αVβ8 integrin targeting to prevent posterior capsular opacification

Fibrotic posterior capsular opacification (PCO), a major complication of cataract surgery, is driven by transforming growth factor–β (TGF-β). Previously, αV integrins were found to be critical for the onset of TGF-β–mediated PCO in vivo; however, the functional heterodimer was unknown. Here, β8 integrin–conditional knockout (β8ITG-cKO) lens epithelial cells (LCs) attenuated their fibrotic responses, while both β5 and β6 integrin–null LCs underwent fibrotic changes similar to WT at 5 days post cataract surgery (PCS). RNA-Seq revealed that β8ITG-cKO LCs attenuated their upregulation of integrins and their ligands, as well as known targets of TGF-β–induced signaling, at 24 hours PCS. Treatment of β8ITG-cKO eyes with active TGF-β1 at the time of surgery rescued the fibrotic response. Treatment of WT mice with an anti-αVβ8 integrin function blocking antibody at the time of surgery ameliorated both canonical TGF-β signaling and LC fibrotic response PCS, and treatment at 5 days PCS, after surgically induced fibrotic responses were established, largely reversed this fibrotic response. These data suggest that αVβ8 integrin is a major regulator of TGF-β activation by LCs PCS and that therapeutics targeting αVβ8 integrin could be effective for fibrotic PCO prevention and treatment.


Introduction
Cataracts, a major cause of blindness (1), are treated by surgical removal of opaque lens cells followed by implantation of an artificial intraocular lens (IOL) (1). However, months to years later, a significant proportion of patients experience an apparent recurrence of their cataract as posterior capsular opacification (PCO) (2,3). PCO occurs when the remnant lens epithelial cells (LCs) left behind post cataract surgery (PCS) migrate into the optical axis and transition into a mixture of myofibroblasts and aberrant lens fiber cells (3). Approximately 25% of adult human and veterinary patients, and almost 100% of pediatric patients who do not receive prophylactic posterior capsulotomy, develop clinically significant PCO within months to years PCS (4). PCO in adults is treated by neodymium:YAG (Nd:YAG) laser capsulotomy (2), although this is often unsuitable/inconvenient for pediatric and veterinary therapy (4). As Nd:YAG laser capsulotomy can also cause side effects, including macular edema and retinal detachment, PCO prevention is desirable (2,4,5). Currently, the only US FDA-approved approach to prevent PCO utilizes prosthetic IOLs, which sequester remnant LCs to the capsular bag periphery, an innovation that delays, but often does not prevent, PCO (4).
Transforming growth factor-β (TGF-β) signaling mediates the epithelial-mesenchymal transition (EMT) of LCs to myofibroblasts (6). While TGF-β concentrations in the aqueous humor are high before surgery, most of this TGF-β is inactive (7). Using a mouse cataract surgery model, we previously demonstrated that canonical TGF-β signaling is not easily detected in LCs until 48 hours PCS, with robust activation initiating at 3 days PCS (8). However, the mechanisms by which cataract surgery results in TGF-β signaling activation are unknown.
Integrins, heterodimeric extracellular matrix (ECM) receptors consisting of 1 α and 1 β subunit, mediate cell/ECM attachment, cell migration, and force transmission (9). Integrins also crosstalk with growth factor signaling (10), including the TGF-β pathway (11,12). Thus, integrins are potential therapeutic targets for PCO prevention and/or treatment (9). Previously, we found that α V integrins are critical for fibrotic PCO (13) consistent with their known roles in latent TGF-β activation (14,15). Notably, the α V integrin subunit heterodimerizes with a variety of β integrins (16), 4 of which (β 1 , β 5 , β 6 , and β 8 ) upregulate in LCs PCS with dynamics similar to α V (13). Since each α V integrin heterodimer binds different ligands, and is inhibited by different compounds (17), the identification of the β subunit that pairs with α V to drive PCO is critical to both the Fibrotic posterior capsular opacification (PCO), a major complication of cataract surgery, is driven by transforming growth factor-β (TGF-β). Previously, α V integrins were found to be critical for the onset of TGF-β-mediated PCO in vivo; however, the functional heterodimer was unknown.
JCI Insight 2021;6(21):e145715 https://doi.org/10.1172/jci.insight.145715 some LCs differentiate into structurally aberrant lens fiber cells during PCO pathogenesis, contributing to "pearl-like" PCO when in the visual axis and Soemmering's ring when restricted to the capsular bag periphery (3). Remnant LCs from WT and β 8 ITG-cKO mice expressed little aquaporin 0 ( Figure 4, A and F) immediately PCS. By 48 hours PCS, WT and β 8 ITG-cKO capsular bags were associated with some aquaporin 0-expressing cells and robustly so by 5 days PCS (Figure 4, A and F; WT P ≤ 0.001; β 8 ITG-cKO P = 0.003), indicating that β 8 ITG-cKO LCs can differentiate into lens fiber-like cells PCS to a similar extent as WT.
Notably, fewer cells were associated with β 8 ITG-cKO capsular bags than WT at 5 days PCS (P ≤ 0.001) ( Figure 4G). Because apoptosis was not detected in the capsular bags of either WT or β 8 ITG-cKO mice at any time PCS (data not shown), LC proliferation was investigated by following the expression of Ki67, a marker of all cell cycle stages except G 0 (24). At 0 hours PCS, few to no remnant LCs were proliferating ( RNA-Seq reveals genes associated with fibrosis and inflammation are differentially expressed in β 8 ITG-cKO LCs PCS. Because phenotypic differences between WT and β 8 ITG-cKO LCs manifest by 48 hours PCS, we performed RNA-Seq on WT and β 8 ITG-cKO LCs at 24 hours PCS to gain insight into this phenotype's proximal cause. This revealed 2312 genes differentially expressed in WT LCs at 24 hours PCS compared with 0 hours PCS (1273 genes upregulated, 1039 genes downregulated under criteria for biologically significant differences; ref. 25). As we previously reported (8,26), the upregulated genes included those that participate in tissue inflammation (Supplemental Table 4) and fibrosis (Supplemental Table 5), while many genes important for lens structure and function were downregulated (Supplemental Table 6).
Comparison between WT and β 8 ITG-cKO LCs at 24 hours PCS revealed that 828 genes were differentially expressed under the biological significance criteria (25) (Supplemental Figure 3). Of these, 97 were upregulated in WT LCs by 24 hours PCS but not in β 8 ITG-cKO LCs (Supplemental Table 7). Consistent with β 8 ITG-cKO LCs' muted fibrotic response PCS, several of these were associated with fibrotic disease while others regulate inflammation (Supplemental Table 8).
Because active TGF-β induces LC conversion to myofibroblasts (6), and β 8 ITG-cKO LCs exhibit reduced TGF-β signaling PCS, we tested whether exogenous active TGF-β could rescue these defects (  Blocking the interaction of TGF-β latency associated peptide with α V β 8 integrin in WT LCs phenocopies the attenuated fibrotic response and TGF-β signaling defects detected in β 8 ITG-cKO LCs PCS. TGF-β is secreted from cells bound to its latency associated peptide (LAP) and latent binding proteins, forming the latent TGF-β complex (14). Upon secretion, the latent TGF-β complex is tethered to the ECM by binding to matrix proteins such as fibronectin (14,23). The release of active TGF-β1 from the latent complex can be accomplished by the interaction of the LAP with integrins such as α V β 8 (12,14). Thus, we next tested whether α V β 8 integrin function-blocking antibody (ADWA-11), which antagonizes LAP binding to α V β 8 integrin (α V β 8 -IBA), thus blocking TGF-β activation (34), can influence the fibrotic response of LCs ( Figure 6, A-F). Systemic treatment of WT mice at surgery with α V β 8 -IBA inhibited canonical TGF-β signaling  were similar to that observed in β 8 ITG-cKO LCs ( Figure 6, A-F), α V β 8 -IBA likely blocks TGF-β activation and subsequent fibrotic response of LCs PCS.
Because 3 experimental approaches revealed that TGF-β activation by α V β 8 integrin is a core mechanism of PCO development, next we characterized the regulatory relationship between α V β 8 integrin-mediated TGF-β activation and 2 other PCO regulators, gremlin-1 and ECM binding integrins.
Induction of α V β 8 integrin is required for LCs to upregulate gremlin-1 levels PCS, though gremlin-1 does not rescue induction of canonical TGF-β signaling and fibrotic gene expression in β 8 ITG-cKO LCs. Previously, we reported that gremlin-1, best known as a BMP antagonist (35), upregulates in LCs by 48 hours PCS and rescues the defects in sustained canonical TGF-β signaling observed in LCs lacking the fibronectin gene (23). Here, RNA-Seq revealed that Grem1 mRNA levels were upregulated 170-fold in WT LCs at 24 hours PCS yet attenuated 3-fold in β 8 ITG-cKO LCs (Supplemental Tables 5 and 8 and Figure 4A). Consistent with these data, gremlin-1 protein levels were sharply upregulated in LCs associated with WT capsular bags by 3 days PCS (Supplemental Gremlin-1 is an agonist of the canonical TGF-β pathway (36,37) and rescues the defect in canonical TGF-β signaling observed in PCS LCs lacking the fibronectin gene (23). However, treating β 8 ITG-cKO mice with recombinant gremlin-1 at cataract surgery did not rescue LC fibrosis (Supplemental Figure 5, A-F). This suggests that gremlin-1-induced LC fibrosis may require the autocrine activation of TGF-β signaling, consistent with studies in other cell types (38,39).
Upregulation of integrin expression and signaling by LCs depends on α V β 8 integrin-mediated TGF-β signaling. Crosstalk between integrins and TGF-β signaling is well documented (12). Notably, LCs elevate the protein levels of α 5 β 1 integrin and several α V integrins in response to lens injury or TGF-β treatment ( Figure 9) (9, 13, 23). Either deletion of the β 8 integrin gene from LCs or treatment of capsular bags with a function blocking antibody against α V β 8 integrin prevents the upregulation of α 5 , β 1 , and α V integrin expression and increased p-FAK, a readout of integrin signaling PCS, at 3 and 5 days PCS (Figures 9 and 10). The addition of active TGF-β1 to β 8 ITG-cKO capsular bags rescued the attenuated integrin expression and p-FAK signaling detected in β 8 ITG-cKO LCs ( Figure 10) 5 days PCS compared with vehicle-treated β 8 ITG-cKO LCs. Our findings indicate that α V β 8 integrin is essential for upregulation of TGF-β signaling in LCs PCS, which drives subsequent upregulation of integrin expression and FAK signaling.

Discussion
Fibrosis-mediated organ damage and failure are among the major causes of natural death worldwide because no effective therapies prevent or treat fibrosis (40). While TGF-β signaling often drives tissue fibrosis (6,14), this pathway is difficult to target due to its complex regulation and diverse roles in normal biology (12). Integrins regulate the TGF-β pathway via their roles in latent TGF-β activation, and ability to mediate TGF-β effects, as integrin expression is often regulated by TGF-β signaling (12,15,41). Thus, integrins are promising therapeutic targets for organ fibrosis, and several integrin blocking agents are undergoing clinical trials (17,42). The α V integrins are particularly promising targets for antifibrotic therapies because blocking this class of integrins can ameliorate fibrosis in several organs (43,44).
Previously, we reported that α V integrin gene deletion from the lens prevents EMT of LCs, potentially due to their inability to initiate TGF-β signaling PCS (13). However, the identity of the β integrin subunit participating with α V integrin was not known as multiple β subunits capable of heterodimerizing with the α V integrin subunit are upregulated by LCs PCS (13). We could not study the role of α V β 1 integrin in the LC response to injury in this study despite its known roles in wound healing (45), as we previously found that the Itgb1 gene is essential for lens development and homeostasis, leading adult mice lacking β 1 integrin expression in the lens to be severely microphthalmic/anophthalmic (46)(47)(48). However, the deletion of neither β 5 , β 6 , nor β 8 integrin JCI Insight 2021;6(21):e145715 https://doi.org/10.1172/jci.insight.145715 from the lens resulted in obvious lens defects, which made it possible to characterize their role in regulating the LC response to cataract surgery in vivo.
As previously reported, mice homozygous for β 5 or β 6 integrin deletions are viable (49,50), while this investigation found that LCs from these mice underwent normal fibrotic responses PCS, indicating that neither α V β 5 nor α V β 6 integrins are critical for fibrotic PCO. While these results were initially surprising as these integrins can participate in latent TGF-β activation (18,51), α V β 5 and α V β 6 integrins' roles in fibrotic disease are tissue and insult specific (43,49,52). Thus, we investigated α V β 8 integrin as (a) its expression rapidly upregulates in mouse LCs PCS and it is found on fibrotic human LCs at extended times PCS; (b) it can regulate tissue fibrosis and inflammation via binding to the RGD sequence present in the LAP of TGF-β1 and TGF-β3 (53), which activates these latent complexes via either formation of a ternary complex with membrane-type matrix metalloproteinase 1 (27), which is upregulated in LCs PCS, or traction-mediated activation (14); and (c) TGF-β1 and TGF-β3 mRNA levels are both upregulated in LCs PCS (8,23).
Integrin α V β 8 adhesion to the LAP of latent TGF-β complexes. TGF-β activation by α V β 8 integrin plays roles in development, fibrosis, inflammation, and wound closure (14,15). Notably, itgav deletion from the lens impedes TGF-β signaling and fibrotic responses PCS (13), and here we show that deletion of itgb8 from the lens (β 8 ITG-cKO) phenocopies this result. As the addition of active TGF-β1 rescued the attenuated TGF-β signaling and fibrotic responses observed in β 8 ITG-cKO capsular bags PCS, this suggests that α V β 8 integrin-mediated activation of latent TGF-β is critical for the development of fibrotic PCO. This hypothesis is supported by the observation that treatment of WT mice with ADWA-11, which specifically inhibits the adhesion of LAP to α V β 8 integrin (34), potently inhibits the activation of TGF-β signaling and subsequent fibrotic responses of LCs to lens fiber cell removal.
α V β 8 integrin regulates both the EMT and LC proliferation in response to lens injury. This study explored 3 potential mechanisms by which the Itgb8 gene deletion from the lens and blocking of α V β 8 integrin interaction with LAP inhibit the upregulation of proteins expressed by myofibroblasts and TGF-β signaling PCS. As LC apoptosis was apparently absent in both WT or β 8 ITG-cKO LCs PCS, the very large decrease in LC proliferation at 2 days and 3 days, and cell number at 5 days, PCS in β 8 ITG-cKO and WT (α V β 8 -IBA) in concert with the reduction of fibrotic marker mRNAs in β 8 ITG-cKO LCs at 24 hours PCS, and reductions in fibrotic marker protein expression later PCS, suggest that the α V β 8 integrin/TGF-β signaling axis regulates both EMT and proliferation of LCs PCS, leading to fibrotic PCO. α V β 8 integrin crosstalk with other integrins, and their signaling, in PCO. Feedforward mechanisms between α V integrins and TGF-β signaling have been previously described (12). Upon activation by α V integrins, the TGF-β homodimer binds to the type II TGF-β receptor to initiate Smad2/3 phosphorylation, leading to increased expression of α V integrins and other fibrotic markers. These newly formed integrins can liberate more TGF-β from latent complexes, sustaining and reinforcing TGF-β-induced fibrosis (12). Indeed, LCs lacking α V β 8 integrin attenuate the upregulation of α V , α 5 , and β 1 integrin expression and FAK phosphorylation PCS while treatment with active TGF-β1 reverses these defects. These findings have 2 implications: 1) targeting α V β 8 integrin could suffice to prevent TGF-β activation in PCO; 2) the resulting attenuation of fibronectin fibril deposition (likely mediated by α 5 β 1 integrin) may contribute to the long-term prevention of fibrotic PCO, as we have previously reported that fibronectin assembly is required for sustained LC fibrosis PCS (23).
α V β 8 integrin as a player in the gene regulatory network driving PCO. Lens injury/cataract surgery results in the EMT of LCs to myofibroblasts that proliferate, migrate, contract the lens capsule, and produce a fibrotic matrix, which all contribute to the degradation of patient vision PCS (3). Most studies of this process start by considering the role of active TGF-β in this phenotypic conversion, though less attention has been paid to how cataract surgery initiates this process (8). Here we report that upregulation of α V β 8 integrin levels on LCs PCS is an important step in their reprogramming into myofibroblasts and suggest that this reprogramming starts 1-2 days before upregulation of detectable canonical TGF-β signaling in these cells, as the upregulation of the mRNAs encoding numerous profibrotic proteins was attenuated in β 8 ITG-cKO LCs 24 hours PCS. As many of these genes have been previously reported to be TGF-β responsive in other cell types (Supplemental Tables 9 and 10), this observation supports our prior model, which proposed that a small upregulation in the protein levels of a TGF-β-activating integrin could lead to a feedforward loop that can rapidly induce autocrine TGF-β-mediated fibrosis (12).
Notably, the profibrotic modulators whose upregulation is attenuated in β 8 ITG-cKO 24 hours PCS included gremlin-1, a BMP signaling antagonist (35), and TGF-β signaling agonist (23,28,36), which usually upregulates sharply by 24 hours PCS. We previously found that LCs lacking the fibronectin gene exhibited greatly attenuated gremlin-1, p-SMAD2/3, and fibrotic marker upregulation PCS, while treatment of capsular bags with exogenous gremlin-1 rescued the ability of fibronectin-null LCs to undergo fibrotic responses (23). However, while both gremlin-1 expression and p-SMAD2/3 signaling are attenuated in β 8 ITG-cKO LCs, treatment of these cells with exogenous gremlin-1 did not rescue either p-SMAD2/3 signaling or the fibrotic response of β 8 ITG-cKO LCs, though exogenous treatment of these cells with active TGF-β1 rescued the fibrotic response. This suggests that gremlin-1 is not eliciting its response directly via the TGF-β receptor and instead may facilitate latent TGF-β activation by an as-yet-unknown mechanism. Alternatively, there could be differences in the requirements for gremlin-1 at different times PCS, with α V β 8 integrin's ability to activate TGF-β signaling at early times PCS kick-starting gremlin-1 expression so that it can influence TGF-β signaling later PCS (23). This concept is supported by work on other cell types suggesting that early activation of endogenous TGF-β is critical for gremlin-1 to exert its profibrotic response later in fibrotic disease (38,39). Further study is required to elucidate the molecular mechanisms and therapeutic potential (if any) of gremlin-1 in PCO.
α V β 8 integrin blockade in halting the onset and progression of fibrotic PCO. The emerging role of α V β 8 integrin in TGF-β1 (and likely TGF-β3) activation has led to intense research into both antibody and small molecule inhibitors of α V β 8 integrin-TGF-β interactions for the treatment (and monitoring) of fibrotic and neoplastic disease (15,16,34,54). In most proof-of-principle experiments for this approach, these drugs prevent the onset of disease in animals, though this would not be clinically efficacious in most cases as fibrotic damage is often extensive by the time clinical symptoms manifest (40). However, treatment before the onset of fibrotic disease could be a viable option for anti-PCO therapy as the initiating insult (cataract surgery) is known, and the site of fibrosis is accessible during surgery, making local administration of the drug feasible. Here we validate this approach in an animal model of PCO as we found that the treatment of mice at the time of surgery with an anti-α V β 8 integrin function blocking antibody prevented the development of the fibrotic sequelae of lens fiber cell removal and thus presumably fibrotic PCO. Our preclinical study supports the idea that this blocking antibody may be useful to halt fibrotic progression in patients who have already developed PCO. However, it should be mentioned that although the levels of p-SMAD3 activation and other fibrotic protein levels reversed in LCs to levels similar to the unoperated lens when mice were treated after fibrosis was established, α-SMA protein levels were still elevated, which may result from its relative stability as the half-life of α-SMA protein in cells is 72 hours (55).
Limitations of this study for clinical translation as a PCO preventative. While data presented here suggest that therapeutics blocking α V β 8 integrin's ability to activate TGF-β have great promise in preventing fibrotic PCO, unanswered questions remain. First, the function blocking antibody therapy used in this study was administered to the animals intravenously as the systemic administration of ADWA-11 showed no evidence of toxicity in prior studies (34,56,57), and a humanized version of ADWA-11 did not elicit notable systemic toxicity in mice and Cynomolgus monkeys treated for over 1 month at doses more than 5 times that used here (56,57). However, attempts to directly inject the ADWA-11 into the mouse capsular bag at surgery did not block the fibrotic transformation of LCs. This may result from the rapid turnover of aqueous humor and the limited amount of antibody we were able to administer into the anterior chamber, leading the local concentration of antibody to quickly drop below the therapeutic dose (58,59). In the future, we envision that function blocking drugs against α V β 8 integrin could be administered locally in the eye at surgery either as a slow-release suspension (60) added to a dropless cataract surgery preparation (61) or by coating or soaking the IOL (62), which would reduce drug costs and opportunity for systemic side effects. Second, the invasive nature of the mouse "cataract surgery" model used in this study makes it difficult to assess drug effects on other ocular structures. Additional studies in rabbit and other animal models of cataract surgery that are more similar to what is performed in humans are necessary to assess ocular toxicity explicitly. Finally, while α V β 8 integrin blockade appeared to reverse fibrotic ECM deposition associated with mouse capsular bags at 5 days PCS, we expect this matrix to still be relatively immature at this time and thus relatively susceptible to turnover. Future work will be needed to determine whether anti-α v β 8 integrin therapeutics can reverse fibrosis in more established fibrotic conditions where the scar tissue has developed abundant amounts of cross-linked collagen.
Summary. This study established that α V β 8 integrin is essential for LCs to transition to myofibroblasts following lens injury, likely through its ability to activate latent TGF-β1 and/or TGF-β3. Blocking α V β 8 integrin binding to ligands via antibody blockade phenocopied the response of β 8 ITG-cKO LCs to lens fiber cell removal, establishing α V β 8 integrin as a potentially novel therapeutic target to prevent PCO. PCO is a prevalent complication of cataract surgery (1,3), especially in children. While there are no FDA-approved pharmacological agents available to prevent PCO, this preclinical study suggests inhibition of α V β 8 integrin is a promising approach to PCO prevention. Further, the reversal of LC fibrosis by α V β 8 integrin blockade suggests that therapeutics targeting α V β 8 integrin have the potential to not just arrest the progression of fibrotic disease but also even reverse it.

Methods
Animals. All mice were maintained under pathogen-free conditions at the University of Delaware animal facility under a 14-hour light/10-hour dark cycle. Animals of both sexes were used in these experiments, and no sex-dependent effects were noted, consistent with our prior report (63).
Human eyes. Transparent lenses (30 ± 2 years of age) were obtained from Lions Eye Bank of Oregon (Portland, Oregon, USA), and aphakic donor eyes were obtained from the Minnesota Lions Eye Bank (Minneapolis, Minnesota, USA) as part of their cadaver eye tissue procurement programs. Intact lenses or lens capsular bag/ IOL implant complexes were isolated, fresh frozen in OCT medium, and prepared for immunofluorescence experiments as described below.
Genotyping and PCR. DNA was isolated from tail snips or whole lenses using the PureGene Tissue and Mouse Tail kit (Gentra Systems) as described (13) and genotyped by PCR using primers described in Supplemental Table 1 (64,65). The deletion of exon 4 of the integrin β 8 gene from the lens was confirmed by PCR analysis of genomic DNA isolated from adult lenses using the primers described in Supplemental Table 1 (65).
Morphological analysis. Lens clarity was determined by viewing isolated lenses using darkfield optics while lens optical properties were assessed by placing lenses on a 200-mesh electron microscopy grid as described previously (66).
Mouse cataract surgery model. Surgical removal of lens fiber cells to mimic human cataract surgery was performed in adult mice as previously described (13,67). Briefly, adult mice were anesthetized, a central corneal incision was made, and the entire lens fiber cell mass removed by a sharp forceps, leaving behind an intact lens capsule. Mice were sacrificed for analysis at time intervals ranging from 24 hours to 10 days PCS (13). RNA-Seq and bioinformatics. Samples from WT (C57BL/6) and β 8 ITG-cKO lenses subjected to cataract surgery (3 biological replicates for each condition, 5 capsules per replicate) were harvested at 0 hours and 24 hours PCS and frozen on dry ice, and RNA was harvested using RNeasy Mini Kit (50) from QIAGEN (catalog 74104) (23). RNA libraries were prepared using the SMARTer Stranded Total RNA-Seq Kit-Pico Input Mammalian (Takara Bio USA, Inc.) and sequenced by DNA Link, USA, on a NovaSeq 6000 (Illumina). Read pairs corresponding to RNA fragments were enumerated as FPKM by Cuffdiff. Biologically significant differentially expressed genes (DEGs) were defined as those exhibiting statistically significant changes (FDR ≤ 0.05), a change in mRNA level greater than 2 FPKM between conditions, fold change greater than 2 in either the positive or negative direction, and expression levels in either condition that were 2 FPKM or greater (8,23,25). Heatmaps were generated using the Morpheus tool (68). RNA-Seq data were submitted to the National Center for Biotechnology Information's Gene Expression Omnibus under accession number GSE145492.
A list of 1603 TGF-β-responsive genes in cultured cells (32) was filtered to consider only those statistically significant (FDR ≤ 0.05) genes with consistent directions of change between treatments, leaving 1390 genes as likely to be directly TGF-β responsive. This filtered list was then compared with the list of genes differentially expressed in WT LCs between 0 and 24 hours PCS and between WT and β 8 ITG-cKO LCs 24 hours PCS (FDR ≤ 0.05) to discover what proportion of the DEG list had the potential to be influenced by TGF-β signaling the first day PCS.
Tail vein injection of α V β 5 , α V β 6 , and α V β 8 integrin function blocking antibodies. Tail vein injection of α V β 5 , α V β 6 , and α V β 8 integrin function blocking antibodies was performed as described (69). Briefly, a single dose of ALULA (α V β 5 integrin blocking) (70), 3G9 (α V β 6 integrin blocking) (71), or ADWA-11 (α V β 8 integrin blocking) (34) was administered to WT mice at 20 mg/kg in PBS via lateral tail vein injection immediately following removal of the lens fiber cells from 1 eye. In another experiment, either a single dose of ADWA-11 (at 5 days PCS) or 2 doses of ADWA-11 (at 5 days PCS and 7.5 days PCS) was administered to WT mice. Control animals were treated with an isotype-matched antibody (anti-human α V β 3 integrin that does not cross-react with the mouse α V β 3 integrin protein) at 20 mg/kg in PBS. All the integrin function blocking antibodies were obtained from UCSF.
Immunofluorescence. The details of sample preparation and immunofluorescence were described previously (23,72). Supplemental Table 2 describes the primary antibodies, blocking buffer compositions, incubation times, and dilutions used in this study, and Supplemental Table 3 lists the secondary antibodies and DNA dyes used in this study. Each experiment/time point was replicated using at least 3 biologically independent specimens (3-5 mice, at least 2 sections per mouse). Fluorescently labeled slides were visualized using Zeiss LSM780 or Zeiss LSM880 confocal microscopes (Carl Zeiss Inc.), and comparisons of images were made between slides imaged using identical imaging parameters (73). In some cases, the brightness and contrast were adjusted to allow viewing on diverse computer screens; however, these adjustments were made identically for all images within a particular time course. Negative controls were prepared and imaged to exclude nonspecific staining by the secondary antibodies or channel bleed-through as previously described (23,72,73).
ImageJ (NIH) quantification and statistics. Immunofluorescence images were quantified by determining the MFI of lens capsule-associated tissue viewed in 3 randomly chosen confocal images from biologically independent samples using ImageJ (v1.52P, NIH) (73). The average number of lens capsule-associated nuclei/ section was analyzed by ImageJ using 6 randomly chosen immunofluorescence images from each PCS time point from at least 3 biologically independent samples as described (73,74).
The diameter of adult lenses was determined by dissecting both lenses from 3 WT and 3 β 8 ITG-cKO mice and photographing them in brightfield using a Zeiss STEMI SV 11 dissecting microscope. The diameter of each lens was measured in 2 perpendicular axes using ImageJ, then averaged for statistical analysis.
All statistics were assessed using either 2-tailed Student's t test (corrected for multiple comparisons using the Holm-Šídák method) or 1-way/2-way ANOVA with Tukey's post hoc test using GraphPad Prism 8.3.0/9.2.0. Data are presented as mean ± SEM, and differences were considered significant at P ≤ 0.05. by the University of Delaware Institutional Animal Care and Use Committee. Experiments using human cadaver-derived lens tissue were reviewed by the University of Delaware Institutional Review Board and were judged as exempt from review.

Author contributions
MHS designed research studies, conducted experiments, analyzed data, and wrote the manuscript; SGN conducted experiments and analyzed data; YW designed research studies and conducted experiments; DS provided reagents and designed studies; AA provided reagents and designed studies; TDA provided reagents and designed studies; NMR conducted experiments; APF analyzed data and designed studies; and MKD designed research studies, analyzed data, and wrote the manuscript.