The local wound environment is a key determinant of the outcome of TGFβ signaling on the fibrotic response of CD44+ leader cells in an ex vivo post-cataract-surgery model

Abstract

The cytokine transforming growth factor beta (TGFβ) has a role in regulating the normal and pathological response to wound healing, yet how it shifts from a pro-repair to a pro-fibrotic function within the wound environment is still unclear. Using a clinically relevant ex vivo post-cataract surgery model that mimics the lens fibrotic disease posterior capsule opacification (PCO), we investigated the influence of two distinct wound environments on shaping the TGFβ-mediated injury response of CD44+ vimentin-rich leader cells. At the leading edge of the wound, the substantial fibrotic response of this cell population to a rigid wound environment required endogenous TGFβ. However, TGFβ was dispensable for the role of leader cells in wound healing in the endogenous basement membrane wound environment, where repair occurs in the absence of a major fibrotic outcome. A difference between leader cell function in these distinct environments was their cell surface expression of the latent TGFβ activator, αvβ3 integrin. This receptor localized exclusively to this CD44+ cell population when they localize to the leading edge of the rigid wound environment. Providing exogenous TGFβ to bypass any differences in the ability of the leader cells to sustain activation of TGFβ in different environments revealed their inherent ability to induce pro-fibrotic reactions on the basement membrane wound environment. Furthermore, exposure of the leader cells in the rigid wound environment to TGFβ led to an accelerated fibrotic response including the earlier appearance of pro-collagen+ cells and alpha smooth muscle actin (αSMA)+ myofibroblasts and increased fibrotic matrix production. Collectively, these findings show the influence of the local wound environment on the extent and severity of TGFβ-induced fibrotic responses, which has important implications for understanding the development of the lens fibrotic disease PCO in response to cataract surgery wounding.

1. Introduction

Shortly after transforming growth factor beta’s (TGFβ) discovery in 1981 (Moses et al., 1981; Roberts et al., 1981; Moses et al., 2016), research revealed roles for this cytokine in regulating the normal and pathological response to wound healing (Sporn et al., 1983; Roberts et al., 1986; Mustoe et al., 1987; Krummel et al., 1988). While there have been major advances in our understanding of TGFβ function since then, how TGFβ shifts from having a pro-repair to a pro-fibrotic role within the wound environment is still unclear. In response to injury, TGFβ signaling is associated with the recruitment of immune cells to the wound site, and promoting the transition of cells such as fibroblasts to a transient myofibroblast phenotype to contract the wound edges for wound healing (Gabbiani, 2003; Pakyari et al., 2013). TGFβ also induces the production of extracellular matrix (ECM) proteins to help stabilize and repair the wound as it heals (Roberts et al., 1986; Ignotz et al., 1987; Pakyari et al., 2013; Xue and Jackson, 2015). Its induction of fibronectin (FN) provides a key provisional matrix that is required for collagen deposition (McDonald et al., 1982; Sottile and Hocking, 2002; Velling et al., 2002). Within the wound microenvironment, TGFβ can become dysregulated, leading to the persistence of myofibroblasts and the accumulation of ECM that alters and disrupts the normal tissue environment, resulting in fibrosis (Hinz and Gabbiani, 2010; Budi et al., 2021). The aberrant accumulation of FN, particularly the EDA splice variant of FN (FN-EDA) and type I collagen are characteristic features of fibrotic disease (Uitto et al., 1982; Serini et al., 1998; Muro et al., 2008; Wight and Potter-Perigo, 2011; Herrera et al., 2018). While TGFβ is a potent driver of fibrotic disease in most tissue types, clinical studies targeting TGFβ to treat fibrosis have met limited success (Gyorfi et al., 2018). TGFβ has remained enigmatic and difficult to target successfully likely due to its pleiotropic nature and ability to function in a highly cell-specific and context-dependent manner (Morikawa et al., 2016).

Importantly, TGFβ exists in an inactive latent complex bound to the surface of cells (Nakamura et al., 2004; Andersson et al., 2008; Stockis et al., 2009a; Stockis et al., 2009b; Tran et al., 2009) or to specific components of the ECM such as FN (Taipale et al., 1994; Hyytiainen et al., 2004; Dallas et al., 2005; Klingberg et al., 2018; Lodyga and Hinz, 2020) to ensure precise control over when and where it is activated within a tissue. Several factors can impact the availability of active TGFβ. The composition of the ECM can influence the amount of latent TGFβ available for activation as well as the presentation of TGFβ ligands because storage of latent TGFβ occurs in the ECM (Robertson and Rifkin, 2016). One of the major ways to activate latent TGFβ is through a conformational-based mechanism involving αv integrins (Munger et al., 1999; Worthington et al., 2011; Hinz, 2013; Lodyga and Hinz, 2020). Integrins bind to the latency-associated peptide (LAP) of the latent TGFβ complex. Through a combination of pulling forces from both the ECM outside the cell and the αv-integrin-linked cytoskeleton, contractile forces are produced inside the cell leading to the release of active TGFβ from the latent complex (Wipff et al., 2007; Lodyga and Hinz, 2020). The availability of TGFβ activators, such as αv integrins, can change the amount of TGFβ ligand accessible for signaling through its cognate receptors. Another critical factor is the mechanical state of the ECM, which through changes in its strain and stiffness can affect cell contractility and influence the ability of a cell to release active TGFβ into the local environment (Wipff et al., 2007; Klingberg et al., 2014; Hinz, 2015).

To determine how the TGFβ-signaling response to injury becomes a major factor in promoting a fibrotic response to wounding, we investigated the effect of two distinct injury microenvironments presented to the wounded cells in a clinically relevant mock-cataract-surgery wound-healing model. Specifically, these studies focused on the TGFβ-signaling response of leader cells, an injury activated CD44+/vimentin-rich cell population that rapidly populates the leading edges of the cataract surgery wound (Walker et al., 2010; Menko et al., 2014; Walker et al., 2018; Menko et al., 2021; Walker and Menko, 2021). This ex vivo mock-cataract-surgery model recapitulates the major features of the lens fibrotic disease posterior capsule opacification (PCO), which enables us to investigate the post-cataract-surgery injury wound healing and fibrotic injury response (Walker et al., 2007; Walker et al., 2010; Walker et al., 2015; Walker et al., 2018). Using this model, we previously identified these CD44+/vimentin-rich leader cells as modulators of the wound-healing response that can serve as progenitors of myofibroblasts, a cell type directly linked to causing fibrotic disease (Walker et al., 2018; Menko et al., 2021). In studies with this model, we previously found that the microenvironment that leader cells encounter profoundly affects their behavior and response to injury. At the wound edge, leader cells direct the wounded epithelium to move across the cell-denuded lens-basement membrane capsule to close the wound by day 3 post-injury (Menko et al., 2014). When these same cells populate the outside wound edge of the cataract surgery explant and move off the lens capsule and onto a rigid tissue-culture substrate, they transition to a myofibroblast phenotype within the short time that the wound on the endogenous lens-basement membrane is closed (Walker et al., 2018).

This paper provides insight into how these two distinct wound microenvironments shape the TGFβ-signaling response and the function of CD44+ leader cells, to influence the outcome of injury. In this study, we determined the requirement for endogenous TGFβ to regulate leader cell-directed collective movement across the basement membrane for wound healing and their fibrotic response to injury. We also investigated how the addition of exogenous TGFβ, to bypass any differences in the ability of the leader cells in these two distinct wound environments to sustain TGFβ, affected their behavior and the outcome to injury. Findings from these studies demonstrate how alterations in the wound microenvironment can shape the severity of the TGFβ-mediated fibrotic response of CD44+ leader cells to injury with important implications to understanding the development of PCO post-cataract surgery wounding.

2. Materials and Methods:

2.1. Ex vivo post-cataract-surgery model and treatments.

Ex vivo post-cataract-surgery lens explants were created from E14 or E15 lenses isolated from chick embryos. Briefly, a mock-cataract surgery procedure was performed on isolated lenses in order to remove the lens fiber cells, leaving behind the tightly attached wounded epithelium to create an ex vivo mock-cataract-surgery explant model in which it is possible to follow both the wound healing and fibrotic response to injury (Walker et al., 2007; Menko et al., 2014; Walker et al., 2015; Walker et al., 2018). Eggs were procured from Poultry futures (Lititz, PA) and Charles River (Wilmington, MA). All animal experiments using chick embryos lenses comply with ARVO guidelines for animals and are approved by the Institutional Animal Care and Use Committee (IACUC) at Thomas Jefferson University (Philadelphia, PA). Ex vivo wound-healing cultures were treated with vehicle (DMSO) or inhibitors to block activation of TGFBR1 using 5μM LY364947 (Selleckchem, Houston, Texas) or 5μM SB525334 (Selleckchem, Houston, Texas). For wound-healing studies, cultures were treated from day 0 through day 3. Phase images were taken on day 0, day 1, day 2 and day 3 using AZ100 Nikon microscope. Wound area was measured using NIS elements software. For fibrosis studies, ex vivo cultures were treated with vehicle, TGFβR1 inhibitors (5μM LY364947 or 5μM SB525334) or SMAD3 was inhibited using 5μM SIS3 (Selleckchem, Houston, Texas) on day 1, after cells began to move onto the rigid wound environment of the ECZ. Inhibitors were replaced each day; day 3 post-injury, cultures were fixed in 4% Formaldehyde or extracted in RIPA buffer (5 mM EDTA, 150 mM NaCl, 1% NP40, 1% sodium deoxycholate, 1% SDS 20% solution, 50 mM Tris-HCl, pH 7.4) with protease/phosphatase inhibitor cocktail (Cell Signaling Technology, Aldrich, Danvers, MA). Phase images of ECZ regions were taken day 3 post-treatment with the Nikon Eclipse Ti microscope and NIS Elements image-analysis software. For exogenous TGFβ studies, ex vivo post-cataract surgery cultures were treated with 10ng/ml TGFβ2 (R&D, Minneapolis, MN) from time 0 post-injury. To assess the effect of TGFβ on proliferation, cultures were pulsed with 10μM EdU for 30 minutes on day 2 in the presence and absence of TGFβ treatment and examined according to manufacturer’s instructions to detect EdU positive cells (Click-iT EdU Imaging Kits; Invitrogen, Carlsbad, CA).

2.2. Immunofluorescence:

Ex vivo mock-cataract surgery cultures were fixed in 4% formaldehyde for 15 minutes, permeabilized in .25% Triton-X-100 for 5 minutes and blocked in 5% goat or donkey serum for 30 minutes. Cultures were incubated with primary antibodies for 30 minutes to 1hr followed by incubation with fluorescent-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA). For immunolabeling we used the following primary antibodies (dilutions for antibodies are listed in Table 1); αSMA (Sigma Aldrich, St. Louis, MO or Abcam, Cambridge, MA), αv integrin (Millipore Burlington, MA), Collagen I alpha (Novus Biologicals, Centennial, CO), Fibronectin (ThermoScientific, Waltham, MA), Fibronectin EDA (Santa Cruz Biotechnology, Santa Cruz, CA), Laminin (Sigma Aldrich, St. Louis, MO), SMAD 2/3 (Cell Signaling, Danvers, MA) and Tenascin-C (Millipore, Burlington, MA). The following primary antibodies were obtained from Developmental Studies Hybridoma Bank, created by the NICHD of the NIH (The University of Iowa, Department of Biology, Iowa City, IA) 1D10 monoclonal antibody to CD44 developed by Halfter, W.M., JG22 monoclonal antibody to β1 integrin developed by Gottlieb, D.I.,and SP1.D8 monoclonal antibody to pro-collagen deposited by Furthmayr, H. Cultures were counterstained with DAPI (Biolegend, San Diego, CA) to identify nuclei or fluorescent-conjugated Phalloidin (Invitrogen, Waltham, MA) to label actin cytoarchitecture. To assess live cell-surface labeling for αvβ3, cultures were incubated day 2 post-injury with αvβ3 antibody (Millipore, Burlington, MA) or isotype control (Jackson ImmunoResearch, West Grove, PA) for 20 min on ice then fixed and processed for immunostaining with fluorescent-conjugated secondary antibodies (Jackson Immunoresearch, West Grove, PA). Images were captured using a confocal Zeiss 510 or Zeiss 800. Z-stacks were collected at .33 μm or .49μm. 3D-structural views and a movie of the 3D structural view were created from z-stack images acquired from Zeiss 800 confocal microscope using Imaris x64.9.5.1 software surface rendering tool. This tool allows the unique ability to view the organization of the extracellular matrix on the lens-posterior capsule; however, is not an indication of fluorescence intensity.

Table 1:

A list of antibodies, companies, catalog numbers, and dilution factors for western blot and immunofluorescence analysis.

Primary Antibody	Company:	Catalog #:	Dilution for western analysis:	Dilution for immunofluorescence analysis:
αSMA (mouse monoclonal)	Sigma Aldrich	a2547	1:500	1:100
αSMA (rabbit polyclonal)	Abcam	ab5694	1:200	1:100
αV Integrin (rabbit polyclonal)	Millipore	ab1930		1:100
αVβ3 (mouse monoclonal)	Millipore	MAB1976Z		1:100 (used for live staining)
β1 integrin (mouse monoclonal)	DSHB	JG22		1:10
CD44 (mouse monoclonal)	DSHB	1D10		1:10
Collagen I (ProCollagen I) (mouse monoclonal)	DSHB	SP1.D8		1:10
Collagen I alpha I (rabbit polyclonal)	Novus Biologicals	NBP1–30054		1:100
FN1 (rabbit polyclonal)	ThermoScientific	PA1–23693		1:100
FN EDA (IST-9) (mouse monoclonal)	Santa Cruz	sc-59826		1:50
GAPDH (mouse monoclonal)	Santa Cruz	sc-47724	1:1000
GAPDH (rabbit polyclonal)	Santa Cruz	sc-25778	1:1000
Laminin (rabbit polyclonal)	Sigma Aldrich	L9393		1:30
SMAD 2/3 (rabbit polyclonal)	Cell Signaling	5678	1:1000	1:100
Phospho-SMAD2 (Ser465/Ser467) (rabbit polyclonal)	Cell Signaling	18338S	1:1000
Tenascin C (rabbit polyclonal	Millipore	ab19013		1:200

2.3. Western blotting:

For SMAD-signaling analysis, ex vivo cultures were extracted 3hr post-injury in Triton X-100/Octyl glucoside (Triton/OG) extraction buffer (44.4 mM n-octyl β-d-glucopyranoside, 1% Triton X-100, 100 mM NaCl, 1 mM MgCl₂, 5 mM EDTA, 10 mM imidazole) with a protease/phosphatase inhibitor cocktail (Cell Signaling Technology, Aldrich, Danvers, MA). To examine the effect of TGFβRI or SMAD inhibitor treatment on αSMA expression, cultures were extracted in RIPA buffer day 3 post-injury. 10–20μg of protein was loaded per lane, separated by 4–12% or 4–16% SDS-PAGE, and transferred at 4 degrees Celsius to PVDF membranes. Membranes were blocked using 5% milk for 1hr and incubated overnight at 4 degrees with the following primary antibodies (dilutions are listed in Table 1): αSMA (Sigma Aldrich St. Louis, MO or Abcam, Cambridge, MA), GAPDH (Santa Cruz Biotechnology, Santa Cruz, CA), pSMAD (Cell Signaling, Danvers, MA) and Total SMAD2/3 (Cell Signaling, Danvers, MA). Following incubation with HRP- conjugated secondary antibodies (Bio-Rad, Hercules, CA) for 1h, membranes were exposed to ECL chemiluminescence-substrate-detection reagent (Thermo Scientific, Waltham, MA) according to manufacturer’s directions, and blots were processed using the ProteinSimple machine.

2.4. Data analysis and Statistics:

Quantification of fluorescent images was performed using Image J by measuring fluorescence intensity. Quantification of western blot data was performed with ProteinSimple software to detect the band mean intensity. Data are presented +/− S.E.M. Unpaired, two-tailed Student’s t-tests were used for comparisons between two groups. One-way ANOVA analysis was used for comparing differences between two or more groups. Statistical analysis was performed using GraphPad Prism software (*, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001, ****, p ≤ 0.0001).

Results

3.1. Leader cells produce FN and pro-collagen I in response to injury in a wound-environment-specific manner

Different types of matrices can be produced by TGFβ in response to wounding that have inherent properties that are either pro-regenerative or pro-fibrotic (Budi et al., 2021). Here, we investigated whether the cells that localize to the leading edge of the wound and that have innate potential to become αSMA+ myofibroblasts (Walker et al., 2018; Menko et al., 2021), are also producers of fibrosis promoting FN/FN EDA and collagen I matrices in the ex vivo post-cataract surgery chick model. Previously we demonstrated that the wound environment is a critical determinant of the fate of these cells and whether the injury outcome is regenerative or fibrotic (Walker et al., 2018). Here, we examined the impact of two distinct wound microenvironments on the induction of matrix production by cells at the leading edge of the mock-cataract-surgery wound. The ex vivo post-cataract surgery procedure creates a star-shaped explant with two major sites of injury and, therefore, two distinct wound environments (Fig.1i). One is at the wound site that requires repair on the basement membrane capsule from where the fiber cells were removed, which is referred to as the central migration zone (CMZ). In this environment, CD44+ cells that locate to the wound edge direct the migration of the wounded-lens epithelial cells across the cell-denuded posterior lens-capsule-basement membrane, a regenerative repair process completed by day 3 post-injury (Fig. 1i,ii (Bleaken et al., 2016)). The other wound environment is located adjacent to the cut outside edge of the surgery explant where the same population of CD44+ cells function to direct movement of the wounded-lens epithelial cells off the lens capsule onto the rigid tissue culture substrate, referred to as the extracapsular zone (ECZ) (Fig. 1i,iii). In this rigid microenvironment these CD44+ leader cells undergo a rapid transition to a αSMA+ myofibroblast phenotype by day 3 post-injury associated with fibrotic repair (Fig. 1iii (Walker et al., 2018)), similar to the response we observed when we provided leader cells with a collagen I substrate with the rigidity (25kPa) of a typical fibrotic microenvironment (Walker et al., 2018). For our studies, we will study how the tissue culture substrate and the lens basement membrane capsule, two distinct, and reductionist microenvironments, impact the behavior of CD44+ leader cells in response to lens injury.

Figure 1: Ex vivo post-cataract surgery model that provides the ability to study how two distinct wound environments influence the outcome to injury.

Ex vivo post-cataract surgery chick model, showing the effect of the wound microenvironment on influencing the pro-regenerative wound healing vs. pro-fibrotic outcome to injury. Mock-cataract surgery produces a star-shaped lens epithelial explant culture with a centrally located fiber-cell-denuded native basement membrane posterior lens capsule (i, asterisk, (Walker et al., 2007; Walker et al., 2015)). Injury-activated CD44+ leader cells (indicated by green lines in i) direct the injury response. On the lens capsule wound environment, CD44+ leader cells (ii) direct the movement of the wounded epithelial cells across the cell-denuded basement membrane into the central migration zone (CMZ) for normal pro-regenerative repair. Wound healing is typically completed by day 3 post-injury. At the cut edge of the lens capsule (black arrows), CD44+ leader cells (Aiii) also direct the wounded epithelial cells off the lens capsule onto the rigid tissue-culture substrate termed the extra capsular zone (ECZ, black arrowhead, i). In the ECZ, by day 3 post-injury there is rapid transition of leader cells into αSMA+ myofibroblasts associated with fibrotic disease.

A cross-sectional confocal view of the mock cataract surgery explant (at 1 day in culture) following immunolabeling of the lens capsule with laminin and labeling of the cells for F-actin and nuclei highlights the laminin-rich basement membrane of the CMZ on which the cells migrate to close the cataract surgery wound (Fig. 2A, arrow). This study also demonstrates that cataract-surgery injury does not damage the native-basement membrane when the fiber cells were removed from their attachment sites on the posterior-lens-capsule by hydroelution. Using immunostaining and confocal imaging, we did not detect expression of FN-EDA (Fig. 2B,C, Day 2) or collagen I (Fig. 2F,G, Day 2) by the cells at the leading edge as they migrate across this basement membrane substrate of the CMZ during the first two days in culture. While at culture day 3, after the wound was closed, FN-EDA was still not detected (Fig. 2D,E), the cells in the center of the closed wound express low levels of collagen I (Fig 2H,I).

Figure 2: Leader cells do not produce significant amounts of FN or collagen I in response to injury on the lens capsule wound environment.

(A) Representative immunofluorescence image at the wound edge, on the lens capsule wound environment. An orthogonal view is shown from a confocal z-stack of a day 1 ex vivo post-cataract-surgery wounded culture labeled for laminin (green, arrow) and counterstained for F-actin (red) and nuclei (blue). The leading edge is indicated by an arrowhead. (B-E) Representative confocal images from the lens capsule wound environment focused at the leading edge of the CMZ on day 1 (B,C), day 2 (F,G) or center of the wound on day 3 (D-I) immunostained for FN EDA (B-E) or collagen I (F-I). Samples were counterstained with DAPI (blue, C,E,G,I) to identify nuclei. Single optical sections presented in (B-E) and projection images in (F-I). Magnification (Mag.) bar = 20μm

Next, we examined matrix production by leader cells that also populate the cut edge of the explants adjacent to the culture substrate and direct the wounded epithelium onto and across the rigid environment of the ECZ. By day 2 post-injury the CD44+ population that locates to the leading edge of the cells migrating across the ECZ have been induced to produce FN (Fig. 3A–F). Imaging of FN EDA labeling of these cells at their basal-substrate facing surfaces showed that the cells at the leading edge also had organized an extracellular FN EDA matrix (Fig. 3G,H), also shown at the leading edge at culture day 3 (Fig. 3I,J(Menko et al., 2020)). While only low levels of collagen I production were expressed by the cells at the leading edge of the ECZ by day 2 in culture (Fig, 3K–M), labeling with a pro-collagen I antibody showed that this matrix protein was highly expressed by day 3 (Fig. 3N–P). In this study the cells were co-labeled for pro-collagen I and αSMA. The results show the coincidence of expression of pro-collagen I with the transition of these cells to an αSMA+ myofibroblast phenotype. However, we also observed pro-collagen I production by certain cells, that appears before the detection of αSMA (Fig. 3P, white arrowheads). By culture day 3, following the organization of a FN-EDA matrix on day 2, we found many of the cells expressing pro-collagen I are αSMA+ myofibroblasts (Fig. 3P, white arrows (Menko et al., 2020)). The expression of collagen and αSMA by these cells is subsequent to their organization of a FN EDA matrix on the substrate surface. These findings demonstrate that in response to wounding, the CD44+ cells at the leading edge of the ECZ that acquire a myofibroblast phenotype also operate as the agents of FN/FN-EDA and collagen I ECM production in response to a rigid local wound microenvironment.

Figure 3: Leader cells produce FN and collagen I in response to injury in the rigid wound-environment.

Representative confocal images focused at the leading edge within the rigid wound environment of the ECZ on day 2 (A-H and K-M) and day 3 (I,J and N-P). Cultures were labeled for FN (C,D,E and F) and CD44 (A,B,E and F) or FN-EDA (G-J). CD44+ cells producing FN in A,C,E in the region of interest indicated by the white box are shown at higher magnification beneath each image in B,D and F (see white arrows). To identify pro-collagen producing myofibroblasts, cultures were labeled for pro-collagen (K,M, N and P) and αSMA (L,O, M and P). White arrows indicate αSMA+ cells co-expressing pro-collagen I (N-P). Pro-collagen+ cells not yet positive for αSMA are indicated by white arrowheads (N-P). Samples were counterstained with fluorescent-conjugated phalloidin to detect F-actin (blue, E,F,J,M and P, blue) or DAPI (blue, H). Single optical sections are presented in (A-J) and projection images in (K-P). Magnification bar = 20μm.

3.2. No requirement for endogenous TGFβ signaling for leader cell-directed movement of the wounded epithelium for wound closure in response to cataract-surgery wounding

Next, we examined whether endogenous TGFβ provides a signaling cue to regulate leader-cell-directed collective migration of the wounded lens epithelium across the posterior lens capsule native basement membrane of the CMZ. Treatment of ex vivo post-cataract surgery cultures 3hrs post-injury with the TGFβR I inhibitors, LY364947 or SB525334 to block endogenous TGFβ signaling prevented activation of SMAD on the wounded lens capsule without any effect on total SMAD 2/3 expression levels (Fig. 4A and B). However, treating ex vivo post-cataract surgery cultures with LY364947 or SB525334 from time 0 through day 3 post-injury to block endogenous TGFβ signaling had no effect on wound closure (Fig. 4C and D). Both vehicle control and TGFβR1-inhibitor (LY364947 and SB525334)-treated ex vivo post-cataract surgery cultures healed by day 3 post-injury (Fig. 4C and D). These findings show that endogenous TGFβ signaling is not required for leader-cell directed collective migration of the wounded epithelium across the cell-denuded basement membrane for wound closure.

Figure 4: Leader cell-directed collective migration for wound closure occurs independently of endogenous TGFβ signaling.

(A,B) Activation of endogenous TGFβ-SMAD signaling was determined 3hr post-wounding in the ex vivo post-cataract surgery cultures in the presence and absence of TGβRI inhibitors LY364947 (LY, 5μM) or SB525334 (SB, 5μM). Immunoblots show expression levels of pSMAD (A), Total SMAD 2/3 (B) and GAPDH (loading control). Graphs depict the relative amount of pSMAD or SMAD to GAPDH with values normalized to control. p-values (****p ≤ .0001, n=3). Error bars represent SEM. (C,D) Ex vivo post-cataract surgery cultures were treated +/− TGFβRI inhibitors LY364947 (LY, 5μM) or SB525334 (SB, 5μM) on day 0-day 3 post-injury. (C) Representative phase images of wound healing shown for day 0 and day 3 post-treatment. (D) Graph depicts the % wound closure on day 0, day 1, day 2, and day 3 +/− TGFβRI inhibitor treatment. Error bars represent SEM. The results represent at least 3 independent experiments.

3.3. Within the rigid wound environment endogenous TGFβ signals leader cells to produce FN-EDA and pro-collagen I

To determine if FN-EDA and pro-collagen I ECM production by leader cells in the ECZ rigid substrate wound environment is under the control of endogenous TGFβ, we treated ex vivo post-cataract surgery cultures with the TGFβR I inhibitors LY364947 and SB52533. First, we examined expression and localization of SMAD2/3, a downstream effector of TGFβ signaling whose localization to the nucleus is an indicator of activation of this pathway. While SMAD2/3 is primarily nuclear in the cells at the leading edge of the ECZ in the vehicle control treated explants, the treatment with either LY364947 or SB52533 appears to lead to an increase in cytoplasmic SMAD2/3, demonstrating an effective suppression of endogenous TGFβ signaling in these cells (Fig. 5A–F). Similar to wound healing on the lens capsule basement membrane wound environment, blocking endogenous TGFβ signaling did not prevent cell movement across the ECZ, nor did it affect the ability of leader cells to extend protrusions at the leading edge (Supplemental Fig. 1). However, inhibiting endogenous TGFβ signaling effectively prevented leader cell expression of both FN-EDA (Fig. 5G–I, M) and pro-collagen I (Fig.5 J–L,N). These findings indicate that injury-activated leader cells are signaled by endogenous TGFβ to produce FN-EDA and pro-collagen I in the rigid wound environment.

Figure 5: Endogenous TGFβ signals FN-EDA and collagen l matrix production by leader cells in the rigid wound environment.

(A-N) Ex vivo post-cataract-surgery cultures were treated with +/− TGFβRI inhibitors LY364947 (LY, 5μM) or SB525334 (SB, 5μM) on day 1 through day 3. (A-F) Representative confocal images of ex vivo cultures treated with +/−TGFβR1 inhibitors immunolabeled on day 3 post-treatment for total SMAD2/3 (white, A-F) and overlaid with DAPI (blue, B,D,F). (G-L) Representative confocal images of ex vivo post-cataract-surgery cultures day 3 +/− TGFβRI inhibitor treatments labeled for FN-EDA (green, G-I) or pro-collagen I (green, J-L) overlaid with DAPI (blue, G-L). (M-N) Graph depicts mean fluorescence intensity change in FN EDA (M) and pro-collagen (N) expression. Values normalized to vehicle control. p-values (****p ≤ .0001, **p≤ .01, n=3). Error bars represent SEM. Magnification bars= 20μm.

3.4. Transition of leader cells to a myofibroblast cell fate in a rigid wound environment is dependent on endogenous TGFβ signaling

We have previously shown that CD44+ leaders are induced to become αSMA+ myofibroblasts when they encounter the rigid environment of the ECZ (Walker et al., 2018). Since TGFβ is a well-known mediator of myofibroblast differentiation (Gabbiani, 2003; Pakyari et al., 2013), we investigated whether endogenous TGFβ signaling regulates this transition of CD44+ leader cells to αSMA+ myofibroblasts. Interestingly, we now show that the presence of CD44 on the surface of the cells at the leading edge is reduced as the cells acquire a myofibroblast phenotype (Fig. 6A–G). Treatment with the TGFβR1 inhibitors LY364947 or SB52533 prevented this decrease in CD44 receptor expression (Fig. 6H–N) and the expression of αSMA+ (Fig. 7A–E), providing evidence that blocking TGFβ signaling resulted in the leader cells maintaining their native phenotype despite their presence in a rigid, pro-fibrotic wound environment. To establish whether the acquisition to a myofibroblast fate is modulated by a canonical TGFβ-SMAD-signaling pathway, we treated the ex vivo post-cataract surgery cultures with the SMAD3 inhibitor SIS3. Paralleling our findings with the TGFβR1 inhibitors (Fig. 7E), we found that SIS3 treatment blocked αSMA expression (Fig. 7F). These findings confirm the role for an endogenous TGFβ-SMAD-signaling pathway in inducing the emergence of αSMA+ myofibroblasts in the rigid ECZ wound environment.

Figure 6: As leader cells transition to αSMA+ myofibroblasts there is a coincident decrease in CD44 expression, which is under the control of TGFβ signaling.

(A-F) Representative confocal images of ECZ regions on day 2 and day 3 labeled for CD44 (A,C,D,F), αSMA (B,C,E,F) and overlaid with DAPI (blue, C,F). (G) Graph depicts mean fluorescence intensity change in CD44 expression on day 2 compared to day 3. Values normalized to day 2. p-value (*p ≤ .05, n=2). (H-N) Representative confocal images of vehicle (DMSO) and TGFβRI inhibitor treated ECZ regions day 3 post-treatment labeled for CD44 (H-M) and overlaid with DAPI (blue, I,K,M). (N) Graph depicts mean fluorescence intensity change in CD44 expression. Values normalized to vehicle control. Error bars represent SEM. p-values (****p ≤ .0001, n=3). Magnification bars= 20μm.

Figure 7: TGFβ-SMAD signaling mediates leader-cell transition into αSMA+ myofibroblasts.

(A-F) Ex vivo post-cataract-surgery cultures were treated with +/− TGFβRI inhibitors LY364947 (LY, 5μM), SB525334 (SB, 5μM) (A-E) or SMAD inhibitor (SIS3, 5μM) from day 1 through day 3 (F). (A-C) Representative confocal images show ECZ regions treated +/− TGFβRI inhibitors day 3 post-treatment labeled for αSMA (white) and overlaid with DAPI (blue). (D) Graph depicts mean fluorescence intensity change in αSMA expression. Values normalized to vehicle control. p-values (**p≤ .01, n=3). (E,F) Immunoblots show expression levels of αSMA and GAPDH +/− TGFβRI inhibitor (E) or +/− SIS3 (F). Graphs depict relative amount of αSMA to GAPDH. Values normalized to control. Error bars represent SEM, p-values (****p ≤ .0001, n=3). Magnification bar = 20μm.

3.5. Exclusive cell-surface expression of the latent TGFβ activator αvβ3 on leader cells when they encounter the rigid environment of the ECZ

We next sought an explanation for why we observe a TGFβ-dependent pro-fibrotic response of leader cells in the rigid wound environment of the ECZ but not on the lens-capsule-wound environment of the CMZ. One possibility could be due to environment specific differences in the ability of leader cells to sustain activation of TGFβ, as a result of variations in the availability of latent TGFβ activators, such as αv integrins on leader-cell populations. To examine this possibility, we live-labeled ex vivo post-cataract surgery cultures on day 2 post-injury with an antibody to the extracellular cell surface epitope of αvβ3 integrin. We found that αvβ3 exclusively labeled the cell surfaces of leader cells that specifically locate to the wound edge in the rigid environment compared to the lens capsule wound environment (Fig. 8A–D). Live labeling with an isotype-control antibody confirmed the specificity of αvβ3 binding to the cell surface of leader cells (Supplemental Fig. 2). Immunolabeling results using an αv-integrin antibody against an epitope in the αv cytoplasmic domain revealed that αv integrin is expressed by the CD44+ cells that locate to the leading edge in both wound environments (rigid vs. lens capsule wound environment) (Fig. 8 E–N), while only present on the cell surface where it can participate in activation of TGFβ in the rigid wound environment. In addition, we noted a distinct organization of αv integrin in the CD44+ leader cells in these two microenvironments. In the rigid wound environment (ECZ) αv integrin localizes to classical focal adhesion complexes at the tips of the CD44+ leader cells as well as along F-actin stress fibers (Fig. 8E–I), consistent with a mechanotransduction function for this integrin. In contrast, in the lens-capsule-wound environment (CMZ), αv integrin had a punctate appearance within the CD44+ leader cells (Fig. 8 J–N), possibly reflecting vesicular trafficking of this integrin (Moreno-Layseca et al., 2019). For comparison, we also examined the localization of β1 integrin, which forms heterodimers with a wide array of α integrin subunits (Hynes and Naba, 2012), in the two wound environments. In contrast to αv and αvβ3, β1 integrin localizes to the cell-cell borders of all leader and follower cells regardless of the microenvironment (CMZ and ECZ) (Fig. 8 O–R). Overall, these findings show important distinctions in the organization of αv and cell-surface expression of αvβ3 that may equip leader cells in the rigid wound environment of the ECZ with ability to activate and sustain TGFβ signaling better than cells at the leading edge of the basement-membrane wound environment of the CMZ.

Figure 8: αvβ3 is expressed on the cell surface of leader cells located in the rigid wound environment.

(A-R) Representative confocal images of leader cells day 2 post injury on the rigid wound environment (ECZ, A, B, E-I, O,P) or the lens-capsule-wound environment (CMZ, C, D, J-N, Q,R) of the ex vivo post-cataract surgery cultures. Cultures were live labeled with an antibody to the extracellular epitope of αvβ3 to identify cell surface expression of αvβ3 integrin (green, A-D) and overlaid with DAPI (blue). (E-N) Representative confocal images labeled for αv integrin (white, E-G, J-L), CD44 (green, F,I,K,N), F-actin (purple, F, H, K, M) and DAPI (blue, E-N). (O-R) Representative confocal images labeled for β1 integrin (O-R) and overlaid with DAPI (blue, P,R). Mag bar for = 20μM.

3.6. Exogenous TGFβ treatment accelerates the transition of CD44+ leader cells to pro-collagen producing myofibroblasts and leads to a worsened fibrotic response within the rigid wound environment of the ECZ

Our findings above show an endogenous TGFβ-dependent fibrotic response by leader cells within the rigid wound environment of the ECZ (Figs. 5–7). Next, we determined whether exposing leader cells to exogenous “activated” TGFβ augments their fibrotic response in the rigid wound environment. First, we examined whether the addition of exogenous TGFβ induces an earlier transition of CD44+ cells to an αSMA+ myofibroblast phenotype prior to day 3 post-injury, at which time this transition typically occurs under endogenous TGFβ conditions (Walker et al., 2018). Confocal imaging in the ECZ region of cultures co-immunolabeled on day 2 post-TGFβ treatment for CD44 and αSMA show decreased expression of CD44 by myofibroblast progenitor cells at the leading edge of the wound that coincided with their accelerated transition into αSMA+ myofibroblasts (Fig. 9D–F, G, H). In contrast, under endogenous TGFβ conditions cells at the leading edge retained their expression of CD44 at day 2 post-wounding and were not yet αSMA+ (Fig.9 A–C, G,H). We also determined if the accelerated differentiation of leader cells to αSMA+ myofibroblasts in the presence of exogenous TGFβ promoted their production of collagen I. Wounded cultures were imaged at the leading edge of the ECZ +/− TGFβ, following co-immunolabeling for pro-collagen I and αSMA. Our findings indicate that exogenous TGFβ induces the expression of collagen by αSMA+ cells, as early as day 2 post-injury (Fig.9 L–N, white arrows). We also observed an increase in pro-collagen producing cells that were not + for αSMA in the presence of exogenous TGFβ (Fig. 9 L–N,O, white arrowheads) compared to endogenous TGFβ treatment (Fig. 9 I–K, O).

Figure 9: Exogenous TGFβ induces earlier pro-collagen production, CD44+ cell proliferation and accelerates CD44+ leader cells transition to an αSMA+ myofibroblast phenotype.

(A-P) Ex vivo post-cataract surgery cultures treated +/− 10ng/ml exogenous TGFβ from time 0 through day 2. (A-F) Representative confocal images of cells within the rigid ECZ wound environment labeled for CD44 (A, C, D, F) and αSMA (B,C,E,F) or pro-collagen (I,K,L,N) and αSMA (J,M,K,N) overlaid with DAPI (blue, C,F and K,N). Graph depicts mean fluorescence intensity change in CD44 (G), αSMA (H), and pro-collagen (O) expression +/− exogenous TGFβ treatment. Values are normalized to exogenous TGFβ treatment. Error bars represent SEM, p-values (*p≤ .05,****p ≤ .0001, CD44 n=6, αSMA n=3, pro-collagen n=3). (P) Graph shows the % EdU+ CD44 cells treated +/− exogenous TGFβ day 2 post-treatment. p-value (*p ≤. .05, n=3). Mag. bar =20μm

Since the number of pro-collagen-producing cells dramatically increased at day 2 post-wounding when cultured in the presence of exogenous, activated TGFβ (Fig. 9 L–O), we examined whether TGFβ had induced proliferation of CD44+ leader cells in the rigid ECZ wound environment. Indeed, exposure to exogenous TGFβ treatment led to a 79% increase in CD44+ leader cell proliferation rate in this wound environment (Fig. 9P). This increase in CD44+ cells with a high susceptibility to becoming myofibroblasts is likely to contribute to the pro-fibrotic wound outcome in this wound environment. Therefore, we compared the effect of exogenous vs. endogenous TGFβ signaling on the extent of the fibrotic reactions produced within the rigid wound environment of the ECZ at day 3 post-wounding. Confocal imaging was performed of the cells in the ECZ at day 3 +/− the addition of exogenous TGFβ following co-immunolabeling for FN and pro-collagen I (Fig. 10 A–F). The results showed an increase in the accumulation and deposition of a fibronectin fibril-rich matrix along with an increase in pro-collagen I-producing cells in the ECZ when the wounded explants were cultured in the presence of exogenous, activated TGFβ (Fig. 10 A–F). Interestingly, while the addition of exogenous TGFβ led to a significant increase in FN and pro-collagen production, it did not necessarily result in a dramatic increase in the transition to αSMA+ stress fiber containing myofibroblasts at culture day 3 (Fig. 10 G–L), the time period when αSMA+ myofibroblasts emerge when exposed only to endogenously produced TGFβ. Similar to our findings at culture day 2 (Fig. 9L–N), in the presence of exogenous TGFβ on culture day 3 there are pro-collagen I producing cells present that are not expressing αSMA (Fig. 10 G–L, white arrowheads). Together, these studies show that the addition of TGFβ to cells in the rigid wound microenvironment leads to an accelerated and exacerbated fibrotic response.

Figure 10: Exogenous TGFβ exacerbates the fibrotic response in the rigid wound environment.

(A-L) Ex vivo post-cataract surgery cultures treated +/− 10ng/ml TGFβ from time 0 through day 3. Representative confocal images of cells within the rigid wound environment (ECZ) regions labeled for FN (A,C,D,F) and pro-collagen (B,C,E,F) or αSMA (G I,J,L) and pro-collagen (H,I,K,L) and overlaid with DAPI (blue, C,F,I,L). Consecutive projected confocal images assembled to create presented images. Mag bar for =20μm

3.7. CD44+ leader cells have an inherent ability to respond to exogenous TGFβ to produce fibrotic reactions on the native basement-membrane wound environment

Our finding that leader cells on the lens capsule basement membrane environment lack cell-surface expression of the TGFβ activator αvβ3 may limit their ability to sustain activation of TGFβ to produce fibrotic reactions in this wound environment. Therefore, to bypass any differences in the availability of active TGFβ to leader cells, we provided exogenous TGFβ to the ex vivo post-cataract surgery explants and determined whether exposure to this activated TGFβ induced a fibrotic response of the leader cells in the wound environment of lens basement membrane capsule (CMZ). Confocal imaging of cultures immunolabeled for CD44 and co-labeled for either TN-C or FN, shows that by culture day 1 the addition of TGFβ induced CD44+ leader cells to produce (Fig. 11 B,D, Supplemental Fig. 3) and deposit (Fig. 11 I–L) a TN-C and FN-EDA-enriched matrix along the basal surfaces of the cells migrating across the lens capsule basement-membrane substrate as they undergo wound healing. The organization of FN-EDA fibrils along the basement membrane was more extensive than that of TN-C (Fig. 11 I–L). In the absence of exogenous TGFβ, we did not detect production (Fig. 11A,C, Supplemental Fig. 3) or deposition (Fig. 11 E–H) of TN-C or FN-EDA by cells at the leading edge of the wound in the basement-membrane wound environment.

Figure 11: Exogenous TGFβ induces TN-C and FN/FN-EDA matrix production by CD44+ leader cells on the endogenous lens-capsule-wound environment.

Ex vivo post cataract surgery cultures treated with and without 10ng/ml TGFβ from day 0 through day 1 (A-L). (A-D) Representative confocal images of day 1 post-injury ex vivo post-cataract surgery explants labeled for CD44 (green, A-D) and Tenascin-C (TN-C, red, A, B) or Fibronectin (FN, red, C, D) overlaid with DAPI (A-D). (E-L) Representative confocal images day 1 +/−TGFβ treatment focused along the basal substrate of the lens basement membrane capsule labeled for FN EDA (green, E,F,H, I, J, L) and TN-C (red, E, G, H, I, K, L) counterstained for DAPI (blue, E,H, I,L) and F-actin (purple, E, H, I, L). Regions boxed in E and I are shown at higher magnification in F-H and J-L, respectively. Mag bar=20μm

Given the extensive deposition of FN-EDA in response to TGFβ treatment in the wound environment of the native lens capsule and the link of this matrix protein to collagen deposition (McDonald et al., 1982; Sottile and Hocking, 2002; Velling et al., 2002) and fibrotic disease (Serini et al., 1998; Muro et al., 2008; Wight and Potter-Perigo, 2011; Herrera et al., 2018), we also examined the presence of a FN-EDA matrix in the ex vivo mock-cataract-surgery cultures at day 3 and day 9 post-injury, times after wound closure was complete, +/− TGFβ. At culture day 3, the CD44+ leader cells that initially located to the wound edge now occupied the center of the healed wound (Fig. 12A, B). In the presence of TGFβ there was a FN-EDA-rich fibrillar matrix specifically within this central area of the now closed wound (Fig. 12 F–H), the same region where the CD44+ cells were accumulated (Fig. 12 A,B). FN-EDA fibrils also were seen radiating out from this central region in thin lines located between the individual sections of the explant created when it is flattened on the substrate surface (Fig 12F, arrow). Accumulation of FN-EDA was notably absent in the region of the capsule where the wounded epithelial cells remain linked to their original sites of attachment (Fig. 12F, arrowhead). FN-EDA was not detected on the central area of the lens capsule in the control cultures (Fig. 12 C–E). Labeling of day 3 cultures +/−TGFβ for F-actin with fluorescently-conjugated phalloidin confirmed that the wounded-lens epithelial cells associated with the basement membrane capsule maintain cortical actin linking their apical domains and their epithelial morphology even in the presence of exogenously added TGFβ (Supplemental Fig. 4).

Figure 12: Exogenous TGFβ induces the assembly of a FN-EDA fibrillar-rich matrix in the central wound area of the posterior lens capsule.

(A-B) Representative confocal images of day 3 post-injury ex vivo post-cataract-surgery explant labeled for CD44 (green) and overlaid with DAPI (blue). Boxed area of the central wound region of the lens capsule (A) is shown at higher magnification in B. (C-O) Ex vivo post-cataract-surgery cultures were treated +/− 10ng/ml TGFβ from time 0 through day 3 (A-H) or day 9 post-injury (I-O). Representative confocal images are labeled for FN EDA (green C-H; white I-O) counterstained for DAPI (blue, C,E,F,H,I,K,L,N,O). Low magnification images of the ex vivo post-cataract-surgery cultures are shown in (A, C, F, I, L). Arrow in F indicates FN-EDA fibrils radiating out from the central wound region. Arrow head in F indicates regions of notable absence of FN-EDA on the lens capsule flap. (O) Imaris 3D-structural rendering of confocal z-stack image is presented in (N). (D,E, G, H, J,K,M, N) projection images. Single optical plane presented in B. Mag. bar for (B, D-H, J-N) =20μm. Mag. bar for (A, C, F, I, L) =500μM. Mag. bar for (O) = 30 μM.

The FN-EDA fibrillar matrix persisted on the posterior lens capsule at day 9 post-wounding in explants cultured in the presence of exogenous TGFβ (Fig. 12 L–N), while even at this later time point no FN-EDA was detected under endogenous TGFβ conditions (Fig. 12 I–K). A 3D-structural view (Fig. 12 O) and movie of the 3D structural view (Supplemental Fig. 5) was constructed from a confocal z-stack of which the image presented in Fig. 12N to highlight the organization of the elaborate fibrillar FN-EDA matrix produced by the leader cells on the center of the posterior lens capsule in the presence of exogenous TGFβ. These results show that in the presence of persistently activated TGFβ, a FN-EDA fibril-rich matrix persisted on the posterior lens capsule, the region where the lens fibrotic disease PCO manifests in cataract surgery patients (Wormstone et al., 2009).

Since dysregulation of type I collagen production is a major indicator of fibrotic disease, we also examined the presence of collagen I-producing cells in response to cataract surgery injury +/− exogenous TGFβ. For these studies we used an antibody that recognizes both collagen I fibrils and pro-collagen I. To identify collagen I production by CD44+ cells, ex vivo post-cataract surgery cultures were co-labeled for collagen I and CD44. Confocal imaging of immunolabeled ex vivo post-cataract surgery cultures at culture day 3, revealed an increase in type I collagen-production by CD44+ cells in response to the presence of exogenous TGFβ (Fig. 13A–F). Next, we determined if exogenous TGFβ treatment promoted leader cell differentiation into αSMA-stress-fiber-expressing myofibroblasts. Confocal imaging of ex vivo post-cataract-surgery cultures day 3 post-wounding +/− TGFβ-treatment showed a small increase in αSMA cytoplasmic expression on the central lens capsule day 3 post-injury compared to untreated controls (Fig. 13 G–J). However, prolonged treatment with TGFβ through day 9 post-injury led to a permissive environment for the formation of αSMA+ stress fiber containing myofibroblasts on the central posterior lens capsule (Fig. 13 N–P), the same region where we find an extensive accumulation of FN-EDA fibrillar matrices (Fig. 12 L–N), a favorable environment for myofibroblast differentiation (Serini et al., 1998). Overall, these findings show that CD44+ leader cells are capable of responding to TGFβ to promote fibrotic reactions in the posterior lens capsule basement-membrane wound environment.

Figure 13: Exogenous TGFβ increases collagen I production by CD44+ leader cells and their transition to αSMA+-stress-fiber-forming myofibroblasts in the center of the lens capsule basement membrane.

(A-P) Ex vivo post-cataract-surgery cultures shown were treated with and without 10ng/ml TGFβ from time 0 through day 3 (A-J) or day 9 post-injury (K-P). (A-F) Representative confocal images focused on the center of the wound capsule of the ex vivo post-cataract-surgery cultures day 3 post-injury co-labeled for collagen I (A,D) and CD44 (B,E) and overlaid with DAPI (blue, C,F). (A-P) Images from day 3 (A-J) or day 9 (K-P) post-cataract-surgery cultures focused on the center of the wound capsule labeled for αSMA (K,-P) and overlaid with DAPI (blue, K, M, N, P). Mag. bar for (A-J, L,M,O,P) = 20μm and Mag. bar for (K,N )= 500μM.

4. Discussion

TGFβ is a key determinant of fibrotic disease across most tissue types (Pohlers et al., 2009; Saika et al., 2009; Biernacka et al., 2011; Frangogiannis, 2020), and its pleiotropic and context-dependent nature have hampered efforts to effectively target this cytokine in fibrosis (Morikawa et al., 2016). Therefore, there is a fundamental need to elucidate how cues from the local microenvironment shape the TGFβ-signaling response and how these cues cause TGFβ signaling to tip away from a normal wound-healing response to promote fibrotic disease. Our present studies, performed in a clinically relevant injury setting, revealed differences in how the local wound microenvironment controls TGFβ signaling to modulate the severity of the fibrotic response of injury-activated CD44+ leader cells.

In the current study, we looked to understand what may account for the wound-environment-specific differences in the endogenous TGFβ-mediated fibrotic response of CD44+ cells that rapidly populate the wound edge following cataract surgery injury. We observed a robust endogenous TGFβ-induced fibrotic response by leader cells in a rigid wound microenvironment, compared to the pro-reparative response of these cells in the wound environment of the lens basement membrane capsule. We speculated that one reason for these distinct leader cell responses at the leading wound edge could be due to differences in the ability of these leader cells to sustain activation of TGFβ on the lens capsule basement membrane vs. the rigid wound environment that surrounds the cataract surgery explant. To overcome any issues that prevent these leader cells to sustain activation of TGFβ, an exogenous source of TGFβ was supplied to the ex vivo post-cataract-surgery cultures. Our findings revealed an inherent ability of leader cells at the leading edge of the cataract surgery wound to respond to TGFβ and produce a fibrotic outcome on the lens capsule. These findings suggest that there is a limited amount of active endogenous TGFβ available to the leader cells in the microenvironment of the native lens capsule basement membrane.

The following factors may help explain why leader cells in the rigid wound environment could sustain activation of endogenous TGFβ that is required to produce a fibrotic response better than when they are present at the leading edge of the cataract surgery wound on the native basement membrane capsule. One is microenvironment-specific differences in leader cell-surface expression of the TGFβ activator αVβ3 integrin. In contrast to leader cells in the lens-capsule-basement-membrane microenvironment (CMZ), αVβ3 integrin localized to the cell-surface of leader cells in the rigid microenvironment of the ECZ. In the ECZ, cell surface expression αVβ3 integrin was specific to the cells at the leading edge of the wound and not detected on the lens epithelial cells that follow them across the rigid wound environment. αV integrins are critical to the activation of TGFβ (Munger et al., 1999; Worthington et al., 2011; Hinz, 2013; Lodyga and Hinz, 2020) and shown to be required for the activation of TGFβ signaling and the fibrotic response to injury in a mouse-cataract-surgery lens model (Mamuya et al., 2014). Rigidity or stiffness of the wound microenvironment is another important factor. The wound environment of the tissue culture substrate of the ECZ is profoundly stiffer than the lens capsule basement membrane. It is well known that rigidity can impact mechanical signaling and exert a strong influence over induction of a fibrotic response (Wells, 2013; Handorf et al., 2015; Tschumperlin and Lagares, 2020). Moreover, the rigidity of the cell’s environment can also affect cell contractility, a crucial aspect of the αV-integrin-mediated mechanism that drives activation of latent TGFβ (Wipff et al., 2007; Wells and Discher, 2008; Wipff and Hinz, 2008; Hinz, 2009; Buscemi et al., 2011; Marinkovic et al., 2012; Hinz, 2015). Our previous studies show wound-environment-specific differences in the actin organization of leader cells that can directly impact leader cell contractility. Leader cells located on the lens capsule basement membrane have low amounts of F-actin compared to leader cells in the rigid wound environment of the ECZ where they exhibit prominent F-actin stress fibers (Menko et al., 2014; Walker et al., 2018) that equip these leader cells with a greater ability for actomyosin contraction. Furthermore, the deposition of a FN-EDA-rich matrix beneath leader cells, located specifically within the rigid wound environment and with a propensity to bind latent TGFβ complexes (Klingberg et al., 2018), provides a local reservoir of latent TGFβ that can be readily activated in a spatiotemporally specific manner by the leader cells. Collectively, these properties provide CD44+ leader cells in the rigid wound environment of the ECZ with an advantage over those on the basement membrane to activate and sustain TGFβ signaling in response to lens injury.

The specific expression of CD44 by leader cells in combination with their location within the rigid wound environment of the ECZ may also contribute to the ability of these cells to activate and sustain TGFβ signaling specifically at these sites of injury. CD44 is expressed on leader cells in both wound environments, however the function of CD44 on cells sensing the rigid wound environment is likely distinct and may provide these cells with a greater potential to activate and modulate TGFβ signaling. This hypothesis is supported by findings that show that CD44 is implicated in sensing changes in ECM stiffness (Razinia et al., 2017) and that CD44 can serve as a docking receptor for matrix metallopeptidase 9 to proteolytically activate latent TGFβ (Yu and Stamenkovic, 1999, 2000). Also, CD44-deficient mice had impaired activation of TGFβ following lung injury (Teder et al., 2002) and CD44KO fibroblasts have defects in TGFβ activation (Acharya et al., 2008), further linking TGFβ activation to CD44 function. It is also known that CD44 can act as a co-receptor for TGFβR (Bourguignon et al., 2002), another way that CD44 could influence TGFβ signaling. There is also the possibility that CD44 and integrins cooperate to drive sustained activation of TGFβ to drive a robust fibrotic response in the rigid wound microenvironment. Future studies are needed to dissect the role of CD44 in modulating the fibrotic response of leader cells to lens injury.

Another factor that could influence the availability of endogenous TGFβ in the wound environment and the fibrotic response is the extent of injury and the degree of damage to the basement membrane caused by injury to the lens. In the cornea, the degree of injury can lead to two different outcomes to wounding. For instance, corneal wounding that leaves the basement membrane intact results in temporary fibrotic hazing that resolves (Zieske et al., 2001; Stramer et al., 2003; Torricelli et al., 2013; Stepp et al., 2014). However, penetrating wounds of the corneal epithelium that disrupt the basement membrane and include the connective tissue underlying the epithelium result in a fibrotic response (Stramer et al., 2003; Torricelli et al., 2013; Stepp et al., 2014). It is believed that the basement membrane provides protective properties to limit the TGFβ response, while in penetrating wounds where this barrier is missing, cells in the stroma can now be exposed to sustained amounts of TGFβ and produce a fibrotic response (Wilson et al., 2017; Wilson et al., 2020). The mock cataract surgery explants used in our study involved the gentle removal of the lens fiber cells, leaving unperturbed the laminin-rich lens-capsule basement membrane across which the cells migrate along to close the wound closure. In this environment there is a reparative wound response that occurs in the absence of a major fibrotic response to the cataract surgery injury. On the other hand, more damage occurs at the cut outside edges of the lens tissue created to produce a flattened explant where leader cells direct the wounded epithelium to move off the lens capsule across the surrounding rigid environment of the ECZ. In this microenvironment the leader cells mount a robust TGFβ-mediated fibrotic response, supporting the notion that the extent of injury sustained by the lens and its basement membrane can influence the cell’s TGFβ-mediated fibrotic response to wounding.

In further support of this concept, our studies also show that the local wound environment of leader cells exhibits control over shaping their TGFβ-mediated fibrotic response to injury even after providing cells with exogenous TGFβ to overcome any deficits in the ability of leader cells to sustain activation of TGFβ. Exogenous TGFβ led to an accelerated and enhanced fibrotic response in the rigid wound environment that resulted in the earlier appearance of pro-collagen-producing cells and transition of CD44+ leader cell to an αSMA+ myofibroblast fate, and a more extensive FN fibrillar matrix. In contrast, on the posterior lens capsule the exposure to exogenous TGFβ induced the appearance of αSMA+ stress-fiber-forming myofibroblasts by culture day 9. The intact lens-capsule basement membrane may offer protection to the lens to limit the TGFβ-mediated fibrotic response, which may help explain differences in the rate and severity of the fibrotic reactions by leader cells seen in these two distinct wound environments. Given the extensive nature of the fibrotic response in the rigid environment, future studies are needed to better understand the contribution of both the CD44+ leader cells and lens epithelial cells to creating this injury-induced response to exogenous TGFβ. Furthermore, consistent with studies using experimental models of fibrosis in the murine lung (Sun et al., 2016; Tsukui et al., 2020), we found that αSMA is an inconsistent marker of collagen-producing cells. Our studies identified both αSMA+ and αSMA- collagen I producing cells. As both these populations can contribute to fibrotic disease future studies are needed to better understand their individual contributions to lens fibrotic disease progression. Intriguingly, following exposure to exogenous TGFβ, fibrotic reactions occurred predominantly in the central wound area of the posterior lens capsule, where PCO develops in post-cataract-surgery patients, these findings indicate a cell- and location-specific response to TGFβ in cataract-surgery wounding.

A strong foundation for our current study lies in findings from many labs that have established the importance of TGFβ as a major factor in promoting fibrotic disease in many tissues, including the lens (Hales et al., 1994; Liu et al., 1994; Srinivasan et al., 1998; Lee and Joo, 1999; Lovicu et al., 2002; Wormstone et al., 2002; de Iongh et al., 2005; Banh et al., 2006; Symonds et al., 2006; Wormstone et al., 2006; Robertson et al., 2007; Saika et al., 2009; Chang and Petrash, 2015; Korol et al., 2016; Raghavan et al., 2016; Taiyab et al., 2016; Boswell et al., 2017; Wang et al., 2017; Kubo et al., 2018; Shu et al., 2019; Taiyab et al., 2019). The findings from this study provide insight into how the local milieu exhibits spatiotemporal control over endogenous TGFβ activity to influence the abundance and extent of active TGFβ available to leader cells in the ex vivo post-cataract surgery cultures. Furthermore, the amount of active TGFβ available within the environment appears to influence the severity of the fibrotic response by leader cells. These findings may help shed light on why only certain (20–30%) adult cataract-surgery patients develop the lens fibrotic disease PCO following cataract surgery, which may relate to the degree of injury to the local microenvironment caused by cataract-surgery wounding to affect the TGFβ-mediated fibrotic response of cells. The ex vivo post-cataract surgery wound healing/fibrosis chick model may reflect a more aggressive response of cells to cataract surgery injury similar to pediatric PCO compared to adult PCO. We propose that the ex vivo post-cataract surgery model is an ideal system to tease apart the molecular constituents of the wound microenvironment that modulate TGFβ signaling and to study the contribution of the injury-activated CD44+cells and wounded-lens epithelial cells to controlling the fibrotic response to lens injury. Such studies have the potential to lead to better understanding about TGFβ’s context-dependent signaling responses to injury and provide new avenues to pursue that may lead to better strategies to target TGFβ signaling to prevent a fibrotic outcome to wounding, such as the development of PCO.

Highlights.

• Endogenous TGFβ signals a leader cell fibrotic response in the rigid wound milieu.

• TGFβ is dispensable for the wound healing response to cataract surgery injury.

• Exogenous TGFβ induced a fibrotic response by leader cells on the lens capsule.

• In the rigid wound milieu, exogenous TGFβ led to an augmented fibrotic response.

• The local wound milieu regulates the extent of the TGFβ-mediated fibrotic response.

Abbreviations:

αSMA

alpha smooth muscle actin

CMZ

central migration zone

ECZ

extracapsular zone

ECM

extracellular matrix

FN

fibronectin

FN-EDA

Fibronectin EDA

PCO

Posterior Capsule Opacification

TN-C

tenascin-C

TGFβ

transforming growth factor beta