Insights into CD154-mediated pathways in ocular hypertensive glaucoma: The role of Müller cells and P2X7 in retinal neuroprotection and therapeutic potential.

An elevation of pathologic intraocular pressure (IOP) is the greatest risk factor for glaucoma. CD154 has been reported to bind to CD40 expressed by orbital fibroblasts and be involved in immune and inflammatory responses. However, the function and mechanism of CD154 in ocular hypertensive glaucoma (OHG) are not fully understood. We isolated and characterized Müller cells and subsequently examined the effect of CD154 on ATP release from those cells. After being cocultured with CD154-pretreated Müller cells, retinal ganglion cells (RGCs) were treated with P2X7 siRNAs or a P2X7 inhibitor. Furthermore, mouse models of glaucoma (GC) were injected with P2X7 shRNA. p21, p53, and P2X7 expression were examined, and cellular senescence and apoptosis were detected by ß-Gal and TUNEL staining, retinal pathology was examined by H&E staining, and CD154 and ß-Gal expression were detected by ELISA. CD154 induced ATP release from Müller cells and accelerated the senescence and apoptosis of RGCs that had been cocultured with Müller cells. We also found that treatment with P2X7 could attenuate the senescence and apoptosis of RGCs mediated by Müller cells pretreated with CD154. In vivo studies in GC model mice verified that P2X7 silencing attenuated pathological damage and prevented the senescence and apoptosis of retinal tissue. The study demonstrates how CD154 accelerates the aging and apoptosis of RGCs by co-cultivating Müller cells pretreated with CD154 in OHG. The research implies that CD154 has the potential to become a new therapeutic target for ocular hypertension glaucoma, providing a new research direction for its treatment.

In vivo confocal microscopy qualitative investigation of the relationships between lattice corneal dystrophy deposition and corneal nerves.

BACKGROUND: To investigate the corneal neurotropic phenomenon in patients with lattice corneal dystrophy (LCD) with in vivo laser scanning confocal microscopy (IVCM). METHODS: IVCM was performed on a total of 15 patients (28 eyes) with LCD annually at a follow-up. A collection of the data was acquired to be analyzed. RESULTS: As indicated by the analysis, the LCD patients' normal corneal stromal nerves (Grade 0) presented a decline with the prolongation of the follow-ups, corresponding to a gradual increase in grade I and II involving amyloid-wrapped nerve fibers, which demonstrated that the growing amount of amyloid deposit due to the corneal nerve invasion increased slowly over time. CONCLUSIONS: The neurotropic phenomenon could increase with its severity in the corneal lesion of the patients with LCD, and also reflect the distribution of the corneal nerves, to some extent. IVCM provides a rapid, noninvasive way to observe the corneal nerves, which can be an efficient means of better understanding the development of LCD.

Lithium chloride (LiCl) induced autophagy and downregulated expression of transforming growth factor ß-induced protein (TGFBI) in granular corneal dystrophy.

This study evaluated whether lithium chloride (LiCl) prevents cytoplasmic accumulation of mutant-transforming growth factor ß-induced protein (Mut-TGFBI) in granular corneal dystrophy (GCD) via activation of the autophagy pathway. Levels of TGFBI and microtubule-associated protein 1A/1B-light chain 3 (LC3) in 3 GCD patients and healthy controls were analyzed by immunohistochemistry (IHC) staining and Western blot. Primary corneal fibroblasts were isolated and transfected with wild type or mutant type TGFBI over-expressed vectors, and then treated with LiCl and/or autophagy inhibitor 3-methyladenine (3-MA). Then, levels of TGFBI, glycogen synthase kinase-3 (GSK-3) and LC3-I/-II were detected. Cell viability and transmission electron microscopy assay were also performed. Levels of TGFBI and LC3 were significantly increased in GCD patients. Over-expression of mutant type TGFBI inhibited cell viability and induced autophagy in corneal fibroblasts. LiCl downregulated the expression of TGFBI in mutant type TGFBI over-expressed cells in a dose- and time-dependent manner. LiCl enhanced autophagy in mutant type TGFBI over-expressed cells and recovered cell viability in those cells. However, the effects of LiCl were partly attenuated when autophagy was suppressed by 3-MA. To summarize, treatment with LiCl inhibited the expression of TGFBI and recovery the inhibitory of mutant type TGFBI in cell viability, at least part through enhancing of autophagy. These data strongly suggest that LiCl may be useful in the treatment of GCD.

Müller cells under hydrostatic pressure modulate retinal cell survival via TRPV1/PLCγ1 complex-mediated calcium influx in experimental glaucoma.

Glaucoma, an irreversible blinding eye disease, is currently unclear whose pathological mechanism is. This study investigated how transient receptor potential cation channel subfamily V member 1 (TRPV1), 1-phosphatidylinositol 4,5-bisphosphate phosphodiesterase gamma-1 (PLCγ1), and P2X purinoceptor 7 (P2X7) modulate the levels of intracellular calcium ions (Ca2+) and adenosine triphosphate (ATP) in Müller cells and retinal ganglion cells (RGCs) under conditions of elevated intraocular pressure (IOP). Müller cells were maintained at hydrostatic pressure (HP). TRPV1- and PLCG1-silenced Müller cells and P2X7-silenced RGCs were constructed by transfection with short interfering RNA (siRNAs). RGCs were cultured with the conditioned media of Müller cells under HP. A mouse model of chronic ocular hypertension (COH) was established and used to investigate the role of TRPV1 in RGCs in vivo. Müller cells and RGCs were analyzed by ATP release assays, intracellular calcium assays, CCK-8 assays, EdU (5-ethynyl-2'-deoxyuridine) staining, TUNEL staining, flow cytometry, and transmission electron microscopy. In vivo changes in inner retinal function were evaluated by hematoxylin and eosin (H&E) staining and TUNEL staining. Western blot analyses were performed to measure the levels of related proteins. Our data showed that HP increased the levels of ATP and Ca2+ influx in Müller cells, and those increases were accompanied by the upregulation of TRPV1 and p-PLCγ1 expression. Suppression of TRPV1 or PLCG1 expression in Müller cells significantly decreased the ATP levels and intracellular Ca2+ accumulation induced by HP. Knockdown of TRPV1, PLCG1, or P2X7 significantly decreased apoptosis and autophagy in RGCs cultured in the conditioned media of HP-treated Müller cells. Moreover, TRPV1 silencing decreased RGC apoptosis and autophagy in the in vivo model of COH. Collectively, inhibition of TRPV1/PLCγ1 and P2X7 expression may be a useful therapeutic strategy for managing RGC death in glaucoma.

Age-related changes in Scheimpflug corneal densitometry and their correlations with corneal topographic measurements in a healthy Chinese population.

Background: Objective measurement of corneal densitometry (CD) values can be used to assess corneal transparency and health status, to investigate corneal diseases, and to review anterior segment surgeries. However, literature regarding the association between CD and corneal parameters in healthy adolescent and older individuals is limited. This study investigated age-related changes in Scheimpflug CD values and their correlations with age, sex, and corneal topographic parameters. Methods: This retrospective cross-sectional observational study included 347 eyes from 181 consecutive healthy Chinese participants aged between 5 and 90 years. They were divided into 9 age groups: 5-9, 10-19, 20-29, 30-39, 40-49, 50-59, 60-69, 70-79, and 80-90 years. CD and corneal topographic measurements were measured using the Oculus Pentacam. To evaluate CD, the cornea was divided into 3 layers according to depth (anterior 120 µm, central, and posterior 60 µm), and into 4 annular regions according to diameter (0-2, 2-6, 6-10, and 10-12 mm). Results: CD across different depths and regions was positively correlated with age (all P values <0.001). Of the 3 layers of corneal thickness, CD was highest in the anterior 120 µm and lowest in the posterior 60 µm (all P values <0.05). Among the 4 annular regions spanning the corneal diameter, the lowest CD values were 6-10, 2-6, and 0-2 mm at 5-29, 30-69, and 70-89 years, respectively. The highest CD values were 10-12 mm at 5-79 years, and 6-10 mm at 80-90 years (all P values <0.05). CD values of 10-12 mm in the anterior 120 µm corneal layer were significantly lower in men than in women (Z=-2.353; P=0.019). CD of 0-10 mm in each layer was not significantly different between sexes (all P values >0.05). Corneal topographic measurements, including flat-axis keratometry (K1), steep-axis keratometry (K2), and spherical aberration, were slightly positively correlated with age and CD (all P values <0.05). However, central corneal thickness (CCT) and thinnest corneal thickness (TCT), and age or CD showed no correlation (all P values >0.05). Conclusions: With age, CD, keratometry, and spherical aberration gradually increased, while the corneal thickness did not change significantly.

The influence of pterygium on corneal densitometry evaluated using the Oculus Pentacam system.

Purpose: To assess the effect of pterygium on corneal densitometry (CD) values. Methods: One hundred and nine patients (155 eyes) with primary pterygium were divided into a severe pterygium group (79 eyes) and a mild-to-moderate pterygium group (76 eyes) according to pterygium severity. Among them, 63 patients had monocular pterygium; and 25 patients (38 eyes) underwent pterygium excision combined with conjunctival autograft follow-up. A Pentacam anterior segment analyzer was used to obtain the CD values and corneal morphological parameters, including central corneal thickness (CCT), flat-axis keratometry (K1), steep-axis keratometry (K2), corneal astigmatism, irregular astigmatism, and spherical aberration. CD was subdivided into four concentric radial regions based on corneal diameter and three layers according to depth. Results: CD values at 0-12 mm of the anterior 120 µm layer, 0-10 mm of the center layer and full thickness, and 2-6 mm of the posterior 60 µm layer were significantly higher in eyes affected by pterygium than in the contralateral unaffected eyes (all P < 0.05). CD values were significantly higher in the severe pterygium group than in the mild to moderate pterygium group (all P < 0.05). Corneal astigmatism, irregular astigmatism, K1, K2, CCT, and spherical aberration correlated with CD values in eyes with pterygium (all P < 0.05). CD values at 6-10, 0-12 mm in the anterior 120 µm layer and full thickness, 10-12 and 0-12 mm in the center layer were significantly decreased 1 month after pterygium surgery compared with those before surgery (all P < 0.05). Conclusion: Patients with pterygium had increased CD values, particularly in the anterior and central layers. CD values were correlated with pterygium severity grading and corneal parameters. Pterygium surgery partially reduced the CD values.

Visual function assessment of posterior-chamber phakic implantable collamer lenses with a central port.

Background: To investigate the short-term visual function of the V4c implantable collamer lens (ICL) for myopia using a Binoptometer 4P. Methods: Eighty eyes from 40 patients (age, 28.75±6.57 years) who underwent ICL V4c implantation at Shenzhen Eye Hospital from August 2020 to June 2021 were enrolled. Uncorrected visual acuity (UCVA), best corrected visual acuity (BCVA), subjective manifest refraction, intraocular pressure (IOP), and corneal endothelial cell density (ECD) were evaluated 1 month after surgery. Corrected distance visual acuity (CDVA), corrected near visual acuity (CNVA), contrast sensitivity (CS), near stereoacuity (NSA), twilight vision, and glare sensitivity were measured using a Binoptometer 4P. Results: The average logMAR (logarithmic minimum angle of resolution) UCVA postoperative was 0.07±0.13, which was significantly better than the preoperative BCVA (P=0.028). At 1 month postoperatively, the mean safety and efficacy indexes were 1.14±0.24 and 0.98±0.16, respectively. The mean spherical equivalent (SE) of all eyes was -8.98±3.20 D preoperatively, and -0.16±0.48 D postoperatively. None of the patients lost one or more lines of BCVA, 62.1% remained unchanged, 10.6% gained one line, and 27.3% gained two or more lines. CDVA, CNVA, CS, and NSA of both eyes were measured using a Binoptometer 4P, which were significantly better than the preoperative values. In addition, some patients had subjective symptoms in the early postoperative period, such as halo, glare, etc. Conclusions: ICL V4c implantation is a safe, effective, and predictable solution for myopia. We found that the short-term visual function of patients was improved significantly. The Binoptometer 4P is an effective and convenient visual testing device for assessing postoperative visual function after ICL V4c implantation.

RNA-seq analysis reveals differentially expressed inflammatory chemokines in a rat retinal degeneration model induced by sodium iodate.

OBJECTIVE: Retinal degeneration (RD) is a group of serious blinding eye diseases characterized by photoreceptor cell apoptosis and progressive degeneration of retinal neurons. However, the underlying mechanism of its pathogenesis remains unclear. METHODS: In this study, retinal tissues from sodium iodate (NaIO3)-induced RD and control rats were collected for transcriptome analysis using RNA-sequencing (RNA-seq). Analysis of white blood cell-related parameters was conducted in patients with retinitis pigmentosa (RP) and age-related cataract (ARC) patients. RESULTS: In total, 334 mRNAs, 77 long non-coding RNAs (lncRNAs), and 20 other RNA types were identified as differentially expressed in the retinas of NaIO3-induced RD rats. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses showed that differentially expressed mRNAs were mainly enriched in signaling pathways related to immune inflammation. Moreover, we found that the neutrophil-to-lymphocyte ratio was significantly higher in RP patients than in ARC patients. CONCLUSION: Overall, this study suggests that multiple chemokines participating in systemic inflammation may contribute to RD pathogenesis.

Case report: Biochemical and clinical phenotypes caused by cysteine substitutions in the epidermal growth factor-like domains of fibrillin-1.

Marfan syndrome, an autosomal dominant disorder of connective tissue, is primarily caused by mutations in the fibrillin-1 (FBN1) gene, which encodes the protein fibrillin-1. The protein is composed of epidermal growth factor-like (EGF-like) domains, transforming growth factor beta-binding protein-like (TB) domains, and hybrid (Hyb) domains and is an important component of elastin-related microfibrils in elastic fiber tissue. In this study, we report a cysteine to tyrosine substitution in two different domains of fibrillin-1, both of which cause Marfan syndrome with ocular abnormalities, in two families. Using protease degradation and liquid chromatography-tandem mass spectrometry analyses, we explored the different effects of substitution of cysteine by tyrosine in an EGF-like and a calcium-binding (cb) EGF-like domain on protein stability. The results showed that cysteine mutations in the EGF domain are more likely to result in altered proteolytic sensitivity and thermostability than those in the cbEGF domain. Furthermore, cysteine mutations can lead to new enzymatic sites exposure or hidden canonical cleavage sites. These results indicate the differential clinical phenotypes and molecular pathogenesis of Marfan syndrome caused by cysteine mutations in different fibrillin-1 domains. These results strongly suggest that failure to form disulfide bonds and abnormal proteolysis of fibrillin-1 caused by cysteine mutations may be an important factor underlying the pathogenesis of diseases caused by fibrillin-1 mutations, such as Marfan syndrome.

Automated detection of retinal exudates and drusen in ultra-widefield fundus images based on deep learning.

BACKGROUND: Retinal exudates and/or drusen (RED) can be signs of many fundus diseases that can lead to irreversible vision loss. Early detection and treatment of these diseases are critical for improving vision prognosis. However, manual RED screening on a large scale is time-consuming and labour-intensive. Here, we aim to develop and assess a deep learning system for automated detection of RED using ultra-widefield fundus (UWF) images. METHODS: A total of 26,409 UWF images from 14,994 subjects were used to develop and evaluate the deep learning system. The Zhongshan Ophthalmic Center (ZOC) dataset was selected to compare the performance of the system to that of retina specialists in RED detection. The saliency map visualization technique was used to understand which areas in the UWF image had the most influence on our deep learning system when detecting RED. RESULTS: The system for RED detection achieved areas under the receiver operating characteristic curve of 0.994 (95% confidence interval [CI]: 0.991-0.996), 0.972 (95% CI: 0.957-0.984), and 0.988 (95% CI: 0.983-0.992) in three independent datasets. The performance of the system in the ZOC dataset was comparable to that of an experienced retina specialist. Regions of RED were highlighted by saliency maps in UWF images. CONCLUSIONS: Our deep learning system is reliable in the automated detection of RED in UWF images. As a screening tool, our system may promote the early diagnosis and management of RED-related fundus diseases.

Intelligent cataract surgery supervision and evaluation via deep learning.

PURPOSE: To assess the performance of a deep learning (DL) algorithm for evaluating and supervising cataract extraction using phacoemulsification with intraocular lens (IOL) implantation based on cataract surgery (CS) videos. MATERIALS AND METHODS: DeepSurgery was trained using 186 standard CS videos to recognize 12 CS steps and was validated in two datasets that contained 50 and 21 CS videos, respectively. A supervision test including 50 CS videos was used to assess the DeepSurgery guidance and alert function. In addition, a real-time test containing 54 CSs was used to compare the DeepSurgery grading performance to an expert panel and residents. RESULTS: DeepSurgery achieved stable performance for all 12 recognition steps, including the duration between two pairs of adjacent steps in internal validation with an ACC of 95.06% and external validations with ACCs of 88.77% and 88.34%. DeepSurgery also recognized the chronology of surgical steps and alerted surgeons to order of incorrect steps. Six main steps are automatically and simultaneously quantified during the evaluation process (centesimal system). In a real-time comparative test, the DeepSurgery step recognition performance was robust (ACC of 90.30%). In addition, DeepSurgery and an expert panel achieved comparable performance when assessing the surgical steps (kappa ranged from 0.58 to 0.77). CONCLUSIONS: DeepSurgery represents a potential approach to provide a real-time supervision and an objective surgical evaluation system for routine CS and to improve surgical outcomes.

Comparative Transcriptomic Analysis to Identify the Important Coding and Non-coding RNAs Involved in the Pathogenesis of Pterygium.

Pterygium is a common ocular surface disease characterized by abnormal fibrovascular proliferation and invasion, similar to tumorigenesis. The formation of tumors is related to a change in the expression of various RNAs; however, whether they are involved in the formation and development of pterygium remains unclear. In this study, transcriptome analysis of messenger RNAs (mRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs) of paired pterygium and normal conjunctiva was performed to explore key genes regulating the development of pterygium. In total, 579 mRNAs, 275 lncRNAs, and 21 circRNAs were differentially expressed (DE) in pterygium compared with paired conjunctival tissues. Functional enrichment analysis indicated that DE RNAs were associated with extracellular matrix organization, blood vessel morphogenesis, and focal adhesion. Furthermore, through protein-protein interaction network and mRNA-lncRNA co-expression network analysis, key mRNAs including FN1, VCAM1, and MMP2, and key lncRNAs including MIR4435-2HG and LINC00968 were screened and might be involved in the pathogenesis of pterygium. In addition, several circRNAs including hsa_circ_0007482 and hsa_circ_001730 were considered to be involved in the pterygium development. This study provides a scientific basis for elucidating the pathogenesis of pterygium and will be beneficial for the development of preventive and therapeutic strategies.

Surgical membranectomy with modified incision and capsulotomy microscissors for persistent pupillary membrane.

BACKGROUND: To evaluate the visual outcome and complications of surgical membranectomy with modified incision and capsulotomy microscissors in patients with persistent pupillary membrane (PPM). METHODS: We enrolled eight eyes with PPM in six patients and performed surgical membranectomy with modified incision located near the limbus and corresponding to the middle of the densest membrane strands. Strands near the collarette of the iris were then cut using capsulotomy microscissors and thick strands were removed with capculorhexis forceps. Complications during or after surgery were evaluated, and uncorrected visual acuity (UCVA) and best corrected visual acuity (BCVA) were compared pre- and post-surgery. RESULTS: The mean age of the patients at surgery was 9.5±3.4 years (range, 5.3 to 13.8 years). Bilateral PPMs were found in two patients, small anterior capsular cataracts not locating on the visual axis in three eyes, and deprivational amblyopia in four eyes. There were no traumatic cataracts, endophthalmitis, corneal opacities, or other complications in patients during or after modified surgical membranectomy. After a mean follow-up period of 5.8±0.4 (range, 5.0 to 6.0) months, UCVA was significantly improved from 0.23±0.14 to 0.36±0.20(P=0.026), and BCVA was also significantly improved from 0.32±0.22 pre-operatively to 0.56±0.25 post-operatively (P=0.006). CONCLUSIONS: Surgical membranectomy with modified incision and capsulotomy microscissors may be a safe approach to clear the visual axis of patients with PPM. However further treatments were needed in amblyopic eyes after surgery.

Polar value analysis of astigmatic change and rotational stability after implantation of V4c toric implantable collamer lens.

BACKGROUND: To evaluate the clinical results and rotational stability of V4c toric implantable collamer lens (TICL, STAAR Surgical Company, Monrovia, CA, USA) in patients with moderate to high myopic astigmatism. Retrospective, interventional case series was performed at Shenzhen Eye Hospital, Shenzhen, Guangdong, China. METHODS: This study enrolled 43 patients (72 eyes) who received TICL implantation to correct myopia and myopic astigmatism. The patients underwent visual and refractive examinations before and 1 month after surgery. Astigmatic changes were estimated using polar value analysis. The difference between the achieved axis and the intended axis at the last follow-up was taken as the rotation of the V4c TICL. RESULTS: At 1 month postoperatively, the mean safety and efficacy indices were 1.17 and 1.13, respectively. A significant reduction of 8.92±2.58 D was observed in the spherical equivalent refraction (SER), which decreased from -9.29±2.41 D preoperatively to -0.37±0.55 D postoperatively. The astigmatic error of treatment in cylinder format was calculated to 0.50±0.41 @ 15.08° relative to the preoperative stronger meridian at 1 month, postoperatively. At 1 month postoperatively, the mean absolute rotation was 8.30±10.00 degrees (median =5.46 degrees; range, 0.00-58.88 degrees). CONCLUSIONS: TICL could achieve good astigmatic outcomes for correcting moderate to high myopic astigmatism. After TICL implantation, corneal astigmatism remained unchanged. To optimize postoperative astigmatic outcomes in TICL, polar value analysis can be used to build a nomogram.

Polymorphisms and Circulating Plasma Protein Levels of Immune Checkpoints (CTLA-4 and PD-1) Are Associated With Posner-Schlossman Syndrome in Southern Chinese.

Cytotoxic T-lymphocyte associated protein 4 (CTLA-4) and programmed cell death 1 (PD-1) are well-known key immune checkpoints that play a crucial dampening effect on regulating T-cell homeostasis and self-tolerance. In this study, we aimed to evaluate the association between immune checkpoints (CTLA-4 and PD-1) and Posner-Schlossman syndrome (PSS) in a southern Chinese population. A total of 137 patients with PSS and 139 healthy controls from a southern Chinese population were recruited. Five single nucleotide polymorphisms (SNPs) of CTLA-4 (rs733618, rs4553808, rs5742909, rs231775, and rs3087243) and five SNPs of PD-1 (rs10204525, rs2227981, rs2227982, rs41386349, and rs36084323) were genotyped by SNaPshot technique. Soluble CTLA-4 (sCTLA-4) and soluble PD-1 (sPD-1) were determined by ELISA and antibody array assay, respectively. The frequencies of T allele at rs733618 and A allele at rs231775 of CTLA-4 were significantly higher in PSS patients than in healthy controls (corrected p (Pc ) = 0.037; Pc = 0.044, respectively). The haplotype frequencies of CACGG haplotype (rs733618-rs4553808-rs5742909-rs231775-rs3087243) of CTLA-4 and TGAGC haplotype (rs10204525-rs2227981-rs2227982-rs41386349-rs36084323) of PD-1 in the PSS group was significantly lower than those in the control group (Pc = 0.015, p = 0.034, respectively). Circulating plasma levels of sCTLA-4 and sPD-1 in PSS patients were significantly higher than those in controls (all p < 0.001). The present study suggests that CTLA-4 and PD-1 genetic polymorphisms are associated with the susceptibility to PSS in a southern Chinese population. The upregulated circulating plasma protein levels of sCTLA-4 and sPD-1 might provide some hints regarding the dysfunction of immune checkpoints in PSS during the active status.

Deep learning for automated glaucomatous optic neuropathy detection from ultra-widefield fundus images.

BACKGROUND/AIMS: To develop a deep learning system for automated glaucomatous optic neuropathy (GON) detection using ultra-widefield fundus (UWF) images. METHODS: We trained, validated and externally evaluated a deep learning system for GON detection based on 22 972 UWF images from 10 590 subjects that were collected at 4 different institutions in China and Japan. The InceptionResNetV2 neural network architecture was used to develop the system. The area under the receiver operating characteristic curve (AUC), sensitivity and specificity were used to assess the performance of detecting GON by the system. The data set from the Zhongshan Ophthalmic Center (ZOC) was selected to compare the performance of the system to that of ophthalmologists who mainly conducted UWF image analysis in clinics. RESULTS: The system for GON detection achieved AUCs of 0.983-0.999 with sensitivities of 97.5-98.2% and specificities of 94.3-98.4% in four independent data sets. The most common reasons for false-negative results were confounding optic disc characteristics caused by high myopia or pathological myopia (n=39 (53%)). The leading cause for false-positive results was having other fundus lesions (n=401 (96%)). The performance of the system in the ZOC data set was comparable to that of an experienced ophthalmologist (p>0.05). CONCLUSION: Our deep learning system can accurately detect GON from UWF images in an automated fashion. It may be used as a screening tool to improve the accessibility of screening and promote the early diagnosis and management of glaucoma.

Knockdown of MALAT1 attenuates high-glucose-induced angiogenesis and inflammation via endoplasmic reticulum stress in human retinal vascular endothelial cells.

Diabetic retinopathy (DR) is one of the most severe complications of diabetes mellitus, and retinal endoplasmic reticulum stress (ERS) plays an important role in the pathogenesis of DR. However, the exact mechanisms by which ERS mediates DR remain unclear. In this study, human retinal vascular endothelial cells (RVECs) were cultured in high-glucose (HG) medium to mimic the environment of DR. The expression of long non-coding RNA (lncRNA)-metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) was determined by quantitative real time PCR. ERS markers (glucose-regulated protein 78 [GRP78] and C/EBP homologous protein [CHOP]) were measured by immunofluorescence and western blotting. Cell viability was analyzed by the CCK-8 assay. The angiogenesis of RVECs was evaluated by tube formation assays. The levels of pro-inflammation cytokines TNF-α and IL-6 in RVECs were determined by ELISA assays. We found that exposure to HG levels upregulated MALAT1 and GRP78 expression in RVECs. While, GRP78 overexpression strengthened CHOP expression, cell proliferation suppression, capillary morphogenesis and inflammation in HG-treated RVECs. Importantly, knockdown of MALAT1 reversed HG-induced cell proliferation suppression, inhibited capillary morphogenesis, and inflammation in RVECs, and those effects were reversed by GRP78 overexpression. These results suggest that MALAT1 promotes HG-induced angiogenesis and inflammation in RVECs by upregulating ER stress, and might be target for treating DR.

Involvement of the JNK signaling in granular corneal dystrophy by modulating TGF-ß-induced TGFBI expression and corneal fibroblast apoptosis.

Granular corneal dystrophy (GCD) is featured by corneal deposits of transforming growth factor beta-induced gene (TGFBI) mediated by the TGF-ß (transforming growth factor-ß)/Smad signaling. However, the roles of c-Jun amino-terminal kinase (JNK) pathway in GCD pathogenesis remains unexplored, which was investigated in this study. JNK signaling activation and inhibition in primary corneal fibroblasts were obtained by treatments with anisomycin and SP600125, respectively. Protein abundance and phosphorylation were detected by immunoblotting. Cell viability and apoptosis were analyzed by CCK-8 and flow cytometry respectively. TGFBI deposit and autophagy progression were assessed by immunofluorescence. The results found that JNK1 expression and phosphorylation were greatly increased in corneal tissues from GCD2 patients. JNK signaling activation impaired the viability and promoted apoptosis and autophagy processes in primary corneal fibroblasts, along with Smad2/3 phosphorylation, TGFBI accumulation and Bcl-2 suppression. Autophagy related proteins, such as ATG5 (autophagy related 5), ATG12 (autophagy related 12) and LC3B (microtubule-associated protein 1 light chain 3 beta), were also increased in anisomycin or TGF-ß1 treated corneal fibroblasts. However, SP600125 effectively reversed the above effect induced by TGF-ß1 treatment in corneal fibroblasts, including the TGF-ß-induced autophagy progression. The results suggested that JNK signaling was activated in GCD2 corneal tissues, and it mediated the TGF-ß-induced TGFBI protein accumulation and apoptosis of corneal fibroblasts during GCD2 pathogenesis.

Deep learning from "passive feeding" to "selective eating" of real-world data.

Artificial intelligence (AI) based on deep learning has shown excellent diagnostic performance in detecting various diseases with good-quality clinical images. Recently, AI diagnostic systems developed from ultra-widefield fundus (UWF) images have become popular standard-of-care tools in screening for ocular fundus diseases. However, in real-world settings, these systems must base their diagnoses on images with uncontrolled quality ("passive feeding"), leading to uncertainty about their performance. Here, using 40,562 UWF images, we develop a deep learning-based image filtering system (DLIFS) for detecting and filtering out poor-quality images in an automated fashion such that only good-quality images are transferred to the subsequent AI diagnostic system ("selective eating"). In three independent datasets from different clinical institutions, the DLIFS performed well with sensitivities of 96.9%, 95.6% and 96.6%, and specificities of 96.6%, 97.9% and 98.8%, respectively. Furthermore, we show that the application of our DLIFS significantly improves the performance of established AI diagnostic systems in real-world settings. Our work demonstrates that "selective eating" of real-world data is necessary and needs to be considered in the development of image-based AI systems.