The experimental workflow for this study is illustrated in Fig. 1.

Fig. 1.

Study workflow. This diagram outlines the entire process from cell culture to data analysis. Initially, human lens epithelial cells (HLE-B3) were cultured and treated with different concentrations of dexamethasone (0, 0.25, and 0.5 µM) for 60 days. Subsequent steps included cell size measurement using phase-contrast microscopy, flow cytometry analysis, and high-throughput transcriptome sequencing. By Figdraw (https://www.figdraw.com/#/).

The HLE-B3 cell line was purchased from American Type Culture Collection (ATCC; Manassas, VA, USA), an internationally recognized institution that provides ethically reviewed cell lines. The ATCC ensures that all cell lines, including the HLE-B3, are obtained and distributed in compliance with ethical standards. An analysis certificate verifying their provenance and quality is in the Supplementary Materials. The HLE-B3 cell line is an immortalized HLEC line frequently used to study the biological properties of HLECs and their responses under various conditions. HLE-B3 cells exhibit stable proliferation and a prolonged lifespan, making them widely used in ophthalmic research, particularly in studies related to lens-associated diseases such as cataracts [19]. HLE-B3 cells were cultured in a nutrient-rich Dulbecco’s Modified Eagle Medium (Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS; Gibco) and the antibiotics penicillin (100 U/mL) and streptomycin (100 µg/mL). These cells were cultured at 37 °C with 5% CO₂. Regular mycoplasma screening was carried out using the MycoAlert Detection Kit (Lonza, Walkersville, MD, USA) to verify the absence of mycoplasma contamination and the test result is negative. Furthermore, the authenticity of the cell line was confirmed through rigorous short tandem repeat (STR) profiling. Authentication and STR profiling were conducted using the Cell Line Authentication Service by ATCC, in accordance with the guidelines of the International Cell Line Authentication Committee (ICLAC) and the ExPASy Cellosaurus database to avoid cross-contamination or misidentification. HLE-B3 cells were divided into three groups: control group with no dexamethasone treatment, 0.25 µmol/L dexamethasone, and 0.5 µmol/L dexamethasone groups. We used HLE-B3 cells treated with 0 µM dexamethasone as the baseline comparison, referred to as the control group, where the cells were treated with the same volume of water used to dissolve the dexamethasone. Each group consist of three samples. All groups were cultured for 60 days, with medium changes and passaging performed weekly.

After 15 min fixation in 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) at room temperature, HLE-B3 cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) for 10 min. Then the cells were blocked in 1% bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA) for 1 h. Next, HLE-B3 cells were incubated overnight at 4 °C with primary antibodies against $\alpha$-crystallin, $\gamma$-crystallin, and vimentin. After three washes with phosphate-buffered saline (PBS) (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), cells were incubated with fluorophore-tagged secondary antibodies for 1 h in the dark at ambient temperature, followed by additional washing. Then nuclei were stained with 4^′,6-diamidino-2-phenylindole (DAPI) for 5 min, followed by coverslip mounting on slides with anti-fade medium. Fluorescence microscopy (Leica, Wetzlar, Germany) was employed to capture the resulting images. Three independent repeat experiments were conducted.

HLE-B3 cells, either untreated or co-cultured with 0.25 µM dexamethasone for 60 days, were fixed for 15 minutes in 4% paraformaldehyde at room temperature. After fixation, cells were permeabilized using 0.1% Triton X-100 for 10 minutes and then blocked with 1% bovine serum albumin (BSA) for 1 hour at room temperature. Cells were incubated overnight at 4 °C with primary antibodies against LPAR1 and phosphorylated S6K1 (p-S6K1). Following three washes with PBS, cells were incubated with appropriate fluorophore-conjugated secondary antibodies for 1 hour in the dark at ambient temperature. Nuclei were stained with DAPI for 5 minutes, and coverslips were mounted on slides using anti-fade mounting medium. Fluorescence images were acquired using a fluorescence microscope, and ImageJ software 1.54f (National Institutes of Health, Bethesda, MD, USA) was employed to quantify the fluorescence intensity of LPAR1 and p-S6K1. Three independent repeat experiments were conducted.

Cell size was measured under a phase-contrast microscope (DMI3000B inverted phase contrast microscope; Leica, Wetzlar, Germany) at 2000$\times$ magnification. Subsequently, ImageJ software (v.1.54k12) was used to acquire images, which were subjected to analysis for quantification of cellular dimensions. GraphPad Prism (version 10.2.2; GraphPad Software, LLC, La Jolla, CA, USA) was used to statistically analyze the data. For each group, 30 measurements were taken for cell length, and 20 measurements were taken for cell area, cytoplasmic area, nuclear area, and nucleus-cytoplasm ratio.

Harvested cells were rinsed with PBS, and then resuspended in a modified PBS solution supplemented with 2% FBS. For detailed analysis of cellular dimensions and granularity, we employed the advanced FACSCalibur flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA). FlowJo™ v.10.10 (Tree Star, Ashland, OR, USA) was used for data processing. Three independent repeat experiments were conducted. Flow cytometry analysis focused on assessing changes in cell size and granularity between the control group and the 0.25 µM dexamethasone-treated group, utilizing forward scatter (FSC) and side scatter (SSC) respectively. FSC measurements provide insight into the relative cell size, as FSC correlates with the cell volume or diameter. SSC, on the other hand, measures the granularity or internal complexity of cells, reflecting aspects such as the presence of internal structures like granules. Given the nature of our investigation into the basic cellular responses to dexamethasone, specific markers were not employed in the flow cytometry analysis. The gating strategy and analysis process, detailed in Supplementary Fig. 1, was based on guidelines from the FlowJo official gating tutorial. We excluded debris and dead cells by setting appropriate thresholds on FSC and SSC parameters, selecting a population that predominantly represents live HLE-B3.

HLE-B3 cells were cultured in vitro and incubated with 0, 0.25, and 0.5 µmol/L dexamethasone for 60 days. Total RNA was extracted from cells with TRIzol® reagent. The integrity and quality of the extracted RNA were assessed utilizing the Agilent 5300 Bioanalyzer (Agilent Technologies, Wilmington, DE, USA). For quantitative analysis, the NanoDrop 2000 spectrophotometer (NanoDrop Technologies Inc., Wilmington, DE, USA) was employed. Only high-quality RNA samples (optical density [OD] 260/280 = 1.8–2.2, OD 260/230 $\ge$ 2.0, RNA integrity number $\ge$6.5, 28S:18S $\ge$1.0, $>$1 µg) were deemed suitable for the subsequent construction of sequencing libraries.

RNA purification, coupled with reverse transcription, library assembly, and ultimately sequencing, were conducted in accordance with the instructions of the manufacturer (Illumina, San Diego, CA, USA). The Illumina Stranded mRNA Prep Ligation Kit was used to construct the RNA sequencing (RNA-seq) transcriptome library, which was accomplished with precise input of 1 µg total RNA, ensuring the highest degree of precision and conformity with academic rigor. Using oligo (dT) beads and polyA selection, mRNA was separated and fragmented using designated buffers. Double-stranded cDNA synthesis was achieved using the SuperScript III First-Strand Synthesis System (Invitrogen, Carlsbad, CA, USA) and random hexamer primers (Illumina). The cDNA was subjected to phosphorylation, ‘A’ tailing, and end repair in accordance with Illumina’s methodology, enhancing its suitability for library construction. The cDNA (300 bp) was separated on a 2% ultra-agarose gel (Thermo Fisher Scientific, Waltham, MA, USA), followed by PCR amplification utilizing phusion DNA polymerase (New England Biolabs, Ipwich, MA, USA) for 15 cycles. Subsequently, the paired-end RNA-seq library was quantified with Qubit 4.0 (Thermo Fisher Scientific, Waltham, MA, USA) and sequenced on the advanced NovaSeq X Plus platform (Illumina). The initial paired-end sequencing data were trimmed and Fastp was adopted for quality control [20], utilizing its standard parameter settings. Next, using HISAT2 v2.2.1 (Daehwan Kim Lab, Johns Hopkins University, Baltimore, MD, USA), clean reads were separately aligned to the reference genome in orientation mode [21]. The StringTie reference-based transcriptome assembler was used to assemble each sample’s mapped readings [22].

To detect differentially expressed genes (DEGs) between different samples, transcripts per million methodology was used to assess transcript abundance. RNA-Seq by Expectation-Maximization was employed for quantitative gene expression profiling [23]. DESeq2 or DEGseq was used for differential expression analysis [24, 25]. DEGs were considered significantly differentially expressed if they had $|$log2FC$|$ $\ge$1 and false discovery rate (FDR) $\le$0.05 (DESeq2) or FDR 0.001 (DEGseq). The study employed functional enrichment analysis, utilizing Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) to determine the DEGs that had high enrichment in GO terms and metabolic pathways. These analyses were conducted with stringent statistical thresholds, applying a Bonferroni-corrected p $\le$ 0.05 against the entire transcriptome as a backdrop. Specifically, GO annotations were enriched using GOATools, while KEGG pathway analyses were conducted with KEGG Orthology-Based Annotation System software (KOBAS 3.0, Peking University, Beijing, China) [26].

To investigate the co-expression network in our cohort, we conducted intricate analysis leveraging the weighted gene co-expression network analysis (WGCNA) framework. Utilizing comprehensive transcriptomic profiles derived from HLECs, we conducted methodical identification of the optimal soft-thresholding parameter, harnessing the pickSoftThreshold functionality within the WGCNA package 1.71. This step ensures adherence to the stringent scale-free topology principle, which is fundamental for constructing robust networks. Subsequently, we computed the adjacency matrix and converted it into the topological overlap matrix (TOM). Employing this refined TOM, we quantified dissimilarity metrics pivotal for hierarchical clustering. To identify distinct, biologically meaningful gene modules exhibiting coordinated expression patterns, we applied the dynamic tree cutting algorithm, implementing a threshold of a minimum module size of 30. To establish the relevance of these gene modules to our study, we assessed the gene significance values and module membership values. This approach enabled us to pinpoint key modules associated with our experimental variables, thereby shedding light on the intricate molecular machinery underlying the phenotypic alterations observed in LECs.

We performed protein–protein interaction (PPI) analyses of DEGs focusing on the phosphoinositide 3-kinase/AKT pathway using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database (https://string-db.org/) with a confidence score threshold of $\ge$0.4. The key nodes and interactions were analyzed to highlight the role of LPAR1 in the pathway.

cDNA synthesis employed the PrimeScript RT Reagent Kit sourced from Takara, Japan. qPCR analysis was conducted on the ABI 7500 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA), utilizing SYBR Premix Ex Taq (Takara, Japan). The levels of expression for LPAR1 and mammalian target of rapamycin (mTOR) were quantified using the 2^{-Δ⁢Δ⁢Ct} method, normalizing to GAPDH as a reference point internally. The primers specific to these genes were procured from Sangon Biotech, Shanghai, China. Three independent repeat experiments were conducted.

The crystal structure of the LPAR1 protein (PDB ID: 7YU4) was obtained from the RCSB PDB database (https://www.rcsb.org/). PyMOL software (version 2.x, Schrödinger, LLC, New York, NY, USA) was used for initial processing, where solvent molecules and small ligands were removed. The protein structure was further processed using ADFRsuite-1.0 (Scripps Research, La Jolla, CA, USA), retaining its native charges, and was converted into the pdbqt format for docking. For the ligand, the small molecule structure of dexamethasone was downloaded from the PubChem database. Similar preprocessing steps were performed using ADFRsuite-1.0 to retain charges and convert it into pdbqt format.

The docking process was carried out using AutoDock Vina 1.2.5 (The Scripps Research Institute, La Jolla, CA, USA). The docking grid for LPAR1 was centered at the coordinates (center_x = 165.337, center_y = 151.376, center_z = 152.603), with a box size of (size_x = 64.36 Å, size_y = 44.56 Å, size_z = 50.91 Å). A grid spacing of 0.375 Å and an exhaustiveness parameter of 32 were used. The prepared pdbqt files of LPAR1 and dexamethasone were input into AutoDock Vina, and docking was executed with these parameters. The resulting docking conformations were ranked by binding energy, and the best-scoring pose was selected for analysis.

This sample size was determined based on our preliminary experimental data and sample size calculation, aiming to achieve a statistical power of 0.8 at a 0.05 significance level. We used SPSS 29.0.2 (IBM Corporation, Armonk, New York, NJ, USA) to conduct statistical analysis and created the statistical graphs by using GraphPad Prism software, version 10.2.2. Depending on the dexamethasone concentration, cells were categorized into 0 µM, 0.25 µM, and 0.5 µM groups. Variables in each group followed a normal distribution, tested by Shapiro-Wilk. For comparison of variables among the three groups, one-way ANOVA was used and test for homogeneity of variance were first performed. The nuclear area of HLE-B3 cells satisfied variance alignment among the groups and further post hoc comparisons were made using Tukey’s test. The groups of HLE-B3 cell length, cell area, cytoplasmic area, and nuclear-cytoplasmic ratio did not satisfy variance alignment, and then corrected for this using Welch’s test with a Tamhane T2 test for post hoc comparison. For direct comparisons between pairs, the student’s t-test was adopted. The comparison of FSC and SSC between the control group and the 0.25 µM dexamethasone group was performed using a t-test, as the data followed a normal distribution and had equal variances. For the PCR experiment, the mRNA expression levels between the control group and the 0.25 µM dexamethasone group were also analyzed using a t-test, with the data meeting the assumptions of normal distribution and equal variances. The average fluorescence intensity of LPAR1 in immunofluorescence staining between the control group and the 0.25 µM dexamethasone group was calculated using a t-test, with the data normally distributed and variances equal. However, for the average fluorescence intensity of S6K1 in immunofluorescence staining between the control group and the 0.25 µM dexamethasone group, where the variances were unequal, Welch’s correction was applied. p 0.05 was considered statistically significant. The results are shown as the mean $\pm$ standard deviation (SD). Additionally, transcriptome datasets underwent rigorous bioinformatic analysis to identify DEGs and potential signaling cascades. In this study, no missing values were found in the dataset.

The concentration of dexamethasone in the blood that leads to GC-induced cataracts is 0.15 µmol/L, with this level of exposure typically occurring over a period of 2 years [27, 28]. A previous study has shown that low-dose dexamethasone (0.1 µM) leads to increased proliferation of HLECs, while exposure to dexamethasone concentrations above 1 µM results in the apoptosis of HLECs [29]. Therefore, we chose 0.25 µM and 0.5 µM concentrations based on their relevance to physiological conditions and clinical scenarios.

We cultured HLE-B3 cells in vitro and treated them with different concentrations of dexamethasone (0, 0.25, and 0.5 µM). To analyze the expression of specific proteins and intermediate filament proteins in HLE-B3 cells, we performed immunofluorescence staining for $\alpha$-crystallin, $\gamma$-crystallin, and vimentin, and found that they were widely expressed in HLE-B3 cells, with strong fluorescent signals observed in the cytoplasm (Fig. 2B). $\alpha$-crystallin and $\gamma$-crystallin are specific to the lens and crucial for its transparency, whereas vimentin supports cell structure and function. The detection of $\alpha$-crystallin, $\gamma$-crystallin, and vimentin proteins via immunofluorescence staining in HLE-B3 cells confirmed their identity as LECs and ensured the reliability of the experimental results on the effects of dexamethasone on HLECs. The cultured cells under a phase-contrast microscope are shown in Fig. 2A.

Fig. 2.

Morphologic effects of different concentrations of dexamethasone on HLE-B3 cells and immunofluorescence staining of HLE-B3 cells. (A) The morphology of HLE-B3 cells is shown for the 0, 0.25, and 0.5 µM groups. The white circles represent the cell contours, whereas the yellow circles represent the cell nuclei. Scale bars represent 100 µm for the first column, 50 µm for the second column, and 20 µm for the third column, respectively. (B) Immunofluorescence staining of HLE-B3 cells for $\alpha$-crystallin, $\gamma$-crystallin, and vimentin. Scale bar = 100 µm. (C) Analysis of cell length among the control, 0.25, and 0.50 µM groups (n = 30). (D) Analysis of cell area among the control, 0.25, and 0.50 µM groups (n = 20). (E) Analysis of the nuclear area among the control group, 0.25, and 0.50 µM groups (n = 20). (F) Analysis of the cytoplasmic area among the control, 0.25, and 0.50 µM groups (n = 20). (G) Analysis of the nuclear-cytoplasmic ratio among the control, 0.25, and 0.50 µM groups (n = 20). The data are shown as the mean $\pm$ SD. All variables follow a normal distribution. The data of the cytoplasmic area among the control, 0.25, and 0.50 µM groups meet the assumption of homogeneity of variance, whereas other data do not. p-values were obtained by ANOVA single factor analysis and are shown in the graphs (* 0.05; *** 0.001; **** 0.0001).

The cell length, cell area, nuclear area, cytoplasmic area, and nuclear-cytoplasmic ratio after 60 days of cell culture are expressed as the mean $\pm$ SD. One-way ANOVA was used to compare these results among the three groups (Supplementary Table 1). Furthermore, for variables that satisfy variance alignment and those that do not, we use Tukey’s test and Tamhane T2 test for post hoc comparisons respectively (Supplementary Table 2). The results showed that the 0.25 and 0.5 µM dexamethasone-treated groups showed an increase in cell length, cell area, and cytoplasmic area, and a decrease in nuclear-cytoplasmic ratio (Fig. 2C–G).

We further analyzed the cell size and intracellular granularity of the control group and 0.25 µM dexamethasone group using flow cytometry. We used the parametric Student’s t-test to compare the cell size and intracellular granularity between the control group and 0.25 µM dexamethasone group. Flow cytometry results indicated that the average cell size was greater in the 0.25 µM group than in the control group (t = 3.528, p = 0.0243; Fig. 3A,B), and there was an increase in cell granularity (t = 24.41, p 0.001; Fig. 3C,D). Increased granularity within LECs may indicate changes in autophagic activity, cellular stress responses, or the aggregation of crystalline [30, 31]. It has been shown that in GIC in vivo, the proliferation of LECs is abnormally regulated and differentiated, producing modest amounts of extracellular granular and fibrillar material as well as increased cell area [32, 33]. Previous studies have indicated that in patients with GIC, the lens fiber cells contain nuclei and organelles, which lead to an increase in intracellular granularity [34, 35]. The results of the experiments showed that dexamethasone treatment resulted in increased cell length and area and increased granularity in HLECs in vitro. This phenomenon may affect the processes of cell proliferation and differentiation and inhibit autophagy by regulating mammalian target of rapamycin (mTOR) activity. The results of the in vitro experiments were similar to the morphological changes observed in HLECs in GIC in vivo.

Fig. 3.

Dexamethasone increases HLE-B3 cell size and granularity. (A) The relative sizes of HLE-B3 cells were compared by measuring forward scatter (FSC) values through fluorescence-activated cell sorting analysis. (B) The mean sizes of HLE-B3 cells were obtained and are shown in a graph (n = 3). (C) The relative cell granularity of HLE-B3 cells was compared using side scatter (SSC) analysis. (D) The mean granularity of the HLE-B3 cells was obtained and are shown in a graph (n = 3). The data are shown as the mean $\pm$ SD. p-values were obtained by the parametric Student’s t-test and are shown in the graphs (* 0.05; **** 0.0001).

To investigate the specific mechanisms by which dexamethasone affects the morphology and function of HLECs, we performed high-throughput transcriptome sequencing on the control, 0.25, and 0.5 µM groups. Analysis of DEGs suggested that there was only one DEG between the 0.25 and 0.5 µM dexamethasone groups (Fig. 4A,B). Therefore, we subsequently chose to compare the control group and 0.25 µM dexamethasone group for further analyses.

Fig. 4.

Differential gene expression analysis in HLE-B3 cells treated with dexamethasone. (A) Statistical chart of DEGs between groups. (B) Venn diagram of DEGs between groups. (C) Volcano plot showing the changes in HLE-B3 genes in the 0.25 µmol/L dexamethasone group (fold change $\ge$2). (D) The heatmap displays gene expression across different groups of DEGs. Colors on the heatmap represent normalized gene expression levels, ranging from high (red) to low (blue). DEGs, Differentially expressed genes.

Using a corrected p 0.05 and fold change $\ge$2 as selection criteria, 427 genes were greatly downregulated and 443 genes were greatly upregulated in the 0.25 mmol/L dexamethasone group compared with the control group (Fig. 4C). Remarkable differences in mRNA expression between the control group and 0.25 µM dexamethasone group were observed, as shown in the heatmap in Fig. 4D. To further investigate the effects of dexamethasone on HLECs, GO enrichment analysis was conducted. The results showed that the DEGs are closely linked to the steroid hormone response, response to GC, regulation of smooth muscle contraction, response to growth factors, regionalization, and positive regulation of cell–cell adhesion (Fig. 5A,C). KEGG enrichment analysis identified pathways related to human diseases, environmental information processing, and cellular processes such as rheumatoid arthritis, malaria, interleukin 17 signaling pathway, cancer signaling pathways, calcium signaling pathway, PI3K/AKT signaling pathway, and tumor necrosis factor signaling pathway, among others (Fig. 5B,D). These enriched pathways further confirmed the accuracy of high-throughput transcriptome sequencing and its similarity to in vivo experiments.

Fig. 5.

GO and KEGG enrichment analyses. (A) Bubble plot illustrating the GO enrichment analysis between the control group and 0.25 µmol/L dexamethasone group. The x-axis represents the rich factor, whereas the y-axis lists the GO terms. The bubble size indicates the number of genes, and the bubble color corresponds to the adjusted p-value (padjust). (B) Bubble plot illustrating KEGG pathway enrichment analysis between the control group and the 0.25 mmol/L dexamethasone group. The x-axis represents the rich factor, whereas the y-axis lists the KEGG pathways. The bubble size indicates the number of genes, and the bubble color corresponds to the padjust. (C) Enriched chordal diagram showing the relationships between the DEGs and enriched GO terms. The different colors represent various GO terms. (D) Enriched chordal diagram showing the relationships between DEGs and enriched KEGG terms. Different colors represent various KEGG terms. DEGs, Differentially expressed genes; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

We identified the top 20 pathways related to the morphological and biological effects of dexamethasone on HLECs using KEGG and GO analyses. Genes within these pathways were compiled into a gene set (Supplementary Table 3). WGCNA analysis was conducted to identify gene modules and central genes within the group. We established a scale-free co-expression network with a soft threshold of 9 and a scale-free index of –0.095, demonstrating good mean connectivity (Fig. 6A,B). The clustering dendrogram is depicted in Fig. 6C, and the genes were ultimately clustered into two modules (Fig. 6D).

Fig. 6.

WGCNA analysis and discovery of potential hub genes. (A) Soft threshold power of WGCNA. (B) Mean connectivity of WGCNA. (C) WGCNA cluster dendrogram. (D) WGCNA clustered modules. (E) Analysis of module significance. WGCNA, Weighted gene co-expression network analysis.

We calculated the correlation between the two modules and the effects of dexamethasone on the size of HLECs. The module significance analysis indicated that the MEblue module was the most significant (Fig. 6E). The MEblue module included 56 genes and was considered important in the effects of dexamethasone on cell size. A network visualization of genes within the MEblue module showed key gene interactions (Fig. 7A). Hub genes identified in this module included atypical chemokine receptor 1, secreted frizzled related protein 1, collagen type IV alpha 4 chain, LPAR1, and growth arrest specific 1, which may play crucial roles in the effects of dexamethasone on cell size. Among these, LPAR1, a subtype of GPCRs, has not been previously reported to interact with dexamethasone and affect cell morphology and function. LPAR1 activates the PI3K/AKT pathway, one of the top 20 KEGG enriched pathways, and is associated with the “response to organophosphorus” term in the top 20 GO enriched pathways. To understand the interactions between LPAR1 and other proteins in the PI3K/AKT pathway, we screened the top 20 KEGG enriched pathways of DEGs, focusing on genes in the PI3K/AKT pathway, and constructed a PPI network diagram using the STRING database (Fig. 7B).

Fig. 7.

Recognition of hub genes and the protein–protein interaction network. (A) Network display of key genes. (B) Each node represents a protein, and the lines represent an association between proteins. COL4A6, Collagen Type IV Alpha 6 Chain; COL6A6, Collagen Type VI Alpha 6 Chain; COL4A4, Collagen Type IV Alpha 4 Chain; COL4A3, Collagen Type IV Alpha 3 Chain; COL6A3, Collagen Type VI Alpha 3 Chain; LAMC2, Laminin Subunit Gamma 2; LAMA4, Laminin Subunit Alpha 4; ITGB4, Integrin Beta 4; ITGA10, Integrin Alpha 10; ITGA2, Integrin Alpha 2; TNC, Tenascin C; COMP, Cartilage Oligomeric Matrix Protein; LPAR1, Lysophosphatidic Acid Receptor 1; THBS2, Thrombospondin 2; PDGA, Platelet-Derived Growth Factor Alpha; PDGFD, Platelet-Derived Growth Factor D; ERBB4, Erb-B2 Receptor Tyrosine Kinase 4; ANGPT1, Angiopoietin 1; AREG, Amphiregulin; FLT1, Fms Related Tyrosine Kinase 1; IL6, Interleukin 6; EFNA1, Ephrin A1; TGFA, Transforming Growth Factor Alpha; IGF2, Insulin-Like Growth Factor 2; IRS1, Insulin Receptor Substrate 1; IL7, Interleukin 7; TILR2, T cell immunoreceptor with Ig and ITIM domains 2.

Moreover, high-throughput transcriptome sequencing revealed differentially expressed autophagy-related genes, such as DDIT4, RPS6KA2, and IRS1 (Supplementary Table 4). KEGG enrichment analysis of these DEGs also indicated the involvement of key autophagy-related pathways, including mTOR, PI3K/AKT, and AMPK signaling.

Based on the changes in the PI3K/AKT signaling pathway indicated by KEGG and GO analyses, we speculate that dexamethasone interacts with LPAR1, triggering a series of signaling responses that affect the PI3K/AKT/mTOR pathway. These effects, in turn, lead to increased LEC length, enhanced crystallin protein synthesis, and autophagy inhibition.

We have conducted PCR validation, which showed significantly increased expression of LPAR1 and mTOR in the 0.25 µM dexamethasone group (t = 9.158, p 0.001; t = 59.854, p 0.001; Fig. 8A). Additionally, we performed immunofluorescence staining to assess the LPAR1 and the phosphorylation of ribosomal protein S6 kinase beta-1 (S6K1), which is a key downstream effector of the mTOR signaling pathway (Fig. 8B,D). Quantification using ImageJ revealed higher immunofluorescence intensities for LPAR1 and phosphorylated S6K1 in the 0.25 µM dexamethasone group (t = 5.085, p = 0.007; t = 3.061, p = 0.038; Fig. 8C,E).

Fig. 8.

Quantitative polymerase chain reaction (qPCR) and immunofluorescent staining of HLE-B3 cells with (0), 0.25 µM dexamethasone treatment. (A) qPCR analysis of the mRNA expression levels of LPAR1 and mTOR. qPCR analysis is presented relative to the GAPDH expression data (n = 3). (B,C) Immunofluorescent staining of LPAR1 (green) merged with DAPI (blue) and relative fluorescence intensity of LPAR1 was quantified (n = 5). (D,E) Immunofluorescent staining of p-S6K1 (green) merged with DAPI (blue) and relative fluorescence intensity of p-S6K1 was quantified (n = 3). Data are shown as mean $\pm$ SD. p-values were obtained by the parametric Student’s t-test and are shown in the graphs (* 0.05, ** 0.01; *** 0.001, **** 0.0001). Scale bar = 100 µm.

In addition, we have performed molecular docking between LPAR1 and dexamethasone shown in Fig. 9, and the results suggest that the binding between the LPAR1 protein and dexamethasone molecule is spontaneous, with a relatively high affinity (binding energy of –8.16 kcal/mol). Interaction analysis revealed multiple types of interactions, including one salt bridge, one electrostatic interaction, four hydrogen bonds, one halogen bond, and six hydrophobic interactions. Key residues involved in these interactions were LYS294 (salt bridge), LYS39 (electrostatic interaction), GLU293, LYS39, and GLN125 (hydrogen bonds), and several leucine residues contributing to hydrophobic interactions. These interactions significantly stabilize the binding of dexamethasone to LPAR1.

Fig. 9.

Molecular docking results of LPAR1 protein with dexamethasone. (A) Surface representation of LPAR1 protein (blue) bound to dexamethasone (sticks in the binding pocket) illustrating the spatial fit and overall surface interactions. (B) 2D interaction diagram highlighting the key residues involved in the interaction between LPAR1 and dexamethasone. The diagram shows hydrogen bonds (green lines), attractive charge interactions (red lines), and salt bridges (orange lines). Involved residues include LYS39, LYS294, GLU293, GLN125, and several hydrophobic interactions (LEU105, LEU297, LEU278, and LEU277). (C) 3D cartoon representation of LPAR1 with a close-up of the binding pocket. The detailed view shows hydrogen bonding and electrostatic interactions between LPAR1 residues (blue sticks) and dexamethasone (beige sticks), with LYS39, GLU293, and LYS294 playing key roles.

These results further provide molecular-level evidence to back the mechanistic claim that LPAR1 interacts with glucocorticoids, impacting the PI3K-AKT-mTOR pathway (Fig. 10), which in turn affects the morphology, area, and autophagic functions of LECs.

Fig. 10.

The LPAR1/PI3K/AKT/mTOR pathway. Dexamethasone interacts with LPAR1, activating the PI3K/AKT/mTOR signaling pathway, which subsequently stimulates cell growth, increases cell size, and potentially enhances cell granularity relative to the group not treated with dexamethasone by inhibiting autophagy. GC, glucocorticoid; LPAR1, Lysophosphatidic acid receptor 1; PI3K, Phosphoinositide 3-kinase; PIP2, phosphatidylinositol-4,5-bisphosphate; PIP3, phosphatidylinositol-3 ,4,5-trisphosphate; AKT, protein kinase B; mTOR, Mammalian target of rapamycin; P, Phosphorylation; 4EBP, 4E-binding protein; S6K, p70 Ribosomal Protein S6 Kinase. By Figdraw (https://www.figdraw.com/#/).