations revealed that mono-layered cuboidal epithelial cells in those clouded lenses turned into disorganized clumps. Elongated nuclei were observed with an aberrant invasion into the underlying lens fibers (Figure 6b). The opacities beneath the anterior capsules were consistent with an accumulation of FN, α-SMA, and N-Cad, and abolished epithelial junction ZO-1 protein (Figure 6c). Similarly, semi-in vivo cultures of human whole lens explants also revealed that the silencing of H19 promoted lens fibrosis, while the overexpression of H19 rescued epithelial features (Figure 7a-e).

H19 Homozygous Knockout Mice Displayed Anterior Polar Cataracts and Delayed Fusion of Anterior Sutures
H19 homozygous knockout (-/-) mice displayed lens opacities with incomplete genetic penetrance. On Postnatal day 21 (P21), H19 (-/-) mice exhibited a delayed fusion of anterior sutures with cortical cataracts at the anterior polar regions (Figure 8a-j), consistent with findings observed in semi-in vivo cultured mouse whole lenses with H19 forcibly silenced. Therefore, it further corroborated that H19 was vital to constructing lens epithelium and maintaining lens transparency.   (Figure 8a-j), consistent with findings observed in semi-in vivo cultured mouse whole lenses with H19 forcibly silenced. Therefore, it further corroborated that H19 was vital to constructing lens epithelium and maintaining lens transparency.

H19 Homozygous Knockout Mice Displayed Anterior Polar Cataracts and Delayed Fusion of Anterior Sutures
H19 homozygous knockout (-/-) mice displayed lens opacities with incomplete genetic penetrance. On Postnatal day 21 (P21), H19 (-/-) mice exhibited a delayed fusion of anterior sutures with cortical cataracts at the anterior polar regions (Figure 8a-j), consistent with findings observed in semi-in vivo cultured mouse whole lenses with H19 forcibly silenced. Therefore, it further corroborated that H19 was vital to constructing lens epithelium and maintaining lens transparency.

Mechanistic Analyses of the Role of H19 in Regulating TGF-β2-Induced EMT in Lens Epithelial Cells
To further understand the potential molecular mechanism, we adopted bulk RNA-Sequencing to mine transcriptomic changes when H19 was overexpressed in LECs by exposure to TGF-β2 ( Figure S4a). Briefly, SRA01/04 cells were infected with H19-overexpressing lentiviruses or negative control for at least 24 h and further treated with TGF-β2 (5 ng/mL, 24 h). Heatmaps displayed DEGs among different groups, including those responsible for TGF-β2-induced EMT in the lens. It was clear that TGF-β2-induced transcriptomic changes were inversely affected by the overexpression of H19. Most genes that TGF-β2 upregulated were reduced with the overexpression of H19. By contrast, those with decreased expression with TGF-β2 treatment were mostly rescued/restored by H19 upregulation (Figure S4b). Gene set enrichment analysis (GSEA) validated that gene signature profiles with TGF-β2 induction in LECs were reversed by H19 overexpression ( Figure S4c). Furthermore, these genes were involved in vital processes strongly related to EMT, including collagen formation and binding, crosslinking of the collagen fibrils, integrin complex and integrinmediated cell adhesion, focal adhesion, ECM proteoglycans, and ECM-receptor interaction ( Figure S4d,e). Apart from this, vital downstream pathways mediating EMT, including classical Smad-dependent TGF-β receptor signaling, were affected by H19 ( Figure S4d,e).
Specifically, the binding of TGF-β ligands to the TGF-β type I and II serine/threonine kinase receptors causes phosphorylation of downstream receptor-activated Smads (R-Smads), such as Smad2 and Smad3. Their phosphorylated forms (p-Smad2/3) show an increased affinity for the core Smad4 and are shuttled to the nucleus as a heteromeric complex with Smad4 [18,42]. We examined the status of these critical mediators in LECs. As expected, exposure of LECs to TGF-β2 led to considerable phosphorylation of Smad2/3 protein. Upregulation of H19 reduced TGF-β2-induced Smad2/3 phosphorylation. In contrast, silencing of H19 in LECs further enhanced p-Smad2 and p-Smad3 levels (Figure 9a-d).
posure to TGF-β2 ( Figure S4a). Briefly, SRA01/04 cells were infected with H19-overex-pressing lentiviruses or negative control for at least 24 h and further treated with TGF-β2 (5 ng/mL, 24 h). Heatmaps displayed DEGs among different groups, including those responsible for TGF-β2-induced EMT in the lens. It was clear that TGF-β2-induced transcriptomic changes were inversely affected by the overexpression of H19. Most genes that TGF-β2 upregulated were reduced with the overexpression of H19. By contrast, those with decreased expression with TGF-β2 treatment were mostly rescued/restored by H19 upregulation ( Figure S4b). Gene set enrichment analysis (GSEA) validated that gene signature profiles with TGF-β2 induction in LECs were reversed by H19 overexpression ( Figure  S4c). Furthermore, these genes were involved in vital processes strongly related to EMT, including collagen formation and binding, crosslinking of the collagen fibrils, integrin complex and integrin-mediated cell adhesion, focal adhesion, ECM proteoglycans, and ECM-receptor interaction ( Figure S4d,e). Apart from this, vital downstream pathways mediating EMT, including classical Smad-dependent TGF-β receptor signaling, were affected by H19 (Figure S4d,e).
Specifically, the binding of TGF-β ligands to the TGF-β type I and II serine/threonine kinase receptors causes phosphorylation of downstream receptor-activated Smads (R-Smads), such as Smad2 and Smad3. Their phosphorylated forms (p-Smad2/3) show an increased affinity for the core Smad4 and are shuttled to the nucleus as a heteromeric complex with Smad4 [18,42]. We examined the status of these critical mediators in LECs. As expected, exposure of LECs to TGF-β2 led to considerable phosphorylation of Smad2/3 protein. Upregulation of H19 reduced TGF-β2-induced Smad2/3 phosphorylation. In contrast, silencing of H19 in LECs further enhanced p-Smad2 and p-Smad3 levels (Figure 9ad).
The nuclear translocation of core Smad4 eventually turns on EMT-inducing transcription factors, such as SNAIL and SLUG, ZEB, and TWIST [18,43,44]. Notably, Smad4 expression was also enhanced by the knocking down of H19 with exposure to TGF-β2, while upregulation of H19 during TGF-β2-induced EMT reduced Smad4 levels (Figure 9a-d). Interestingly, the 3 -UTR (untranslated region) of SMAD4 contained predicted binding sites of miR-675-3p, which was embedded in the first exon of H19. Dual-luciferase assay confirmed that miR-675-3p could only bind wild-type SMAD4 3 -UTR, not the mutant one ( Figure S5a,b), highlighting the potential of the H19/miR-675-3p/SMAD4 axis in regulating lens fibrosis.

Discussion
Given that there are currently limited options for lens fibrosis, developing novel strategies or unveiling new targets is of great clinical importance. The prerequisite is gaining an in-depth insight into the molecular mechanisms governing this pro-fibrotic EMT process. This study deciphers the role of lncRNA H19 in maintaining LEC features by suppressing EMT. We report that the upregulation of H19 can prevent lens fibrosis induced by TGF-β2 through a blockade of the downstream Smad-dependent signaling.
Less than 3% of genes encode functional proteins in the human genome, while more than 95% are non-coding RNAs (ncRNAs) [45]. Increasing numbers of ncRNAs have been identified, but their functions in biological development and pathological diseases are largely unknown [46,47]. The oncofetal H19 lncRNA, with relatively high evolutionary conservancy across species, is most abundantly enriched during embryonic stages and in cancer tissues [28,31]. A growing body of studies was conducted on the orchestrating role of H19 in various cancers in multiple settings, such as initiation, progression, and metastasis [31]. However, controversies remain concerning whether it serves as an oncogene or a tumor suppressor [48][49][50]. Likewise, no consensus has been reached on its promoting or inhibitory role in the pathogenesis of fibrotic diseases [30]. Intriguingly, Raveh proposed a plausible unifying theory that can explain the seemingly contradictory and versatile functions of H19, which rely on the actual context [51]. Likely, the exact effect of H19 on EMT in cancers and fibrotic diseases is tissue-and time-dependent.
We observed for the first time that H19 was highly-enriched in LECs but downregulated during TGF-β2-induced EMT. Both loss-and gain-of-function studies of H19 were performed in human lens epithelial explants and SRA01/04 cells, consistently demonstrating that the silencing of H19 exacerbated EMT. At the same time, the overexpression of H19 reversed this phenotypic transformation. Noteworthily, knocking down H19 itself could lead to EMT-like changes, even without TGF-β2, implying the indispensability of H19 in maintaining LEC integrity. RNA-Seq data further validated that H19 overexpression could reverse TGF-β2-induced transcriptomic alterations to restore the epithelial phenotypes, indicating the importance of H19 in maintaining the distinctive properties of LECs.
LECs underlying the anterior lens capsules are drawing increasing attention for their importance in establishing and maintaining lens transparency. Many studies have been conducted to explore the potential therapeutic options for preventing or reversing lens fibrosis using diverse experimental models. To better validate their roles in lens transparency maintenance, in recent years, apart from in vitro studies using lens epithelial cell lines, semi-in vivo epithelial explants, the whole lens culturing, and in vivo animal models have been more widely used, including injury-induced ASC models, mock lens cataract surgery (Extracapsular Lens Extraction, ECLE) for PCO models, and transgenic/genetic knockout mice [19,[52][53][54][55][56][57]. It is generally accepted that during TGF-β2-induced EMT, loss of LEC integrity is accompanied by the abnormal secretion of ECM and aberrant cell migration, resulting in fibrotic plaques and lens opacification. Our semi-in vivo whole lens culture studies corroborated that the downregulation of H19 led to significant EMT of LECs, contributing to clouding of the lens with or without exposure to TGF-β2. On the contrary, the upregulation of H19 maintained lens epithelial properties and restored lens clarity. Similar lens opacities were observed in H19 homozygous knockout mice, indicating the indispensable role of H19 in establishing lens transparency through the maintenance of LEC phenotypes. Apart from fibrotic opacities formed in the lens, EMT of LECs also led to abnormal differentiation, resulting in defective lens fibers with the disordered arrangement. Upregulation of H19 further avoids the abnormality of lens fibers by restoring LECs, which also contributes to lens transparency. Therefore, H19 is vital for maintaining LEC phenotypes and sustaining lens clarity.
As a long non-coding RNA, H19 can interact with DNA/chromatin, RNA, and proteins, regulating EMT at transcriptional, post-transcriptional, and post-translational levels [58][59][60]. Notably, competing endogenous RNA networks (ceRNETs) are gaining increasing attention: H19 may also execute a decoy for sequestrating miRNAs [60][61][62]. Alternatively, H19 also serves as a reservoir for producing miRNAs [63]. The first exon of the H19 gene itself can be transcribed into miRNA-675 as a by-product which targets functional proteins involved in various biological and pathological processes, such as EMT [63][64][65][66][67]. Apart from the predicted binding from the online database, we have confirmed the binding of miR-675-3p to 3 -UTR of SMAD4 directly in LECs with luciferase reporter assays ( Figure S5a,b). Smad4 is the core member belonging to the Smads family. It transduces signals from TGF-β2 signaling and activates crucial downstream effector genes via nuclear translocation [8]. Admittedly, when H19 was upregulated by transfection in LECs, the changes in miRNA-675 levels were much more minor and negligible compared to H19 ( Figure S5c). Therefore, alternative explanations are warranted for the regulatory role of H19 in lens fibrosis.
Cheng reported an increase in H19 expression in human LECs upon exposure to UV-B, contributing to the oxidative damage of LECs [68]. This study observed the considerable downregulation of H19 in human and mouse LECs with TGF-β2 administration. Previous studies implied the possible involvement of both canonical Smad-dependent and Smadindependent pathways, such as mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK)1/2, phosphatidylinositol-3-kinase (PI3K)/Akt, and the mammalian target of rapamycin (mTOR) signaling during lens fibrosis [18,42,61]. RNA-Seq of LECs suggested their activation by TGF-β2, whereas H19 overexpression blocked such changes ( Figure S4e). Immunoblotting confirmed that upregulation of H19 reduced phosphorylation of R-Smads (Smad2/3) and inhibited core Smad4 expression directly. Our RNA-seq data also revealed the involvement of essential processes, such as collagen formation and crosslinking, integrin-mediated cell adhesion, focal adhesion, and ECM-receptor interaction. Therefore, future work could be directed at the crosstalks of these critical signaling pathways or biological processes.
In particular, H19 is unique in its genetic linkage with its neighboring gene, insulinlike growth factor 2 (IGF2), where both genes are expressed reciprocally or parentalspecifically [69]. This genetic imprinting at the H19/IGF2 locus was reported to be strictly controlled by an upstream differentially methylated domain (DMD), also known as the imprinting control region (ICR) [38,70,71]. However, our current study can temporarily exclude the effects of genetic imprinting, as RNA analysis by qPCR revealed a similar increase in IGF2 mRNA levels when H19 was overexpressed ( Figure S5d). We have already obtained genetic animals with H19-DMD/ICR deletions and found congenital lens developmental defects in heterozygous knockout mice ( Figure S6a-d). Functional studies of these genetic mice are required to clarify the specific role of H19/IGF2 balance in lens development.
Admittedly, apart from H19, alternative epigenetic regulatory ncRNAs may also be involved in lens fibrosis [32,[72][73][74]. Nevertheless, our current study presented the first evidence that H19 serves a protective role during TGF-β2-induced EMT in the lens. Noteworthily, we demonstrated the close relationship between the integrity of LECs and the maintenance of lens transparency. In conclusion, although further investigations are needed, our findings open up possibilities for H19 as a novel and potential therapeutic target for treating lens fibrosis.

Conclusions
This study showed for the first time that lncRNA H19 could prevent TGF-β2-induced lens fibrosis by suppressing the EMT-mediating canonical Smad-dependent pathways. Overexpression of H19 could partially restore lens transparency by reversing EMT and maintaining lens epithelial features. Thus, H19 could be a potential therapeutic option for fibrotic cataracts.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11162559/s1, Figure S1: H19 identified by RNA sequencing in human lens epithelial explants; Figure S2: Duration-dependent analysis of H19 expression in TGFβ2-treated SRA01/04 cells; Figure S3: Digital droplet PCR of human lens epithelial explants and mouse whole lenses; Figure S4: Bioinformatic analyses of the potential role of H19 in regulating TGF-β2-induced EMT in lens epithelial cells; Figure S5: Validation of the potential H19/miR675-3p/Smad4 axis and temporary exclusion of imprinting effects; Figure S6: H19-ICR heterozygote mice displayed lens dysmorphogenesis or aplasia.

Informed Consent Statement:
We have obtained written informed consent from all subjects involved in the study.

Data Availability Statement:
The datasets generated in this study are not publicly available but can be accessed from the corresponding author upon reasonable request.