Intracapsular metformin bioactivity and its role in posterior capsule opacification

Abstract

Background/aims

Posterior capsule opacification (PCO) is the most frequent long-term complication after cataract surgery, caused by proliferation of residual lens epithelial cells (LECs). Metformin, a common antidiabetic drug with antifibrotic properties, may act as a pharmacological modulator of this process. The purpose was to assess whether systemically administered metformin reaches the human lens capsule and inhibits LEC proliferation at physiologically relevant concentrations.

Methods

Metformin concentrations were quantified in serum and lens capsule tissue of patients with type 2 diabetes mellitus undergoing cataract surgery using high performance liquid chromatography-tandem mass spectrometry. Patients were stratified into low-dose (≤1000 mg/day) and high-dose (>1000 mg/day) groups. Additionally, anterior lens capsules from non-diabetic donors were cultured with or without metformin (0.75 pg/µL), and LEC proliferation was monitored for 14 days by live-cell imaging.

Results

Metformin was detectable in both serum and capsule specimens, with a significant correlation between compartments (r=0.553, p=0.011). Capsule concentrations were similar across dose groups, while serum levels were higher in the low-dose group. In vitro, metformin significantly reduced LEC proliferation compared with controls (p<0.001).

Conclusion

Systemically administered metformin reaches the human lens capsule in vivo and suppresses LEC proliferation in vitro at clinically relevant concentrations. These findings provide first clinical and experimental evidence supporting metformin as a potential pharmacological adjunct to current strategies for PCO prevention.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Posterior capsule opacification (PCO) is the most common long-term complication after cataract surgery. Metformin has antifibrotic and antiproliferative properties, but its presence and activity in the human lens capsule were unknown.

WHAT THIS STUDY ADDS

• We show that systemic metformin reaches the human lens capsule and, at these concentrations, significantly reduces lens epithelial cell proliferation in vitro.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Metformin may contribute to pharmacological prevention of PCO and supports further translational and clinical studies in this field.

Introduction

Cataract remains a leading cause of blindness worldwide, particularly in ageing populations. In 2020, more than 15 million individuals over the age of 50 were affected globally, with age-related cataract representing the most prevalent subtype.^{1 2} Although pharmacological approaches have been explored,^{3 4} phacoemulsification remains the treatment gold standard.⁵ This surgical technique removes the opacified lens while preserving the capsular bag to hold an intraocular lens (IOL). Driven by demographic change and longer life expectancy, cataract surgeries now exceed 4.7 million annually in the European Union.⁵

Despite its safety and efficacy, cataract surgery is often complicated by posterior capsule opacification (PCO), the most frequent long-term issue. PCO causes visual deterioration in about 20%–30% of patients within 5 years.^6–8 Visually significant PCO is treated with neodymium-doped yttrium aluminum garnet (Nd:YAG) laser capsulotomy, rapid and effective, but not without risks such as elevated intraocular pressure (IOP), cystoid macular edema (CME) or retinal detachment.

The pathophysiology of PCO involves residual lens epithelial cells (LECs) proliferating, migrating and undergoing epithelial-mesenchymal transition (EMT), ultimately leading to fibrotic or regenerative capsular changes.^9–14 Fibrotic variants are characterised by matrix contraction and EMT, while regenerative types manifest as Elschnig pearls.

PCO risk factors include age, ocular inflammation and systemic disease.¹⁵ A particular focus has been on type 2 diabetes mellitus (T2DM). Evidence is inconsistent: some studies suggest diabetes may protect against PCO,¹⁶ others report higher risk with longer duration.^{15 17} These discrepancies may be due to the lack of subgroup analyses with respect to antidiabetic therapy. In particular, metformin use might act as a protective factor, whereas diabetes without metformin treatment could increase PCO risk. The absence of stratification by medication may thus explain the heterogeneous findings. Beyond PCO, diabetes is linked to multiple ocular comorbidities including retinopathy,¹⁸ ocular surface disease¹⁹ and a higher overall risk for cataract formation.²⁰

Metformin is the most widely prescribed oral antidiabetic drug^{21 22} and exerts pleiotropic effects beyond glucose control. It has demonstrated antifibrotic, antiproliferative and anti-senescent properties in various tissues. Preclinical studies indicate that metformin suppresses LEC senescence and proliferation,²³ and recent clinical data link metformin use to reduced rates of Nd:YAG laser capsulotomy.²⁴ Systemically, metformin modulates fibrotic and proliferative responses in hepatic,²⁵ renal, cardiovascular and neoplastic tissues²⁶ primarily via activation of AMP-activated protein kinase (AMPK),²⁷ mechanistic target of rapamycin (mTOR) inhibition and reduced oxidative stress and inflammation.^{28 29}

While metformin has been detected in aqueous humour,³⁰ its presence and concentration in the human lens capsule remain unknown. Moreover, it is unclear whether intraocular metformin levels are sufficient to affect LEC proliferation—an essential driver of PCO pathogenesis. Previous studies have not established a link between systemic metformin exposure and functional intraocular bioactivity, nor used live-cell tracking to directly assess its effect on LEC behaviour.

In addition to systemic and pharmacological factors, IOL design also influences PCO development. A sharp square-edge design acts as a mechanical barrier to LEC migration and remains the current standard in PCO prevention.^{10 31} Hydrophobic acrylic materials are associated with lower PCO rates compared with hydrophilic or polymethylmethacrylate (PMMA) optics, although none completely prevent it.^32–35

To address these gaps, we conducted a prospective single-centre study with two primary aims:

1. To quantify metformin concentrations in serum and lens capsule tissue of patients with T2DM undergoing cataract surgery using high-performance liquid chromatography (HPLC), stratified by dosage.

2. To assess the effect of metformin on human LEC (hLEC) proliferation in vitro, using explanted anterior lens capsules cultured with or without metformin, and monitored by live-cell imaging.

Through this integrative pharmacokinetic and functional approach, we aim to determine whether systemically administered metformin reaches the lens capsule at biologically active concentrations and may influence PCO-related cellular behaviour.

Materials and methods

Patient recruitment and sample collection

Anterior lens capsule specimens were obtained from patients undergoing routine cataract surgery at the Department of Ophthalmology, Hanusch Hospital, Vienna, Austria.

For the concentration analysis, eligible patients had age-related cataract and a confirmed diagnosis of T2DM treated with oral metformin. Inclusion criteria were: age-related cataract, confirmed T2DM, ongoing metformin therapy, age ≥21 years and informed consent before surgery. Exclusion criteria were: type 1 diabetes, previous relevant ocular surgery or trauma and pregnancy (tested in women of reproductive age). Only one eye per patient was included, randomly selected by surgical scheduling. For the concentration analysis, a total of 20 capsules of 20 eyes were included (10 in the low-dose group, 10 in the high-dose group).

To ensure systemic exposure, all patients confirmed metformin intake on the day of surgery. Blood samples were collected in the morning after patients had taken their regular metformin dose. Time of sampling was noted.

Patients were stratified into a low-dose group (≤1000 mg/day) and a high-dose group (>1000 mg/day). Capsulorhexis was performed using a continuous curvilinear technique (5.0–5.5 mm). The central anterior capsule was extracted with forceps, diluted in 20 µL distilled water, supplemented with 80 µL methanol (MeOH), shock-frozen and stored at −80°C until further analyses by liquid chromatography-tandem mass spectrometry (LC-MS/MS).

On the same day, 5 mL of venous blood was drawn prior to surgery. After centrifugation (3000 rpm, 10 min), 20 µL serum was mixed with 80 µL MeOH and stored at −80°C until LC-MS/MS analysis.

For the proliferation experiments, another 20 capsules were initially obtained, of which only 16 could be analysed due to a software malfunction. Only capsules from non-diabetic donors were included. These donors had no diabetes or relevant systemic medication (antidiabetic, immunosuppressive or antiproliferative). All underwent routine age-related cataract surgery; 40 capsules from 40 patients were obtained for both experiments.

Quantification of metformin via LC-MS/MS

Sample preparation

Metformin concentrations in serum and lens capsule samples were quantified at the Vienna Bio Center Core Facilities (VBCF Metabolomics Platform) following a standard LC-MS/MS protocol. Capsule specimens were diluted in 20 µL distilled water and supplemented with 80 µL MeOH, while serum samples were prepared by mixing 20 µL serum with 80 µL MeOH.

Chromatographic conditions and detection

Analyses were performed on an Ultimate 3000 HPLC system (Dionex, Thermo Fisher Scientific) coupled via electrospray ionisation to a TSQ Altis triple quadrupole mass spectrometer (Thermo Fisher Scientific). One microlitre of each extract was injected onto a Kinetex C18 column (100 Å, 150×2.1 mm; Phenomenex) with a guard column. Separation was achieved using an 8 min linear gradient from 100% solvent A (1% acetonitrile, 0.1% formic acid in water) to 80% solvent B (0.1% formic acid in acetonitrile) at a flow rate of 80 µL/min.

Metformin was quantified in positive ion mode using selected reaction monitoring with the transitions m/z 130.1→60.1 (collision energy 13.5 eV) and m/z 130.1→71.1 (collision energy 21.9 eV). For validation, tryptophan was monitored via the transition m/z 205→188. Retention times and collision energies for metformin were optimised using authentic standards.

Calibration and quality control

Quantification was based on external calibration curves prepared from authentic standards and validated by standard addition. Quality control samples at low, medium and high concentrations were included in each analytical batch. Final concentrations were expressed as pg/µL and normalised to extraction volume.

In vitro proliferation assay of hLECs

Capsule culture and treatment

Anterior lens capsules from non-diabetic, non-metformin-treated donors were transferred into RPMI-1640 medium (Roswell Park Memorial Institute) supplemented with 10% fetal calf serum, 100 U/mL penicillin, 0.1 mg/mL streptomycin and 2 mM L-glutamine (all purchased from Thermo Fisher Scientific, Waltham, Massachusetts, USA). Each capsule was placed into a well of a sterile 12 well plate with glass-like polymer bottom (Cellvis, Mountain View, California, USA) containing 2 mL of medium per well. Capsules were allowed to settle by gravity, where they adhered and stabilised once initial proliferation commenced. Capsules were randomly assigned to a control group or to a treatment group with metformin (0.75 pg/µL). Metformin hydrochloride (Sigma-Aldrich, now part of Merck KGaA, Darmstadt, Germany; catalogue no. 317240; purity ≥97%) was stored as a solid according to the manufacturer’s recommendations and freshly dissolved in distilled water before each experiment. The resulting solution was diluted to the desired final concentration immediately prior to use. The concentration was chosen based on the results of our pharmacokinetic measurements. Cultures were maintained at 37°C, 5% CO₂ and >95% humidity. To preserve initial adhesion, the first medium change was at day 5, followed by every 3 days thereafter.

Live-cell imaging and quantification

Capsule handling and seeding were performed under sterile conditions within a laminar air flow hood. Each anterior lens capsule was gently manipulated at the rim only, using fine forceps, to avoid possible damage to the epithelial cell layer. The capsule was carefully transferred into the well. Subsequently, the well was filled very slowly with the medium as described above from the edge toward the centre, ensuring minimal turbulence. Before incubation, it was confirmed microscopically that the capsule had settled at the bottom of the well.

Capsule proliferation was monitored using the CytoSMART Omni-FL system (Axion Biosystems, Atlanta, USA). Images were captured every 8 hours over 14 days. The cell-covered area within a standardised 5 mm diameter region (≈19.63 mm²) was quantified using the integrated Omni-FL analysis software.

Statistical analysis and software

All statistical analyses were performed using IBM SPSS Statistics (V.29.0, IBM). Graphs and visualisations were generated with GraphPad Prism (V.10.0, GraphPad Software, San Diego, California, USA), Microsoft Excel (V.16.77, Microsoft Corp) and Matplotlib (Python 3.11, Python Software Foundation), and finalised in Microsoft PowerPoint (V.16.77, Microsoft). Imaging data were analysed with the integrated CytoSMART Omni-FL software (V.1.3.2, Axion Biosystems, Atlanta, USA).

Normality of continuous variables was assessed using Shapiro-Wilk tests. Descriptive statistics were reported as mean, SD and range. Inter-group comparisons (low-dose vs high-dose) were performed using the Mann-Whitney U test, and correlations between clinical parameters and metformin concentrations were assessed by Spearman’s rank correlation.

For the in vitro proliferation assays, a linear mixed-effects model was applied with time as a repeated measure, treatment group as a fixed effect and capsule ID as a random intercept. Interaction terms (time×treatment) and covariates (age, gender) were included in the model. Estimated marginal means with 95% CIs were reported. The significance of fixed effects was assessed using Type III Wald χ² tests. A p value <0.05 was considered statistically significant.

Results

Metformin concentrations were measured in lens capsule tissue and serum from 20 patients with T2DM undergoing cataract surgery. Lens capsule concentrations were similar between groups (low-dose: 0.77 pg/µL±0.79; high-dose: 0.73 pg/µL±0.53; p=0.821), with no significant difference (p=0.821) (figure 1a, online supplemental table 1). In serum, however, the inverse pattern was observed: the low-dose group exhibited higher concentrations (mean 890.2 pg/µL±683.0) compared with the high-dose group (mean 303.8 pg/µL±328.3), without statistical significance (p=0.096) (figure 1b, online supplemental table 1). Blood sampling times did not differ between groups (p=0.199).

Figure 1. Boxplots comparing metformin concentrations between dosage groups (low-dose ≤1000 mg/day, high-dose >1000 mg/day). The central line within each box represents the median. (a) Capsule metformin concentration (pg/µL): no statistically significant difference between groups (p=0.821). (b) Serum metformin concentration (pg/µL): a trend towards higher levels in the low-dose group compared with the high-dose group was observed (p=0.096).

Clinical characteristics differed between groups. Patients in the high-dose cohort were significantly younger (mean 70.3 vs 75.9 years; p=0.027) and had a longer duration of diabetes (21.1 vs 11.2 years; p=0.021) (online supplemental figure 1a, online supplemental table 1). HbA1c (%) values did not differ significantly (p=0.970) (online supplemental figure 1b, online supplemental table 1), and visual acuity was also comparable (p=0.361). Regression analysis identified no predictors of capsule metformin concentration, though serum levels showed a trend (p=0.115). Correlation analysis confirmed a significant positive association between capsule and serum concentrations (r=0.553, p=0.011) (figure 2a, online supplemental table 2) but not between capsule concentration and HbA1c (r=0.383, p=0.095) (figure 2b, online supplemental table 2).

Figure 2. Correlations of capsule metformin concentrations. (a) Relationship between capsule and serum metformin concentrations: significant positive correlation (r=0.553, p=0.011). (b) Relationship between capsule metformin concentration and HbA1c (%): no statistically significant positive correlation (r=0.383, p=0.095).

To examine whether these intracapsular concentrations exert biological effects, an in vitro proliferation assay was performed using 16 anterior lens capsule explants from non-diabetic donors. Explants were cultured for 14 days in either standard medium (control) or medium supplemented with metformin at 0.75 pg/µL, corresponding to the mean concentration detected in patient capsules in the previous experiments. Proliferation was monitored continuously by live-cell imaging at 8-hour intervals, and growth was quantified as the increase in cell-covered area within a standardised region of interest of 5 mm diameter (≈19.63 mm²).

Linear mixed-effects modelling revealed a highly significant effect of time (p<0.001), consistent with progressive proliferation in both groups. The interaction between time and treatment was also significant (p<0.001), indicating that metformin reduced proliferation compared with controls (figure 3). Growth in the treated group was attenuated but not arrested. Random intercepts for capsule ID were included to account for baseline differences between individual donor capsules, thereby ensuring robust modelling across biological replicates.

Figure 3. Proliferation of human lens epithelial cells over time in control and metformin-treated samples, shown as estimated marginal means±95% CIs. Cultures were monitored for 14 days with live-cell imaging at 8-hour intervals. A significant effect of time (p<0.001) and a significant time×treatment interaction (p<0.001) were observed, indicating reduced proliferation in the metformin group.

Representative microscopy images illustrated these dynamics. In central capsule regions, untreated controls showed continuous expansion at 12 hours, 48 hours and 96 hours (figure 4a). At the capsule rim, LECs migrated outward, forming spindle-like extensions resembling structures observed clinically in PCO (figure 4b). In contrast, metformin-treated capsules exhibited reduced expansion and shorter spindle-like outgrowths appearing less differentiated (figure 4c). Automated surface coverage analysis confirmed the inhibitory effect of metformin, showing consistently reduced proliferative spread in the treated group (figure 5a, b).

Figure 4. Representative phase-contrast microscopy images of human lens capsules, shown at three consecutive time points (12 hours, 48 hours, 96 hours) to illustrate proliferation dynamics. (a) Proliferation of LECs in central capsule regions, demonstrating progressive cell expansion beyond the rim area over time (white arrowhead marks proliferation in an area distant from the rim). (b) Proliferation and migration of LECs from the capsule edge, forming spindle-like outgrowths that extend over 96 hours and resemble structures clinically observed in posterior capsule opacification (black arrow marks a representative spindle-like outgrowth). (c) Proliferation and migration of metformin-treated LECs, showing visibly reduced proliferation with fewer peripheral extensions; spindle-like outgrowths appeared shorter, broader and less differentiated compared with untreated controls. Scale bars represent 100 µm. LEC, lens epithelial cell.

Figure 5. Representative image from live-cell proliferation assay, illustrating automated quantification of capsule surface coverage in (a) treatment and (b) control group after 60 hours (top row) and 280 hours (bottom row). Scale bars represent 1000 µm.

Discussion

This study provides novel insights into the intraocular pharmacokinetics of systemically administered metformin and its potential anti-proliferative effects on hLECs relevant to PCO. We demonstrated that metformin is detectable in the human lens capsule at physiologically relevant concentrations and that these levels are capable of attenuating hLEC proliferation in vitro. To our knowledge, this is the first study to provide such evidence.

In our study, capsule metformin concentrations were similar between the low-dose (≤1000 mg/day) and high-dose (>1000 mg/day) groups, with no statistically significant difference. By contrast, low-dose patients exhibited higher serum concentrations than high-dose patients; however, this difference did not reach statistical significance. Metformin is known to display considerable inter-individual variability in absorption, bioavailability, tissue distribution and renal clearance.^{21 22} In our cohort, patients receiving higher daily doses were younger and had a longer duration of diabetes, factors that may influence pharmacokinetics. Moreover, although all patients confirmed metformin intake on the day of surgery and blood sampling times did not differ significantly between groups, indicating that blood was drawn under comparable conditions across both groups, heterogeneity in the timing of the last dose cannot be fully excluded. Importantly, metformin concentrations in the lens capsule were comparable between dose groups, and a significant positive correlation between serum and capsule concentrations supports the notion that systemic exposure influences ocular tissue availability irrespective of prescribed dose.

Collectively, these findings demonstrate that metformin reaches the lens capsule at measurable concentrations linked to systemic levels, possibly influenced by age and metabolic activity.

The significant positive correlation between serum and capsule concentrations nevertheless supports the notion that systemic exposure influences local ocular drug availability.

Our in vitro experiments provide functional evidence that even low intracapsular concentrations of metformin, comparable to those measured in patients, substantially suppress hLEC proliferation. Continuous live-cell imaging with mixed-effects modelling revealed a sustained anti-proliferative effect, consistent with previous findings of metformin-mediated inhibition of LEC senescence,²³ reduction of fibrotic gene expression and attenuation of proliferative activity in other tissues including vascular smooth muscle, hepatic stellate cells and cancer cell lines.^24–29 These results support the hypothesis that metformin interferes with pathways driving secondary cataract formation.

Mechanistically, the observed anti-proliferative effect aligns with metformin’s known actions on intracellular signalling, particularly activation of AMPK, inhibition of mTOR signalling and modulation of oxidative stress and transforming growth factor-β pathways.^{18 22 25–30} These pathways are implicated in fibrosis, aberrant proliferation and wound healing and are likely relevant to PCO pathogenesis.

The clinical implications of these findings are noteworthy. If metformin administered systemically reaches the lens capsule and reduces the proliferative drive of residual LECs, it could contribute to a reduced incidence of PCO after cataract surgery. This hypothesis is consistent with recent epidemiological data linking metformin use to lower Nd:YAG capsulotomy rates.²⁴ Importantly, this observation may help explain the contradictory evidence regarding diabetes as a risk or protective factor for PCO. While some studies suggested a protective role of diabetes,¹⁷ others indicated an increased risk with longer disease duration.^{15 16} These discrepancies may result from a failure to stratify patients by antidiabetic medication. Our results support the notion that metformin treatment, rather than diabetes itself, could be a key determinant influencing PCO risk.

Beyond systemic factors, IOL design remains the established cornerstone of PCO prevention. Sharp square-edge optic designs provide a mechanical barrier to LEC migration and remain the clinical standard.^{9 10 31} Variations in IOL materials such as hydrophobic acrylic versus hydrophilic acrylic or PMMA have demonstrated differences in PCO incidence, though no material has eliminated the problem entirely.^{25 33–35} Previous approaches, including heparin-coated IOLs, aimed at reducing cell adhesion showed limited long-term efficacy.¹⁰ Our findings suggest that combining mechanical barrier designs with pharmacological strategies may offer synergistic benefit. Concepts such as incorporating metformin into IOL surfaces or slow-release reservoirs may be envisioned, but remain hypothetical and will require rigorous preclinical and clinical evaluation.

Strengths of our study include the dual approach of combining pharmacokinetic measurements with functional cell proliferation assays, the use of sensitive LC-MS/MS analysis and continuous live-cell imaging with robust statistical modelling. These methods allowed us to capture both systemic drug distribution and its direct cellular consequences. Limitations include the modest sample size typical of a pilot study, limiting statistical power for regression analyses. Larger cohorts with detailed pharmacokinetic profiling will be required to validate our findings and assess inter-individual variability in greater depth.

Together, these results provide evidence that systemically administered metformin accumulates in the human lens capsule at physiologically relevant concentrations and, at these levels, significantly suppresses LEC proliferation in vitro. These complementary pharmacokinetic and functional findings support the hypothesis that metformin may mitigate the cellular processes underlying PCO.

A limitation of the present study is the absence of direct intracellular metformin quantification in cultured LECs. While such measurements could provide additional mechanistic insight, the present study focused on assessing tissue-level exposure and functional effects on LEC behaviour. Importantly, metformin was detectable in human lens capsule tissue in vivo despite systemic metabolism and clearance, and corresponding concentrations were associated with a statistically significant attenuation of proliferation in vitro. Future studies integrating intracellular uptake measurements with pathway-specific analyses will be essential to further elucidate the underlying mechanisms.

In summary, to the best of our knowledge, this study provides the first evidence that systemically administered metformin reaches the human lens capsule and suppresses hLEC proliferation at physiologically relevant concentrations. These results support the hypothesis that metformin may serve as a pharmacological adjunct to current mechanical strategies for PCO prevention. Future research should expand patient cohorts, investigate long-term clinical outcomes and explore innovative applications such as drug-eluting IOLs to translate these findings into clinical practice.

Data availability statement

Data are available upon reasonable request.