GDF-15 Attenuates the Epithelium–Mesenchymal Transition and Alleviates TGFβ2-Induced Lens Opacity

Abstract

Purpose: We sought to evaluate the efficacy of growth differentiation factor (GDF)-15 treatment for suppressing epithelial–mesenchymal transition (EMT) and alleviating transforming growth factor β2 (TGFβ2)-induced lens opacity.

Methods: To test whether GDF-15 is a molecule that prevents EMT, we pretreated the culture with GDF-15 in neural progenitor cells, retinal pigment epithelial cells, and lens epithelial cells and then treated with factors that promote EMT, GDF-11, and TGFβ2, respectively. To further investigate the efficacy of GDF-15 on alleviating lens opacity, we used mouse lens explant culture to mimic secondary cataracts. We pretreated the lens culture with GDF-15 and then added TGFβ2 to develop lens opacity (n = 3 for each group). Western blot and quantitative reverse transcription polymerase chain reaction (qRT-PCR) were used to measure EMT protein and gene expression, respectively.

Results: In cell culture, GDF-15 pretreatment significantly attenuated EMT marker expression in cultured cells induced by treatment with GDF-11 or TGFβ2. In the lens explant culture, GDF-15 pretreatment also reduced mouse lens opacity induced by exposure to TGFβ2.

Conclusions: Our results indicate that GDF-15 could alleviate TGFβ2-induced EMT and is a potential therapeutic agent to slow or prevent posterior capsular opacification (PCO) progression after cataract surgery.

Translational Relevance: Cataracts are the leading cause of blindness worldwide, with the only current treatment involving surgical removal of the lens and replacement with an artificial lens. However, PCO, also known as secondary cataract, is a common complication after cataract surgery. The development of an adjuvant that slows the progression of PCO will be beneficial to the field of anterior complications.

Introduction

Cataracts are a major cause of vision impairment and blindness worldwide.^1,2 A 2010 study showed that cataracts led to an estimated 51% of the blindness that affected 10.8 million people worldwide.^3,4 Aging and diabetes are major risk factors for cataracts.⁵ With current long life expectancies and a large diabetic population, the number of people blinded by cataracts is estimated to reach 40 million by 2025.^3,6 

Currently, the main treatment for cataracts is surgical removal of the opaque lens and replacement with a synthetic lens.⁷ However, complications can still occur after cataract surgery. The most common is secondary cataract, also known as posterior capsular opacification (PCO).^8,9 PCO occurs in ∼20% to 50% of adults after cataract surgery.^10,11 No matter how skillfully performed, cataract surgery disrupts the integrity of the eye and triggers a wound-healing response^12,13 that elevates levels of transforming growth factor beta (TGFβ) in the eye.^10,11 TGFβ can induce an epithelial–mesenchymal transition (EMT) by modulating transcription factors^14,15 in residual lens epithelial cells, which then proliferate and migrate to the posterior capsule instead of differentiating into normal lens fiber cells.^16–19 This abnormal differentiation leads to re-formation of cataracts in PCO.²⁰ 

EMT plays an important role in the transformation of lens epithelial cells to develop cataracts.²¹ In EMT, epithelial cells undergo a multitude of molecular and morphologic changes to transition into mesenchymal cells.²² This process includes specific activation of transcription factors, expression of cell surface proteins, reorganization of the cytoskeleton, progressive loss of epithelial cell markers, and gain of mesenchymal cells markers.^23,24 

The TGFβ superfamily of proteins is involved in regulating EMT. This family binds to TGFβ type II receptors, leading to formation of a heterotetrameric complex of TGFβ type I and type II receptors and subsequent phosphorylation and activation of Smad proteins.¹⁴ Among TGFβ superfamily proteins, TGFβ2 and growth differentiation factor (GDF)-11 both bind to TGFβ type II receptors to activate Smad2/3^25,26 and induce EMT in epithelial cells.^27,28 However, GDF-15, another member of the TGFβ superfamily, inhibits GDF-11–induced Smad2 phosphorylation.²⁹ Thus, we hypothesized that GDF-15 could also inhibit TGFβ2-induced Smad2 phosphorylation. We also hypothesized that this would alleviate the effects of TGFβ2-induced EMT on the lens and potentially inhibit the formation of PCO in eyes after cataract surgery. Here, we used differentiated human neural progenitor cells (NPCs), human adult retinal pigment epithelial cells (RPEs), human fetal lens epithelial cells (LECs), and mouse lenses to test these hypotheses. 

Methods and Materials

Materials and Cell Culture

Recombinant human TGFβ2 protein (#302-B2) was purchased from R&D Systems (Minneapolis, MN). GDF-11 (#120-11) and GDF-15 (#120-28C) were obtained from PeproTech (Cranbury, NJ). NPCs were differentiated from H9 embryonic stem cells, a gift from Don Zack's laboratory at Johns Hopkins University, following a previous protocol,³⁰ and the characterization of NPC was confirmed by immunofluorescence staining of NPC markers. Human adult RPEs (ARPE-19, #CRL-2302; American Type Culture Collection, Manassas, VA) and fetal human lens epithelial cells (FHL124, kindly provided by Ram H. Nagaraj, PhD, at the University of Colorado) were cultured in Gibco Dulbecco's Modified Eagle Medium (DMEM, #11885084; Thermo Fisher Scientific, Waltham, MA) supplemented with 10% and 5% fetal bovine serum, respectively, and 100 units/mL of Gibco penicillin–streptomycin (#15140122; Thermo Fisher Scientific). Cells were cultured in a humidified incubator with 5% CO₂ at 37°C. 

Mouse Lens Explant Cultivation

This research was conducted in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. We used 8- to 12-week-old C57BL/6 wild-type mice (both male and female; The Jackson Laboratory, Bar Harbor, ME). The animals were kept in microisolator cages on a 12-hour day/night cycle. Mice were euthanized by isoflurane liquid inhalant (#66794-017-25; Medline, Northfield, IL) and cervical dislocation (IACUC #21018446). Mouse lenses were isolated and cultured in DMEM supplemented with 0.1% bovine serum albumin, and 100 units/mL of penicillin–streptomycin in a humidified incubator with 5% CO₂ at 37°C. 

Ex Vivo Mouse Lens Culture

To compare lens opacity, lenses were pretreated with phosphate buffered saline (PBS, #BP39950; Fisher Scientific, Pittsburgh, PA) or GDF-15 (50 ng/mL) for 1 hour and then treated with TGFβ2 (5 or 10 ng/mL) for 7 days. Fresh medium with GDF-15 and TGFβ2 was replaced at day 3. Lenses were observed above a metal grid (spacing interval of 250 µm), and lens transparency was assessed using a fixed light intensity and exposure time with an EVOS M5000 microscope (Thermo Fisher Scientific). The increased darkness and thickness of the ring within the lens indicate greater opacity, obstructing the passage of light through the lens. 

Cell Assays

To investigate the effects of GDF-15 on EMT in the cell cultures, embryonic stem cell-derived NPCs were plated in 6-well plates with Matrigel pre-coating (#CLS354277; Sigma-Aldrich, St. Louis, MO), and FHL124 cells were plated in 24-well plates. Cells (at 70% confluence) were pretreated with dimethyl sulfoxide (DMSO, 0.1%, #D8418; Sigma-Aldrich) or Smad2 inhibitor SB431542 (10 µM, #S1067; Selleck Chemicals, Houston, TX) or PBS or GDF-15 (50 ng/mL) for 1 hour followed by the addition of PBS or GDF-11 (16.7 ng/mL) treatment for 2 days (FHL124) or 7 days (NPCs). In the NPC culture, GDF-15 was added again on day 3, whereas SB431542 and GDF-11 were only added once at the beginning. Expression of nucleus and fibronectin was detected by immunofluorescence staining. 

To determine the protein expression levels of Smad2 and phosphorylated Smad2 (p-Smad2), and fibronectin, ARPE-19 cells (at 70% confluence) were plated in six-well plates. Cells were pretreated with PBS or GDF-15 (50 ng/mL) for 1 hour and then treated with PBS or TGFβ2 (1 ng/mL) for 1 hour or 2 days, respectively. Expression of Smad2, p-Smad2, and fibronectin was detected by western blotting. 

Immunofluorescence Staining

NPCs and FHL124 were fixed in 4% paraformaldehyde (#50-980-488; Fisher Scientific) for 10 minutes and then washed with PBS for 5 minutes three times. Cells were blocked for 1 hour using cell blocking buffer: Invitrogen 5% normal goat serum (#10000C; Thermo Fisher Scientific) and 0.2% Triton X-100 (#X100; Sigma-Aldrich) in PBS. NPCs and FHL124 were incubated with an anti-fibronectin antibody (1:500, #ab2413; Abcam, Cambridge, UK) overnight at 4°C. After being washed with PBS for 5 minutes three times, the cells were incubated with Invitrogen Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488 (1:500, #A-11034; Thermo Fisher Scientific), and counterstained with Invitrogen 4′,6-diamidino-2-phenylindole, dihydrochloride (DAPI, 30-nM, #D1306; Thermo Fisher Scientific) for 1 hour at room temperature. Fluorescence images were acquired using an Olympus IX83 microscope (Olympus Corporation, Tokyo, Japan). 

Western Blot

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (#89900; Thermo Scientific) with Halt protease and phosphatase inhibitor cocktail (1:100, #78440; Thermo Scientific) followed by sonication and centrifugation at 12,800 rpm for 10 minutes at 4°C. Supernatant was collected and protein concentration was quantified using a bicinchoninic acid (BCA) assay (#23225; Thermo Scientific). Protein lysates were subject to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE, 4%–20% Mini-PROTEAN TGX precast protein gels, #4561094; Bio-Rad Laboratories, Hercules, CA), and then transferred to polyvinylidene difluoride (PVDF) membranes (#1704272; Bio-Rad Laboratories) using a semi-dry Trans-Blot Turbo Transfer System (Bio-Rad Laboratories). After they were blocked with blocking buffer (diluted 1:1 in H₂O, #927-60001; LI-COR Biosciences, Lincoln, NE) for 1 hour, membranes were probed with primary antibodies against Smad2 (1:1000, #3103S; Cell Signaling Technology, Danvers, MA), p-Smad2 (1:1000, #18338S; Cell Signaling Technology), fibronectin (1:1000, #ab2413; Abcam), and glyceraldehyde 3-phosphate dehydrogenase (GADPH, 1:1000, #2118S; Cell Signaling Technology) overnight at 4°C. Membranes were washed with TRIS-buffered saline (TBS, #1706435; Bio-Rad Laboratories) containing 1% Tween 20 (#P9416; Sigma-Aldrich) and then probed with fluorescence-conjugated secondary antibodies: IRDye 800CW Goat anti-Rabbit IgG Secondary Antibody (#926-32211; LI-COR Biosciences) and IRDye 680RD Donkey anti-Mouse IgG Secondary Antibody (#926-68072; LI-COR Biosciences). Protein expression was visualized using chemiluminescence detection reagent and a luminescence imaging system (Odyssey M Imaging System; LI-COR Biosciences). The results were analyzed using ImageJ (National Institutes of Health, Bethesda, MD). 

Quantitative Reverse Transcription Polymerase Chain Reaction

To determine the gene expression levels of EMT markers (α smooth muscle actin 2, fibronectin 1, vimentin), lenses were pretreated with GDF-15 (50 ng/mL) for 1 hour and then treated with TGFβ2 (1 ng/mL) for 3 days. In the FHL124 culture, cells were treated with GDF-15 (50 ng/mL) for 1 hour followed by the addition of TGFβ2 (1 ng/mL) treatment for 24 hours. After collection, RNA was extracted using an RNeasy Plus Mini Kit (#74134; Qiagen, Hilden, Germany) and converted into cDNA by an iScript cDNA synthesis kit (#1708891; Bio-Rad Laboratories). Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed to determine gene expression using iTaq Universal SYBR Green Supermix (#1725121; Bio-Rad Laboratories) and primer sets (OriGene, Rockville, MD). Quantitative PCR reactions were performed using a QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific). The comparative cycle threshold method was used for data analysis, and the relative fold change was the expression level of the specific group compared to PBS only. GAPDH was selected as an internal control for each qRT-PCR analysis. Primer sequences used for qRT-PCR included the following: mouse fibronectin 1 (Fn1), forward 5′-CCCTATCTCTGATACCGT TGTCC-3′, reverse 5′-TGCCGCAACTACTGTGATTCGG-3′; mouse alpha smooth muscle actin 2 (Acta2), forward 5′-TGCTGACAGAGGCACCACTGAA-3′, reverse 5′-CAGTTGTACGTCCAGAGGCATAG-3′; mouse vimentin (Vim), forward 5′-CGGAAAGTGGAATCCTTGCAGG-3′, reverse 5′-AGCAGTGAGGTCAGGCTTGGAA-3′; mouse GAPDH, forward 5′-CATCACT GCCACCCAGAAGACTG-3′, reverse 5′-ATGCCAGTGAGCTTCCCGTTCAG-3′, human fibronectin 1 (Fn1), forward 5′-ACAACACCGAGGTGACTGAGAC-3′, reverse 5′-GGACACAACGATGCTTCCTGAG-3′; human alpha smooth muscle actin 2 (Acta2), forward 5′-CTATGCCTCTGGACGCACAACT-3′, reverse 5′-CAGATCCAGACGCATGATGGCA-3′; human vimentin (VIM), forward 5′- AGGCAAAGCAGGAGTCCACTGA-3′, reverse 5′-ATCTGGCGTTCCAGGGACTCAT-3′; and human GAPDH, forward 5′-GTCTCCTCTGACTTCAACAGCG-3′, reverse 5′-ACCACCCTGTTGCTGTAGCCAA-3′. 

Statistical Analysis

All experiments were independently conducted at least three times. Continuous variables are presented as the mean ± standard error of the mean (SEM). Prior to analyzing group differences, the F-test for equal variance was conducted. Student's t-test was used for comparing two groups, and one-way analysis of variance (ANOVA) was employed for multiple groups. Post hoc analysis involved the application of Tukey’s multiple-comparison post hoc test to correct for multiple comparisons. Target protein expression measured by immunoblotting was determined via densitometry and is expressed relative to an internal control or as phosphorylated protein relative to total protein expression. Statistical significance was defined as P < 0.05. All analyses were performed using Prism 10 (GraphPad Software, Boston, MA). 

Results

GDF-15 Inhibits GDF-11–Induced EMT in NPCs

To determine whether GDF-11 induces EMT in NPCs, we pretreated cells with DMSO or Smad2 inhibitor SB431542, followed by treatment with PBS or GDF-11. Immunofluorescence showed that fibronectin protein was significantly elevated following GDF-11 treatment and was suppressed by SB431542 pretreatment (Fig. 1A), demonstrating that GDF-11 induces EMT via Smad2 activation. We next investigated whether GDF-15 could inhibit GDF-11–induced EMT in NPCs. We pretreated cells with PBS or GDF-15, followed by treatment with PBS or GDF-11. GDF-15 pretreatment suppressed the GDF-11–induced increase in fibronectin expression, indicating that GDF-15 alleviates GDF-11–induced EMT (Fig. 1B). FHL124 cells were treated under the same conditions for 2 days, and the results showed that pretreatment with SB431542 (Fig. 1C) or GDF-15 (Fig. 1D) significantly reduced GDF-11–induced fibronectin expression. 

Figure 1.

 

Immunofluorescence staining of fibronectin (green) and DAPI (blue) in NPCs (A, B) and FHL124 (C, D). GDF-11–induced fibronectin expression was attenuated by either Smad2 inhibitor (SB431542) (A, C) or GDF-15 (B, D) pretreatment. Int Den, integrated density. Scale bar: 100 µm.

GDF-15 Reduces TGFβ2-Induced Smad2 Phosphorylation and EMT in RPEs

TGFβ2 also induces EMT in RPEs.²⁷ To determine whether GDF-15 can inhibit TGFβ2-induced EMT in human ARPE-19 cells, we pretreated cells with PBS or GDF-15, followed by treatment with TGFβ2. TGFβ2 treatment increased p-Smad2 (Fig. 2A) and fibronectin (Fig. 2B), and these effects were inhibited by GDF-15 pretreatment. We further investigated the expression of Fn1 in FHL124. Treatment with GDF-15 significantly reduced TGFβ2-stimulated Fn1 expression (Fig. 2C). These results indicate that GDF-15 is capable of reducing TGFβ2-induced Smad2 phosphorylation and its downstream signaling, including the expression of fibronectin. 

Figure 2.

 

ARPE-19 was treated with TGFβ2 alone or in combination with GDF-15 for 24 hours. (A, B) Typical western blotting of p-Smad2 and Smad2 (A) and fibronectin (B) and quantification of the images. Simultaneous culture with GDF-15 reduced the protein expression of TGFβ2-induced upregulation of p-Smad2 and fibronectin. (C) FHL124 was treated with TGFβ2 alone or in combination with GDF-15 for 24 hours. Treatment with GDF-15 significantly reduced TGFβ2-stimulated Fn1 expression. Protein and gene expressions were normalized to GAPDH. Relative fold changes are represented as 2^–ΔΔCt compared to the PBS-only group. Data are expressed as mean ± SEM (n = 3 per group). *P < 0.05, ***P < 0.001 by Student's t-test.

GDF-15 Reduces the Effects of TGFβ2-Induced EMT in Mice Lenses

To determine if TGFβ2 could induce EMT in mice lenses, extracted lenses were cultured and treated with different concentrations of TGFβ2. As shown in the images, greater opacity (less light transmission) appears darker. After 7 days of treatment, the opacity of the lens increased with the concentration of TGFβ2 (Fig. 3A). To investigate the effects of GDF-15 on lens opacity, extracted mouse lenses were pretreated with PBS or GDF-15 for 1 hour followed by the addition of TGFβ2 treatment for 7 days. Data showed that TGFβ2-induced lens opacity was attenuated by GDF-15 pretreatment (Fig. 3B). In addition, qPCR revealed elevated expression of Acta2, Fn1, and Vim^23,31 in lenses treated with TGFβ2 compared to PBS controls, and GDF-15 pretreatment significantly inhibited this effect (Fig. 3C). These results suggest that GDF-15 could alleviate the effects of TGFβ2-induced EMT and opacity in the lens. 

Figure 3.

 

(A) Mouse lenses treated with increasing concentrations of TGFβ2 showed increased formation of opacity. (B) A significant difference could be seen in the appearance of the lenses treated with PBS, TGFβ2 (10 ng/mL), and TGFβ2 with GDF-15 pretreatment. (C) Treatment with 1-ng/mL TGFβ2 in lenses elevated the expression levels of EMT markers, Acta2, Fn1, and Vim. Pretreatment with GDF-15 for 1 hour alleviated the TGFβ2-induced EMT marker elevation. Scale bar: 300 µm. Relative fold changes are represented as 2^–ΔΔCt compared to the PBS-only group. Data are expressed as mean ± SEM (n = 3 per group). *P < 0.05, **P < 0.01 by one-way ANOVA and post hoc t-test with Tukey’s correction.

Discussion

Currently, the main treatment for cataracts is surgical removal of the opaque lens and replacement with a synthetic lens.⁷ However, complications may occur after surgery, most commonly PCO.^8,9 The standard treatment for PCO is neodymium-doped yttrium aluminum garnet (Nd:YAG) laser capsulotomy,³² a safe and effective treatment.³³ Nonetheless, many complications can still occur after laser capsulotomy, such as retinal edema and retinal detachment.³⁴ Development of novel therapeutics to treat or prevent PCO could further avoid the complications of laser capsulotomy. Here, we showed that GDF-15 inhibits Smad2 phosphorylation induced by TGFβ2 in ARPE-19 cells (Fig. 1). Additionally, GDF-15 reduces TGFβ2-induced fibronectin expression in both ARPE-19 cells and LECs (Fig. 2). The potential mechanism for slowing the progression of PCO is depicted in Figure 4. Accurate differentiation and stable functionality of RPEs and retinal progenitor cells (RPCs) are crucial for retinal development and the treatment of ocular diseases.³⁵ EMT plays a pivotal role in neural differentiation. Neural stem/progenitor cells undergo EMT, characterized by the loss of epithelial traits, cell instability, and increased migration and invasion. For example, alterations in epithelial polarity by neural stem/progenitor cells can lead to impaired cell polarity, contributing to neurodevelopmental disorders.³⁶ RPE dysfunction is associated with retinal diseases, presenting as a loss of RPE barrier function and disrupted RPE polarization. The persistent inflammation and wound healing result in RPE cells undergoing EMT, losing differentiation, and increasing migratory abilities, which lead to the formation and contraction of PVR membranes, contributing to proliferative vitreoretinopathy (PVR).³⁷ 

Figure 4.

 

GDF-15 alleviates TGFβ2-induced EMT by impeding Smad2 phosphorylation, potentially inhibiting the opacity in the lenses.

GDF-15, also known as macrophage inhibitory cytokine-1 (MIC-1), is a member of the TGFβ superfamily and plays an important role in development, differentiation, and repair in the human body.³⁸ Some studies have shown that GDF-15 is involved in the development and progression of diseases such as diabetes, coronary heart disease, and cancer.³⁹ Other studies have shown that GDF-15 is widely distributed in the nervous system, can be positively associated with stress response, and increases neurogenesis.^38,40 GDF-15 also is reported to help promote retinal ganglion cell differentiation.²⁹ 

Here, we found that GDF-15 could be a potential therapeutic agent for PCO. Specifically, GDF-15 inhibited TGFβ2-induced Smad2 phosphorylation in ARPE-19 cells and attenuated TGFβ2-induced fibronectin expression in both ARPE-19 cells and LECs, thereby slowing the progression of lens opacity. Thus, it is possible that GDF-15 also could be applied in other diseases caused by EMT, including idiopathic pulmonary fibrosis^41,42 and migration and metastasis of breast cancer.⁴³ However, recent studies suggest that GDF-15 plays a role in disease progression, including contributing to lung fibrosis and cell senescence.⁴⁴ Therefore, further research is needed to determine if GDF-15 is helpful or harmful to other organs. In addition, further studies are needed to identify any potential complications of long-term GDF-15 treatment in the eye. We previously reported that metformin exposure reduces the risk for requiring early surgical correction of PCO, possibly acting through a mechanism to reduce TGFβ2-induced EMT of lens epithelial cells.⁴⁵ It is of interest that metformin exposure was recently reported to induce circulating levels of GDF-15 in mice.⁴⁶ Further studies will be required to test for the possibility that the effect of metformin in reducing PCO following cataract extraction may be acting through induction of increased levels of circulating GDF-15. 

Despite its strengths, there are some limitations to this study. The biostability of GDF-15 after application to the eyes is unknown. Further, little is known about the retention time of GDF-15 in the eyes.⁴⁷ It also is unknown if GDF-15 can pass through the cornea, as the easiest drug delivery is eye drops. Thus, determining the permeability of GDF-15 in the eyes is a necessary next step toward preclinical study. All in all, this study shows that GDF-15 treatment could prevent PCO or delay its progression, warranting further study of this potentially valuable therapeutic target to prevent vision impairment. 

Acknowledgments

The authors thank Jeffrey Goldberg (Stanford University) for the scientific discussion and Hui-Chun Cheng and Chandler Meadows for their technical assistance. 

Supported by grants from the National Institutes of Health (P30-EY008098 to University of Pittsburgh, 1R01EY028147 to JMP); by a grant from the Henry L. Hillman Foundation (Eye & Ear Foundation of Pittsburgh); by unrestricted research grants from Research to Prevent Blindness (Departments of Ophthalmology at the University of Pittsburgh and University of Colorado); and by a Research to Prevent Blindness career development award (K-CC). 

Author Contributions: SW, CYC and KCC designed the experiments. SW, CYC, DS, CCL, and MR performed the experiments. JMP provides significant materials. SW, CYC and K-CC wrote and revised the manuscript. All authors read and approved the final manuscript. 

Disclosure: S. Wang, None; C.-Y. Chen, None; C.-C. Liu, None; D. Stavropoulos, None; M. Rao, None; J.M. Petrash, None; K.-C. Chang, Stanford University (P)